V O L U M E 8 1 · N o.

9 · S E P T E M B E R 2 0 1 5
MINERVA ANESTESIOLOGICA
ITALIAN JOURNAL OF ANESTHESIOLOGY AND ANALGESIA
MONTHLY JOURNAL FOUNDED IN 1935 BY A. M. DOGLIOTTI
OFFICIAL JOURNAL OF ITALIAN SOCIETY OF ANESTHESIOLOGY, ANALGESIA,
RESUSCITATION AND INTENSIVE CARE (S.I.A.A.R.T.I.)

Vol. 81 September 2015 No. 9

OFFICIAL JOURNAL OF ITALIAN SOCIETY OF ANESTHESIOLOGY, ANALGESIA,

RESUSCITATION AND INTENSIVE CARE (SIAARTI)

CONTENTS

937 951
EDITORIALS Hyperglycemia and ambulatory surgery
Lights and shadows of palliative sedation in Italy: Polderman J. A., Van Velzen L., Wasmoeth L. G., Eshuis
the role of anesthetists J. H., Houweling P. L., Hollmann M. W., Devries J. H.,
Preckel B., Hermanides J.
Orsi L.

940
Changing epidemiology of pathogens in ICU bactere-
mia: should we take the path of least resistance?
Opdam M. H., Van Zanten A. R. H.
960
Preoperative adherence to continuous positive airway
pressure among obstructive sleep apnea patients
Deflandre E., Degey S., Bonhomme V., Donneau A. F.,
Poirrier R., Brichant J. F., Hans P.

943
Localization of central venous catheter by vascular 968
ultrasound and transthoracic echocardiography: easy Careful monitoring of the use of sedative drugs
and accurate? at the end of life: the role of Epidemiology. The
Bolliger D., Seeberger M. D. ITAELD study
Miccinesi G., Caraceni A., Raho J. A., Paci E., Bulli F.,
Van Den Block L., Giannini A.

946
ORIGINAL ARTICLES
Randomized controlled pilot trial of the rigid and flex-
ing laryngoscope versus the fiberoptic bronchoscope
for intubation of potentially difficult airway
Alvis B. D., King A. B., Hester D., Hughes C. G., Higgins
M. S.
980
Changed epidemiology of ICU acquired bloodstream
infections over 12 years in an Italian teaching hospital
Orsi G. B., Giuliano S., Franchi C., Ciorba V., Protano C.,
Giordano A., Rocco M., Venditti M.

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA I

CONTENTS

989 1031
Confirmation of correct central venous catheter posi- Safe mechanical ventilation in patients without acute
tion in the preoperative setting by echocardiographic respiratory distress syndrome (ARDS)
“bubble-test” Pannu S. R., Hubmayr R. D.
Meggiolaro M., Scatto A., Zorzi A., Roman-Pognuz E.,
Lauro A., Passarella C., Bonaccorso G.
1041
LETTERS TO THE EDITOR
1001 Non-adherence to CPAP and prevention of postop-
EXPERT OPINION erative complications in OSA: what are the limits of
Continuous regional anesthesia and inflammation: a anesthesia consultation?
new target Esquinas A. M., Insalaco G.
Grosu I., Lavand’homme P.

1042
1010 Obstructive sleep apnea and detection of non-adher-
Blood glucose amplitude variability in critically ill ence to CPAP in OSA: limits of the preanesthesia visit
patients Deflandre E., Brichant J. F., Bonhomme V.
Meyfroidt G.

1044
1019 An additional tip to facilitate glidescope intubation
REVIEW Turkstra T., Rachinsky M., Batohi P.
Colloids versus crystalloids in the prevention of hypo-
tension induced by spinal anesthesia in elective cesar-
ean section. A systematic review and meta-analysis
Ripollés Melchor J., Espinosa Á., Martínez Hurtado E.,
1045
Casans Francés R., Navarro Pérez R., Abad Gurumeta A., TOP 50 MINERVA ANESTESIOLOGICA
Calvo Vecino J. M. REVIEWERS

V O L U M E 8 1 · N o. 9 · S E P T E M B E R 2 0 1 5

About the cover: the cover shows a bubble test performed in a patient with malposition. For more
information, see article by Meggiolaro M. et al. beginning on page 989.

II MINERVA ANESTESIOLOGICA September 2015

MINERVA ANESTESIOLOGICA
ITALIAN JOURNAL OF ANESTHESIOLOGY AND ANALGESIA
MONTHLY JOURNAL FOUNDED IN 1935 BY A. M. DOGLIOTTI
OFFICIAL JOURNAL OF ITALIAN SOCIETY OF ANESTHESIOLOGY, ANALGESIA,
RESUSCITATION AND INTENSIVE CARE (S.I.A.A.R.T.I.)

EDITORIAL BOARD

EDITOR IN CHIEF
F. Cavaliere
Roma, Italy
CRITICAL CARE ANESTHESIA

General Critical Care General Anesthesia

Associate Editor Associate Editor
G. M. Albaiceta (Oviedo, Spain) M. Rossi (Roma, Italy)
Section Editor Section Editor
G. Biancofiore (Pisa, Italy) E. Cohen (New York, USA)
E. De Robertis (Napoli, Italy) P. Di Marco (Roma, Italy)
J. T. Knape (Utrecht, The Netherlands)
O. Langeron (Paris, France)
P. M. Spieth (Dresden, Germany)
Circulation Critical Care Pediatric Anesthesia
Section Editor Section Editor
S. Scolletta (Siena, Italy) M. Piastra (Roma, Italy)
E. Bignami (Milano, Italy),
Respiration Critical Care Obstetric Anesthesia
Section Editor Section Editor
S. Grasso (Bari, Italy) E. Calderini (Milano, Italy)
P. Terragni (Sassari, Italy),
Neurocritical Care Regional Anesthesia
Section Editor Section Editor
F. S. Taccone (Brussels, Belgium) A. Apan (Giresun, Turkey)
M. Carassiti (Roma, Italy)

ETHICS PAIN
Section Editor Section Editor
A. Giannini (Milano, Italy) M. Allegri (Parma, Italy)
F. Coluzzi (Roma, Italy)
MEDICAL STATISTIC
Section Editor
B. M. Cesana (Brescia, Italy)

MANAGING EDITOR
A. Oliaro
Torino, Italy

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA III

 GENERAL INFORMATION

This journal is PEER REVIEWED and is quoted in:

Current Contents, SciSearch, PubMed/MEDLINE, EMBASE, Scopus

Impact factor: 2,134

Published by
Edizioni Minerva Medica - Corso Bramante 83-85 - I-10126 Turin (Italy)
Tel. +39 (011) 678282 - Fax +39 (011) 674502
Website: www.minervamedica.it
Editorial office: minerva.anestesiologica@minervamediac.it - Subscriptions: subscriptions.dept@minervamedica.it
Advertising: journals3.dept@minervamedica.it
Presidential Office: Istituto di Anestesiologia e Rianimazione, Policlinico Universitario A. Gemelli, Università Cattolica del
Sacro Cuore, Largo A. Gemelli 8, 00168 Roma, Italy - Tel. 06 30153226.
Technical Secretariat: S.I.A.A.R.T.I., Viale dell’Università 11, 00185 Roma, Italy - Tel. 06 4452816
E-mail: segreteria@siaarti.it - Website: www.siaarti.it
Administrative Secretariat S.I.A.A.R.T.I. - Edizioni Minerva Medica - Corso Bramante 83-85 - 10126 Torino - Tel. +39
(011) 678282 - Fax +39 (011) 674502 - E-mail: minervamedica@minervamedica.it

ANNUAL SUBSCRIPTIONS

OUTSIDE THE
ITALY EUROPEAN UNION EUROPEAN UNION

Individual: Print € 130,00 Individual: Print € 225,00 Individual: Print € 250,00

Print+Online € 135,00 Print+Online € 235,00 Print+Online € 260,00
Online € 140,00 Online € 245,00 Online € 270,00

Institutional: Print € 180,00 Institutional: Print € 300,00 Institutional: Print € 335,00

Online Small € 315,00
Medium € 350,00
Large € 407,00
Extra Large € 429,00
Print+Online Small € 325,00
Medium € 367,00
Large € 420,00
Extra Large € 446,00
Online Small € 315,00
Medium € 351,00
Large € 409,00
Extra Large € 429,00
Print+Online Small € 336,00
Medium € 378,00
Large € 430,00
Extra Large € 456,00
Online Small € 351,00
Medium € 383,00
Large € 446,00
Extra Large € 462,00
Print+Online Small € 372,00
Medium € 404,00
Large € 467,00
Extra Large € 483,00

Subscribers: Payment to be made in Italy: a) by check; b) by bank transfer to: Edizioni Minerva Medica, INTESA SANPAOLO Branch
no. 18 Torino. IBAN: IT45 K030 6909 2191 0000 0002 917 c) through postal account no. 00279109 in the name of Edizioni Minerva
Medica, Corso Bramante 83-85, 10126 Torino; d) by credit card Diners Club International, Master Card, VISA, American Express.
Foreign countries: a) by check; b) by bank transfer to: Edizioni Minerva Medica, INTESA SANPAOLO Branch no. 18 Torino. IBAN:
IT45 K030 6909 2191 0000 0002 917; BIC: BCITITMM c) by credit card Diners Club International, Master Card, VISA, American
Express. Members: for payment please contact the Administrative Secretariat SIAARTI.
Notification of changes to mailing addresses, e-mail addresses or any other subscription information must be received in good time.
Notification can be made by sending the new and old information by mail, fax or e-mail or directly through the website www.minerva-
medica.it at the section “Your subscriptions - Contact subscriptions department”.
Complaints regarding missing issues must be made within six months of the issue’s publication date.
Prices for back issues and years are available upon request.
© Edizioni Minerva Medica - Torino 2015. All rights reserved. No part of this publication may be reproduced, stored or transmitted in
any form or by any means, without the prior permission of the copyright owner.
Monthly publication. Authorized by Turin Court no. 281 of 24-1-1953.

This Journal is associated with This Journal complies with the Code of Self-Discipline of Medical/Scientific Publishers associated with FARMAMEDIA and may accept advertising

IV MINERVA ANESTESIOLOGICA September 2015


EDITORIAL
Lights and shadows of palliative sedation in Italy:
the role of anesthetists
L. ORSI
Palliative Care Unit, Carlo Poma Hospital, Mantova, Italy
I n Italy, the proportion of expected or non-
sudden deaths among patients subjected to
PS has increased from 8.5% (2001-2002) 2 to
18.2% (2007).1 This last percentage is similar to
that observed in other countries in Europe (15%
Belgium; 18.7% United Kingdom) 3, 4 that have
a sound palliative care culture.
This upward trend may be considered as a
positive outcome. However, statistics highlights
that nowadays in Italy the number of patients
who do not receive any appropriate PS and ad-
equate symptoms’ control during their last days
of life, in any health facilities, is still high. In fact,
it is not by chance that in Italy, in 2012, the opi-
oids consumption in terms of morphine equiva-
lence (mg/capita) was 144.2 vs. 513.4 in Austria,
395.1 in Switzerland, 256.7 in Germany, 245.4
in the UK, 239 in Spain, 212.8 in France, and
102.2 in Greece.5
Thus, also taking into consideration the up-
ward trend reported by Miccinesi et al.1 in this
issue of Minerva Anestesiologica, on the whole
the need for PS of dying or suffering patients is
far from being fulfilled.
In this regard, all Italian physicians, and not
only palliativists, have to strive to ensure a bet-
ter quality of death in line with the Italian Code
of Medical Ethics (art. 39) 6 and the guidelines
recently published by the Council of Europe.7
There is a second positive outcome: the rough
number of deceased patients assisted by anesthe-
tists and intensivists (A-I) represents a small per-
Comment on p. 968.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 937
centage of all deaths in the study’s sample, but
the proportional incidence of dead patients in
this group is quite high (40 out 73 deaths) when
compared with the incidence of deaths in other
groups of patients assisted by different special-
ists.
These data show that A-I, thanks to their sci-
entific and ethical background,8 could play a
new and relevant role with regard to the end-of-
life palliative care, at least in a hospital setting.9
It is crucial that A-I, even at academic level,
have a full awareness of this new perspective,
opening a window onto unexplored care path-
ways based on a close collaboration with the
palliative care team and aimed at widening and
improving palliative care and PS at the end of
life.
In home care settings the role of general prac-
titioners (GPs) is emerging in connection with
the palliative care team. GPs were the least likely
to report PS but they performed the procedure
in more than half the cases (51.8%), delivering
palliative/pain therapy in the last month of life
in 2 out of 3 cases.
These data convey the third good news: Ital-
ian GPs are actively involved in end-of-life care
within the Local Palliative Care Network 10, 11
and palliative care specialist supervision prevents
wrong sedative use as experienced in Belgium 12
or the Netherlands.13
From an ethical point of view, this study shows
that a positive/negative attitude towards physi-
cian-assisted death is not related to PS practice
because these two issues seem to be independent.
ORSI Lights and shadows of palliative sedation in Italy:
This finding is important to reassure health care
professionals or lay people (caregivers or family
members) about the ethical validity of this last-
resort therapy for refractory symptoms.14, 15
According to the ethical principles of PS
pointed out in SICP 16 and EAPC 17 recom-
mendations, in 81% of cases of PS, proportional
increase of sedative drugs and presence of refrac-
tory symptoms were respectively the main way
and reason to use PS.
By contrast, in only 7.9% of cases an opioid’s
dose higher than necessary was reported. This re-
sult, however, should be cautiously interpreted
because self-evaluation by the physician may be
influenced by a potential bias,1 especially when
the “European framework of end-of-life deci-
sions” (which include those medical decisions
that take into account the possibility that death
is hastened) is used. This frame work includes a
too wide (and likely confusing) range of practic-
es in the same group such as forgoing futile treat-
ments, palliative sedation, and hastening death.
Clinical practice and educational experience
show the harmful confusion still existing today
with regard to these issues. Accordingly, PS is
unexpectedly considered by health care profes-
sionals as a procedure borderline with euthanasia
(ethical “gray” zone), despite the clear evidence
that a proportional increase of sedatives do not
hasten death.18, 19
In Western countries the increasing number
of elderly people led in recent years to a signifi-
cant proportion of patients with degenerative,
chronic diseases. Despite the new resources pro-
vided by biomedical sciences, the natural evolu-
tion of these diseases do not change and when
the “end-stage” gets close, clinicians have to face
end-of-life situations that challenge their profi-
ciency in turning the cure into care.
In this regard, Miccinesi’s study shows that,
due to an incorrect pharmacological approach, a
high number of patients were sedated only with
opioids (45.6%). The high prevalence (74.4%)
of artificial nutrition/hydration during the last
days is consistent with these data.
According to us, in the era of modern medi-
cine, and especially in academic settings, all ef-
forts should be carried out with the intent to
make palliative care medicine and related end-
938 MINERVA ANESTESIOLOGICA September 2015
of-life care practices an integral part of medical
education and clinical training programs.
Thus, considering the above-mentioned ethi-
cal and scientific backgrounds, A-I and palliative
care teams should now be encouraged to collab-
orate in education programs dedicated to train-
ing other clinicians serving in all care settings.
In conclusion, two questions are demanding
a compelling answer:20 will Intensive Care Unit
and palliative care teams meet these new chal-
lenges? Will academic institutions include basic
end-of-life care in the curricula of doctors and
nurses?
References
 1. Miccinesi G, Caraceni A, Raho JA, Paci E, Bulli F, Van
den Block L et al. Careful
��������������������������������������
monitoring of the use of seda-
tive drugs at the end of life: the role of Epidemiology. The
ITAELD study. Minerva Anestesiol 2015;81:968-79.
  2. Miccinesi G, Rietjens JA, Deliens L, Paci E, Bosshard G,
Nilstun T et al. Continuous deep sedation: physicians’ ex-
periences in six European countries. J Pain Symptom Man-
age 2006;31:122-9.
  3. Van den Block L, Deschepper R, Bilsen J, Bossuyt N, Van
Casteren V, Deliens L. Euthanasia and other end-of-life
decisions: a mortality follow-back study in Belgium. BMC
Public Health 2009;9:79.
  4. Van der Heide A, Onwuteaka-Philipsen BD, Rurup ML,
Buiting HM, van Delden JJ, Hanssen-de Wolf JE et al.
End-of-life practices in the Netherlands under the Eutha-
nasia Act. N Engl J Med 2007;356:1957-65.
  5. Pain and Policy Studies Group [Internet]. [cited 2015 Janu-
ary 20]. Available from: http://www.painpolicy.wisc.edu/
country/profile/italy
 6. Code of Medical Ethics (art. 39) [Internet]. [cited 2015
January 20]. Available from: http://www.fnomceo.
it/fnomceo/Codice+di+Deontologia+Medica+2014.
html?t=a&id=115184
  7. Guide on the decision-making process regarding medical
treatment in end-of-life situations [Internet]. [cited 2015
January 20]. Available from: http://undirittogentile.files.
wordpress.com/2014/05/guida_finevita.pdf
  8. Gruppo di Studio Bioetica SIAARTI. End-of-life care and
the intensivist: Italian Society of Anaesthesia Analgesia and
Intensive Care Medicine (SIAARTI) Recommendations on
the Management of the Dying Patient. Minerva Anestesiol
2006;72:927-63.
  9. Randall Curtis J, Vincent JL. Ethics and end-of-life care for
adults in the intensive care unit. Lancet 2010;375:1347-53.
10. Hilton AK, Jones D, Bellomo R. Clinical review: the role of
the intensivist and the rapid response team in nosocomial
end-of-life care. Crit Care 2013;17:224-35.
10. Parlamento Italiano Legge n. 38 del 15 marzo 2010 “Di-
sposizioni per garantire l’accesso alle cure palliative e alla te-
rapia del dolore” [Internet]. [cited
�������������������������������
2015 January 20]. Avail-
able from: http://www.sicp.it/web/procedure/protocollo.cf
m?List=WsIdEvento,WsPageNameCaller,WsIdRisposta,W
sRelease&c1=NORMSICP&c2=%2Fweb%2Feventi%2F
CPMEET%2Findex%2Ecfm&c3=8&c4=1
11. Intesa del 25 luglio 2012- repertorio atti n. 151 ‑ Conferen-
za Stato-Regioni Definizione requisiti minimi e modalità
organizzative per accreditamento delle strutture di assisten-
za ai malati in fase terminale e delle unità di cure palliative
Lights and shadows of palliative sedation in Italy: ORSI

e terapia del dolore. [Internet]. [cited 2015 January 20]. collo.cfm?List=WsIdEvento,WsPageNameCaller,WsIdRisp

Available from: http://www.sicp.it/web/procedure/proto-
collo.cfm?List=WsIdEvento,WsPageNameCaller,WsIdRisp
osta,WsRelease&c1=NORMSICP&c2=%2Fweb%2Feven
ti%2FCPMEET%2Findex%2Ecfm&c3=7&c4=1
12. Anquinet L, Rietjens JA, Van den Block L, Bossuyt N, De-
liens L. General practitioners’ report of continuous deep se-
dation until death for patients dying at home: a descriptive
study from Belgium. Eur J Gen Pract 2011;17:5-13.
13. Verkerk M, van Wijlick E, Legemaate J, de Graeff A. A na-
tional guideline for palliative sedation in the Netherlands. J
Pain Symptom Manage 2007;34:666-70.
14. Maltoni M, Scarpi E, Nanni O. Palliative sedation in end-
of-life care. Curr Opin Oncol 2013;25:360-7.
15. Olsen ML, Swetz KM, Mueller PS. Ethical decision making
with end-of-life care: palliative sedation and withholding
or withdrawing life-sustaining treatments. Mayo Clin Proc
2010;85:949-54.
16. SICP- gruppo di studio su etica e cultura al termine della
vita. Raccomandazioni della SICP sulla sedazione termina-
le/sedazione palliativa. [Internet]. [cited 2015 January 20].
Available from: http://www.sicp.it/web/procedure/proto-
osta,WsRelease&c1=DOCSICP&c2=%2Fweb%2Feventi
%2FCPMEET%2Findex%2Ecfm&c3=7&c4=1
17. EAPC recommended framework for the use of sedation
in palliative care Palliative Medicine 2009;23(7):581-593.
Relevance: 3018 [Internet]. [cited 2015 January 20]. Avail-
able from:
http://www.eapcnet.eu/Corporate/AbouttheEAPC/EAP-
Cpublications/EAPCrecommendations/tabid/1616/Arti-
cleid/133/mod/3090/Default.aspx - 19/05/2014 17:54:29
18. Morita T, Chinone Y, Ikenaga M, Miyoshi M, Nakaho T,
Nishitateno K et al. ��Efficacy
�������������������������������������
and safety of palliative seda-
tion therapy: a multicenter, prospective, observational study
conducted on specialized palliative care units in Japan. J
Pain Synptom Manage 2005;30:320-8.
19. Sykes N, Thorns A. Sedative use in the last week of life and
the implications for end-of-life decision making. Arch Inter
Med 2003;163:341-4.
20. Kompanje EJO. The worst is yet to come. Many elderly
patients with chronic terminal illnesses will eventually
die in the emergency department. Intensive Care Med
2010;36:732-4.

Conflict of interest.—The author certifies that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on January 7, 2015. - Accepted for publication on January 16, 2015. - Epub ahead of print on January 23, 2015.
Corresponding author: L. Orsi, Unito f Palliative Care, Carlo Poma Hospital, via Lago Paiolo, 46100 Mantova, Italy.
E-mail: luciano.orsi@aopoma.it

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 939


EDITORIAL
Changing epidemiology of pathogens in ICU
bacteremia: should we take the path of least resistance?
M. H. OPDAM 1, A. R. H. VAN ZANTEN 2
Department of Intensive Care, Gelderse Vallei Hospital, Ede, The Netherlands
S evere sepsis and septic shock are the lead-
ing causes of mortality among critically ill
patients.1 The definition of sepsis is based on
the combination of both an inflammatory re-
sponse and a proven or suspected infection. In
the situation that sepsis is complicated by organ
dysfunction by definition it will be classified as
“severe sepsis”. Related complications such as
hyperlactataemia or hypotension refractory to
fluid resuscitation are classified as “septic shock”
with mortality rates ranging from 50% to 80%,
respectively.2 Severe sepsis is due to bacterial in-
fections in the majority of cases, however in only
one third of the cases blood cultures are posi-
tive. Moreover, in 30% of cases cultures from all
sites are negative.3 Culture-positive sepsis is as-
sociated with significantly higher mortality rates
compared to culture-negative sepsis.4
The cornerstones of successful treatment of
(severe) sepsis comprise early recognition, often
based on a probable diagnosis, administration of
broad-spectrum antibiotics within “ the golden
hour” after diagnosing severe sepsis or septic
shock and infection source control within 12
hours of diagnosis if necessary.5 Blood cultures
must be obtained prior to the administration of
broad-spectrum antibiotics and daily reassess-
ment of antimicrobial therapy in order to pro-
mote de-escalation should be performed.5
Worldwide emergence of multidrug resistance
(MDR) to widely used antibiotics is observed,
especially among gram-negative bacteria.6 The
Comment on p. 980.
940 MINERVA ANESTESIOLOGICA September 2015
increasing use of broad-spectrum antibiotics and
lack of good stewardship have contributed to the
increase in these MDR organisms.7
The reduction of gram-positive bloodstream
infections (BSI) may be due to the awareness of
line sepsis and programs to prevent central line
associated BSI (CLABSI).8 As a consequence
proportionally more gram-negative infections
will be encountered.
In this issue of Minerva Anestesiologica Orsi et
al. compared the aetiology of ICU acquired BSIs
over time.9 Their results clearly demonstrate a
shift from gram-positive to gram-negative ICU-
acquired bacterial bloodstream pathogens from
2000-2007 to 2010-2012 associated with an in-
crease of in-hospital mortality. In addition, they
observed an increase in incidence of multidrug
resistance among the Enterobacteriaceae, in par-
ticular for Klebsiella Pneumoniae.
Zilberberg et al. conducted a retrospective co-
hort study among adult ICU patients with bac-
teraemia and severe sepsis/septic shock caused by
gram-negative bacteria.10 Hospital mortality was
33% among patients with gram-negative bac-
teraemia. Inadequate initial antibiotic therapy
was encountered more frequent in non-survivors
compared to survivors (43.4% versus 14.6%,
P<0.001). Failure to institute initial adequate
antibiotic therapy was an independent predictor
of hospital mortality (adjusted OR 3.87, 95%
CI 2.77 to 5.41). A strong association between
MDR pathogens and inadequate initial antibi-
otic therapy was found.
Therefore, the administration of antimi-
CHANGING EPIDEMIOLOGY OF PATHOGENS IN ICU BACTEREMIA OPDAM
crobial therapy should be both early and ade-
quate.11 This may be increasingly challenging in
the context of emerging BSIs caused by MDR
pathogens. As a result more antibiotics, espe-
cially broad-spectrum antibiotics with MDR
coverage will be prescribed.
To circumvent this problem, alternative strat-
egies to prevent BSIs and to advocate proper use
of antimicrobials are of paramount importance.
Antibiotic stewardship should play a pivotal
role in the treatment of infections in the criti-
cally ill.
Selective decontamination of the digestive
tract and selective oropharyngeal decontamina-
tion in critically ill patients have been shown
to prevent MDR gram-negative infections.12
Furthermore, preventive measures to improve
hand hygiene and aseptic conditions during
catheter insertion are successful to decrease the
incidence of CLABSI.13, 14
New developments in rapid molecular test-
ing of pathogens and antibiotic resistance may
reduce the duration of pre-emptive therapy
(without specific information on the pathogen)
in the near future.15, 16 Meanwhile, bedside
clinical scores can be used as a prediction tool
to select high-risk patients with MDR pathogen
infections, that will likely benefit from broad-
spectrum antibiotics with MDR coverage.17, 18
Therapeutic strategies to combine multiple
simultaneously administered and complimen-
tary spectrum antimicrobial medications must
be designed.19 At present a cluster-randomized
crossover trial studying the effects of mixing and
cycling broad-spectrum antibiotics in the ICU
to reduce the incidence of MDR gram-negative
infections is ongoing.20
With the emergence of MDR pathogens in
ICU acquired infections we should carefully
select our empiric broad-spectrum antimicrobi-
als with MDR coverage. In case a non-MDR
pathogen is cultured immediate de-escalation
is mandatory. The combination of rapid di-
agnostic techniques, preventive measures, ev-
idence-based therapeutic strategies and most
importantly antibiotic stewardship are essential
to preserve current antibiotics, prevent further
development of resistance to be able to improve
patient outcomes now and in the future.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 941
Therefore, although it may seem attractive to
take the path of least resistance and increase the
use of antibiotics with MDR-coverage parallel
with the emerging gram-negative MDR inci-
dence, we should apply principles of antibiotic
stewardship and be very cautious not to overuse
these last resort medications. Of course, all our
patients deserve early and adequate antibiotic
therapy, however, the path of least resistance in
this context may finally lead to the most resistance
and the risk to arrive in the post-antibiotic era.
References
  1. Brun-Buisson C, Meshaka P, Pinton P, Vallet B; EPISEPSIS
Study Group. EPISEPSIS: a reappraisal of the epidemiol-
ogy and outcome of severe sepsis in French intensive care
units. Intensive Care Med 2004;30:580-8.
  2. Jawad I, Lukšić I, Rafnsson SB. Assessing
����������������������������
available informa-
tion on the burden of sepsis: global estimates of incidence,
prevalence and mortality. J Glob Health 2012;2:010404.
  3. Angus DC, Van der Poll T. Severe sepsis and septic shock. N
Engl J Med 2013;369:840-51.
  4. Phua J, Ngerng W, See K, Tay C, Kiong T, Lim H et al.
Characteristics and outcomes of culture-negative versus
culture-positive severe sepsis. Crit Care 2013;17:R202.
  5. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach
H, Opal SM et al. Surviving sepsis campaign: international
guidelines for management of severe sepsis and septic shock:
2012. Crit Care Med 2013;41:580-637.
  6. Tabah A, Koulenti D, Laupland K, Misset B, Valles J, Bru-
zzi de Carvalho F et al. Characteristics and determinants
of outcome of hospital-acquired bloodstream infections in
intensive care units: the EUROBACT International Cohort
Study. Intensive Care Med 2012;38:1930-45.
  7. Kunz AN1, Brook I. Emerging resistant Gram-negative aer-
obic bacilli in hospital-acquired infections. Chemotherapy
2010;56:492-500.
  8. Pronovost P, Needham D, Berenholtz S, Sinopoli D, Chu
H, Cosgrove S et al. An intervention to decrease catheter-
related bloodstream infections in the ICU. N Engl J Med
2006;355:2725-32.
 9. Orsi GB, Giuliano S, Franchi C, Ciorba V, Protano C,
Giordano A et al. Changing epidemiology of ICU acquired
bloodstream infections over 12 years in an Italian teaching
hospital. Minerva Anestesiol 2015;81:980-8.
10. Zilberberg MD, Shorr AF, Micek ST, Vazquez-Guillamet
C, Kollef MH. Multi-drug resistance, inappropriate initial
antibiotic therapy and mortality in Gram- negative severe
sepsis and septic shock: A retrospective cohort study. Criti-
cal Care 2014;18:596.
11. Van Zanten AR. The golden hour of antibiotic administra-
tion in severe sepsis: avoid a false start striving for gold. Crit
Care Med 2014;42:1931-2.
12. De Smet AM, Kluytmans JA, Cooper BS, Mascini EM,
Benus RF, van der Werf TS et al. Decontamination of the
digestive tract and oropharynx in ICU patients. N Engl J
Med 2009;360:20-31.
13. Johnson L, Grueber S, Schlotzhauer C, Phillips E, Bullock
P, Basnett J, Hahn-Cover K. A multifactorial action plan
improves hand hygiene adherence and significantly reduces
central line-associated bloodstream infections. Am J Infect
Control 2014;42:1146-51.
14. Gonzales M, Rocher I, Fortin E, Fontela P, Kaouache M,
Tremblay C et al. A survey of preventive measures used and
OPDAM CHANGING EPIDEMIOLOGY OF PATHOGENS IN ICU BACTEREMIA

their impact on central line-associated bloodstream infec-
tions (CLABSI) in intensive care units (SPIN-BACC).
BMC Infect Dis 2013;13:562.
15. Hindiyeh MY, Smollan G, Gefen-Halevi S, Mendelson
E, Keller N. Molecular detection of antibiotic resistance
genes from positive blood cultures. Methods Mol Biol
2015;1237:97-108.
16. Van Zanten AR. Real-time polymerase chain reaction to
evaluate antibiotic appropriateness: should we spread the
news to multiply it? Crit Care Med 2012;40:2492-3.
17. Vasudevan A, Mukhopadhyay A, Li J, Yuen E, Tambyah
P. A
��������������������������������������������������
prediction tool for nosocomial multi-drug resis-
tant gram-negative bacilli infections in critically ill pa-
tients - prospective observational study. BMC Infect Dis
2014;14:615.
18. Neulier C, Birgand G, Ruppé É, Armand-Lefèvre L, Lolom
I, Yazdanpanah Y et al. Enterobacteriaceae bacteremia: risk
factors for ESBLPE. Med Mal Infect 2014;44:32-8.
19. Brooks BD, Brooks AE. Therapeutic strategies to combat
antibiotic resistance. Adv Drug Deliv Rev 2014;S0169-
409X(14)00235-X.
20. Van Duijn PJ, Bonten MJ. ���������������������������
Antibiotic rotation strate-
gies to reduce antimicrobial resistance in Gram-negative
bacteria in European intensive careunits: study protocol
for a cluster-randomized crossover controlled trial. Trials
2014;15:277.

Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on December 7, 2014. - Accepted for publication on January 22, 2015. - Epub ahead of print on January 30, 2015.
Corresponding author: A. R. H. van Zanten, Medical Manager Care Division, Department of Intensive Care, Gelderse Vallei Hospital,
Willy Brandtlaan 10, 6716 RP Ede, The Netherlands. E-mail: zantena@zgv.nl

942 MINERVA ANESTESIOLOGICA September 2015


EDITORIAL
Localization of central venous catheter
by vascular ultrasound and transthoracic
echocardiography: easy and accurate?
D. BOLLIGER 1, M. D. SEEBERGER 2, 3
1Department of Anesthesia, Surgical Intensive Care, Prehospital Emergency Medicine and Pain Therapy, University
of Basel Hospital, Basel, Switzerland; 2Medical Faculty of the University of Basel, Basel, Switzerland; 3Institute of
Anesthesia, Klinik Hirslanden, Zurich, Switzerland
I nsertion of a central venous catheter (CVC)
is one of the most common procedures per-
formed by physicians for many indications
including hemodynamic monitoring, fluid re-
placement, hemodialysis, total parenteral nutri-
tion, and administration of drugs such as vaso-
pressors, chemotherapy, and antibiotics.1 It is
estimated that more than 5 million CVC are in-
serted per year in the USA in intensive care units
alone.2 However, CVC insertion is associated
with potential complications, the incidences of
which depend on the puncture site, e.g., punc-
ture of the subclavian vs. the internal jugular
vein.3 The most frequent problem is malposition
of the CVC occurring in 4-14% of patients.3, 4
More severe complications such as arterial punc-
ture, pneumothorax, and hematothorax are less
common.5-7 However, CVC malposition must
be avoided because it is associated with rare but
potentially lethal complications including peri-
cardial tamponade, cardiac and vena cava per-
foration.1
In this issue of Minerva Anestesiologica, Meg-
giolaro et al.8 report on the diagnostic accuracy
and reproducibility of an ultrasound method to
detect CVC malposition in the pre-operative
setting. Using transthoracic echocardiography in
105 surgical patients, the authors aimed at di-
rectly visualizing the catheter tip as well as tracing
Comment on p. 989.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 943
opacization of the right atrium after injection of
agitated saline (“bubble test”). In addition, the
authors identified the “push-to-bubbles” time as
a fast and accurate tool for verifying the correct
CVC position. The authors should be congratu-
lated for designing and performing this elegant
study. However, the question remains whether
the proposed method is sufficiently simple, tech-
nically mature, and accurate to be recommended
for routine use.
Several points need to be considered when ap-
praising the proposed method.8 First, the “gold
standard” method for testing correct CVC po-
sition, i.e. conventional chest radiography, also
has specific limitations. While it can accurately
locate the catheter tip in the subclavian, axillary,
or jugular vein, it is less precise in differentiat-
ing catheter positioning in the vena cava vs. the
right atrium.1, 9 This limitation especially applies
to surgical patients in whom chest radiography
is frequently performed in bed and only in the
antero-posterior view. Various radiographic
landmarks have been described to ensure ex-
trapericardial location of the CVC tip.10-12 The
most accepted landmark is the radiologically
well-depicted carina,13, 14 because the pericardial
sac ends below or on this level. However, a tip
position below the carina does not necessarily re-
sult in an intra-atrial position. Potentially, only
computed tomography or MRI can exclude an
intra-atrial position.
BOLLIGER LOCALIZATION OF CENTRAL VENOUS CATHETER BY ULTRASOUND
Second, while ultrasound is advantageous for
venous puncture and CVC insertion,5-7 it is of
limited value for reliably identifying catheter tip
position. In particular, ultrasound is less suit-
able for imaging the superior vena cava, and
customarily used probes may fail to visualize
the catheter tip in this vessel.1 Most recently, a
study by Kim et al. suggested that this limita-
tion might be eliminated using a microconvex
probe for locating the catheter tip.15 Further,
the “CVC sono” has been described as bedside
ultrasound technique to determine CVC tip
position and to exclude mechanical complica-
tion of CVC insertion using different thoracic,
vascular, and cardiac ultrasound views.16 “CVC
sono” has been suggested to eliminate the need
for chest radiography in about 70% of patients
with CVC insertion on a surgical intensive care
unit.17 Similar to the study by Meggiolaro et al.,8
only one anesthesiologist/intensivist specialized
in ultrasound examination performed CVC tip
location. In another study, three intensivists
performed “CVC sono” after a 2-hour specific
training course and three supervised examina-
tions. In this study, the investigators identified
9 of 10 malpositions and 1 of 1 pneumothorax
by “CVC sono”.18 Similar results were reported
in patients with CVC insertion in the emer-
gency department.9 In that study, ultrasound
examinations were performed by emergency
department physicians after a 24-hour specific
training followed by 25 supervised ultrasound
examinations.9 Despite the encouraging reports
on “CVC sono”, the number of reports and
examiners in those studies is limited. Intense
training is required to accurately and quickly
perform ultrasound examinations. This require-
ment questions whether “CVC sono” should be
taught to every anesthesiologist.
Third, the intra-atrial position is most likely
after full insertion of the 20-cm CVC in the
right-sided subclavian and jugular vein,9, 19 as
was done in the study by Meggiolaro et al.8 Use
of a shorter catheter and/or ECG-controlled
CVC placement might relevantly reduce repo-
sitioning procedures and, therefore, increase pa-
tient safety.20 ECG-guided CVC placement is
easy to perform and seems to save resources.20 In
patients with sinus rhythm, this technique seems
944 MINERVA ANESTESIOLOGICA September 2015
to be more recommendable than “CVC sono” or
the method proposed by Meggiolaro et al.8
Finally, the authors measured the “push-to-
bubble” time and found significant differences
in patients with malpositions compared with pa-
tients with correct CVC positions. “Push-to-bub-
ble” time was also slightly but significantly shorter
when the tip of the intra-atrial CVC was in the
atrium compared to when it was correctly posi-
tioned in the superior vena cava. The “push-to-
bubble” time had a very good specificity and sen-
sitivity for differentiating correct position of the
CVC tip from malposition. However, it remains
questionable whether this short difference in time
is large enough to clearly differentiate between po-
sitions in the superior vena cava vs. in the atrium.
A shortcoming is that the authors had to use a
self-constructed electric switch and an ECG-lead
technique for defining the start of the time inter-
val. Both techniques produced electric artifacts
during ultrasound acquisition. Although these
self-made techniques worked for the authors, we
think that the method can only be recommended
after a commercial production of such a device.
To conclude, the “push-to-bubble” method
had a very good sensitivity and specificity for
identification of malposition of the CVC tip
(including intra-atrial position), whereas the
ultrasound-guided localization of the CVC tip
showed very good sensitivity but insufficient
specificity. In the hands of ultrasound experts,
both techniques might be of clinical value, espe-
cially when the CVC is inserted via an approach
other than the right internal jugular vein. Never-
theless, the proposed method must be tested by
larger groups of physicians, and a commercially
available device for measuring the time for the
“push-to-bubble” method should be available
before this technique can be generally recom-
mended for perioperative use.
References
  1. Tan PL, Gibson M. Central venous catheters: the role of
radiology. Clin Radiol 2006;61:13-22.
 2. Taylor RW, Palagiri AV. Central venous catheterization.
Crit Care Med 2007;35:1390-6.
  3. Ruesch S, Walder B, Tramer MR. Complications of central
venous catheters: internal jugular versus subclavian access-
-a systematic review. Crit Care Med 2002;30:454-60.
LOCALIZATION OF CENTRAL VENOUS CATHETER BY ULTRASOUND BOLLIGER

  4. Pikwer A, Baath L, Davidson B, Perstoft I, Akeson J. The
incidence and risk of central venous catheter malposition-
ing: a prospective cohort study in 1619 patients. Anaesth
Intensive Care 2008;36:30-7.
  5. Hind D, Calvert N, McWilliams R, Davidson A, Paisley S,
Beverley C et al. Ultrasonic locating devices for central ve-
nous cannulation: meta-analysis. Br Med J 2003;327:361.
  6. Keenan SP. Use of ultrasound to place central lines. J Crit
Care 2002;17:126-37.
  7. Randolph AG, Cook DJ, Gonzales CA, Pribble CG. Ul-
trasound guidance for placement of central venous cath-
eters: a meta-analysis of the literature. Crit Care Med
1996;24:2053-8.
 8. Meggiolaro M, Scatto A, Zorzi A, Roman-Pognuz E,
Lauro A, Passarella C et al. Confirmation of correct cen-
tral venous catheter position in the pre-operative setting by
echocardiographic “bubble test”. Minerva Anestesiologica
2015;81:989-99.
 9. Zanobetti M, Coppa A, Bulletti F, Piazza S, Nazerian P,
Conti A et al. Verification of correct central venous cath-
eter placement in the emergency department: comparison
between ultrasonography and chest radiography. Intern
Emerg Med 2013;8:173-80.
10. Collier PE, Goodman GB. Cardiac tamponade caused by
central venous catheter perforation of the heart: a prevent-
able complication. J Am Coll Surg 1995;181:459-63.
11. Greenall MJ, Blewitt RW, McMahon MJ. Cardiac tampon-
ade and central venous catheters. Br Med J 1975;2:595-
7.
12. Rutherford JS, Merry AF, Occleshaw CJ. Depth of central
venous catheterization: an audit of practice in a cardiac sur-
gical unit. Anaesth Intensive Care 1994;22:267-71.
13. Caruso LJ, Gravenstein N, Layon AJ, Peters K, Gabrielli A.
A better landmark for positioning a central venous catheter.
J Clin Monit Comput 2002;17:331-4.
14. Schuster M, Nave H, Piepenbrock S, Pabst R, Panning B.
The carina as a landmark in central venous catheter place-
ment. Br J Anaesth 2000;85:192-4.
15. Kim SC, Heinze I, Schmiedel A, Baumgarten G, Knuefer-
mann P, Hoeft A et al. Ultrasound confirmation of central
venous catheter position via a right supraclavicular fossa
view using a microconvex probe: An observational pilot
study. Eur J Anaesthesiol 2015;32:29-36.
16. Matsushima K, Frankel HL. Detection of central venous
catheter insertion-related complication using bedside ultra-
sound: the CVC sono. J Trauma 2011;70:1561-3.
17. Matsushima K, Frankel HL. Bedside ultrasound can safely
eliminate the need for chest radiographs after central ve-
nous catheter placement: CVC sono in the surgical ICU
(SICU). J Surg Res 2010;163:155-61.
18. Maury E, Guglielminotti J, Alzieu M, Guidet B, Offenstadt
G. Ultrasonic examination: an alternative to chest radiog-
raphy after central venous catheter insertion? Am J Respir
Crit Care Med 2001;164:403-5.
19. McGee WT, Moriarty KP. Accurate placement of central
venous catheters using a 16-cm catheter. J Intensive Care
Med 1996;11:19-22.
20. Gebhard RE, Szmuk P, Pivalizza EG, Melnikov V, Vogt C,
Warters RD. The accuracy of electrocardiogram-controlled
central line placement. Anesth Analg 2007;104:65-70.

Acknowledgements.—The authors thank Allison Dwileski, BS Department of Anesthesia, Surgical Intensive Care, Prehospital Emergency
Medicine, and Pain Therapy, University Hospital Basel, Switzerland, for editorial assistance.
Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on December 6, 2014. - Accepted for publication on December 11, 2014. - Epub ahead of print on December 12, 2014.
Corresponding author: D. Bolliger, Department of Anesthesia, Surgical Intensive Care, Prehospital Emergency Medicine, and Pain Ther-
apy, University Hospital Basel, Spitalstrasse 21, CH-4031 Basel, Switzerland. E-mail: daniel.bolliger@usb.ch

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 945



O R I G I N A L A RT I C L E

Randomized controlled pilot trial of the rigid and

flexing laryngoscope versus the fiberoptic bronchoscope
for intubation of potentially difficult airway
B. D. ALVIS 1, A. B. KING 1, D. HESTER 2, C. G. HUGHES 1, M. S. HIGGINS 2

1Department Of Anesthesiology, Vanderbilt University School Of Medicine, Nashville, TN, USA; 2Division Of Multi-
Specialty Anesthesiology, Department Of Anesthesiology, Vanderbilt University School Of Medicine, Nashville, TN,
USA

ABSTRACT
Background. The flexible fiberoptic bronchoscope (FOB) is viewed as the gold standard device for awake intuba-
tion in the difficult airway. The newer rigid flexible laryngoscope (RIFL) was developed for similar indications. In
this study we compare these two devices for management of potentially difficult airways after induction of general
anesthesia.
Methods. Adult surgical patients requiring endotracheal intubation and having a predicted difficult airway based on
airway examination, BMI≥35, and/or history of prior difficult intubation were randomized to undergo endotracheal
intubation with either the RIFL or FOB. Induction was performed in usual manner, and intubation was performed
by providers proficient with both airway devices after induction of general anesthesia. The primary outcomes meas-
ured were intubation success, time to intubation, number of attempts, and the need for airway assist maneuvers. The
lowest observed oxygen saturation and airway trauma were also recorded.
Results. A total of 41 patients were enrolled, with 20 randomized to each group and 1 withdrawal. Intubation was
successful in all patients with both devices. The median time for successful intubation was significantly shorter in
the RIFL group compared to the FOB group (49 vs. 64 seconds; P=0.048). Airway assist maneuvers were required
in 2 (10%) intubations with the RIFL compared to 16 (80%) intubations with the FOB (P<0.001). There were no
significant differences in lowest oxygen saturation or airway trauma.
Conclusion. The RIFL required significantly less time and fewer airway assist maneuvers for successful endotracheal
intubation compared to FOB when used by experienced providers in patients with anticipated difficult airways.
(Minerva Anestesiol 2015;81:946-50)
Key words: Intubation - Laryngoscopes - Airway management.

F ailure to successfully manage the airway is

among the principal causes of anesthetic-
related morbidity and mortality.1 Up to 30% of
all deaths during general anesthesia are related
to difficult airway management, and 85% of all
respiratory-related malpractice claims involve
brain damage or death.2 Thus, many advanced
airway devices have been developed to assist the
clinician in the management of difficult endotra-
cheal intubation.
The flexible fiberoptic bronchoscope (FOB)
is generally regarded as the gold standard ad-
vanced airway tool for management of difficult
intubation, especially for awake intubation.3-6
However, the FOB requires significant train-
ing to obtain proficiency3 and often requires
airway maneuvers from a second provider dur-
ing airway management to successfully perform
endotracheal intubation. The rigid and flexing
laryngoscope (RIFL, AI Medical, MI) is an air-

946 MINERVA ANESTESIOLOGICA September 2015

RIFL VERSUS FOB FOR INTUBATION OF DIFFICULT AIRWAYs ALVIS
way management tool with a rigid stylet-based
laryngoscope, a portable inline video display,
and a lever that controls flexion of the distal tip
up to 135 degrees.1 The RIFL combines the de-
sirable features of a FOB with the simplicity and
portability of the video laryngoscope.1 Also, the
addition of the stylet allows the provider to ma-
nipulate the device with a single hand, freeing
the second hand to perform airway manipula-
tion to assist in the intubation procedure and
potentially reducing the time to intubation and
the need for assistance from another provider
during the procedure.
To date, no study has prospectively compared
the use of the RIFL versus the FOB for airway
management in patients with expected difficult
airways undergoing general anesthesia. The main
objective of this study, therefore, is to evaluate
the RIFL versus the FOB in anticipated difficult
intubations with respect to successful intubation
rate, time to successful intubation, number of at-
tempts requiring assist maneuvers (ARAM) from
a second provider, lowest oxygen saturation, and
number of traumatic airway events.
Materials and methods
Study design and population
This study was approved by the institutional
review board of Vanderbilt University (Nashville,
TN) and written informed consent was obtained
from all study participants. We enrolled adult
patients with anticipated difficult airways un-
dergoing general anesthesia requiring endotra-
cheal intubation and neuromuscular blockade at
Vanderbilt University Medical Center from June
2013 to July 2013. Anticipated difficult airway
was defined as patients who met one or more of
the following criteria: past history of documented
difficult intubation, Mallampati Score ≥3, and a
body mass index (BMI)≥35.7 Exclusion criteria
included the inability to obtain consent and need
for a rapid sequence or awake fiberoptic intuba-
tion. Patients were randomized to intubation with
the RIFL or the FOB via computer generated al-
location in blocks of 10 and placed in sealed en-
velopes by an author (M. S. H.) who was blinded
to patient recruitment and data collection.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 947
Airway management
Selection and dosing of induction medica-
tions were at the discretion of the attending
anesthesiologist. The attending anesthesiolo-
gists were instructed to provide assist maneu-
vers during airway management if necessary
for successful intubation. After induction of
anesthesia, confirmation of muscle relaxation
(absence of twitches on a train-of-four moni-
tor), and adequate preoxygenation (end tidal
oxygen concentration greater than 80%), en-
dotracheal intubation was performed with the
randomized device by one of the investigators
(B.D.A., A.B.K.) who were experienced users
of both devices. The RIFL was used with the
standard RIFL blade1 in patients randomized
to airway management with the RIFL. The Ber-
man oral airway (Mainline Medical Airway,
Norcross, GA, USA) and a FOB (Karl STORZ
Endoscope, Culver City, CA, USA) were used
in patients randomized to undergo endotracheal
intubation with the FOB. A Mallinckrodt Hi-
Lo Oral/Nasal Tracheal Tube Cuffed 7.5 – 8.0
mm I.D. (Covidien, Minneapolis, MN, USA)
was used for each intubation as the discretion
of the attending anesthesiologist. An independ-
ent observer recorded data during laryngoscopy
attempts.
Statistical analysis
The primary outcomes measured were the in-
tubation success, time to successful intubation,
and number of ARAM provided by the attending
anesthesiologist. For study purposes, a successful
intubation was defined as the ability to view the
glottis and successfully pass an endotracheal tube
through the glottis within 100 seconds, based
on a consensus from literature review. Time to
successful intubation was defined as the period
from when the tip of the RIFL or FOB passed
the incisors until withdrawal past that same
point after successful intubation. ARAM was
defined as lifting of the jaw at the mandibular
angle, guided laryngeal pressure, and/or manual
tongue protrusion. Lowest oxygen saturation on
pulse oximetry during airway management and
oropharyngeal injuries (e.g., dental injury, lip
ALVIS RIFL VERSUS FOB FOR INTUBATION OF DIFFICULT AIRWAYs
laceration, pharyngeal, laryngeal, or tracheal in-
jury) were also recorded.
Study data were collected and managed using
the REDCap (Research Electronic Data Cap-
ture) database hosted at Vanderbilt University.8
In order to detect a clinically significant differ-
ence of 30 seconds for successful intubation with
80% power and two-sided alpha level of 0.05 in
this randomized control trial, 20 patients were
required in each group. The time of 30 seconds
was decided upon based on consensus of our re-
search groups’ review of the literature. Descrip-
tive statistics were calculated on patient demo-
graphics and physical exam variables. Time to
intubation analysis utilized a Cox proportional
hazards regression. Number of ARAM was as-
sessed using a Wilcoxon rank sum test. Binary
outcomes: successful first attempt, airway trau-
ma, and whether there was an assistance ma-
neuver were evaluated with Fisher exact test or
logistic regression. Statistical analyses were per-
formed using SOFA 3.0.
Results
Forty-one patients were enrolled in this study
(21 in the RIFL group and 20 in the FOB group)
with median age of 49 years and median BMI of
37 kg/m2. One patient in the in the RIFL group
withdrew prior to induction of general anesthe-
sia. The baseline characteristics of the patients in
Table I.—Patient characteristics.
RIFL
Baseline Characteristics (N.=20)
Age (years; median) 50
Gender
Males 10 (50%)
Females 10 (50%)
BMI (kg m2; median) 38.35
# with BM ≥35 (kg·m-2) 14 (70%)
Mallampati Class (N)
1 3 (15%)
2 2 (10%)
3 6 (30%)
4 9 (45%)
History of Difficult Airway
Yes 9 (45%)
No 11(55%)
RIFL: rigid and flexing laryngoscope; FOB: flexible fiberoptic bronchoscope; BMI: body mass index.
948 MINERVA ANESTESIOLOGICA September 2015
both groups were similar and are summarized in
Table I.
Intubation was successful in all patients in
both groups (20/20 vs. 20/20 attempts, Table II).
The median time for successful intubation was
significantly shorter in the RIFL group vs. the
FOB group (49 vs. 64 seconds; P=0.048, Table
II). The number of subjects that required airway
assist maneuvers in the RIFL group was 2 (10%)
vs. 16 (80%) in the FOB group (P<0.001, Table
II, Figure 1). Assist maneuvers included: lifting
of the jaw at the mandibular angle (0 patients in
RIFL group vs.16 patients in the FOB group),
laryngeal pressure (2 patients in the RIFL group
vs. 4 patients in the FOB group), and manual
tongue protrusion (0 patients in RIFL group vs.
1 patient in the FOB group). There were no sig-
nificant differences in the lowest oxygen satura-
tion or airway trauma between groups.
Discussion
This study demonstrated that when used by
experienced providers, the RIFL performs favo-
rably compared to the FOB for post-induction
endotracheal intubation in patients with poten-
tially difficult airways. While there was no statis-
tical difference in rate of successful intubation,
the RIFL was significantly faster and required
fewer airway assist maneuvers compared to the
FOB.
FOB P-value
(N.=20)
49
0.34
7 (35%)
13 (65%)
35.95 0.27
13 (65%)
0.74
1 (5%)
3 (15%)
7 (35%)
9 (45%)
6 (30%) 0.33
14(70%)
RIFL VERSUS FOB FOR INTUBATION OF DIFFICULT AIRWAYs ALVIS
Table II.—Comparison of RIFL vs. FOB.
RIFL
(N.=20)
Successful Intubation 20 (100%)
Median time to Intubation(s) (IQR) 49 seconds (49)
Assistance Required
Yes 2 (10%)
No 18 (90%)
Lowest pulse oximetry saturation 100%
Number of documented airway trauma events 2 (10%)
RIFL: rigid and flexing laryngoscope; FOB: flexible fiberoptic bronchoscope.
Comparison of attempts that required assistance and time to intubation. Values are the number with percentage of total, unless otherwise stated.
Airway trauma events included any damage/laceration of lips, gums, teeth or mucosa secondary to the intubation attempt.
Figure 1.—Airway assist maneuvers required with the RIFL vs.
the FOB. RIFL: rigid and flexing laryngoscope; FOB: flexible
fiberoptic bronchoscope.
The RIFL group had a median time to suc-
cessful intubation that was 15 seconds shorter
than the time required for intubation with the
FOB. This may potentially have clinical signifi-
cance for patients at high risk for desaturation
or aspiration. However, no statistically signifi-
cant difference in oxygen saturation occurred
between the two groups in our small pilot study.
In addition, the RIFL group required fewer air-
way assist maneuvers for successful endotracheal
intubation whereas the FOB group required
these maneuvers on the majority of airway man-
agement attempts. This assistance of an “expe-
rienced second pair of hands” may not always
be present in some healthcare institutions or
scenarios, especially in out-of-operating room
locations or emergent airways. Both these de-
vices require experience to maneuver efficiently
and effectively. Having two clinicians with the
needed experience present in these scenarios may
be challenging. In these situations, therefore, it
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 949
FOB P-value
(N.=20)
20 (100%) 1
64 seconds (80) 0.048
16 (80%) <0.001
4 (20%)
100% 0.79
3 (15%) 0.63
may be easier for a solo clinician to use the RIFL
to intubate an anticipated difficult airway rather
than the FOB.
There are several limitations of this study.
While we randomized patients to the RIFL vs.
the FOB, it was impossible to blind either the
investigators performing airway management
or the independent observer. Also, although we
specifically included patients with either prior
difficult intubation or anticipated difficult in-
tubation based upon published criteria from
prior investigations,9-15 obesity and oropharyn-
geal class are not perfect predictors of difficult
intubation and, therefore, we may have recruited
some patients that would not be difficult to intu-
bate into our study. Finally, predictors of difficult
direct laryngoscopy may not reliably predict dif-
ficult intubation with flexible fiberoptic devices
which were used in this investigation, although
there is some evidence to suggest that the criteria
we chose are commonly used by clinicians as cri-
teria for choosing fiberoptic intubation.16
Conclusions
While both the RIFL and FOB demonstrated
high rates of successful endotracheal intubation,
the RIFL required shorter time to intubation
and fewer airway assist maneuvers compared to
the FOB when intubating adult patients with
predicted difficult airways. Future investigations
should explore device comparisons in emergency
airway management scenarios, intubations per-
formed by less experienced laryngoscopists, in
novice clinicians, and non-operating room en-
vironments.
ALVIS RIFL VERSUS FOB FOR INTUBATION OF DIFFICULT AIRWAYs

on Management of the Difficult Airway. Anesthesiology

Key messages
—— The RIFL may be an alternative to the
FOB in patients with potential difficult air-
ways not requiring awake intubation.
—— Less assist maneuvers are likely re-
quired for successful intubation with the
RIFL versus the FOB.
—— Oxygen saturation and incidence of
airway trauma are similar in the RIFL and
FOB.
References
1. Setty H, Rawlings JL, Dubin S. Video RIFL: a rigid flex-
ible laryngoscope to facilitate airway management. J Clin
Anesth 2010;22:642-7.
2. Mihai R, Blair E, Kay H, Cook TM. A quantitative review
and meta-analysis of performance of non-standard laryn-
goscopes and rigid fibreoptic intubation aids. Anaesthesia
2008;63:745-60.
3. Kim SH, Woo SJ, Kim JH. A comparison of Bonfils intuba-
tion fiberscopy and fiberoptic bronchoscopy in difficult air-
ways assisted with direct laryngoscopy. Korean J Anesthesiol
2010;58:249-55.
4. Rosenblatt WH, Wagner PJ, Ovassapian A, Kain ZN. Prac-
tice patterns in managing the difficult airway by anesthesi-
ologists in the United States. Anesth Analg 1998;87:153-7.
5. Apfelbaum JL, Hagberg CA, Caplan RA, Blitt CD, Con-
nis RT, Nickinovich DG et al. Practice guidelines for
management of the difficult airway: an updated report
by the American Society of Anesthesiologists Task Force
2013;118:251-70.
6. Petrini F, Accorsi A, Adrario E, Agrò F, Amicucci G, An-
tonelli M et al. Recommendations for airway control
and difficult airway management. Minerva anestesiol
2005;71:617-57.
7. Cattano D, Panicucci E, Paolicchi A, Forfori F, Giunta
F, Hagberg C. Risk factors assessment of the difficult air-
way: an italian survey of 1956 patients. Anesth Analg
2004;99:1774-9, table of contents.
8. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N,
Conde JG. Research electronic data capture (REDCap)--a
metadata-driven methodology and workflow process for
providing translational research informatics support. J Bi-
omed Inform 2009;42:377-81.
9. Benumof JL. Obstructive sleep apnea in the adult obese pa-
tient: Implications for airway management. J Clin Anesth
2001;13:144-56.
10. Brodsky JB, Lemmens HJ, Brock-Utne JG, Vierra M, Said-
man LJ. Morbid obesity and tracheal intubation. Anesth
Analg 2002;94:732-6; table of contents.
11. Wilson ME, Spiegelhalter D, Robertson JA, Lesser P. Pre-
dicting Difficult Intubation. Brit J Anaesth 1988;61:211-6.
12. Rose DK, Cohen MM. The Airway - Problems and Predic-
tions in 18,500 Patients. Can J Anaesth 1994;41:372-83.
13. Dargin JM, Emlet LL, Guyette FX. The effect of body mass
index on intubation success rates and complications during
emergency airway management. Internal and emergency
medicine 2013;8:75-82.
14. Adams JP, Murphy PG. Obesity in anaesthesia and inten-
sive care. Brit J Anaesth 2000;85:91-108.
15. Hagberg CA, Vogt-Harenkamp C, Kamal J. A retrospective
analysis of airway management in obese patients at a teach-
ing institution. J Clin Anesth 2009;21:348-51.
16. Lundstrom LH, Moller AM, Rosenstock C, Astrup G, Wet-
terslev J. High Body Mass Index Is a Weak Predictor for
Difficult and Failed Tracheal Intubation A Cohort Study
of 91,332 Consecutive Patients Scheduled for Direct La-
ryngoscopy Registered in the Danish Anesthesia Database.
Anesthesiology 2009;110:266-74.

Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on March 20, 2014. - Accepted for publication on October 1, 2014. - Epub ahead of print on October 3, 2014.
Corresponding author: B. D. Alvis, Department Of Anesthesiology, Vanderbilt University School Of Medicine, 1301 Medical Center
Drive, TVC 4648, Nashville, TN, USA. E-mail: bret.d.alvis@vanderbilt.edu

950 MINERVA ANESTESIOLOGICA September 2015



O R I G I N A L A RT I C L E

Hyperglycemia and ambulatory surgery

J. A. POLDERMAN 1, L. VAN VELZEN 1, L. G. WASMOETH 1
J. H. ESHUIS 1, P. L. HOUWELING 2, M. W. HOLLMANN 3, J. H. DEVRIES 4
B. PRECKEL 3, J. HERMANIDES 1

1Department of Anesthesiology, Academic Medical Centre, Amsterdam, The Netherlands; 2Department of Anesthesiology,

Diakonessenhuis, Utrecht, The Netherlands; 3Department of Anesthesiology and Laboratory of Experimental Intensive
Care and Anesthesiology L.E.I.C.A., Academic Medical Centre, Amsterdam, The Netherlands; 4Department of Internal
Medicine, Academic Medical Centre, Amsterdam, The Netherlands

ABSTRACT
Background. Perioperative hyperglycemia is associated with postoperative complications after major surgery. How-
ever, more than 50% of surgical procedures are performed in an ambulatory setting, where glucose is not routinely
measured. The objectives of this study were to investigate the change in capillary glucose during ambulatory surgery,
to identify patients at risk for perioperative increasing glucose and to evaluate whether hyperglycemia predisposes
for complications after ambulatory surgery.
Methods. In this prospective multicenter cohort study, adult patients planned for ambulatory surgery, were included
and capillary glucose was measured 1 hour before and 1 hour after surgery. Patients were contacted 90 days after
surgery to determine the occurrence of postoperative complications.
Results. Nine hundred and nine patients were included, 48 (5.3%) patients had diabetes mellitus (DM). Overall
median glucose increased from 5.4 mmol L-1 preoperatively to 5.6 mmol L-1 postoperatively (P<0.001). Hyperg-
lycemia, glucose ≥7.8 mmol L-1, occurred in 8.8% of the patients. Dexamethasone administration (given in 406
[44.7%] patients) was a risk factor for glucose increase (P<0.001). Hyperglycemia was not a risk factor for post-
operative complications (OR 1.19, 95%CI 0.57-2.48, P=0.646). However, prediagnosed DM was a risk factor for
postoperative complications, independent of hyperglycemia (OR 2.56, 95%CI 1.10-5.97, P=0.030).
Conclusion. Minor ambulatory surgery is not associated with a clinically relevant increase in glucose. The very
small glucose increase we observed could be attributed to the administration of dexamethasone for PONV
prophylaxis. Hyperglycemia during ambulatory surgery is not associated with complications after discharge.
(Minerva Anestesiol 2015;81:951-9)
Key words: Ambulatory surgical procedures, complications - Postoperative period – Dexamethasone - Diabetes
mellitus - Hyperglycaemia.

I n-hospital hyperglycemia occurs frequently,

both in patients with and without diagnosed
diabetes mellitus (DM). Often, hyperglycemia
during hospital admission resolves spontane-
ously after hospital discharge and is referred to
as “stress hyperglycemia”.1 However, 15-46% of
these patients have undiagnosed DM.1, 2
In several types of major surgery, such as pan-
creatoduodenectomy, liver transplantation and
cardiac surgery, perioperative hyperglycemia is
associated with postoperative complications.3-5
However, the majority of surgical patients un-
dergo minor surgery and are operated on in an
ambulatory setting.6, 7 Glucose is not routinely
measured in patients without DM undergoing
ambulatory surgery,8 thus hyperglycemia due to
stress or undiagnosed DM is easily missed.
Although patients undergoing ambulatory
surgery are the largest surgical population at risk
for stress hyperglycemia, information on the

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 951

POLDERMAN HYPERGLYCEMIA AND AMBULATORY SURGERY
incidence and clinical consequences of hyperg-
lycemia during ambulatory surgery is scarce. If
we were to know the incidence of hyperglycemia
during ambulatory surgery and whether this is
associated with postoperative complications, the
efficacy and cost effectiveness of glucose control
during ambulatory surgery can be assessed.
Therefore, the objectives of this study were to
investigate capillary glucose change during am-
bulatory surgery in patients with and without
DM, to identify patients at risk for perioperative
increasing glucose and to evaluate whether hy-
perglycemia predisposes for complications after
ambulatory surgery.
Materials and methods
The study protocol NL37311.018.11 was ap-
proved by the Medical Ethical Committee of the
Academic Medical Centre and the local ethical
committee of the Diakonessenhuis. We included
patients between October 1st 2011 and July 31st
2012. Written informed consent was obtained
from every patient. This prospective multicenter
observational cohort study was conducted in the
ambulatory surgery departments of two hospitals
(Academic Medical Centre (AMC), Amsterdam
and Diakonessenhuis Zeist and Utrecht, The
Netherlands). The AMC is an academic center
with a separate ambulatory surgery department,
where predominantly orthopedic and minor
general surgery procedures are performed. The
Diakonessenhuis is a non-academic center where
ambulatory surgery is predominantly focused on
orthopedic and laparoscopic inguinal hernia re-
pair procedures. Reporting was done according
to the STROBE-guideline.
Patients and study outline
Patients older than 18 years of age scheduled
for ambulatory surgery with or without a history
of DM were included. Both patients with DM
type 1 and type 2 were eligible. After written in-
formed consent, a capillary glucose sample was
taken one hour before and one hour after surgery
via finger stick. Capillary glucose was measured
using the Accu-Chek Inform (Roche Diagnostics,
Indianapolis, IN, USA). The Accu-Chek Inform
952 MINERVA ANESTESIOLOGICA September 2015
is frequently used in studies as well as in clinical
settings, with only 2.5% of inaccurate readings
according to the International Organization for
Standardization (ISO) criteria when used in a
non-ICU setting.9 The time since last carbohy-
drate intake was recorded. Patients without DM
and preoperative glucose ≥7.8 mmol L-1 were ad-
vised to see their general practitioner postopera-
tively. Patients with DM had to withhold insulin
and glucose lowering tablets on the morning of
surgery. In patients with DM, glucose >10 mmol
L-1 was treated at the discretion of the attending
anesthesiologist (e.g. a bolus of intravenous short
acting insulin), according to the local guidelines
of each hospital. Patients without DM were not
treated for hyperglycemia.
Before surgery a brief questionnaire was com-
pleted, focused on patient characteristics and
history of DM. Patients received dexamethasone
for prophylaxis of nausea and vomiting at the
discretion of the attending anesthesiologist. No
additional sources of glucose were administered
during the perioperative period, and saline was
used to dilute drugs. All details about the anes-
thetic regime and the type of surgery were noted.
We determined the occurrence of postoperative
complications for the first 90 days after surgery.
Therefore, we reviewed patients charts where
available and all patients were contacted >90
days postoperatively for a short questionnaire by
telephone. Patients were called on three different
dates and times before they were considered as
lost to follow-up.
Outcome measures
The primary outcome was median glucose
change during ambulatory surgery as compared
to preoperative fasting glucose. We identified
risk factors for perioperative increase in glucose.
Also the association of hyperglycemia during
ambulatory surgery and postoperative complica-
tions was assessed. Postoperative complications
were defined as: death, hospital readmission,
postoperative infection (e.g. wound infection,
pulmonary infection or cystitis/urinary tract
infection), wound bleeding, delirium, thrombo-
embolic complications (e.g. myocardial infarc-
tion, stroke, deep venous thrombosis and pul-
HYPERGLYCEMIA AND AMBULATORY SURGERY POLDERMAN
monary embolism) and other complications.
Postoperative infection was scored for events
ranging from redness of the skin for which medi-
cal help was sought to antibiotic treatment for
a postoperative fever with suspected infection.
Prophylactic antibiotic treatment was not scored
as a postoperative infection.
As secondary outcome measures, the propor-
tion of patients with hyperglycemia (glucose ≥7.8
mmol L-1) preoperatively, new postoperatively
and during admission were calculated. The risk
factors for pre- and new postoperative hypergly-
cemia were assessed. Additionally the risk factors
for postoperative complications were determined.
The used cut-off value for hyperglycemia was 7.8
mmol L-1 (140 mg dL-1). This is in agreement
with a recently performed trial which showed that
complications occurred more frequently above a
glucose of 7.8 mmol L-1 and with the American
Diabetes Association (ADA) cut-off value for
DM during an oral glucose tolerance test.3, 10
Statistical analysis
No data about glucose change during ambula-
tory surgery was available, thus a formal power
analysis was not possible. To ensure adequate
power of the predictive model for the risk factors
for hyperglycemia, we aimed for the inclusion of
1000 patients.
Glucose change during surgery was analyzed
using the Wilcoxon signed-rank test and glucose
change between subgroups (no DM without
dexamethasone, no DM with dexamethasone,
DM without dexamethasone and DM with
dexamethasone) was analyzed using the Mann-
Whitney U test. The overall glucose change was
calculated by subtracting the preoperative glu-
cose value from the postoperative glucose value.
A multivariate linear regression analysis was per-
formed to identify risk factors for perioperative
increase in glucose. The difference in complica-
tion rate between patients with and without hy-
perglycemia during ambulatory surgery was cal-
culated with the Chi-square test. The odds ratios
(ORs) for risk factors for preoperative hypergly-
cemia, new postoperative hyperglycemia and hy-
perglycemia at any time during admission were
calculated with multivariate logistic regression
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 953
analyses. In addition the ORs for risk factors of
postoperative complications for the whole study
population were calculated. In all regression
analyses we adjusted for age, sex, BMI, ASA clas-
sification, dexamethasone use, DM and duration
of surgery (divided in quartiles). Also, postoper-
ative glucose was analyzed by quartiles, to assess
if the higher quartiles were associated with com-
plications when compared to the lowest quartile.
Statistical analyses were done using SPSS version
20.0 (SPSS Inc., Chicago, IL, USA).
Results
Study population
One thousand five hundred twenty-nine pa-
tients were eligible for this study, of which 951
patients gave written informed consent. Four
patients were excluded due to cancellation of
surgery or conversion to local anesthesia and one
patient withdrew informed consent. One patient
was excluded from the study because he received
30 mg prednisolone per day for 5 days due to
COPD GOLD classification II. Thirty-six pa-
tients were excluded because only one glucose
measurement was available, leaving 909 patients
for final analyses.
The patient characteristics are displayed in Ta-
ble I. With regard to patients with DM, in 43
patients their glucose was measured before and
after surgery, but no additional insulin was ad-
ministered; In three patients, a glucose-insulin
infusion was started during surgery; In two pa-
tients, their own subcutaneous insulin pump
was continued during surgery.
Follow up was completed for 670 patients
(73.7%), of whom 40 had known DM (6.1%).
The remaining patients could not be contacted.
Any of the predefined complications occurred
in 15.8% of patients. The main postoperative
complication was postoperative infection (8.1%,
Table II).
Perioperative glucose values and hyperglycemia
Overall median glucose increased from 5.4
(IQR 5.1-5.9) preoperatively to 5.6 mmol L-1
(IQR 5.1-6.4) postoperatively (P<0.001). In the
POLDERMAN HYPERGLYCEMIA AND AMBULATORY SURGERY

Table I.—Patient characteristics (N.=909).

Male, N. (%) 465 (51.2)
Mean age (years) 47.2 (15.3)
Mean BMI 25.9 (4.7)
Diabetes Mellitus, N. (%) 48 (5.3)
Type 1 7 (0.7)
Type 2 treated with diet 1 (0.1)
Type 2 treated with tablets 31 (3.4)
Type 2 insulin dependent 9 (1.0)
Dexamethasone, N. (%) 406 (44.7)
Intake of carbohydrate containing liquids <2 h before surgery, N. (%) 7 (0.8)
ASA-classification, N. (%)
ASA 1 533 (58.6)
ASA 2 357 (39.3)
ASA 3 19 (2.1)
Type of anesthesia, N. (%)
General anesthesia 668 (73.5)
Spinal anesthesia 178 (19.6)
Peripheral nerve block 63 (6.9)
Median duration of surgery (min) 28 (18-46)
Treating specialty, N. (%)
Orthopedics 307 (33.8)
General surgery 273 (30)
Gynecology 82 (9)
Urology 20 (2.2)
ENT-surgery 74 (8.1)
Neurosurgery 4 (0.4)
Plastic surgery 69 (7.6)
Radiotherapy 13 (1.4)
Ophthalmology 47 (5.2)
Maxillofacial surgery 19 (2.1)
Values are mean (SD), median (IQR) or number (proportion).

Table II.—Postoperative complications.

All (N.=670) Normal range (N.=597) Hyperglycemia (N.=73)
Any complication 106 (15.8) 87 (14.6) 19 (26.0)*
Death 0 (0.0) 0 (0.0) 0 (0.0)
Re-admission 27 (4.0) 20 (3.4) 7 (9.6)*
Wound bleeding 25 (3.7) 24 (4.0) 1 (1.4)
Postoperative infection 54 (8.1) 46 (7.7) 8 (11.0)
Postoperative delirium 11 (1.6) 7 (1.2) 4 (5.5)*
Thrombo-embolic complications 5 (0.7) 5 (0.8) 0 (0.0)
Other 9 (1.4) 7 (1.2) 2 (2.8)
Values are N. (%).
*χ2 test P<0.05.

subgroup of patients with DM, glucose changed
from 8.0 (IQR 6.4-10.1) to 8.2 mmol L-1 (IQR
6.3-9.7, P=0.123). Although patients with DM
had higher glucose values compared to patients
without DM, the increase in glucose in patients
with known DM was not statistically significant.
A total of 80 (8.8%) patients had hyperglycemia
(glucose ≥7.8 mmol L-1) at any time during admis-
sion (Table III), while 49 (5.7%) had no previous
diagnosis of DM. Two patients were subsequently
diagnosed with DM; one patient with DM type
1 and one patient with DM type 2. Forty (4.4%)
patients developed hyperglycemia during surgery.
In total, 398 patients without DM received
a single dose of dexamethasone at a mean dos-
age of 4.2 mg (SD 2.0) during surgery. In these

954 MINERVA ANESTESIOLOGICA September 2015

HYPERGLYCEMIA AND AMBULATORY SURGERY POLDERMAN

Table III.—Hyperglycemia during admission in patients with and without diabetes.

All (N.=909) Non DM (N.=861)* DM (N.=48)
Preoperative hyperglycemia 40 (4.4) 13 (1.5) 27 (56.3)
Postoperative hyperglycemia 65 (7.2) 40 (4.6) 25 (52.1)
New postoperative hyperglycemia 40 (4.4) 36 (4.2) 4 (8.3)
Hyperglycemia during admission 80 (8.8) 49 (5.7) 31 (64.6)
Values are N. (%); DM: diabetes mellitus.

Table IV.—Median glucose pre- and post-operative values in patients with and without diabetes.
Non DM (N.=861) DM (N.=48)
All (N.=909)
No Dexa (N.=461) Dexa (N.=398) No Dexa (N.=40) Dexa (N.=8)
Preoperative 5.4 (5.1-6.0) 5.4 (5.0-5.9) 5.4 (5.1-5.9) 8.5 (6.7-11.0) 6.6 (5.7-9.3)
Postoperative 5.6 (5.1-6.4) 5.3 (4.8-6.0) 5.8 (5.3-6.5) 8.1 (5.9 –9.7) 8.3 (7.0-9.5)
WSR-test P<0.001 P=0.262 P<0.001 P=0.006 P=0.093
Median difference 0.2 (-0.3-0.6) -0.1 (-0.5-0.4) 0.4 (0.0-0.9) -0.3 (-1.8-0.3) 1.5 (0.3-2.6)
MWU-test n/a P<0.001 P=0.002
Values are median (IQR) stated in mmol l-1. DM: diabetes mellitus; Dexa: dexamethasone; WSR: Wilcoxon signed rank test; MWU: Mann
Whitney-U test.

patients, glucose significantly increased from
5.4 mmol L-1 (IQR 5.1-5.9) preoperatively to
5.8 mmol L-1 (IQR 5.3-6.5) postoperatively
(P<0.001). In patients without DM, who did
not receive dexamethasone, median glucose
slightly decreased from 5.4 mmol L-1 pre- to 5.3
mmol L-1 postoperatively (P=0.052, Table IV).
Factors associated with an increase in glucose
during ambulatory surgery were administration
of dexamethasone (β 0.53, 95% CI 0.40-0.65,
P<0.001) and duration of surgery (β 0.13, 95%
CI 0.07- 0.18, P<0.001), adjusted for age, sex,
BMI, ASA classification and DM.
In the logistic regression analysis, DM was a
risk factor for hyperglycemia at any time dur-
ing admission (OR 26.81, 95% CI 11.08-64.89,
P<0.001), after adjustment for age, sex, BMI,
ASA classification, dexamethasone use and dura-
tion of surgery. DM was also the only risk factor
for preoperative hyperglycemia (OR 53.29, 95%
CI 21.44-132.48, P<0.001) when adjusted for
age, sex, BMI and ASA classification.
Dexamethasone was a risk factor for new
postoperative hyperglycemia (OR 3.96, 95% CI
1.80-8.68, P=0.001), when adjusted for age, sex,
BMI, ASA classification, DM and duration of
surgery. Also, age (OR 1.06, 95% CI 1.03-1.09,
P<0.001), BMI (OR 1.11, 95% CI 1.03-1.19,
P=0.004) and duration of surgery (3rd quar-
tile >28 min, OR 9.46, 95% CI 1.15-78.18,
P=0.037 and 4th quartile >46 min, OR 29.21,
95% CI 3.88-223.99, P=0.001) were associated
with new postoperative hyperglycemia.
Glucose and postoperative complications
The incidence of postoperative complications
was 26% in the patients with hyperglycemia and
14.6% in the patients without hyperglycemia
(P=0.012, Table II). In the logistic regression
analyses, perioperative increase in glucose was not
a risk factor for postoperative complications (OR
1.07, 95% CI 0.88-1.32, P=0.498) when ad-
justed for age, sex, BMI, ASA classification, DM,
dexamethasone use and duration of surgery. Also,
neither hyperglycemia at any time during admis-
sion (OR 1.19, 95% CI 0.57-2.48, P=0.646,
Figure 1), nor new postoperative hyperglycemia
(OR 0.96, 95% CI 0.37-2.50, P=0.930) were as-
sociated with postoperative complications. Fur-
thermore, the upper three quartiles of postopera-
tive glucose (glucose >5.1; >5.6 and >6.4 mmol
L-1) were not associated with postoperative com-
plications, when compared to the lowest quartile
(glucose <5.1 mmol L-1, P=0.679, P=0.465 and
P=0.958, respectively).
DM in itself was a significant risk factor for
postoperative complications (OR 2.56, 95% CI
1.10-5.97, P=0.030, Figure 1) when adjusted for
age, sex, BMI, ASA classification, dexametha-

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 955

POLDERMAN HYPERGLYCEMIA AND AMBULATORY SURGERY

Figure 1.—Risk factors for postoperative complications. Covariates in included in the analysis were: gender, age, BMI, ASA-
classification, diabetes mellitus (DM), dexamethasone use, length of surgery and hyperglycemia at any time during admission.

sone use, hyperglycemia and duration of surgery.
Also, ASA 3 classification (OR 4.36, 95% CI
1.19-16.02, P=0.027) and duration of surgery
(4th quartile >46 min, OR 2.52, 95% CI 1.29-
4.91, P=0.007) showed a significant association
with postoperative complications. The use of
dexamethasone was not associated with postop-
erative complications (OR 0.76, 95% CI 0.49
- 1.19, P=0.236).
Discussion
In this study we found a significant, but small
increase of the median glucose from 5.4 to 5.6
mmol L-1 during ambulatory surgery. Plasma
glucose of ≥7.8 mmol L-1 at any time during ad-
mission occurred in 8.8% of the patients. Risk
factors for a perioperative increase in glucose are
a longer duration of surgery and the administra-
tion of dexamethasone during ambulatory sur-
gery. Postoperative complications were increased
in patients with hyperglycemia; however after ad-
justment for confounders neither an increase in
glucose during surgery, nor hyperglycemia dur-
ing admission was associated with postoperative
complications. DM was associated with postop-
erative complications independent of hypergly-
cemia.
The estimated prevalence of in-hospital hy-
perglycemia in medical patients is 4-12%.11 In
the present study, 8.8% of patients had hyper-
glycemia during admission. After exclusion of
patients with DM the incidence was 5.7%. This
can be attributed to either undiagnosed DM, to
a moderate/severe stress response or to the use
of dexamethasone. When considering the ADA
treatment guidelines for hyperglycemia, 7 pa-
tients (0.8%) had a glucose >10 mmol L-1. Thus,
in minor ambulatory surgery it doesn’t seem
worthwhile to take additional measures with re-
gard to blood glucose regulation. Furthermore
screening for the sole purpose of diagnosing hy-
perglycemia does not appear to be clinically rel-
evant nor cost effective in patients without DM.
The risk factors for hyperglycemia we found
in our study, are comparable to the 9 variables
proposed by the ADA to identify patients who
should be screened for DM.12 As the stress re-
sponse due to ambulatory surgery seems to be
minimal in the vast majority of the patients,
the preoperative fasting state could be used as a
screening opportunity for latent DM.13, 14
In patients without DM, a single dose of dex-
amethasone as prophylaxis for nausea and vomit-
ing gives a significant but small and clinically ir-
relevant increase in median glucose in our study.
The increase from baseline glucose is smaller as
compared to studies in non-ambulatory surgery,

956 MINERVA ANESTESIOLOGICA September 2015

HYPERGLYCEMIA AND AMBULATORY SURGERY POLDERMAN
which might be explained by the more invasive
nature and longer duration of surgery.15-17
Patients with DM had a significantly higher
fasting glucose than patients without DM.
However this was stable during surgery when
no dexamethasone was administered and well
below the target range of 10 mmol L-1 advo-
cated by the ADA.18 Nonetheless, one has to
keep in mind that in patients with DM glucose
increased by 1.5 mmol l-1 after administration
of dexamethasone. Although our sample size
is limited, a similar increase in glucose in pa-
tients with DM who received a single dose of
dexamethasone during surgery has been seen in
a previous study.19 In our study, only DM was
associated with postoperative complications
independent of dexamethasone and hypergly-
cemia, thus the clinical consequences of dexa-
methasone induced hyperglycemia in patients
with DM could be questioned. We would sug-
gest alternative postoperative nausea and vomit-
ing (PONV) prophylaxis in patients with DM.
However, if multimodal PONV prophylaxis is
indicated, we would not withhold dexametha-
sone in patients with DM, as long as glucose is
monitored postoperatively and hyperglycaemia
treated promptly. However, further research on
this topic is needed.
A postoperative infection (8.1%) was the
predominant postoperative complication in the
study population. We used a broad definition of
postoperative infection, ranging from seeking
medical help for redness of the skin to initiation
of antibiotics for postoperative fever. This may
lead to an overestimation of surgical site infec-
tions and because of the self-reporting nature of
the assessment, may be subjected to recall bias.
Van Boxel et al. reported a similar rate of 7.6%
of surgical site infections in patients with and
without DM after laparoscopic cholecystecto-
my.20 In addition, 33% of these patients sought
medical help within 30 days after discharge, al-
though a postoperative complication was noted
in just 15% of the patients,20 which is compa-
rable to the 16% postoperative complications
found in our study. The reported incidence of
unexpected admissions after day-case surgery
was between 2.7% and 6.7%.21-23 This is in ac-
cordance with our study population, of which
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 957
4.0% of the patients needed to be admitted or
re-admitted within three months of surgery.
Interestingly, hyperglycemia was not associ-
ated with postoperative complications, contrary
to studies in major surgery.3-5 Because the inci-
dence of hyperglycemia was lower than expected,
future studies are needed to confirm our find-
ings. However, to illustrate the importance of the
other different risk factors found in our study,
we have calculated the theoretical risk of postop-
erative complications after ambulatory surgery in
three different hypothetical male patients with
a mean age and BMI, without administration
of dexamethasone and with hyperglycemia, us-
ing our multivariate regression model. An ASA
1 patient without DM and duration of surgery
<18 minutes has a calculated 10.6% risk of de-
veloping a postoperative complication. An ASA
3 patient without DM with a duration of surgery
>46 minutes has a hypothetical 53.7% risk of any
complication and the ASA 3 patient with DM
and a duration of surgery >46 min has a 74.8%
risk of developing a postoperative complication.
This illustrates the possible clinical impact of
duration of surgery, ASA classification and DM
during ambulatory surgery.
This study has several limitations. First, we
had a follow-up time of 90 days. Clinical tri-
als investigating intensive insulin therapy range
in follow-time from 30 days to 6 months.24 Al-
though increasing the follow-up period might
introduce recall bias, we also believe that it de-
creases the possibility of missing late complica-
tions. When available, we used chart reviews to
minimize recall bias. Second, the loss to follow
up was 26.3%. Patients were lost to follow up
because of incorrect phone numbers or simply
because patients did not answer their phone.
However, this patient group was not significant-
ly different from patients who have been fol-
lowed with regard to their perioperative glucose
control and only including patients with com-
plete follow up in the analysis did not change
the results. Finally, the number of patients with
DM was limited (N.=7 DM1 and N.=41 DM2,
respectively). Excluding patients who received
insulin during surgery, did not change the re-
sults. Furthermore, this reflects the ambulatory
surgery population in the Netherlands. Due to
POLDERMAN HYPERGLYCEMIA AND AMBULATORY SURGERY
the low number of patients with DM, we were
cautious with the interpretation of these results.
Another potential source of bias was that seven
patients (0.8%) drank glucose containing bever-
ages 2 hours before glucose measurement, but
excluding these patients from the analyses did
not alter the results. This is the first study on
stress hyperglycemia during ambulatory sur-
gery in a large patient population. Due to the
prospective design of the study, measurement
bias, e.g. measuring glucose on indication, was
prevented. Furthermore, the characteristics of
the study population are comparable to other
ambulatory surgery studies in Europe with re-
gard to age, ASA-classification and duration of
surgery.25, 26 The multicenter nature of the study
makes it relevant for both academic and non-
academic ambulatory settings.
Conclusions
Minor ambulatory surgery has no clinically
relevant influence on overall glucose change in
patients with and without diabetes. The glu-
cose increase that does occur can be attributed
to dexamethasone administration during surgery.
Furthermore, glucose change and hyperglycemia
during ambulatory surgery are not associated
with complications after discharge. However,
patients with DM are at increased risk for post-
operative complications after ambulatory surgery
and the treating physician should be aware of this
during follow-up visits.
Key messages:
—— Minor ambulatory surgery has no sig-
nificant effect on blood glucose.
—— Dexamethasone causes a small, but
clinically irrelevant rise in blood glucose in
patients without DM.
—— Hyperglycemia during ambulatory
surgery does not seem to be associated with
postoperative complications.
—— Patients with DM are at increased risk
for postoperative complications after ambu-
latory surgery, irrespective of preoperative
hyperglycemia.
958 MINERVA ANESTESIOLOGICA September 2015
References
  1. Dungan KM, Braithwaite SS, Preiser JC. Stress hypergly-
caemia. Lancet 2009;373:1798-807.
  2. Smith FG, Sheehy AM, Vincent JL, Coursin DB. Critical
illness-induced dysglycaemia: diabetes and beyond. Crit
Care 2010;14:327.
  3. Eshuis WJ, Hermanides J, van Dalen JW, van SG, Busch
OR, van Gulik TM et al. Early postoperative hyperglyc-
emia is associated with postoperative complications after
pancreatoduodenectomy. Ann Surg 2011;253(4):739-
44.
  4. Park C, Hsu C, Neelakanta G, Nourmand H, Braunfeld M,
Wray C et al. Severe intraoperative hyperglycemia is inde-
pendently associated with surgical site infection after liver
transplantation. Transplantation 2009;87:1031-6.
  5. Pasternak JJ, McGregor DG, Schroeder DR, Lanier WL,
Shi Q, Hindman BJ et al. Hyperglycemia in patients under-
going cerebral aneurysm surgery: its association with long-
term gross neurologic and neuropsychological function.
Mayo Clin Proc 2008;83:406-17.
 6. Statistiek Cbvd. Ziekenhuisopnamen [Internet] Available
from: http://www.cbs.nl/nl-NL/menu/themas/gezondheid-
welzijn/cijfers/extra/2010-ziekenhuisopname.htm [cited
July 11, 2014].
  7. Russo A, Elixhauser A, Steiner C, Wier L. Hospital-Based
Ambulatory Surgery, 2007: Statistical Brief #86. 2006.
  8. Joshi GP, Chung F, Vann MA, Ahmad S, Gan TJ, Goul-
son DT et al. Society for Ambulatory Anesthesia consensus
statement on perioperative blood glucose management in
diabetic patients undergoing ambulatory surgery. Anesth
Analg 2010;111:1378-87.
 9. Hoedemaekers CW, Klein Gunnewiek JM, Prinsen MA,
Willems JL, Van der Hoeven JG. Accuracy of bedside glu-
cose measurement from three glucometers in critically ill
patients. Crit Care Med 2008;36:3062-6.
10. American Diabetes A. Diagnosis and classification of diabe-
tes mellitus. Diabetes Care. 2014;37(Suppl 1):S81-90.
11. Umpierrez GE, Isaacs SD, Bazargan N, You X, Thaler LM,
Kitabchi AE. Hyperglycemia: an independent marker of in-
hospital mortality in patients with undiagnosed diabetes. J
Clin Endocrinol Metab 2002;87:978-82.
12. Sheehy AM, Flood GE, Tuan WJ, Liou JI, Coursin DB,
Smith MA. Analysis of guidelines for screening diabetes
mellitus in an ambulatory population. Mayo Clin Proc
2010;85:27-35.
13. Grek S, Gravenstein N, Morey TE, Rice MJ. A cost-effec-
tive screening method for preoperative hyperglycemia. An-
esth Analg 2009;109:1622-4.
14. Sheehy AM, Benca J, Glinberg SL, Li Z, Nautiyal A, An-
derson PA et al. Preoperative “NPO” as an opportunity for
diabetes screening. J Hosp Med 2012;7:611-6.
15. Abdelmalak BB, Bonilla AM, Yang D, Chowdary HT,
Gottlieb A, Lyden SP et al. The hyperglycemic response to
major noncardiac surgery and the added effect of steroid ad-
ministration in patients with and without diabetes. Anesth
Analg 2013;116:1116-22.
16. Hans P, Vanthuyne A, Dewandre PY, Brichant JF, Bon-
homme V. Blood glucose concentration profile after 10 mg
dexamethasone in non-diabetic and type 2 diabetic patients
undergoing abdominal surgery. Br J Anaesth 2006;97:164-
70.
17. Murphy GS, Szokol JW, Avram MJ, Greenberg SB, Shear T,
Vender JS et al. The effect of single low-dose dexamethasone
on blood glucose concentrations in the perioperative peri-
od: a randomized, placebo-controlled investigation in gyne-
cologic surgical patients. Anesth Analg. 2014;118:1204-
12.
18. Standards of medical care in diabetes--2009. Diabetes Care
2009;32(Suppl 1):S13-S61
19. Nazar CE, Echevarria GC, Lacassie HJ, Flores RA, Munoz
HYPERGLYCEMIA AND AMBULATORY SURGERY POLDERMAN

HR. [Effects on blood glucose of prophylactic dexametha-
sone for postoperative nausea and vomiting in diabetics and
non-diabetics]. Rev Med Chil 2011;139:755-61.
20. van Boxel GI, Hart M, Kiszely A, Appleton S. Elective day-
case laparoscopic cholecystectomy: a formal assessment of
the need for outpatient follow-up. Ann R Coll Surg Engl
2013;95:e142-e6.
21. Khan M, Ahmed A, Abdullah L, Nizar A, Fareed A, Khan
FA. Unanticipated hospital admission after ambulatory sur-
gery. J Pak Med Assoc 2005;55:251-2.
22. Tewfik MA, Frenkiel S, Gasparrini R, Zeitouni A, Dan-
iel SJ, Dolev Y et al. Factors affecting unanticipated hos-
pital admission following otolaryngologic day surgery. J
Otolaryngol 2006;35:235-41.
23. Whippey A, Kostandoff G, Paul J, Ma J, Thabane L, Ma
HK. Predictors of unanticipated admission following am-
bulatory surgery: a retrospective case-control study. Can J
Anaesth 2013;60:675-83.
24. Kao LS, Meeks D, Moyer VA, Lally KP. Peri-opera-
tive glycaemic control regimens for preventing surgi-
cal site infections in adults. Cochrane Database Syst Rev
2009:CD006806.
25. Martin-Ferrero MA, Faour-Martin O, Simon-Perez C, Pe-
rez-Herrero M, de Pedro-Moro JA. Ambulatory surgery in
orthopedics: experience of over 10,000 patients. J Orthop
Sci 2014;19:332-8.
26. Mattila K, Toivonen J, Janhunen L, Rosenberg PH, Hy-
nynen M. Postdischarge symptoms after ambulatory sur-
gery: first-week incidence, intensity, and risk factors. Anesth
Analg 2005;101:1643-50.

Preliminary data for this study were presented as a poster presentation at the European Society of Anaesthesiology (ESA) Euroanaesthesia,
June 12th 2012, Paris and at the European Society of Anaesthesiology (ESA) Euroanaesthesia 2014, Stockholm, June 3rd 2014.
Acknowledgments.—We would like to acknowledge the following people for their help during patient recruitment and data collection: W.
ten Hoope, R.H.A. Soetens, E. Pariama, N.C.J. Schellekens, F. de Groot from the Department of Anesthesiology, Academic Medical Cen-
tre, Amsterdam, The Netherlands and E. Molmans from the Department of Anesthesiology, Diakonessenhuis, Utrecht, The Netherlands.
We acknowledge Roche Diagnostics Nederland BV, who provided the Accu-Chek Inform and test strips free of charge during the inclusion
period of the study.
Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on September 2, 2014. - Accepted for publication on January 14, 2015. - Epub ahead of print on January 16, 2015.
Corresponding author: B. Preckel, Department of Anesthesiology, Academic Medical Centre, Postbus 22660, 1100 DD Amsterdam, The
Netherlands. E-mail: b.preckel@amc.uva.nl

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 959



O R I G I N A L A RT I C L E

Preoperative adherence to continuous positive airway

pressure among obstructive sleep apnea patients
E. DEFLANDRE 1, S. DEGEY 2, V. BONHOMME 3, A. F. DONNEAU 4
R. POIRRIER 5, J. F. BRICHANT 6, P. HANS 7

1Department of Anesthesia, Clinique Saint-Luc, Bouge, Belgium & Cabinet Medical ASTES, Jambes, Belgium;
2Cabinet Medical ASTES, Jambes, Belgium; 3University Department of Anesthesia and ICM, CHR Citadelle and CHU
Liege, Liege, Belgium; 4Medical Informatics and Biostatistics, University of Liege, Liege, Belgium; 5Department of
Neurology, CHU Liege, Liege, Belgium; 6Department of Anesthesia and ICM, CHU Liege, Liege, Belgium; 7University
Department of Anesthesia and ICM, CHR Citadelle and CHU Liege, Liege, Belgium

ABSTRACT
Background. Obstructive Sleep Apnea (OSA) increases the perioperative risk of complications. Chronic use of Con-
tinuous Positive Airway Pressure (CPAP) by patients decreases the importance of comorbidities caused by the OSA.
However, many patients do not adhere to the treatment. Given the postoperative complications, it is important for
the anesthesiologist to identify non-adherent patients. This prospective study was designed to identify factors that
would predict patient adherence.
Methods. Ninety patients who were treated by CPAP for more than one year were recruited. Among them, and
based on objective criteria such as length of use of CPAP during the night, 75 were considered as being adherent
to CPAP, while the other 15 were not. Sixty-two potential causes of non-adherence were investigated (some have
not been tested before), and further divided into five categories. Those categories included cultural, intellectual, or
economic factors, OSA comorbidities, patient belief about health, ENT-related problems, and pathophysiological
features estimating the degree of improvement afforded by CPAP introduction.
Results. Multivariate binary logistic regression analysis identified one criterion of non-adherence to treatment,
namely the feeling of breathlessness, and three criteria of adherence, namely awareness of the risk of complications,
awareness of treatment efficacy, and feeling of being less tired with CPAP therapy.
Conclusions. These four new criteria should preoperatively be sought, in order to detect non-adherent patients
more efficiently. (Minerva Anestesiol 2015;81:960-7)
Key Words: Sleep apnea - Obstructive - Continuous positive airway pressure - Patient compliance - Anesthesia -
Perioperative period.

T he obstructive sleep apnea (OSA) syndrome

is characterized by recurrent interruptions
(apnea) or reductions (hypopnea) of flow in the
airway. The prevalence of this syndrome is esti-
mated to range between 3 and 7% in the general
population.1 OSA is associated with multiple
comorbidities (coronary artery disease, chroni-
cally elevated arterial blood pressure, increased
risk of stroke, and higher rates of postoperative
complications such as respiratory failure or car-
diac arrhythmia).2-4 All these reasons contribute
to consider OSA as an important risk factor of
perioperative morbidity and mortality.5-9 It is
also considered as a contra-indication for surgery
on an outpatient basis.10 Recent studies demon-
strate the utility of identifying apneic patients
in the preoperative period. Indeed the incidence
of these patients can reach 13% in the surgical
population.11, 12
Continuous Positive Airway Pressure (CPAP),

960 MINERVA ANESTESIOLOGICA September 2015

PREOPERATIVE ADHERENCE TO CPAP AMONG OSA PATIENTS DEFLANDRE
first described by Sullivan in 1981,13 is the most
successful treatment of OSA.14-19 CPAP appli-
cation has been demonstrated significantly to
reduce the risk of both apnea and cardiovascu-
lar complications.14-19 However, many patients
do not adhere to this treatment for several rea-
sons.20, 21 It is reasonable to assume that patients
with poor adherence to CPAP have complica-
tion risks that overlap the risks of non-treated
OSA patients. Those patients should be carefully
monitored during the postoperative period, and
should, therefore, not be eligible for outpatient
surgery. Measures for improving treatment ad-
hesion have been investigated.20, 22-25 The most
often cited reasons of non-adhesion include
worsening of sleepiness, lack of improvement in
daily functioning, nasal congestion, and high-
baseline AHI (Apnea Hypopnoea Index, which
is defined by the number of apnea and hypopnea
per hour).20, 21, 23, 26 Several previous studies have
been conducted with older noisy machines.
The anesthesiologists can have a false sense
of security if they take care of non-adherent pa-
tients. This study aimed at identifying criteria
associated with poor adherence to current CPAP
devices.
Materials and methods
Following Institutional Ethics Committee
approval and patient’s informed consent, 90 pa-
tients were enrolled (in the CHU of Liege). The
study was recorded by ClinicalTrials.gov under
the number NCT02055027. All of the 90 pa-
tients had received CPAP for at least one year
and were visiting the Sleep Center of the Univer-
sity Hospital of Liege for the annual checking of
the CPAP machine. Upon that visit, all patients
spent an overnight polysomnography (OPS)
and had an oral interview described below. Data
from the OPS, the oral interview, and the CPAP
microchip were collected.
The investigator in charge of the interview
was not aware of the patient’s status (adherent
Table I.—The psychometric Lickert’s scale.
1 2
Strongly disagree Disagree Neither agree nor disagree
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 961
or non-adherent). The interview was based on
a questionnaire that was partly defined, accord-
ing to pertinent criteria found in the literature.
Original criteria were also added. As a result, 62
criteria were defined and divided into 5 axes of
interest: 1) demographic data and socioeconom-
ic status; 2) CPAP-related problems; 3) ENT-re-
lated problems; 4) patient belief about healt; and
5) pathophysiological features. For questions
requiring a patient choice, and related to his/
her own experience, the psychometric Lickert’s
scale was used to rate the different criteria (Table
I).27-28 The 62 criteria (divided into 5 axes) are
listed in Table II.
Classification of patients into adherent and
non-adherent to CPAP treatment was per-
formed according to the number of hours of ef-
fective CPAP use during the night, using data
recorded on the CPAP microchip. Patients with
less than four hours of CPAP use were consid-
ered as non-adherent while patients with more
than four hours were considered as adherent.29
Results were expressed as means ± standard
deviations (SD) for quantitative variables, and
as counts and proportions (%) for categori-
cal variables, unless otherwise indicated. Mean
values between the two groups were compared
using Student t-tests or Wilcoxon tests. Propor-
tions were compared using χ2 tests. Adherents
and non-adherents were compared in each of the
five axes using multivariate binary logistic regres-
sion analysis. Results were expressed in terms of
odds ratio (OR) together with 95% confidence
intervals (CI). All results were considered to be
significant at the 5% critical level. Statistical
analyses were carried out using SAS (version 9.3
for Windows) and S-Plus (version 8.1) statistical
packages.
Results
None of the 90 enrolled patients was ex-
cluded. According to the analysis of the CPAP
microchip,29 75 patients (83.3 %) could be con-
3 4 5
Agree Strongly agree
DEFLANDRE PREOPERATIVE ADHERENCE TO CPAP AMONG OSA PATIENTS

Table II.—Criteria analysed in the five axis (demographic data and socioeconomic status).
First axis Second axis Fourth axis
(demographic data and socioeconomic (CPAP -related problems) (patient beliefs about health)
status) All were graded using the Lickert’s scale. All were graded using the Lickert’s scale.
Noise-related problems: Patient’s perception of a threat to his
Demographic data: –– CPAP noise level as described by the health:
–– Age patient –– Patient is convinced that OSA is a
–– Gender –– CPAP noise level as described by the serious disease
–– Weight cohabitant –– Patient knows the risk of OSA
–– Height –– Importance of air leakage with the complications without CPAP use
–– Length of CPAP therapy CPAP mask
–– Level of pressure applied by the CPAP Patient’s level of belief in the efficacy of its
–– Model of CPAP Patient’s perceived disadvantages: treatment:
–– Hours of CPAP use as recorded by the –– Allergies to the CPAP mask –– Patient knows that CPAP is an effective
microchip –– Wounds to the face caused by the CPAP treatment for the OSA syndrome
–– Hours of CPAP use as said by the mask –– There are more advantages than
patient –– Feeling of breathlessness with the CPAP disadvantages to CPAP therapy
–– Others comorbidities? mask –– Number of CPAP advantages
–– Hypertension spontaneously cited by the patient
–– Coronary artery diseases (cad) Patient’s aesthetic considerations: –– Number of CPAP disadvantages
–– Need for coronary artery bypass –– Aesthetic discomfort felt by the patient spontaneously cited by the patient
grafting –– Teasing from the cohabitant
–– Diabetes –– Causes discomfort in the privacy of the Fifth axis
–– Hypercholesterolemia couple (pathophysiological features axis)
–– Obesity The subjective items (with exception of
–– Others Third axis the Epworth’s scale) were graded using
(ent-related problems) the Lickert’s scale. The others are numeric
Socioeconomic status: All were graded using the Lickert’s scale. values. Explanations in the text.
–– What is your nationality? Ent-related problem due to the CPAP
–– What is the level of your highest mask: Subjective approach:
diploma? –– Vasomotor rhinitis: –– Fatigue is the patient main symptom
–– What is your profession? –– Rhinorrhea –– The patient is less tired with CPAP use
–– What is your current level of activity? –– Sneezing fits –– The patient is less drowsy with CPAP
–– Active –– Nasal obstruction: use
–– Inactive: –– Stuffy feeling –– Epworth’s scale before CPAP treatment
–– Retired –– Inability to breathe through the nose –– Epworth’s scale at the day of
–– Sick leave consultation (under CPAP treatment)
–– Unemployed Ent-surgical status:
–– None Objective approach:
–– Ent-surgery: –– Ahi, before and after CPAP therapy
–– Nasal septum –– Vai, before and after CPAP therapy
–– Sinus –– Odi, before and after CPAP therapy
–– Nasal polypectomy –– Sqi, before and after CPAP therapy
–– Uvulopalatopharyngoplasty –– Sei, before and after CPAP therapy
–– Others –– Sci, before and after CPAP therapy

sidered as adherent to CPAP therapy and 15 pa-
tients (16.7%) as non-adherent. Demographic
data are summarized in the Table III. The two
groups were comparable in terms of demo-
graphic data and socioeconomic status listed in
the Table II. The mean value of CPAP length of
use was 7.08±1.35 hours per night for adherent
patients, and 0.4±1.06 hours per night for the
non-adherent patients. There was no between-
group difference regarding the number of years
of CPAP therapy (6.09±3.04 years for adher-
ent and 6.03±2.09 for non-adherent group,
P=0.56), the level of CPAP pressure (10.6±2.05
cm of water for adherent and 10.5±2.73 cm of
water for non-adherent group, P=0.78), and the
model of CPAP device (Model PV10: P=0.92,
Model PV100: P=0.7, Model Sleep 20: P=0.64,
Model Somnotron: P=0.06, Model PC 90:
P=0.52, Model Sullivan: P=0.65).
Significant results of between-group compari-
sons of each recorded criteria are listed in Table
IV (univariate analysis). Significant results of the

962 MINERVA ANESTESIOLOGICA September 2015

PREOPERATIVE ADHERENCE TO CPAP AMONG OSA PATIENTS DEFLANDRE

Table III.—Demographic data in both groups of patients.

Adherent Non-adherent
Variable P value
N.=75 N.=15
Age (y), mean (standard deviation) 59.0 (9.0) 59.1 (13.0) 0.76
Gender m/f (%) 66/9 (88.0/12.0) 13/2 (86.7/13.3) 0.88
BMI (kg m-2), mean (standard deviation) 34.5 (6.5) 33.6 (7.2) 0.63

Table IV.—Significant results of the between-group comparisons (univariate analysis). Data are expressed as mean (SD -
standard derivation). Statistics were performed using student t-tests for unrelated samples. A two-tailed P-value <0.05 was
considered as significant.
Adherent Non-adherent
Variable P value
N.=75 N.=15
CPAP -related problems
CPAP noise level as described by the patient 1.21 (0.83) 1.93 (1.67) 0.001
Importance of air leakage with the CPAP mask 1.92 (1.35) 3.00 (2.00) 0.001
Allergies to the CPAP mask 1.13 (0.68) 1.67 (1.45) 0.03
Feeling of breathlessness with the CPAP mask 1.13 (0.60) 2.33 (1.95) <0.0001
Patient’s belief about health
Patient is convinced that OSA is a serious disease 4.48 (0.98) 3.67 (1.50) 0.009
Patient knows the risk of OSA complications without CPAP use 4.67 (0.64) 3.33 (1.88) <0.0001
Patient knows that CPAP is an effective treatment for the OSA syndrome 4.76 (0.63) 3.93 (1.28) 0.0003
There are more advantages than disadvantages to CPAP therapy 4.76 (0.67) 4.07 (1.44) 0.005
Number of CPAP advantages spontaneously cited by the patient 2.27 (1.00) 1.47 (1.06) 0.006
Number of CPAP disadvantages spontaneously cited by the patient 1.47 (1.18) 2.87 (1.24) <0.0001
Pathophysiological features
The patient is less tired with CPAP use 4.36 (1.26) 3.33 (1.72) 0.008
The patient is less drowsy with CPAP use 4.32 (1.38) 2.93 (1.94) 0.01

Table V.—Significant results of the multivariate analysis adjusted for other items. Data are expressed as odds ratio and 95%
CI (confidence interval).
Variable Odds ratio 95% CI
CPAP -related problems
Feeling of breathlessness with the CPAP mask 0.52 0.31-0.86
Patient’s belief about health
Awareness of the risk of OSA complications without CPAP use 2.54 1.36-4.75
Awareness of CPAP is an effective treatment for the OSA syndrome 3.19 1.20-8.46
Pathophysiological features
Feeling to be less tired with CPAP use 1.74 1.2-2.52

multivariate logistic regression analysis adjusted CPAP-related problems

for other items are listed in the Table V and il-
lustrated in Figure 1. All results are detailed here-
after.
Demographic data and socioeconomic status
Demographic data (age: P=0.76, sex: P=0.88,
nationality: P=0.64, BMI: P=0.63 and comor-
bidities: P=0.74) and socioeconomic status were
comparable between groups.
As shown in Table IV, the level of discomfort
induced by the CPAP noise was significantly
higher in the non-adherent group (P=0.001), as
was the frequency of facial pressure lesions (due
to the mask) (P=0.03). However, these criteria
did not reveal to increase the odds of non-adher-
ence significantly when going through the multi-
variate analysis (OR was respectively 0.85 [95%
CI: 0.48-1.49] and 0.82 [95% CI: 0.54-1.25]).
The sensation of choking when using CPAP was

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 963

DEFLANDRE PREOPERATIVE ADHERENCE TO CPAP AMONG OSA PATIENTS

Figure 1.—Significant results from the multivariate analysis. Data are expressed in terms of Odds Ratio and 95% CI (Confidence
Interval).

significantly higher in the non-adherent group
as compared to the other group (P<0.0001). The
presence of that sensation significantly decreases
the odds of adherence by a factor of 0.52 (95%
CI: 0.3-0.86).
ENT-related problems
As evidenced by the multivariate analysis, dif-
ficulties of breathing through the nose under
CPAP decrease the odds of adherence by a factor
of 0.62 (95% CI: 0.38-1.01). However, this ef-
fect was not statistically significant. For the other
criteria of this axis, statistical analysis did not re-
veal any significant between-group differences.
Patient’s beliefs about health
From a univariate standpoint, there was a
significant between-group difference for each
of the four following tested factors: in the ad-
herent group, patients were more convinced
that OSA is a serious disease (P=0.009), they
better apprehended the risk of OSA complica-
tions without CPAP use (P<0.0001), and that
CPAP is an effective treatment for the OSA
syndrome (P=0.0003). They were also more
convinced that there are more advantages than
disadvantages to CPAP therapy (P=0.005). The
multivariate analysis (Table V) revealed that
awareness of OSA complications without CPAP
use increases the probability of adherence by
a factor of 2.54 (95% CI: 1.36-4.75), as did
a knowledge of the CPAP efficiency at treat-
ing OSA syndrome (OR: 3.19 [95% CI: 1.2-
8.46]). Other factors did not show a significant
result in the multivariate analysis: 1) patients
were convinced that OSA is a serious disease
(OR: 1.04 [95% CI: 0.58-1.87]); 2) patients
were convinced that there are more advantages
than disadvantages to CPAP therapy (OR: 0.72
[95% CI: 0.29-1.83]).

964 MINERVA ANESTESIOLOGICA September 2015

PREOPERATIVE ADHERENCE TO CPAP AMONG OSA PATIENTS DEFLANDRE
Pathophysiological features
Adherent patients were significantly less tired
(P=0.008) and less drowsy (P=0.01) than non-
adherent patients. The multivariate analysis
showed that only the feeling of being less tired
had a significant positive effect on the odds of
adherence to CPAP therapy (OR: 1.74 [95% CI:
1.2-2.52]).
Discussion
This study revealed one criterion of non-ad-
herence to treatment (feeling of breathlessness),
and three criteria of adherence (awareness of the
risk of complications, awareness of treatment
efficacy and feeling to be less tired with CPAP
therapy).
As other authors have already highlighted, de-
mographic variables and comorbidities do not
influence treatment adherence.30, 31 We found
no difference between the CPAP brands used in
our service. In this study, the feeling of suffoca-
tion was a significant risk factor of non-adher-
ence to treatment. This finding is in agreement
with previous studies.21, 32
Several authors have demonstrated that the
initial phase of treatment with CPAP (one week
to three months) is crucial for treatment ad-
herence.33 Therefore, patient education should
begin from the first days of treatment. The an-
esthesia consultation may be the perfect time to
remind the patient the possible complications
of untreated OSA, and especially postoperative
complications he/she may face.34 The fear of de-
veloping these complications makes the preop-
erative period a privileged moment for perceived
benefits from CPAP treatment that medical staff
wants to restore. This is why Shi and Warner de-
scribed (for tobacco cessation) the preoperative
period as a “teachable moment”.35
Our results show that patients who felt less
tired with CPAP were more adherent to treat-
ment than the others, which is in agreement with
already published data.36, 37 Other authors have
also highlighted the fact that patients with a low
AHI were less adherent to treatment with CPAP.38
This is entirely consistent with our results, insofar
as AHI and fatigue are closely related.39
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 965
In this study, adherent and non-adherent pa-
tients were arbitrarily differentiated according
to a threshold value of nocturnal CPAP length
of use. Many of these thresholds can be found
in the literature.40 We selected a reasonable and
clinically relevant threshold of four hours, which
has been indisputably validated by Kribbs.29
In the present study, the rate of non-adherence
(16.7%) is lower than the one that is usually re-
ported in the literature. Some authors published
values ranging from 25 to 50%.41 Such a differ-
ence could be explained by the fact that we only
included patients with CPAP therapy for a mini-
mum of one year, which could be considered as
a primary selection.
From our study, a patient should be consid-
ered as being at high risk of non-adherence to
CPAP therapy if one or more of the following
items are encountered: 1) the patient has a feel-
ing of breathlessness when he/she is using CPAP;
2) CPAP does not improve his/her feeling of fa-
tigue; 3) the patient is not aware of the risk of
complications with OSA; 4) he/she is not aware
that CPAP therapy is effective to reduce this risk.
Although this has not been specifically
studied in our study, in certain circumstances
(ENT-surgery) the postoperative CPAP use
may be contraindicated, and this for many rea-
sons. First, there may be a mechanical impedi-
ment to the application of CPAP due to the
surgical nasal packing. Secondly, in subjects
with normal Eustachian tube function, CPAP
use (even at minimal levels) increases middle
ear pressure above supraphysiologic levels.42
These data can make the use of CPAP inap-
propriate after some surgeries of the ear. Fi-
nally, CPAP use may increase the risk of bleed-
ing after a trans-oral or a pharyngeal surgery.
However, there are few data in the literature
to address this problem. Authors can refer to
the recent report of the American Society of
Anesthesiologists which suggests administering
supplemental oxygen in these patients.43 This
attitude does not reduce the AHI, but reduces
the frequency and severity of desaturation. On
the other hand, recent published data suggest
that a previous nasal surgery in patients with
nasal surgery can reduce the level of therapeu-
tic CPAP pressure 44 and improve CPAP adher-
DEFLANDRE PREOPERATIVE ADHERENCE TO CPAP AMONG OSA PATIENTS
ence.45 This topic has not been addressed in
our research.
A patient non-adherent to CPAP therapy
should be considered as a patient at high risk of
postoperative complications, due to the instabil-
ity of his/her upper airway. In this patient, it will
be necessary to minimize the doses of hypnotic
and opioid agents. The neuromuscular func-
tion should also be monitored very carefully. If
in doubt about the safety of the upper airway
for the first postoperative night, those patients
should be followed in a PostAnesthesia Care
Unit for an intensive monitoring. The eligibility
for surgery on an outpatient basis must be recon-
sidered.10 As soon as possible, the patient should
be referred to a patient education program tar-
geting obstructive sleep apnea and CPAP thera-
py problems. However, this should not be at the
expense of delay for performing surgery.
Conclusions
Our study highlights one criterion of non-ad-
herence to treatment (feeling of breathlessness),
and three criteria of adherence (awareness of the
risk of complications, awareness of treatment
efficacy and feeling to be less tired with CPAP
therapy). These four criteria should preopera-
tively be sought, in order to detect non-adherent
patients more efficiently. Patient non-adherent
to CPAP therapy should not be cancelled before
surgery, but the anesthesiologists must take this
into account for the management of postopera-
tive patient risk. This adherence shall be consid-
ered for eligibility in a day-care center.
Key messages
—— Non-adherence to CPAP therapy can
raise up to 50%.
—— It is important for anesthesiologists to
detect preoperative non-adherence to CPAP
therapy.
—— Our study found four criteria of ad-
herence/non-adherence that every anesthesi-
ologist should seek before surgery.
—— This adherence shall be considered for
eligibility in a day-care center.
966 MINERVA ANESTESIOLOGICA September 2015
References
  1. Punjabi NM. The Epidemiology of Adult Obstructive Sleep
Apnea. Proc Am Thorac Soc 2008;5:136-43.
 2. Budhiraja R, Budhiraja P, Quan S. Sleep-disordered
breathing and cardiovascular disorders. Respir Care
2010;55:1322-32; discussion 30-2.
 3. Shafazand S. Perioperative management of obstructive
sleep apnea: Ready for prime time? Cleve Clin J Med
2009;76:S98-S103.
  4. Malhotra A, White DP. Obstructive sleep apnoea. The Lan-
cet 2002;360:237-45.
 5. Gross J, Bachenberg K, Benumof J, Caplan R, Connis
R, Cote C et al. Practice Guidelines for the Perioperative
Management of Patients with Obstructive Sleep Apnea: A
Report by the American Society of Anesthesiologists Task
Force on Perioperative Management of Patients with Ob-
structive Sleep Apnea. Anesthesiology 2006;104:1081-93.
  6. Adesanya A, Lee W, Greilich N, Joshi G. Perioperative man-
agement of obstructive sleep apnea. Chest 2010;138:1489-
98.
  7. Chung SA, Yuan H, Chung F. A Systemic Review of Ob-
structive Sleep Apnea and Its Implications for Anesthesiolo-
gists. Anesth Analg 2008;107:1543-63.
  8. Chung F, Elsaid H. Screening for obstructive sleep apnea
before surgery: why is it important? Curr Opin Anaesthesiol
2009;22:405-11.
  9. Hwang D, Shakir N, Limann B, Sison C, Kalra S, Shulman
L et al. Association of Sleep-Disordered Breathing With
Postoperative Complications. Chest 2008;133:1128-34.
10. Sabers C, Plevak DJ, Schroeder DR, Warner DO. The
Diagnosis of Obstructive Sleep Apnea as a Risk Factor for
Unanticipated Admissions in Outpatient Surgery. Anesth
Analg 2003;96:1328-35.
11. Corso RM, Petrini F, Buccioli M, Nanni O, Carretta E,
Trolio A et al. Clinical utility of preoperative screening with
STOP-Bang questionnaire in elective surgery. Minerva An-
estesiol 2014;80:877-84.
12. Deiner S, Osborn IP. Stop! Bang the drum slowly, proceed
with caution when assessing obstructive sleep apnea. Min-
erva Anestesiol 2014;80:867-8.
13. Sullivan C, Issa F, Berthon-Jones M, Eves L. Reversal of ob-
structive sleep apnoea by continuous positive airway pres-
sure applied through the nares. Lancet 1981;1:862-5.
14. Loube DI, Gay PC, Strohl KP, Pack AI, White DP, Col-
lop NA. Indications for Positive Airway Pressure Treat-
ment of Adult Obstructive Sleep Apnea Patients. Chest
1999;115:863-6.
15. Verse T, Pirsig W, Stuck B, Hormann K, Maurer J. Recent
developments in the treatment of obstructive sleep apnea.
Am J Respir Med 2003;2:157-68.
16. Stephen G, Eichling P, Quan S. Treatment of sleep disor-
dered breathing and obstructive sleep apnea. Minerva Med
2004;95:323-36.
17. Kushida C, Littner M, Hirshkowitz M, Morgenthaler T,
Alessi C, Bailey D et al. Practice parameters for the use of
continuous and bilevel positive airway pressure devices to
treat adult patients with sleep-related breathing disorders.
Sleep 2006;29:375-80.
18. Gay P, Weaver T, Loube D, Iber C. Evaluation of positive
airway pressure treatment for sleep related breathing disor-
ders in adults. Sleep 2006;29:381-401.
19. Giles T, Lasserson T, Smith B, White J, Wright J, Cates
C. Continuous positive airways pressure for obstruc-
tive sleep apnoea in adults. Cochrane Database Syst Rev
2006;3:CD001106.
20. Kakkar RK, Berry RB. Positive Airway Pressure Treatment
for Obstructive Sleep Apnea. Chest 2007;132:1057-72.
21. Weaver TE, Grunstein RR. Adherence to Continuous Posi-
tive Airway Pressure Therapy: The Challenge to Effective
Treatment. Proc Am Thorac Soc 2008;5:173-8.
PREOPERATIVE ADHERENCE TO CPAP AMONG OSA PATIENTS DEFLANDRE

22. Hui DSC, Chan JKW, Choy DKL, Ko FWS, Li TST, Leung
RCC et al. Effects of Augmented Continuous Positive Air-
way Pressure Education and Support on Compliance and
Outcome in a Chinese Population. Chest 2000;117:1410-
6.
23. Hui DSC, Choy DKL, Li TST, Ko FWS, Wong KK, Chan
JKW et al. Determinants of Continuous Positive Airway
Pressure Compliance in a Group of Chinese Patients With
Obstructive Sleep Apnea. Chest 2001;120:170-6.
24. Haniffa M, Lasserson T, Smith I. Interventions to improve
compliance with continuous positive airway pressure for
obstructive sleep apnoea. Cochrane Database Syst Rev
2004:CD003531.
25. Smith I, Nadig V, Lasserson T. Educational, supportive
and behavioural interventions to improve usage of con-
tinuous positive airway pressure machines for adults with
obstructive sleep apnoea. Cochrane Database Syst Rev
2009:CD007736.
26. Weaver T, Sawyer A. Adherence to continuous positive
airway pressure treatment for obstructive sleep apnoea:
implications for future interventions. Indian J Med Res
2010;131:245-58.
27. Drinkwater B. A comparison of the direction-of-perception
technique with the Likert method in the measurement of
attitudes. J Soc Psychol 1965;67:189-96.
28. Lickert R. A Technique for the Measurement of Attitudes.
Archives of Psychology 1932;140:1-55.
29. Kribbs N, Pack A, Kline L, Smith P, Schwartz A, Schubert
N et al. Objective measurement of patterns of nasal CPAP
use by patients with obstructive sleep apnea. Am Rev Respir
Dis 1993;147:887-95.
30. Budhiraja R, Parthasarathy S, Drake C, Roth T, Sharief I,
Budhiraja P et al. Early CPAP use identifies subsequent ad-
herence to CPAP therapy. Sleep 2007;30:320-4.
31. Scharf SM, Seiden L, DeMore J, Carter-Pokras O. Ra-
cial differences in clinical presentation of patients with
sleep-disordered breathing. Sleep Breath 2004;8:173-
83.
32. Weaver T, Chasens E. Continuous positive airway pressure
treatment for sleep apnea in older adults. Sleep Med Rev
2007;11:99-111.
33. Collen J, Lettieri C, Kelly W, Roop S. Clinical and polys-
omnographic predictors of short-term continuous positive
airway pressure compliance. Chest 2009;135:704-9.
34. Bollig SM. Encouraging CPAP adherence: it is everyone’s
job. Respir Care 2010;55:1230-9.
35. Shi Y, Warner D. Surgery as a teachable moment for smok-
ing cessation. Anesthesiology 2010;112:102-7.
36. Wells R, Freedland K, Carney R, Duntley S, Stepanski E.
Adherence, reports of benefits, and depression among pa-
tients treated with continuous positive airway pressure. Psy-
chosom Med 2007;69:449-54.
37. Barbe F, Mayoralas L, Duran J, Masa J, Maimo A, Montser-
rat J et al. Treatment with continuous positive airway pres-
sure is not effective in patients with sleep apnea but no day-
time sleepiness. a randomized, controlled trial. Ann Intern
Med 2001;134:1015-23.
38. Lin H, Prasad A, Pan C, Rowley J. Factors associated with
noncompliance to treatment with positive airway pressure.
Arch Otolaryngol Head Neck Surg 2007;133:69-72.
39. Kapur V. Obstructive sleep apnea: diagnosis, epidemiology,
and economics. Respir Care 2010;55:1155-67.
40. Lewis KE, Seale L, Bartle IE, Watkins AJ, Ebden P. Early
predictors of CPAP use for the treatment of obstructive
sleep apnea. Sleep 2004;27:134-8.
41. Zozula R, Rosen R. Compliance with continuous positive
airway pressure therapy: assessing and improving treatment
outcomes. Curr Opin Pulm Med 2001;7:391-8.
42. Lin FY, Gurgel RK, Popelka GR, Capasso R. The effect of
continuous positive airway pressure on middle ear pressure.
The Laryngoscope 2012;122:688-90.
43. Practice Guidelines for the Perioperative Management of
Patients with Obstructive Sleep Apnea: An Updated Report
by the American Society of Anesthesiologists Task Force on
Perioperative Management of Patients with Obstructive
Sleep Apnea. Anesthesiology 2014;120:268-86.
44. Camacho M, Riaz M, Capasso R, Ruoff CM, Guil-
leminault C, Kushida CA et al. The Effect of Nasal Surgery
on Continuous Positive Airway Pressure Device Use and
Therapeutic Treatment Pressures: A Systematic Review and
Meta-Analysis. Sleep 2015;38:279-86.
45. Poirier J, George C, Rotenberg B. The effect of nasal surgery
on nasal continuous positive airway pressure compliance.
The Laryngoscope 2014;124:317-9.

Presented at the following Congresses: First World ACCP-CHEST Congress, Madrid, Spain, March 21-24, 2014; Forth Annual Meeting
of the Society of Anesthesia and Sleep Medicine (SASM), New Orleans, October 9-10, 2014; Annual Meeting of the American Society of
Anesthesiologists (ASA), New Orleans, October 11-15, 2014.
Acknowledgements.—The authors want to acknowledge Laurent Cambron as participating investigator.
Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on August 1, 2014. - Accepted for publication on November 27, 2014. - Epub ahead of print on December 5, 2014.
Corresponding author: E. Deflandre, Chaussee de Tongres, 29,B – 4000 Liege, Belgium. E-mail: eric.deflandre@gmail.com

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 967



O R I G I N A L A RT I C L E

Careful monitoring of the use of sedative drugs

at the end of life: the role of Epidemiology.
The ITAELD study
G. MICCINESI 1, A. CARACENI 2, J. A. RAHO 3, E. PACI 1, F. BULLI 1
L. VAN DEN BLOCK 4, A. GIANNINI 5

1Clinicaland Descriptive Epidemiology Unit, Cancer Prevention and Research Institute-ISPO, Florence, Italy;2Palliative
Care Unit (Pain Therapy-Rehabilitation), National Cancer Institute, Milan, Italy;3Department of Philosophy,
University of Pisa, Pisa, Italy;4End-of-Life Care Research Group, Vrije Universiteit Brussel (VUB) and Ghent University,
Department of Family Medicine and Chronic Care, Brussels, Belgium;5Pediatric Intensive Care Unit, Fondazione
IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, Milan, Italy

ABSTRACT
Background. Sedative drugs are often used at the end of life for different clinical indications, and sometimes seda-
tion is not interrupted until the patient dies. The aim of this study was to estimate the prevalence of patients who
died while deeply sedated in Italy in 2007.
Methods. Cross-sectional survey which asked physicians about the last death that occurred among their assisted pa-
tients during the last year, and about their attitudes towards end-of-life decisions. All general practitioners (N=5,710)
and a random sample of hospital physicians (N=8,950) from 14 Italian provinces were invited to participate.
Results. The response rate was 20%. Among 1855 reported deaths, 1466 (79.2%) were classified by physicians
as expected or non-sudden; 18.2% of these expected or non-sudden deaths occurred while the patient was
deeply sedated. GPs were the least likely to report deep sedation, whereas anesthetists were the most likely. In
8% of cases, sedation occurred along with an abrupt increase in the dosage of opioids during the last day of
life, reaching a dosage considered higher than necessary by the doctor. No association with positive attitudes
of the physician towards physician assisted death was found, whereas reporting sedation was associated with a
positive attitude towards respecting the choice of relatives to forgo life-sustaining treatment in the case of an
incompetent patient.
Conclusion. Our study confirms the high prevalence of patients in Italy who die while being deeply sedated
and shows that different practices may converge under the same label. Careful descriptive language is needed.
(Minerva Anestesiol 2015;81:968-79)
Key words: Deep sedation - Terminal care - Epidemiology.

E uropean palliative care specialists recently

described palliative sedation as a medical
procedure that deserves continuous quality mon-
itoring.1 The ethically-relevant dimensions of se-
dation also must be taken into account, such as
the motivation to perform sedation and the in-
Comment on p. 937.
tention ‑ i.e., the plan of action ‑ that guides the
clinician.2 This is particularly important when
evaluating sedation that induces a deep state of
unconsciousness, which is then not interrupted
before the death of the patient. It has been ar-
gued that the relevant ethical dimensions are as
many as seventeen.3 This suggests that sedation
at the end of life may bring clinicians into an

968 MINERVA ANESTESIOLOGICA September 2015

CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE MICCINESI
ethical “gray zone”; sedation may be an appro-
priate medical response to refractory symptoms,
which honors the palliative intent of neither
������������
has-
tening nor postponing natural death, or sedation
may be used intentionally to hasten death.4, 5
Close monitoring of the quality of care of pa-
tients in this context is important for research
in palliative care and helps ensure that sedation
practices are ethical. An epidemiological per-
spective is also necessary to search for temporal
trends and differences between geographical ar-
eas. These two perspectives taken together ‑ the
clinical and the epidemiological ‑ overlap and are
complementary.
A consistent and widely shared methodology
must be applied in order to achieve the epide-
miological aim of depicting the practice at the
population level. Different options are avail-
able—from prospective or retrospective studies
on clinical series 6-10 to population-based studies.
Regarding the latter, the����������������������
methodological frame-
work used for end-of-life medical decisions ‑ i.e.,
those medical decisions that take into account
the possibility that death is hastened ‑ can be ap-
plied.11, 12 Although it may be criticized for rely-
ing on the retrospective opinions of physicians,
this framework has afforded an opportunity to
reliably compare different European countries
regarding relevant end-of-life issues.13, 14
In 2007, Italy began its first (controversial)
public debate about advance directives. At the
time, the ITAELD observational study ��������
was�����
con-
ducted to offer further empirical grounding for
the public debate; it showed a trend in Italy to-
wards growing awareness and acceptability of
end-of-life medical decisions.13, 15, 16
Three items of the ITAELD questionnaire
referred to the use of drugs to deeply sedate
patients who afterwards died. We made use of
these items in our current study. Our aim was to:
1. present an updated population-based es-
timate of the prevalence of patients who died
while being deeply sedated in Italy;
2. evaluate the association between sedation
and patient/physician characteristics, including
physicians’ attitudes towards end-of-life medical
decisions; and
3. describe how deep sedation at the end of
life was performed.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 969
Materials and methods
The ITAELD survey was conducted in 2007.
This study asked physicians about the end-of-life
care provided to their last deceased patient and
about physicians’ attitudes towards end-of-life
medical decisions. The research protocol was re-
viewed by an independent board of the medical
professional organization which represents all
Italian physicians.
In this study 14 (of 110) Italian provinces were
selected from throughout the country. All general
practitioners (N.=5710) and a random sample of
hospital physicians (N.=8950) from the 14 prov-
inces were invited to participate. Provinces were
enrolled throughout the country to homogene-
ously represent the four Italian statistical areas,
including both metropolitan and non-metro-
politan areas. Specifically, the selected provinces
were the following: in the North-West, Bergamo
(1163 physicians invited) and Turin (1770); in
the North-East, Padua (771), Trento (863), and
Forlì-Cesena (773); in Central Italy, Florence
(945) and Ancona (783); and in the South/Is-
lands areas, Naples (1994), Catanzaro (769),
Lecce (1082), Pescara (738), Palermo (1507),
Ragusa (692), and Sassari (860). Physicians were
asked about the last death that occurred during
the past 12 months among their patients.17 These
physicians were also asked about their attitudes
towards end-of-life medical decisions. ��������
Respond-
ents who did not have any patients die during the
last year were asked to complete only those items
pertaining to attitudes. The questionnaire was
sent by mail with a prepaid envelope for anony-
mous return. No reminder was sent.
Physicians ����������������������������������
were provided���������������������
one single question-
naire, including 13 items on physicians’ attitudes
(agreement on a 5-point Likert Scale), derived
from the European survey conducted in 2002,
and 22 items regarding practice, adapted from
the Dutch survey conducted in 2005.15, 18
Use of sedative drugs
1. “Was the patient kept continuously in deep
sedation or coma until death?”
2. “If yes, which drugs (benzodiazepines, opi-
oids, other drugs) were used?”
MICCINESI CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE

3. “During sedation, were artificial nutrition
and hydration given?”
Use of opioids
1. “Did the patient receive morphine or a de-
rivative during the last 24 hours?”
2. “If yes, which drugs were used?”
3. “Was a dosage of opioids used that was
higher than necessary to control pain and other
symptoms?”
4. “Which of the following patterns (stable,
gradual increase, abrupt increase in the last day
of life) better describes the administration of
opioids in the last three days of life?”
Statistical analysis
Only “expected or non-sudden” deaths where
information about sedation was not missing
were included in the analysis. The total number
of severe symptoms (score >3 on a 5-point scale)
was computed. Concerning opioids, we com-
bined “dosage higher than necessary” (yes/no)
and “pattern in the last three days” (collected as
a nominal variable with three categories) which
resulted in 6 categories. χ2 tests for categorical
data and unpaired t-tests for continuous data
were performed, according to the nature of the
measured variable. Multivariable logistic regres-
sion models were fitted to adjust for confound-
ing. Analysis was performed using the statistical
package STATA 12.19
Results
The flow chart of the selected sample is pre-
sented in Figure 1.
A total of 14,660 questionnaires were sent to
selected physicians and 590 questionnaires were
returned due to an incorrect mailing address;
2818 physicians returned the questionnaire.

Figure 1.—Flow chart of the selected sample from the ITAELD study.

970 MINERVA ANESTESIOLOGICA September 2015

CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE MICCINESI

Among 1,850 last deaths in the year reported,
1466 (79.2%) were classified by physicians as
expected or non-sudden; information on the
use of sedative drugs was available for 1376 of
these deaths. A total of 251 patients died while
being deeply sedated (18.2%; 95% C.I. 16.2-
20.4%).
Table I examines the characteristics of the
deceased patients, according to whether or not
they were deeply sedated. Sedated patients were
younger than non-sedated patients, more often
died of cancer, had 3 or more severe symptoms,
received palliative care or pain therapy during
the last month, and their place of death was
more often in hospice or hospital rather than at
home or in a nursing home. These associations
remained statistically significant after adjust-
ing for other patients’ characteristics, except for
cause of death.
Table II ���������������������������������
provides information on ���������
the char-
acteristics of physicians, according to whether
deep sedation was reported for their last assisted
patient; 14.8% of general practitioners, 35.1%
of oncologists, and 54.8% of physicians working
in anesthesia or intensive care units reported use
of sedation for their last assisted patient. Among
sedated patients assisted by GPs, 84/130 (65%)
received palliative care or pain therapy during
their last month of life, whereas this proportion
was quite lower for anesthetists (22%) (data not
shown in table). Younger physicians, female phy-
sicians, physicians who assisted a larger number
of dying patients�������������������������������
���������������������������������������
during the last year, and phy-
sicians who received training in palliative care

Table I.—Characteristics of deceased patients, by sedation.

Not deeply sedated Deeply sedated Percentage of P value
N. Column % N. Column % Sedated

Age at death (years) <0.001*

18-64 196 18.0 77 32.3 28.2
65-79 396 36.4 113 47.5 22.2
80+ 495 45.6 48 20.2 8.8
Sex 0.497
Female 513 47.8 108 45.4 17.4
Male 560 52.2 130 54.6 18.8
Cause of death <0.001
Malignancies 565 52.6 168 70.9 22.9
Cardiovascular 252 23.4 24 10.1 8.7
Respiratory 80 7.4 12 5.1 13.0
Nervous system 59 5.5 9 3.8 13.0
Other 119 11.1 24 10.1 16.8
Place of death <0.001*
Hospital 437 39.9 124 50.2 22.1
Home 580 53.0 105 42.5 15.3
Nursing Home 39 3.6 3 1.2 7.1
Hospice 25 2.3 15 6.1 37.5
Other 13 1.2 0 0 0
Palliative care or pain therapy received <0.001*
during the last month
Yes 279 24.8 130 51.8 31.8
No 846 75.2 121 48.2 12.5
Severe symptoms <0.001*
during the last 24 hours
0 74 6.7 5 2.0 6.3
1 188 16.9 37 14.8 16.4
2 249 22.4 48 19.2 16.2
3+ 599 54.0 160 64.0 21.1
*Association retains statistical significance (P<0.05) after adjustment (Multivariable logistic regression) for the other patient characteristics presented
in table.
Missing information: patient age=51, patient sex=65, cause of death=64, place of death=35, symptoms=16.

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 971

MICCINESI CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE

Table II.—Characteristics of physicians, by reported sedation.

Not reporting deep Percentage
Reporting deep sedation
sedation reporting P value
N. Column % N. Column % Sedation

Clinical field <0.001*

General Practitioner 747 66.4 130 51.8 14.8
Internal Medicine 65 5.8 17 6.8 20.7
Anesthesia/ICU 33 2.9 40 15.9 54.8
Surgery 42 3.7 16 6.4 27.6
Oncology 24 2.1 13 5.2 35.1
Neurology 22 2.0 1 1.4 4.4
Other 115 11.1 18 6.2 13.5
Age (years) <0.001
<40 52 4.8 24 9.6 31.6
40-50 339 31.2 97 38.8 22.3
>50 696 64.0 129 51.6 15.6
Sex 0.014*
Female 320 28.6 91 36.5 22.1
Male 798 71.4 158 63.5 16.5
Training in Palliative Care 0.002*
Yes 455 40.4 129 51.4 22.1
No 670 59.6 122 48.6 15.4
Religious beliefs or philosophical life stance relevant 0.873
for end-of-life medical decisions
Very important 320 28.6 73 29.2 18.6
Important 456 40.8 106 42.4 18.9
Somewhat important 227 20.3 45 18.0 16.5
Not at all important 115 10.3 26 10.4 18.4
Mean (sd) number of deaths occurred during the last <0.001*
12 months
Mean and (SD) 10.1 (11.5) 14.3 (17.4)
*Association retains statistical significance (P<0.05) after adjustment (Multivariable logistic regression) for the other physician characteristics pre-
sented in the table.
Missing information: age=39, sex=9, religious/philosophical beliefs=8.

more often reported sedation. These associations
remained statistically significant after adjusting
for other physicians’ characteristics, except for
age.
After adjusting for physicians’ characteristics,
Table III shows how physicians who reported
sedation more often agreed to allow relatives to
make non-treatment decisions for incompetent
patients. No statistically significant ‑ not even
borderline ‑ results were obtained for items that
concerned physician assisted death (statements
5, 6, 8, 9, 10, 11, 12).
Finally, Table IV �����������������������������
offers�����������������������
details on how the se-
dation itself was performed. Artificial nutrition/
hydration was administered in 74% of cases;
“opioids only” were used in 46% of cases. In
8% of cases, sedation occurred together with an
abrupt increase in the dosage of opioids during
the last day of life, reaching a dosage considered
higher than necessary by the doctor. Supple-
mentary analysis through logistic regression
modelling showed no statistically significant as-
sociation between this behaviour and any of the
reported attitudes (data not shown).
Discussion
Italian physicians reported that 18.2% of ex-
pected or non-sudden deaths among the adult
population in 2007 occurred while the patient
was deeply sedated. Those more likely to die
while sedated �������������������������������
included hospital/hospice
����������������������
inpa-
tients who were younger than 80 years of age,
presenting with at least 3 severe symptoms, and
who had received palliative care or pain therapy
during the last month of life. GPs were the least
likely to report deep sedation;�������������������
by contrast, �����
anes-
thetists were the most likely group to report this

972 MINERVA ANESTESIOLOGICA September 2015

CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE MICCINESI

Table III.—Agreement of physicians with statements on end-of-life medical decisions, by reported sedation. Row percentages.
%Strongly %Strongly
Level of agreement with statements %Agree %Neutral %Disagree
agree disagree P value
Reporting deep sedation for their last assisted patient No Yes No Yes No Yes No Yes No Yes
  1. Physicians should comply with a patient’s request 20 30 44 43 9 11 18 11 8 5 .002
to withhold or withdraw life-sustaining treatment
 2. If a patient is incompetent, relatives should be 6 7 30 39 17 13 30 28 17 12 .044*
allowed to decide whether or not to withhold or
withdraw life-sustaining treatment
  3. Decisions to intensify the alleviation of pain and/ 27 26 57 55 6 7 7 11 2 1 .237
or symptoms by using potentially life-shortening
drugs should be discussed with the patient
  4. If necessary, a terminally-ill patient should receive 41 53 51 40 4 3 4 3 1 1 .012
drugs to relieve pain and suffering, even if these
drugs may hasten the end of the patient’s life
  5. A person should have the right to decide whether 19 24 29 29 16 17 22 17 14 13 .344
or not to hasten the end of his or her life
 6. Sufficient availability of high-quality palliative 24 30 47 44 14 12 13 10 2 3 .258
care prevents almost all requests for euthanasia or
assisted suicide
  7. Every person should be allowed to appoint anoth- 21 29 41 40 13 10 17 15 8 7 .060
er person to be legally entitled to make end-of-life
decisions on his or her behalf in the case of incom-
petence
  8. In all circumstances physicians should aim at pre- 16 13 31 28 15 17 30 31 9 12 .358
serving the lives of their patients, even if patients
ask for the hastening of the end of their lives
  9. Permitting the use of drugs in lethal doses at the 9 11 28 23 18 17 33 36 12 13 .363
explicit request of the patient will gradually lead
to an increase in the use of drugs in lethal doses
without a request of the pt.
10. The use of drugs in lethal doses at the explicit re- 15 20 33 36 13 12 28 24 10 9 .356
quest of the patient is acceptable for patients with
a terminal illness with extreme uncontrollable pain
or other distress
11. If a terminally-ill patient is suffering unbearably 9 10 15 14 16 19 33 31 17 16 .840
and is not capable of making decisions, the phy-
sician should be allowed to administer drugs in
lethal doses
12. Permitting the use of drugs in lethal doses at the 6 7 19 15 24 24 40 44 11 10 .539
explicit request of the patient will harm the rela-
tionship between patients and physicians
13. Clear wishes on withholding or withdrawing life- 14 14 40 40 14 14 25 21 7 6 .467
sustaining treatment of an incompetent patient as
expressed in an advance directive must always be
respected, even if this could hasten the end of the
patient’s life
*Association retains statistical significance (P<0.05) after adjustment (multivariable logistic regression) for physician characteristics, using agree/
strongly agree vs. other categories.
Missing information: min 32 (item 4), max 45 (item 2).

intervention. Palliative care training, more expe- tion found between reporting deep sedation and
rience with dying patients, and being a female positive attitudes of the physician towards physi-
doctor increased the likelihood that deep seda- cian assisted death; however, reporting sedation
tion would be reported. There was no associa- was associated with a positive attitude towards

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 973

MICCINESI CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE

Table IV.—Characteristics of deep sedation.

N. %
Artificial nutrition/hydration during deep sedation
Yes 183 74.4
No 63 25.6
Drugs administered for deep sedation
Benzodiazepines only 12 4.8
Opioids only 114 45.6
Other drugs only 6 2.4
Benzodiazepines+opioids 88 35.2
Benzodiazepines+opioids+other drugs 14 5.6
Benzodiazepines+other drugs 3 1.2
Opioids+other drugs 13 5.2
Administration of opioids in the last three days of life
Stable, proportionate 78 36.3
Stable, higher than necessary 7 3.3
Gradual increase, proportionate 74 34.4
Gradual increase, higher than necessary 17 7.9
Abrupt increase in the last day, proportionate 22 10.2
Abrupt increase in the last day, higher than necessary 17 7.9
Missing information: artificial nutrition/hydration=5, drugs used=1, dosage of opioids during the last three days=14.

allowing relatives to decide to forgo life-sustain-
ing treatment for an incompetent patient. In 8%
of sedated patients, there had been an abrupt in-
crease in the dosage of opioids during the last
day of life, defined as a dosage considered higher
than necessary by the doctor.
We aimed to examine sedation performed
on patients close to death, mainly with pallia-
tive intent, rather than address the larger con-
text in which sedation may be used (e.g., to treat
agitation, anxiety, delirium, and suffering).20 P���
al-
liative sedation has been recently identified as a
cornerstone of end-of-life care in the Intensive
Care Units (ICU)������������������������������
by t�������������������������
he Italian Society of An-
esthesia and Intensive Care.21 The definition
proposed by Morita more than ten years ago is
still helpful in clarifying what palliative sedation
should be: “the use of sedative medications to
relieve intolerable and refractory distress by the
reduction in patient consciousness”.22 In fur-
ther specifying this complex medical procedure,
Morita also suggested that we should distinguish
between different types of sedation ‑ i.e., accord-
ing to degree (mild/deep) and duration (con-
tinuous/intermittent). Furthermore, agreement
should be sought on what is meant by “refractory
symptoms”. According to Cherny and Portenoy,
refractory symptoms are those that “cannot be
adequately controlled despite aggressive efforts
to identify a tolerable therapy that does not com-
promise consciousness”. In suggesting how to as-
sess the tolerability of therapy, it is noteworthy
that Cherny and Portenoy also considered the
patient’s personal wishes and whether relief can
be achieved within a limited time frame.23 Fi-
nally, other criteria have been suggested to make
palliative sedation an appropriate medical proce-
dure: the patient has an irreversible and advanced
illness; death is imminent; there are clear goals
of care; informed consent was obtained; there
has been corroborative consultation, including
the involvement of staff and family; there is full
documentation of the patient’s clinical condition
and the medication(s) used; lastly, there is agree-
ment that cardiopulmonary resuscitation should
not be initiated.24
The particular intervention described in this
paper has been also commonly referred to in the
literature as “continuous deep sedation” or “deep
sedation until death” which,
��������������������������
taking a������������
retrospec-
tive point of view, refers to deep sedation that
was not in fact interrupted before death. The
literature recommends paying attention to the
various meanings that the term has gained in
practice, especially with reference to: the question
of palliative intent; the use of sedation as a last
resort; the administration of fluids if clinically
requested; the presence of refractory symptoms;

974 MINERVA ANESTESIOLOGICA September 2015

CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE MICCINESI
and the criterion of imminent death.3, 25, 26 We
agreed to use a descriptive ‑ i.e., a pre-evaluative
‑ label when describing epidemiological reports
on sedation. This epidemiological study cannot
address the specific clinical practice of palliative
sedation within the context of providing appro-
priate palliative care, as described in a number
of well-developed clinical guidelines 27-29 and in
some empirical observations.6-10 Nevertheless, it
does allow us to gain some general insights into
clinical practice.
In Italy, the percentage of patients who died
while being deeply sedated in 2007 is higher
than that previously reported (8.5%) for the ��� en-
tire Italian population in the EURELD study on
end-of-life medical decisions from 2000-2001.30
This difference remains even after adjusting for
the different denominators (non-sudden deaths
vs. all deaths), estimating 14.4% in 2007 among
all deaths (18.2%x0.792, i.e. the proportion of
expected deaths). Among non-sudden deaths
only, a follow-back study conducted in Belgium
using the sentinel network of GPs showed lower
estimates than ours (8% of patients in the Dutch-
speaking community and 15% of patients in the
French-speaking community received ����������
deep seda-
tion).31
Our study resembles the findings from recent
research performed in the United Kingdom con-
cerning the high percentage of deeply sedated
patients (18.7%).17 Physicians in that study were
asked to report on their last patient who died.
The well-known selection bias associated with
this kind of end-of-life study (as physicians tend
to report about the last ‘problematic’ death, in-
stead of the last patient who died) is illustrated in
our study by the low percentage of sudden and
totally unexpected deaths (21% instead of the ex-
pected 29%-34%, according to the EURELD 13
and EUROSENTIMELC 32 studies) and by the
distribution of age and cause of death (patients
younger than expected, and who died of cancer
more often than expected).16 Even granting a
likely slight overestimate, our study confirms the
wide use of sedation in Italy.
In line with other studies,33, 34 sedation seems
to be a medical practice that requires a palliative
care context: sedated patients more often had re-
ceived pain therapy or palliative care in the last
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 975
month of life, were younger, and suffered from a
higher number of severe symptoms. Additionally,
physicians more often received palliative training.
It is a priority for palliative care associations to
offer clinical guidelines and consider educating
professionals in non-specialist palliative care
���������
set-
����
tings (both in hospital and at home).
Deaths assisted by anesthetists are a small per-
centage of all deaths in the sample, but propor-
tionally the incidence of sedation is quite higher
than in other settings: of 73 deceased patients re-
ported by anesthetists, 40 were sedated. Further
research is needed on the specific characteristics
of the use of sedation at the end of life by this
group of clinicians. Concerning Italian anesthe-
tists, we must remember that, at the time of the
study, they were often involved in the end-of-life
care of patients dying in the hospital, given that
many Italian hospitals were (and still are) lacking
in palliative care services for inpatients. Finally,
training in end-of-life care is still lacking in the
curriculum of Italian anesthetists.35
The reverse is true for GPs: whereas being a
GP reduced the likelihood that a patient would
be sedated, more than half of the cases of seda-
tion reported in the study occurred among these
practitioners. This highlights the importance of
assuring quality in performing sedation. During
the last month of life, palliative/pain therapy was
used in 2 out of 3 cases reported by GPs, making
it likely that a palliative care specialist had been
involved in the sedation as well. Characteristics of
sedation performed at the end of life by GPs have
been recently studied in the Netherlands, where
guidelines had been published in 2005 and up-
dated in 2009.27, 36 These
������������������������������
guidelines professional-
�������������
ized palliative care and, according to one study,
made it possible to perform “palliative sedation as
a best practice under clearly defined conditions”.8
That study revealed that there is not an increas-
ing trend toward palliative sedation. Moreover,
there is neither a lack of clinical indications for
commencing sedation nor some “vague border
between euthanasia and palliative sedation”.
Nonetheless, according to data collected by a sen-
tinel network of GPs during 2005-2006, among
non-sudden deaths at home in Belgium more
than one third of the sedated patients awoke
largely because of “insufficient medication”.37 In
MICCINESI CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE
the Italian context, where end-of-life discussions
with patients are����������������������������������
�������������������������������������
quite unsatisfactory and the pal-
liative care system is growing but not yet well-
established,38, 39 we would suggest that the use of
sedative drugs at the end of life in non-specialist
settings should be closely monitored.
To our knowledge, this is the only survey re-
ported in the international literature that asked
physicians simultaneously about practice and at-
titudes on end-of-life care, which allowed us to
study their association at the individual level.
Two different results stand out: first, agreement
with the acceptability of non-treatment decisions
is positively associated with having reported seda-
tion in the last assisted patient; second, attitudes
towards physician assisted death are not associ-
ated with sedation at all.
Allowing relatives to ������������������������
make �������������������
non����������������
-���������������
treatment �����
deci-
sions for incompetent patients, although difficult
(the sample was half split on this item), was more
acceptable for physicians who had reported using
sedation. This highlights how important non-
treatment decisions are in performing sedation,
even if both should be considered separately.1, 4
Our study shows that, on average, there was
no association between positive/negative atti-
tudes towards physician assisted death and deep
sedation. Rather, these appear to be independent
issues, at least in Italy. The possibility that deep
sedation ‑ particularly when associated with the
withholding/withdrawal of artificial nutrition/
hydration ‑ is a kind of stealth euthanasia has
been vigorously debated internationally.5, 40, 41
This is certainly a possibility in the Netherlands,
where the issue ��������������������������
has been extensively stud-
ied.18, 42, 43 R�������������������������������������
ecent debate�������������������������
�������������������������������
has focused on the ethi-
cal significance of maintaining consciousness be-
fore death,21 with some commentators suggesting
that deep sedation is a kind of social death 20 or
even functionally equivalent to euthanasia.46 One
way to clarify the issue is to define and specify
the necessary indications for commencing pallia-
tive sedation.1, 4, 1, 27-29 One might also argue that
deep sedation i������������������������������������
s�����������������������������������
a����������������������������������
proportional
��������������������������������
therapy of
�����������
last re-
sort that does not entirely abolish consciousness;
according to this perspective, whether sedation
is irreversible is a contingent matter, dependent
on the physician’s response to refractory symp-
toms.47
976 MINERVA ANESTESIOLOGICA September 2015
Knowing how sedatives were used at the end
of life is an important aspect for understand-
ing clinical practice. In our study, 81% of se-
dated patients received a proportionate dosage
of opioids during the last day of life. The EAPC
framework 1, 4 states that palliative sedation
should not include an intention to hasten death.
This condition seems to have been satisfied in
the majority of cases of sedation in our study. It
is worth noting, however, that there is an impor-
tant minority of physicians who reported using a
dosage of opioids that was higher than necessary
to treat the symptoms. For 17 sedated patients,
this occurred via an abrupt increase of opioids
during the last 24 hours of life. Self-evaluation
by physicians may have led to potential bias.46
Nevertheless, whether this finding is present in
other countries should be examined in much
more detail. Any ambiguity on this sensitive is-
sue could undermine the common acceptance
of palliative sedation ‑ a medical procedure that
has become “best practice” in well-defined situa-
tions of last resort.
Furthermore, 114 patients (45.6%) were
“deeply sedated” with opioids only. This pro-
portion goes down to 7.5% among physicians
who work in Anesthesia/ICU settings (data not
shown in table). In light of the high risk of cog-
nitive impairment and delirium induced by the
opioids ‑ above all, morphine ‑ 47 there is room
for considering palliative sedation inadequately
performed in those cases.
It has been recently shown that light seda-
tion, in comparison with deep sedation, reduces
ICU stay and duration of ventilation without
negatively affecting subsequent mental health.48
Similarly, the influential document of the Ital-
ian Society for Palliative Care (SICP), published
in 2007, focused on palliative sedation in the
last days of life; it highlighted the importance
of providing a proportionate intensity of seda-
tion to control refractory symptoms, rather
than aiming at a predefined depth of sedation.49
These aspects were not considered in our epi-
demiological study, which focused on assessing
the intention of physicians in using opioids in
deeply sedated patients rather than the propor-
tionality of sedation itself; this is in accordance
with the international end-of-life medical deci-
CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE MICCINESI
sions study from which the questionnaire we
used was derived.
In order to name the practice described in our
study, both in the questionnaire and in this pa-
per, we decided not to use the term “palliative/
terminal”. This approach was recommended by
the SICP document, mentioned above. Indeed,
in order to avoid reporting bias when dealing
with a population-based study, we still consider
it advisable to ask for the depth of sedation,
whether it was present at the moment of death,
and which drugs were used separately, so as to
avoid using any potentially value-laden label
of the effective medical behavior. As far as this
paper is concerned, using the term “palliative/
terminal” to indicate palliative sedation in the
last days of life would have risked reigniting a
debate from twenty years ago;50 in this earlier
debate, the adjective “terminal” was interpreted
by clinicians and the public as “hastening the
end of life” instead of “occurring in the last days
of life”.
The most important limitation of this study
is the low response rate, which suggests cau-
tion in generalizing the data on the prevalence
of sedation.�����������������������������������
However, we believe that the asso-
ciation between practice and attitudes is much
less exposed to selection bias. Low response rates
to questionnaires are �������������������������
common�������������������
in Italy, particu-
larly regarding sensitive topics. For example, the
previous study on end-of-life practices had an
Italian response rate of only 44% after sending
two reminders ‑ approximately 20 percentage
points less than in the other participating Eu-
ropean countries.13 The parallel study on atti-
tudes towards end-of-life medical decisions had
a response rate in Italy of 39%.15 The conflict
between assuring anonymity and the foreseen
need to raise the response rate was resolved for
the ITAELD study by avoiding reminders alto-
gether, since alternatives (e.g., providing a clear-
ing house for the questionnaires) were too ex-
pensive.
Another limitation to be taken into consider-
ation is that the data were collected seven years
ago, just before important modifications were
made to end-of-life care in Italy. The National
Law 38/2010 51 began a process of professional-
ization and extension of palliative care in differ-
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 977
ent settings throughout Italy. This process is still
growing and also likely to modify our under-
standing of the use of sedative drugs at the end
of life. New population-based studies are needed
in Italy on the topic.
Conclusions
Our study confirms the high prevalence of
deep sedation at the end of life and the conver-
gence of different practices under the same label.
Careful descriptive language is needed.
Although the ethical discussion should be
reserved for professional organizations, policy-
makers, ethicists, and the public, our data indi-
cate a need for more training in palliative care
among Italian physicians.
Key messages
—— Among the population that died in
Italy in 2001, 8.5% were deeply sedated.
Among expected deaths, this proportion
was 12.0%. Six years later, among expected
deaths, the proportion of people who died
while being deeply sedated increased to
18.2%. This indicates that sedation per-
formed in the last days of life is a common
part of end-of-life care.
—— Half of all deaths preceded by seda-
tion were assisted by GPs. This underscores
the importance of providing specific training
to GPs on the use of sedative drugs at the
end of life.
—— Palliative sedation today is character-
ized by specific clinical indications and the
proportionate use of sedatives to treat refrac-
tory symptoms, as established in numerous
guidelines. However, as revealed in Italy in
2007, almost half of those who died while
being deeply sedated had received “opioids
only,” and there was an intention to hasten
death in a minority of cases. This highlights
the need for ethical debate among profes-
sionals and the public, as well as new epide-
miological studies to evaluate any change in
practice after the promulgation of Italian law
38/2010
MICCINESI CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE
References
  1. Cherny NI, Radbruch L, Board of the European Associa-
tion for Palliative Care. European Association for Palliative
Care (EAPC) recommended framework for the use of seda-
tion in palliative care. Palliat Med 2009;23:581-93.
 2. Schildmann J, Schildmann E. Clinical and ethical chal-
lenges of palliative sedation therapy. The need for clear
guidance and professional competencies. Int J Clin Pract
2013;67:1086-8.
  3. Jones DA. Death by equivocation: a manifold definition of
terminal sedation. In: Sterckx S, Raus K, Mortier F, editors.
Continuous Sedation at the End of Life. Ethical, Clinical
and Legal Perspectives. Cambridge: Cambridge University
Press; 2013. p. 47-64.
  4. Materstvedt LJ, Clark D, Ellershaw J, Førde R, Gravgaard
AM, Müller-Busch HC et al. Euthanasia and physician-
assisted suicide: a view from an EAPC Ethics Task Force.
Palliat Med 2003;17:97-101.
 5. Ten Have H, Welie JV. Palliative sedation versus eutha-
nasia: an ethical assessment. J Pain Symptom Manage
2014;47:123-36.
  6. Maltoni M, Miccinesi G, Morino P, Scarpi E, Bulli F, Mar-
tini F et al. Prospective observational Italian study on pallia-
tive sedation in two hospice settings: differences in casemix-
es and clinical care. Support Care Cancer 2012;20:2829-36.
 7. Caraceni A, Zecca E, Martini C, Gorni G, Campa T,
Brunelli C et al. Palliative sedation at the end of life at a
tertiary cancer center. Support Care Cancer 2012;20:1299-
307.
  8. Donker GA, Slotman FG, Spreeuwenberg P, Francke AL.
Palliative sedation in Dutch general practice from 2005 to
2011: a dynamic cohort study of trends and reasons. Br J
Gen Pract 2013;63:e669-675.
  9. Mercadante S, Porzio G, Valle A, Aielli F, Casuccio A. Pal-
liative sedation in patients with advanced cancer followed
at home: a prospective study. J Pain Symptom Manage
2014;47:860-6.
10. Jaspers B, Nauck F, Lindena G, Elsner F, Ostgathe C, Rad-
bruch L. Palliative sedation in Germany: how much do we
know? A prospective survey. J Palliat Med 2012;15:672-
80.
11. Van der Maas PJ, Van der Wal G, Haverkate I, de Graaff
CL, Kester JG, Onwuteaka-Philipsen BD et al. Euthana-
sia, physician-assisted suicide, and other medical practices
involving the end of life in the Netherlands, 1990-1995. N
Engl J Med 1996;335:1699-705.
12. Weissman DE. Decision-making at a time of crisis near the
end of life. JAMA 2004;292:1738-43.
13. Van der Heide A, Deliens L, Faisst K, Nilstun T, Norup M,
Paci E et al. End-of-life decision-making in six European
countries: descriptive study. Lancet 2003;362:3453-50.
14. Bosshard G, Nilstun T, Bilsen J, Norup M, Miccinesi G,
van Delden JJ et al. European End-of-Life Consortium.
Forgoing treatment at the end of life in 6 European coun-
tries. Arch Intern Med 2005;165:401-7.
15. Miccinesi G, Fischer S, Paci E, Onwuteaka-Philipsen BD,
Cartwright C, van der Heide A et al. Physicians’ attitudes
towards end-of-life decisions: a comparison between seven
countries. Soc Sci Med 2005;60:1961-74.
16. Miccinesi G, Puliti D, Paci E. End-of-life care and end-of-
life medical decisions: the ITAELD study. Epidemiol Prev
2011;35:178-87.
17. Seale, C. Continuous deep sedation in medical practice: a
descriptive study. J Pain Symptom Manage 2010;39:44-53.
18. Van der Heide A, Onwuteaka-Philipsen BD, Rurup ML,
Buiting HM, van Delden JJ, Hanssen-de Wolf JE et al.
End-of-life practices in the Netherlands under the Eutha-
nasia Act. N Engl J Med 2007;356:1957-65.
19. StataCorp. (Stata Statistical Software Release 12). College
Station, TX: StataCorp LP; 2011.
978 MINERVA ANESTESIOLOGICA September 2015
20. Barr J, Fraser GL, Puntillo K, Ely EW, Gélinas C, Dasta JF
et al. American College of Critical Care Medicine. Clinical
practice guidelines for the management of pain, agitation,
and delirium in adult patients in the intensive care unit.
Crit Care Med 2013;41:263-306.
21. Italian Society of Anaesthesia Analgesia Resuscitation
and Intensive Care-Bioethical Board. Le cure di fine vita
e l’anestesista-rianimatore: raccomandazioni SIAARTI
per l’approccio al malato morente. Minerva Anestesiol
2006;72:927-63.
22. Morita T, Tsuneto S, Shima Y. Definition of sedation for
symptom relief: a systematic literature review and a pro-
posal of operational criteria. J Pain Symptom Manage
2002;24:447-53.
23. Cherney NI, Portenoy RK. Sedation in the management
of refractory symptoms: guidelines for evaluation and treat-
ment. J Palliat Care 1994;10:31-8.
24. Wein S. Sedation in the imminently dying patient. Oncol-
ogy 2000;14:585-92.
25. Van Delden JJ. The ethical evalutation of continuous seda-
tion at the end of the life. In: Sterckx S, Raus K, Mortier
F, editors. Continuous Sedation at the End of Life. Ethical,
Clinical and Legal Perspectives. Cambridge: Cambridge
University Press; 2013. p. 218-27.
26. Sterckx S, Raus K, Mortier F, editors. Continuous sedation
at the end of life. Ethical, Clinical and Legal Perspectives.
Cambridge: Cambridge University Press; 2013.
27. Royal Dutch Medical Association. Guideline for palliative
sedation [Internet]. Available from: http://knmg.artsennet.
nl/Publicaties/KNMGpublicatie/66978/Guideline-for-pal-
liative-sedation-2009.htm [cited 2014, Jul 16].
28. Report of the Council on Ethical and Judicial Affairs. Se-
dation to Unconsciousness in End-of-Life Care [Internet].
Available from: www.ama-assn.org/resources/doc/code-
medical-ethics/2201a.pdf [cited 2014, Mar 18].
29. Dean MM, Cellarius V, Henry B, Oneschuk D. Librach
Canadian Society Of Palliative Care Physicians Taskforce
SL. Framework for continuous palliative sedation therapy
in Canada. J Palliat Med 2012;15:870-9.
30. Miccinesi G, Rietjens JA, Deliens L, Paci E, Bosshard G,
Nilstun T et al. Continuous deep sedation:
�������������������������
physicians’ ex-
periences in six European countries. J Pain Symptom Man-
age 2006;31:122-9.
31. Van den Block L, Deschepper R, Bilsen J, Bossuyt N, Van
Casteren V, Deliens L. Euthanasia and other end-of-life
decisions: a mortality follow-back study in Belgium. BMC
Public Health 2009;9:79.
32. Van den Block L, Onwuteaka-Philipsen B, Meeussen K,
Donker G, Giusti F, Miccinesi G et al. Nationwide continu-
ous monitoring of end-of-life care via representative net-
works of general practitioners in Europe. BMC Fam Pract
2013;14:73.
33. Anquinet L, Rietjens JA, Seale C, Seymour J, Deliens L,
van der Heide A. The practice of continuous deep sedation
until death in Flanders (Belgium), the Netherlands, and
the U.K.: a comparative study. J Pain Symtpom Manage
2012;44:33-43.
34. Van Deijck RH, Hasselaar JG, Verhagen SC, Vissers KC,
Koopmans RT. Determinants of the administration of con-
tinuous palliative sedation a systematic review. J Palliat Med
2013;16:1624-32.
35. Giannini A. Ethics and end-of-life care in the new training
curriculum for ICU physicians in Italy. Intensive Care Med
2007;33:1043-5.
36. Verkerk M, van Wijlick E, Legemaate J, de Graeff A. A na-
tional guideline for palliative sedation in the Netherlands. J
Pain Symptom Manage 2007;34:666-70.
37. Anquinet L, Rietjens JA, Van den Block L, Bossuyt N, De-
liens L. General practitioners’ report of continuous deep se-
dation until death for patients dying at home: a descriptive
study from Belgium. Eur J Gen Pract 2011;17:5-13.
CAREFUL MONITORING OF THE USE OF SEDATIVE DRUGS AT THE END OF LIFE MICCINESI

38. Beccaro M, Costantini M, Giorgi Rossi P, Miccinesi G,
Grimaldi M, Bruzzi P. Actual and preferred place of death
of cancer patients. Results from the Italian survey of the
dying of cancer (ISDOC). J Epidemiol Community Health
2006;60:412-6.
39. Centeno C, Lynch T, Donea O, Rocafort J, Clark D. EAPC
Atlas of Palliative Care in Europe 2013. Full Edition. Mi-
lan: EAPC Press; 2013.
40. Billings JA, Block SD. Slow euthanasia. J Palliat Care
1996;12:21-30.
41. Tännsjö T, editor. Terminal sedation: euthanasia in dis-
guise? Dordrecht: Kluwer; 2004.
42. Rietjens JA, Van Delden JJ, van der Heide A, Vrakking AM,
Onwuteaka-Philipsen BD, van der Maas PJ et al. Terminal
sedation and euthanasia: a comparison of clinical practices.
Arch Intern Med 2006;166:749-53.
43. Onwuteaka-Philipsen BD, Brinkman-Stoppelenburg A,
Penning C, de Jong-Krul GJ, van Delden JJ, van der Heide
A. Trends in end-of-life practices before and after the en-
actment of the euthanasia law in the Netherlands from
1990 to 2010: a repeated cross-sectional survey. Lancet
2012;380:908-15.
44. LiPuma S. Continuous sedation until death as physician-
assisted suicide/euthanasia: a conceptual analysis. J Med
Philos 2013;38:190-204.
45. Raho JA, Miccinesi G. Contesting the equivalency of con-
tinuous sedation until death and physician-assisted suicide/
euthanasia: a commentary on LiPuma. J Med Philos 2015
(in press).
46. Bilsen J, Norup M, Deliens L, Miccinesi G, van der Wal
G, Löfmark R et al. Drugs used to alleviate symptoms
with life shortening as a possible side effect: end-of-life
care in six European countries. J Pain Symptom Manage
2006;31:111-21.
47. Demeure MJ, Fain MJ. The elderly surgical patient and
post-operative delirium. J Am Coll Surg 2006; 203:752-7.
48. Treggiari MM, Romand JA, Yanez D, Deem SA, Goldberg
J, Hudson L et al. Randomized trial of light versus deep se-
dation on mental health after critical illness. Crit Care Med
2009;37:2527-34.
49. SICP- gruppo di studio su etica e cultura al termine della
vita. Raccomandazioni della SICP sulla sedazione termi-
nale/sedazione palliativa [Internet]. Available from: http://
www.sicp.it/web/procedure/protocollo.cfm?List=WsIdEve
nto,WsPageNameCaller,WsIdRisposta,WsRelease&c1=D
OCSICP&c2=%2Fweb%2Feventi%2FNEPHROMEET
%2Findex%2Ecfm&c3=7&c4=1 [cited 2014, Nov 7].
50. Mount B. Morphine drips, terminal sedation, and slow eu-
thanasia: definitions and facts, not anecdotes. J Palliat Care
1996;12:31-7.
51. Parlamento Italiano. Legge 15 Marzo 2010, n. 38. Dispo-
sizioni per garantire l’accesso alle cure palliative e alla tera-
pia del dolore [Internet]. Available
��������������������������������
from http://www.parla-
mento.it/parlam/leggi/10038l.htm [cited 2014, Jul 16].

The only oral presentation of the general results of the ITAELD study took place in Udine, July 2007.
The only publication on a peer-reviewed journal of the general results of the ITAELD study was reported among the references (n° 16).
The results on deep sedation have not been extensively presented yet.
Funding.—This research received reimbursement for mailing and optical reading of the questionnaires from Federazione Nazionale Ordini
dei Medici Chirurghi e Odontoiatri-FNOMCeO.
Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on July 17, 2014. - Accepted for publication on December 2, 2014. - Epub ahead of print on December 5, 2014.
Corresponding author: G. Miccinesi, SC Epidemiologia Clinica e Descrittiva, ISPO, Via Oblate 2, 50141, Florence, Italy.
E-mail: g.miccinesi@ispo.toscana.it

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 979



O R I G I N A L A RT I C L E

Changed epidemiology of ICU acquired bloodstream

infections over 12 years in an Italian teaching hospital
G. B. ORSI 1, S. GIULIANO 1, C. FRANCHI 2, V. CIORBA 3, C. PROTANO 1, A. GIORDANO 1,
M. ROCCO 4, M. VENDITTI 1

1Department of Public Health and Infectious Diseases, “Sapienza” University of Rome, Rome, Italy; 2Department of
Clinical Medicine, “Sapienza” University of Rome, Rome, Italy; 3Department of Biomedical, Biotechnological and
Translational Sciences (SBiBiT), University of Parma, Parma, Italy; 4Anesthesiology and Intensive Care, “Sapienza”
University of Rome, Rome, Italy

ABSTRACT
Background. We compared the etiology of 203 ICU-acquired laboratory confirmed bloodstream infections (LC-
BSI) prospectively collected between January 2000-December 2007 (first period) with 83 LC-BSI recorded between
January 2010-December 2012 (second period), after the diffusion in 2008 of K. pneumoniae expressing carbapenem-
resistance due to K. pneumoniae carbapenemases production (KPC-CR-Kp).
Methods. In the ���������������������������������������������������������������������������������������������
general ICU of teaching hospital “Umberto I” in Rome, all ICU-acquired LC-BSI episodes occur-
ring in patients admitted to ICU≥48h were included. Baseline characteristics, clinical features, antimicrobial resist-
ance and outcome were recorded. All isolated strains multidrug resistance (MDR) were evaluated according to the
European Centre for Disease Control (ECDC) guidelines.
Results. Overall the study included 329 isolates, 214 in 2000-2007 and 115 in 2010-2012. In the second period
we registered a Gram-positive reduction (55.1% vs. 26.9%; P<0.01) and Gram-negative increase (40.2% vs. 69.6%;
P<0.01). In 2000-2007 staphylococci were responsible for 45.8% LC-BSI’s, whereas 18.3% during 2010-2012.
Enterobacteriaceae increased dramatically (15.4% vs. 39.2%; P<0.01), especially Klebsiella spp. (5.6% vs. 31.3%;
P<0.01). LC-BSI associated mortality decreased among Gram-positive (56.8% vs. 51.6%), but increased in Gram-
negative (41.9% vs. 60.0%; P<0.03), especially in Enterobacteriaceae (RR 2.13; 95% CI 1.21 – 3.73; P<0.01). MDR
increased remarkably among Enterobacetriaceae (51.5% vs. 73.3%). The study highlighted the associated mortality
for Enterobacteriaceae when comparing MDR to non-MDR microorganisms.
Conclusion. ICU-acquired LC-BSI etiology shifted from Gram-positive to Gram-negative during the study period
in our ICU. Also associated mortality decreased among the former, whereas it increased in the latter. Last MDR
increased enormously among Enterobacteriaceae with the diffusion of KPC (75% of strains), adding significantly to
associated mortality (RR 2.17; 1.16-4.05; P<0.01). (Minerva Anestesiol 2015;81:980-8)
Key words: Infection, blood - Etiology - Epidemiology - Intensive care units.

I CU-acquired laboratory confirmed blood-

stream infection (LC-BSI) is an important
cause of morbidity and mortality in intensive
care units (ICU), where it affects more than 10%
of patients, determining high associated mortal-
ity >20% 1-3 and adding costs.4
Appropriate empiric therapy is considered to
Comment on p. 940.
be the most important factor for patients out-
come, and is associated with reduced mortality
and length of stay.3, 5, 6 For this reason it is im-
portant to know the pathogens causing infec-
tion and their antimicrobial resistance pattern to
guide appropriate antimicrobial treatment.
Unfortunately the epidemiology and antimi-
crobial profile of microorganisms responsible
for bloodstream infection varies between insti-

980 MINERVA ANESTESIOLOGICA September 2015

ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI
tutions, and among ICUs within the hospitals.
Also, as the rates of antimicrobial resistance in
pathogens principally among Gram-negatives
are increasing, every center should be familiar
with its local trends in order to target a more
appropriate empirical therapy.
In our hospital during the years 2008 and
2009 we documented the appearance of a first
Klebsiella pneumoniae clone (ST37) expressing
ertapenem resistance by modification of the
outer membrane permeability (Porin-ER-Kp).7
Subsequently in 2010, we observed the disap-
pearance of Porin-ER Kp strains and the appear-
ance and rapid spread of a new clone of Klebsiella
pneumoniae expressing carbapenem resistance
due to K. pneumoniae carbapenemases produc-
tion (KPC-CR-Kp), especially in high risk areas
such as ICUs.8 As the two carbapenem resistant
K. pneumoniae clones were different, we decided
to compare the earlier period before the diffusion
of carbapenem resistance in Enterobacteriaceae
(2000-2007), to our current epidemiological
situation (2010-2012), excluding the intermedi-
ate period 2008-2009.
Therefore to
���������������������������������
understand the changing etiol-
ogy and antimicrobial profile of microorganisms
responsible for ICU-acquired LC-BSI in our
ICU, we carried out an epidemiological study
in order to compare the ICU-acquired LC-BSI
etiology of the two periods 2000-2007 (precar-
bapenemases period) versus 2010-2012 (carbap-
enemases period).
Materials and methods
Setting
We conducted a prospective observational
study at the 13 bed general ICU of the 1.300
bed University hospital “Policlinico Umberto I”
of Rome. From January 2010 to December 2012
all episodes of ICU-acquired LC-BSI occurring
in patients admitted to the ward ≥48h were in-
cluded. Information on baseline characteristics,
clinical features, antimicrobial resistance and
outcome were recorded. These episodes (2010-
2012) were compared with those prospectively
collected from January 2000 to December 2007
at the same institution (2000-2007). Data re-
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 981
garding the first period has been previously de-
scribed.9
Definitions
For the purpose of the study, only ICU-ac-
quired LC-BSI were taken into account, diag-
nosed in patients at least 48 hours after ICU
admission. ICU-acquired LC-BSI was defined
as the isolation of one or more microorganisms
from a blood culture in a patient with two or
more of the following: temperature >38° C or
<36° C, heart rate (HR) >90 beats/min, respi-
ratory rate >20 breaths/min, WBC>12,000/
mm3 or >10% immature neutrophils. Coagulase
negative staphylococcus LC-BSI was defined as
≥2 blood cultures demonstrating the same phe-
notype on separate occasions within a 48 hours
period.10
When blood cultures were collected, dupli-
cate isolates were excluded from the analysis.
Source of ICU-acquired LC-BSI was deter-
mined on the basis of the isolation of the mi-
croorganisms from the presumed portal of en-
try and clinical evaluation. When no link with
a primary source could be found, LC-BSI was
considered primary in origin.
Data collection
During the 12 years survey an infection con-
trol team (ICT), composed by one physician
specialized in intensive care, two in infectious
diseases and one epidemiologist, actively partici-
pated to the surveillance.
Data were collected prospectively by two phy-
sicians especially trained, using a specific data-
base oriented software (Epi-info Version 2011,
CDC). The following information was recorded:
demographic characteristics (i.e. sex, age, etc.),
date of admission and discharge, patient origin
(i.e. emergency, operating rooms wards, other
ICU), admission diagnosis (principal diagnosis
leading to ICU admission), severity score (SAPS
II), underlying diseases presence (diabetes mel-
litus, chronic renal failure, cirrhosis, chronic
obstructive pulmonary disease) and final ICU
outcome. Invasive procedures are associated to
BSI’s in ICU patients, therefore surveillance also
ORSI ICU ACQUIRED BLOODSTREAM INFECTIONS
included central venous catheter (CVC), me-
chanical ventilation and urinary tract catheter
exposure and duration.11-14
All ICU-acquired LC-BSI isolated microor-
ganisms and their antibiotic susceptibility were
screened and recorded.
Microbiological methods
All patient samples were taken for culture ac-
cording to the general principles of specimen
collection and transport.15 The species iden-
tification and the antimicrobial susceptibility
testing were performed on the isolated strains
by using the VITEK system and performed a
phenotypic confirmation test of carbapenemase
production in Enterobacteriaceae isolates having
a minimum inhibitory concentration >0.5 mcg/
mL for meropenem and/or imipenem. Carbap-
enemase production was confirmed by first with
Hodge-Test modified and after with disc diffu-
sion synergy test including meropenem and two
carbapenemase inhibiting compounds (dipico-
linic acid and boronic acid).16 From 2000 to
2011 antimicrobial susceptibility was referred to
the most recent CLSI breakpoints. Since 2012
Breakpoints were interpreted in accordance with
the European Committee on Antimicrobial Sus-
ceptibility Testing (EUCAST) guidelines.17, 18
Among microorganisms the multi-drug re-
sistance (MDR) was defined according to the
ECDC guidelines.18
We defined early mortality as death occurring
≤7 days after ICU-acquired LC-BSI.20, 21
Data Analysis
Statistical analyses were carried out using
SPSS ����������������������������������������
software (version 14.0 for Windows, Chi-
cago, IL, USA). The χ2 test was used to exam-
ine differences between groups. Statistical sig-
nificance was defined as a P value of less then
0.05. When the observed frequencies were >5
χ2 with Yates correction was used to compare
the proportions observed in the two groups.
When observed frequencies were less or equal to
5 Fisher’s exact test was used. The normality of
quantitative data (age, SAPS II score, length of
stay) was assessed by non-parametric technique
982 MINERVA ANESTESIOLOGICA September 2015
(Kolmogorov-Smirnov test). When data result-
ed normally distributed, t-test for independent
samples was used, otherwise Mann-Whitney test
was applied. The univariate relationship between
infection and death was tested using relative risk
and its 95% confidence interval (CI95).
Results
Between 2000-2007 and 2010-2012 respec-
tively 1741 and 1165 patients, with length of
stay >48 hours, were surveyed and evaluated for
ICU-acquired LC-BSI in the ICU. During the
first period (2000-2007) 167 (9.6%) patients
developed 203 (11.7%) ICU-acquired LC-BSI,
whereas in the second period (2010-2012) 83
(7.1%) patients developed 101 (8.7%) ICU-
acquired LC-BSI.
Overall the study included 329 isolates. In
total, 214 microorganisms isolated from 203
LC-BSI in 167 patients during the first period
(2000-2007) were compared with 115 micro-
organisms isolated from 98 LC-BSI in 83 pa-
tients during the second period (2010-2012).
As reported in Table I, the two patient clusters
showed similar demographic and clinical char-
acteristics, also for the SAPS II score which was
46.5±15.7 vs. 43.0±12.0; P=0.087.
Table II shows the causative microorgan-
isms of all ICU-acquired LC-BSI compared by
study periods. Between the first period (2000-
2007) before KPC appearance and the second
(2010-2012) with the diffusion of KPC’s in
the enterobacteriaceae we registered significant
Gram-positive reduction (55.1% vs. 26.9%;
P<0.01), Gram-negative increase (40.2% vs.
69.6%; P<0.01) and a minimum variation of
fungi (4.7% vs. 3.5%). In particular, during
the first period (2000-2007) staphylococci were
responsible for 45.8% of all LC-BSI’s, whereas
only 18.3% during 2010-2012. On the contrary
enterobacteriaceae increased dramatically (15.4%
vs. 39.2%; P<0.01), especially Klebsiella spp.
(5.6% vs. 31.3%; P<0.01) and also Acinetobacter
baumannii (7.5% vs. 20.0%) became more com-
mon.
We also considered microorganisms according
to ICU-acquired LC-BSI source, and found that
primary and secondary strains were respectively
ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI

Table I.—Characteristic of patients with ICU-acquired laboratory confirmed bloodstream infection in the two periods.
Characteristics 2000-2007 Period (167 patients) 2010-2012 Period (83 patients) P
Age (SD) 52.3±19.5 56.8±17.3 0.144
Male 108 (64.8%) 53 (63.8%) 1.004
Female 59 (35.2%) 30 (36.2%) 1.004
SAPS II (SD) 46.5±15.7 43.0±12.0 0.094
Length of stay (days) 36.2±32.6 34.3±26.7 0.734
ICU Crude mortality 606 (34.8%) 383 (32.9%) 0.304
Admission diagnosis
Medical 38 (22.7%) 18 (21.7%) 0.984
Surgical 79 (47.3%) 33 (39.8%) 0.324
Traumatic 50 (29.9%) 32 (38.5%) 0.224
Underlying conditions
Neurologic disorders 22 (13.2%) 4 (4.8%) 0.048
Cardiovascular disorder 41 (24.5%) 11 (13.2%) 0.064
Hypertension 24 (14.4%) 25 (30.1%) 0.005
COPD 33 (19.8%) 11 (13.2%) 0.274
Renal disorders 7 (4.2%) 6 (7.2%) 0.474
Liver cirrhosis 5 (3.0%) 1 (1.2%) *0.67*4
Diabetes mellitus 16 (9.6%) 11 (13.2%) 0.514
Invasive procedures
Central venous catheter 159 (95.2%) 83 (100.0%) 0.104
Mechanical ventilation 164 (98.2%) 83 (100.0%) 0.544
Urinary catheter 167 (100.0%) 83 (100.0%) -
P values are obtained using the χ2 corrected test except when indicated by (*) which are obtained from the Fisher exact test. COPD: Chronic ob-
structive pulmonary disease.

Table II.—Distribution of microorganisms responsible for ICU-acquired laboratory confirmed bloodstream infection.
2000-2007 Period 2010-2012 Period Total
Microorganism P
Isolates % Isolates % Isolates %
MRSA 32 (14.9%) 2 (1.7%) 34 (10.3%) <0.0001*
MSSA 5 (2.3%) 2 (1.7%) 7 (2.1%) 1.00*
CNS-MR 56 (26.3%) 17 (14.9%) 73 (22.2%) 0.02
CNS-MS 5 (2.3%) - - 5 (1.5%) -
Enterococcus spp. 2 (0.9%) 3 (2.6%) 5 (1.5%) 0.35*
E. faecalis 13 (6.1%) 3 (2.6%) 16 (4.9%) 0.19*
E. faecium 5 (2.3%) 4 (3.5%) 9 (2.7%) 0.72*
Total Gram+ 118 (55.1%) 31 (26.9%) 149 (45.3%) <0.0001
A. baumannii 16 (7.5%) 23 (20.0%) 39 (11.8%) <0.001
Aeromonas spp. 1 (0.5%) - 1 (0.3%) -
Citrobacter freundi † - 1 (0.9%) 1 (0.3%) -
Enterobacter spp. † 5 (2.3%) 3 (2.6%) 8 (2.4%) 1.00*
E. coli † 5 (2.3%) 1 (0.9%) 6 (1.8%) 0.70*
Klebsiella spp. † 12 (5.6%) 36 (31.3%) 48 (14.6%) <0.0001
(KPC-CR-Kp) # 31# (27.0%)# 31# (9.4%)# -
Morganella spp. † 1 (0.5%) 1 (0.9%) 2 (0.6%) 1.00*
Proteus spp. † 2 (0.9%) 3 (2.6%) 5 (1.5%) 0.35*
Providencia spp. † 2 (0.9%) - - 2 (0.6%) -
P. aeruginosa 29 (13.6%) 12 (10.4%) 41 (12.5%) 0.41
S. maltophilia 7 (3.3%) - - 7 (2.1%) -
S. marcescens † 6 (2.8%) - - 6 (1.8%) -
All Enterobacteriaceae 33 (15.4%) 45 (39.1%) 78 (23.7%) <0.0001
Total Gram- 86 (40.2%) 80 (69.6%) 166 (50.5%) <0.0001
Candida spp. 10 (4.7%) 4 (3.5%) 14 (4.2%) 0.78*
TOTAL 2144 100,0% 115 100,0% 329 100,0%
MRSA: Methicillin resistant S. aureus; MSSA: Methicillin susceptible S. aureus; CNS-MR: methicillin resistant coagulase negative staphylococci;
CNS-MS: Methicillin susceptible coagulase negative staphylococci; † Enterobacteriaceae; #KPC-CR-Kp strains are a subgroup of the overall Kleb-
siella.
P values are obtained using the χ2 corrected test except when indicated by (*) which are obtained from the Fisher exact test.

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 983

ORSI ICU ACQUIRED BLOODSTREAM INFECTIONS

47.1% and 52.9%. The latter 31.3% (CVC),
14.3% (respiratory tract), 4.6% (surgical site)
and 2.7% (UTI).
Results showed that onset time (days) be-
tween ICU admission and ICU-acquired LC-
BSI was higher (P<0.01) among Gram-nega-
tive (21.1±18.0) compared to Gram-positive
(15.5±16.0), whereas for fungi were 18.8±22.7.
The distribution of microorganisms associated
mortality rate in the two periods is illustrated
in Table III. Overall we observed a higher LC-
BSI associated mortality in the second period
(51.4% vs. 59.1%), with a modest reduction
among Gram-positive (56.8% vs. 51.6%) but
significant increase in Gram-negative (41.9%
vs. 60.0%; P<0.03). However during the second
period among Gram-negative LC-BSI associated
mortality was lower in non-fermentative micro-
organisms as A. baumanni (75.0% vs. 65.2%)
and P. aeruginosa (48.3% vs. 33.3%), whereas it
increased over two-fold in enterobacteriaceae (RR
2.13; 95% CI 1.21 – 3.73; P<0.01).
Antimicrobial susceptibility testing showed
that staphylococci expressed along time a high
resistance to methicillin (>80%). In enterobac-
teriaceae, from the first to the second period we
observed a significant increase (P<0.01) in anti-
microbial resistance to a wide range of antibiot-
ics: imipenem (6.7% vs. 68.3%), meropenem
(10.0% vs. 73.2%), amikacin (10.0% vs. 65.9%),
ceftazidime (37.9% vs. 82.5%), cefepime (26.7%
vs. 75.6%), cefotaxime (10.0% vs. 80.5%), cypro-
floxacin (34.5% vs. 82.5%), levofloxacin (33.3%
vs. 78.4%) and piperacillin/tazobactam (27.2%
vs. 74.3%). Only during the second period (2010-
2012) resistance to colistin resulted 28.9%.
All isolated strains multidrug resistance
(MDR) were evaluated according to the ECDC
guidelines.19 MDR rate and evolution along the
two periods is shown in Table IV.

Table III.—Distribution of microorganisms associated mortality rate (2010-2012 vs. 2000-2007).

Early Death (≤7 days) Final associated mortality
Microorganism P* RR (95%CI) P
2010-2012 Period 2000-2007 Period 2010-2012 Period 2000-2007 Period
S. aureus 1/4 (25.0%) 11/37 (29.7%) 1.00* 2/4 (50.0%) 26/37 (70.3%) 0.71; 0.26-1.94 0.75*
CNS 3/17 (17.6%) 10/61 (16.4%) 1.00* 9/17 (52.9%) 31/61 (50.8%) 1.04; 0.62-1.74 0.90
Enterococci 0/10 (0.0%) 2/20 (10.0%) 0.88* 5/10 (50.0%) 10/20 (50.0%) 1.00; 0.47-2.14 1.00*
All Gram+ 4/31 (12.9%) 23/118 (19.5%) 0.57 16/31 (51.6%) 67/118 (56.8%) 0.91; 0.62-1.32 0.10
A. baumannii 6/23 (26.1%) 6/16 (37.5%) 0.68 15/23 (65.2%) 12/16 (75.0%) 0.87; 0.58-1.31 0.77
Enterobacteriaceae 12/45 (26.7%) 6/33 (18.2%) 0.54 29/45 (64.4%) 10/33 (30.3%) 2.13; 1.21-3.73 0.0056
(KPC-CR-Kp) # - - - 24/31 (77.4%) - - -
P. aeruginosa 2/12 (16.7%) 7/29 (24.1%) 0.94* 4/12 (33.3%) 14/29 (48.3%) 0.69; 0.28-1.67 0.60*
All Gram- 20/80 (25.0%) 19/86 (22.1%) 0.80 48/80 (60.0%) 36/86 (41.9%) 1.43; 1.05-1.95 0.029
Candida spp. 1/4 (25.0%) 4/10 (40.0%) 1.00* 4/4 (100.0%) 7/10 (70.0%) 1.43; 0.95-2.14 0.66*

TOTAL 25/115 (21.7%) 46/214 (21.5%) 0.93 68/115 (59.1%) 110/214 (51.4%) 1.15; 0.94-1.40 0.22
#KPC-CR-Kp strains are a subgroup of the overall Klebsiella spp.
P values are obtained using the χ2 corrected test except when indicated by (*) which are obtained from the Fisher exact test.

Table IV.—Evolution of multidrug resistance in microorganisms responsible for ICU-acquired LC-BSI (2010-2012 vs. 2000-
2007)
Multidrug resistance
Microorganism RR (95%CI) P
2010-2012 Period 2000-2007 Period
S. aureus 2/4 (50.0%) 32/37 (86.5%) 0.58; 0.21-1.55 0.26*
Enterococci 3/10 (30.0%) 2/20 (10.0%) 3.00; 0.59-15.2 0.39*
A. baumannii 20/23 (87.0%) 11/16 (68.8%) 1.26; 0.88-1.82 0.33
Enterobacteriaceae 33/45 (73.3%) 17/33 (51.5%) 1.42; 0.98-2.07 0.081
P. aeruginosa 5/12 (41.7%) 16/29 (55.2%) 0.75; 0.36-1.59 0.66*
P values are obtained using the χ2 corrected test except when indicated by (*) which are obtained from the Fisher exact test.

984 MINERVA ANESTESIOLOGICA September 2015

ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI
Table V.—Associated mortality in MDR vs. non MDR microorganisms (data cumulative of periods 2000-2007 and 2010-
2012).
Associated mortality
Microorganism
MDR
S. aureus 24/34 (70.6%)
Enterococci 2/5 (40.0%)
A. baumannii 22/31 (71.0%)
Enterobacteriaceae 31/50 (62.0%)
(KPC-CR-Kp) # 24/31 (77.4%)
P. aeruginosa 16/21 (57.1%)
#KPC-CR-Kp strains are a subgroup of the overall Klebsiella spp.
P values are obtained using the χ2 corrected test except when indicated by (*) which are obtained from the Fisher exact test.
Among Gram-positive S. aeruginosa
������������������
bauman-
nii expressed always high methicillin resistance
(≥50%), although in the period 2010-2012
staphylococci isolation was much less com-
mon. On the contrary no vancomycin resistant
enterococci were isolated during study periods.
During the study period there was an impressive
increase of MDR in enterobacteriaceae, whereas
among non-fermentative bacilli MDR increased
slightly in A. baumannii, but diminished in P.
aeruginosa. Table V shows the associated mortal-
ity comparing MDR to non-MDR microorgan-
isms, highlighting the growing impact of En-
terobacteriaceae (RR 2.17; 1.16-4.05; P<0.01),
and particularly K. pneumoniae carbapenemases
producing strains (RR 2.63; 1.23-5.63; P<0.01).
We also considered strains antimicrobial
susceptibility according to the sources of LC-
BSI. Results showed that strains MDR rates
were respectively 70.2% (secondary to respira-
tory tract), 44.4% (UTI), 46.7% (surgical site),
40.6% (unknown) and 33.0% (CVC).
Discussion
The main purpose of surveillance on a specific
disease etiology such as LC-BSI is to detect shifts
in antimicrobial susceptibility of the involved
bacteria in order to help improving the choice
of empiric therapy. The strength of such studies
on LC-BSI relies on clear and standardized clini-
cal diagnostic criteria which make data reliable
and realistic, avoiding confounding colonizing
agents not directly related to clinical disease (as
it may occur when ventilator associated pneu-
monia are considered).
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 985
RR (95%CI) P*
Non MDR
4/7 (57.1%) 1.23; 0.63-2.43 0.78*
13/25 (52.0%) 0.77; 0.25-2.40 1.00*
5/8 (62.5%) 1.13; 0.63-2.03 0.94*
8/28 (28.6%) 2.17; 1.16-4.05 0.0094
5/17 (29.4%) 2.63; 1.23-5.63 0.0032*
6/20 (30.0%) 1.90; 0.89-4.09 0.0080
The decision whether to use χ2 with Yates
correction rather than Fisher’s exact test to
compare proportions was made by arbitrarily
using a cut off observed frequency of >5. Some
authorities recommend other cut-offs also
based on the expected frequency. In fact, we
have used both tests for all comparisons and we
found that these are not important differences
in the results.
In order to reduce bias we decided to compare
the earlier period before the diffusion of carbap-
enem resistance in Enterobacteriaceae (2000-
2007), to our current epidemiological situation
(2010-2012), excluding the intermediate pe-
riod 2008-2009, because as we explained in the
introduction the two carbapenem resistant K.
pneumoniae clones were different.
Demographically there were limited differ-
ences between the two period patient popula-
tions which appeared homogeneous by age,
gender and SAPS II score (Table I). Major
differences were represented by an increase, in
proportion, of trauma admissions and a reduc-
tion of patients with neurologic disorders as un-
derlying conditions in the second period. This
because a separate specialistic neurosurgical
ICU was enlarged and the opening of new beds
in it explains the shift in percentages between
the two periods. Especially in a general ICU,
the case-mix is a very important parameter to
evaluate and compare the clusters, therefore it
was encouraging to note that in 2000-2007 and
2010-2012 the principal diagnosis at admission
as the underlying diseases had remained mostly
unchanged.22
Main finding was a dramatic change in
ORSI ICU ACQUIRED BLOODSTREAM INFECTIONS
ICU-acquired LC-BSI etiology between the
two study periods, with a marked reduction of
Gram-positive microorganisms and a fierce in-
crease in isolation of Gram-negative (Table II).
Among Gram-positive we observed mainly a
decrease of staphylococci, particularly MRSA.
In the last decade Gram-negative pathogens
have emerged as dominant contributors for
bloodstream infection in various institutions of
developed countries. The study results appear
consistent with these trends and other reports
from USA,23 Brazil,24 Taiwan,25 and Europe 26
also have shown a decreasing trend in the inci-
dence of MRSA associated to BSI.
Overall the reasons for this are probably mul-
tifactorial and still partially unclear. Hand hy-
giene reinforcement 27 and CVC management
policy, which became more selective in our ICU
at the beginning of the survey 2000/2001,22
may have contributed. Also as a consequence
of the MDR microorganisms diffusion, since
2011 a senior infection disease physician (MV)
was directly involved in the routine ICU assis-
tance. As this is an observational study, the rea-
sons for MRSA decrease in our institution are
still partially unclear, and further analysis from
various ICU’s carrying out relationship between
infection control measures is needed.
Our observations support the notion that
S. aureus BSI is decreasing, however, although
this should be interpreted cautiously since the
MRSA disease burden remains high.28, 29
During the second period a significant in-
crease in incidence was noted for various of the
most common Gram-negative organisms asso-
ciated to LC-BSI, but the most remarkable in-
crease was noted for Enterobacteriaceae particu-
larly K. pneumoniae. This striking change of the
etiology with a predominance of Gram-negative
was also seen in the large European multicentre
EPIC II study although considering all bacterial
isolates from different body sites.14
The second study notable finding was related
to mortality. During the second period we ob-
served a reduction associated both to all Gram-
positive and Gram-negative non-fermentative
microorganisms (A. baumannii and P. aerugino-
sa), whereas we registered a fierce over two-fold
increase of mortality associated to Enterobacte-
986 MINERVA ANESTESIOLOGICA September 2015
riaceae. This trend was similar when consider-
ing separately only the early death (Table III).
The rise in resistance for all Enterobacteriace-
ae but particularly K. pneumoniae may had an
impact on the increased mortality rate. Gram-
negative LC-BSI has been associated with high
mortality and the risk is further increased if
appropriate antibiotic treatment is delayed
because of antimicrobial resistance.30 Several
case-control studies reported a worse outcome
associated with infections due to a resistant
strain.31-34
Although negative outcome was not a prima-
ry objective of our study and may be influenced
by several factors, undoubtedly the inversion
in mortality trend between Gram-positive and
Gram-negative appears influenced by MDR
among Enterobacteriaceae (Table IV).
As in the past many different definitions for
MDR microorganisms have been used in the
medical literature to characterize the different
patterns of resistance found in healthcare asso-
ciated bacteria, limiting the possibility to com-
pare the surveillance data, we decided to adopt
the ECDC MDR standard definitions.19
Actually the third finding in our study was an
increase of Gram-negative MDR in the second
period. This is consistent with those of recent
studies, which report an increase in antibiotic
resistance among Gram-negative in immuno-
competent and immunocompromised hosts.26
As resistance to multiple antibiotics increases
the chances for inappropriate empiric therapy,
which has been shown to be an independent
risk factor for adverse outcome among bactere-
mic patients,6 it was not surprising a strict re-
lationship between mortality and MDR P. aer-
uginosa and Enterobacteriaceae (Table V). In the
latter adverse outcome was even higher when
considering separately K. pneumoniae ��������
KPC pro-
ducing strains (77.4%).
In relation to the source of infection it was
not surprising that the highest antimicrobial
susceptibility rate was associated to LC-BSI sec-
ondary to respiratory tract infections.
We are aware of some limitations to our
study. First it was carried out in a single center,
secondly the data were collected for the purpos-
es of infection control surveillance and conse-
ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI
quently some clinical information may be lim-
ited, third mortality was reported as all-cause
mortality, and not as infection related death.
However, although the study did not focus on
antimicrobial treatment, as in the second pe-
riod (2010-2012) many strains were MDR we
can assume that also many patients could not
receive an early adequate therapy before micro-
biological results (i.e. colistin ± tigecycline for
KPC-CR-Kp) affecting the final outcome.
We are also aware that in a long period study
(12 years) the adoption of different breakpoints
might lead to some time bias effect on antimi-
crobial susceptibility, but surely not on the dif-
ferent bacterial species isolated from blood.
The study adopted a clear and standardized
clinical diagnostic criteria which avoided con-
founding colonizing agents not directly related
to clinical disease, excluding blood culture du-
plicates was important,�����������������������
as inclusion of dupli-
cates can result in overestimation of some mi-
croorganisms. Data was collected prospectively  
by an infection control team using the same
methodology along all the study period, and
in order to improve surveillance data compari-  
son, ECDC MDR standard definitions were
adopted. 19  
Conclusions
 
The present study provided some useful in-
sight into the LC-BSI pathogens epidemiology
at our institution. Following the appearance  
and spread of carbapenem resistance in Entero-
bacteriaceae, over the last years in our ICU the
LC-BSI etiology shifted from Gram-positive to
 
Gram-negative. It also associated mortality de-
creased among the former, whereas it increased
in the latter, especially among Enterobacte-
riaceae with the diffusion of KPC producing K.  
pneumoniae.
Our results underline the importance of
surveillance, particularly of MDR organisms
 
which add significantly to mortality. Since early
appropriate empirical treatment improves sur-
vival in patients affected by serious infections,
specific studies are needed to identify risk fac-  
tors for MDR organisms isolation in ICU pa-
tients.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 987
Key messages
—— Over the last decade in our institu-
tion microorganisms responsible for ICU-
acquired LC-BSI shifted from Gram-positive
to Gram-negative.
—— During the study period ��������������
also the asso-
ciated mortality decreased among the former,
whereas it increased in the latter; mainly as a
consequence for the diffusion of enterobacte-
riaceae expressing carbapenem-resistance due
to K. pneumoniae carbapenemases produc-
tion.
—— In the second period we observed an
increase of Gram-negative MDR and adverse
outcome was even higher when considering
separately K. pneumoniae KPC producing
strains.
References
1. Sligl W, Taylor G, Brindley PG. Five years of nosocomial
gram-negative bacteremia in a general intensive care unit:
epidemiology, antimicrobial susceptibility patterns, and
outcomes. Int J Infect Dis 2006;10:320-5.
2. Vincent JL, Sakr Y, Sprung CL, Ranieri VM, Reinhart K,
Gerlach H et al. Sepsis in european intensive care units: re-
sult of the SOAP study. Crit Care Med 2006;34: 344-53.
3. Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel
RP, Edmond MB. Nosocomial bloodstream infections in
US hospitals: analysis of 24,179 cases from a prospective na-
tionwide surveillance study. Clin Infect Dis 2004;39:309-
17.
4. Al-Rawajfah OM, Hewitt JB, Stetzer F, Cheema J. Length
of stay and charges associated with health care-acquired
bloodstream infections. Am J Infect Control 2012;40:227-
32.
5. Kang CI, Kin SH, Park WB, Lee KD, Kim HB, Kim EC
et al. Bloodstream infections caused by antibiotic-resistant
gram-negative bacilli: risk factors for mortality and impact
of inappropriate initial antimicrobial therapy on outcome.
Antimicrob Agents Chemother 2005;49:760-6.
6. Marchaim D, Zaidenstein R, Lazarovitch T, Karpuch Y, Ziv
T, Weinberger M. Epidemiology of bacteremia episodes in
a single center: increase in Gram-negative isolates, antibiot-
ics resistance and patients age. Eur J Microbiol Infect Dis
2008;27:1045-51.
7. Orsi GB, García-Fernández A, Giordano A, Venditti C,
Bencardino A, Gianfreda R et al. �����������������������
Risk factors and clini-
cal significance of endemic ertapenem resistant Klebsiella
pneumoniae isolates in hospitalized patients. J Hosp Infect
2011;78:54-8.
8. Orsi GB, Bencardino A, Vena A, Carattoli A, Venditti C,
Falcone M et al. Patient risk factors for outer membrane
permeability and KPC producing carbapenem-resistant
Klebsiella pneumoniae isolation: results of a double case-
control study. Infection 2013;41:61-7.
9. Orsi GB, Franchi C, Marrone R, Giordano A, Rocco M,
Venditti M. Laboratory confirmed bloodstream infection
etiology in an intensive care unit: eight years study. Ann Ig
2012;24:269-78.
ORSI ICU ACQUIRED BLOODSTREAM INFECTIONS
10. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM.
CDC definitions for nosocomial infections. In: Olmsted
R.N., editors. APIC Infection Control and Applied Epide-
miology: Principles and Practice. St Louis: Mosby; 1996. p.
A-1-A-20.
11. Maki DG, Kluger DM, Crnich CJ. The risk of bloodstream
infection in adults with different intravascular devices: a sys-
tematic review of 200 published prospective studies. Mayo
Clin Proc 2006;81:1159-71.
12. Mermel LA. Prevention of intravascular catheter-related in-
fections. Ann Intern Med 2000;132:391-402.
13. Polderman KH, Girbes AR. Central venous catheter use.
Part 2: infectious complications. Intensive Care Med
2002;28:18-28.
14. Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Mar-
tin CD et al. International study of the prevalence and
outcomes of infection in intensive care units. JAMA
2009;302:2323-9.
15. Wilson ML. General principles of specimen collection and
transport. Clin Infect Dis 1996;22:776-7.
16. Cohen Stuart J, Leverstein-Van Hall MA, Dutch working
Party on the detection of highly resistant microorganisms.
Guideline for phenotypic screening and confirmation of
carbapenemases in Enterobacteriaceae. Int J Antimicob
Agents 2010;36:205-10.
17. Clinical and Laboratory Standards Institute. Performance
standards for antimicrobial susceptibility testing. CLSI
document. Wayne, PA
18. EUCAST breakpoint tables for interpretation of MIC’s and
zone diameters. [Internet]. Available from: http://www.eu-
cast.org/clinical_breakpoints/ [cited on 2015, July 21].
19. Magiorakos AP, Srinivasan A, Carey RB, Carmeli Y, Fala-
gas ME, Giske CG et al. Multidrug-resistant, extensively
drug-resistant and pandrug-resistant becteria: an interna-
tional expert proposal for interim standard definitions for
acquired resistance. Clin Microbiol Infect 2012;18:268-81.
20. Almirante B, Rodrìguez D, Park BJ, Cuenca-Estrella M,
Planes AM, Almela M et al. Epidemiology and predictors of
mortality in cases of Candida bloodstream infection: results
from population-based surveillance, Barcelona, Spain, from
2002 to 2003. J Clin Microbiol 2005;43:1829-35.
21. Velasco E, Portugal RD, Salluh JIF. A simple score to pre-
dict early death in adult cancer patients with bloodstream
infections. J Infect 2009;59:332-6.
22. Orsi GB, Raponi M, Franchi C, Rocco M, Mancini C, Ven-
ditti M. Surveillance and infection control in an intensive
care unit. Infect Control Hosp Epidemiol 2005;26:321-5.
23. Burton DC, Edwards JR, Horan TC, Jernigan JA, Fridkin
SK. Methicillin-resistant Staphylococcus aureus central
Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on April 16, 2014. - Accepted for publication on November 18, 2014. - Epub ahead of print on November 20, 2014.
Corresponding author: GB Orsi, Dipartimento di Sanità Pubblica e Malattie Infettive, “Sapienza” Università di Roma, P.le Aldo Moro 5,
00185 Roma, Italy. E-mail giovanni.orsi@uniroma1.it
988 MINERVA ANESTESIOLOGICA September 2015
line –associated bloodstream infections in US intensive care
units, 1997-2007. JAMA 2009;301:727-36.
24. Marra AR, Camargo LF, Pignatari AC, Sukiennik T, Behar
PR, Medeiros EA et al. Nosocomial bloodstream infections
in brazilian hospitals: analysis of 2,563 cases from a pro-
spective nationwide surveillance study. J Clin Microbiol
2011;49:1866-71.
25. Liu CY, Liao CH, Chen YC, Chang SC. Changing epide-
miology of nosocomial bloodstream infections in 11 teach-
ing hospitals in Taiwan between 1993 and 2006. ��������
J Micro-
biol Immunol Infect 2010;43:416-29.
26. Gudiol C, Bodro M, Simonetti A, Tubau F, Gonzalez-Barca
E, Cisnal M et al. �����������������������������������������
Changing etiology, clinical features, an-
timicrobial resistance and outcomes of bloodstream infec-
tion in neutropenic cancer patients. Clin Microbiol Infect
2013;19:474-9.
27. Jain R, Kralovic SM, Evans ME, Ambrose M, Simbartl
LA, Obrosky DS et al. Veterans affairs initiative to prevent
methicillin-resistant Staphylococcus aureus infections. N
Engl J Med 2011;364:1419-30.
28. Klein E, Smith DL, Laxminarayan R. Hospitalizations
and deaths caused by methicillin-resistant Staphylococ-
cus aureus, United States, 1999-2005. Emerg Infect Dis
2007;13:1840-6.
29. Klevens RM, Morrison MA, Nadle J, Petit S, Gersham K,
Ray S et al. Invasive methicillin-resistant staphylococcus au-
reus infections in the United States. JAMA 2007;298:1763-
71.
30. Orsi GB, Falcone M, Venditti C. Surveillance and manage-
ment of multi-drug resistant microrganisms. Expert Rev
Anti Infect Ther 2011;9:653-79.
31. Borer A, Saidel-Odes L, Riesenberg K, Eskira S, Peled N,
Nativ R et al. Attributable mortality rate for carbapenem-
resistant Klebsiella pneumoniae bacteremia. Infect Control
Hosp Epidemiol 2009;30:972-6.
32. Gasink LB, Adelstein PH, Lautenbach E, Synnestvedt M,
Fishman NO. Risk factors and clinical impact of Klebsiella
pneumoniae carbapenemase-producing K. pneumoniae.
Infect Control Hosp Epidemiol 2009;30:1180-5.
33. Patel G, Huprikar S, Factor SH, Jenkins SG, Calfee DP.
Outcomes of carbapenem-resistant Klebsiella pneumoniae
infection and the impact of antimicrobial and adjunctive
therapies. Infect Control Hosp Epidemiol 2008;29:1099-
106.
34. Schwaber MJ, Klarfeld-Lidji S, Navon-Venezia S, Schwartz
D, Leavitt A, Carmeli Y. Predictors of carbapenem-resist-
ant Klebsiella pneumoniae acquisition among hospitalized
adults and effect of acquisition on mortality. Antimicrob
Agents Chemother 2008;52:1028-33.


O R I G I N A L A RT I C L E

Confirmation of correct central venous

catheter position in the preoperative setting
by echocardiographic “bubble-test”
M. MEGGIOLARO 1, A. SCATTO 1, A. ZORZI 2, E. ROMAN-POGNUZ 1
A. LAURO 3, C. PASSARELLA 1, G. BONACCORSO 1

1Division of Anaesthesiology and Intensive Care, University Hospital of Padua, Padua, Italy; 2Division of Cardiology,
Deparment of Cardiac, Thoracic and Vascular Sciences, University of Padua, Padua, Italy; 3Division of Radiology,
University Hospital of Padua, Padua, Italy

ABSTRACT
Background. Verification of the central venous catheters (CVCs) position by chest X-ray (CXR) is usually per-
formed in the postoperative period with the risk related to possible malposition. This prospective observational study
aimed to assess the diagnostic accuracy and reproducibility of ultrasound (US) and “bubble test” to detect malposi-
tions of CVC in the preoperative setting.
Methods. The study included 105 patients undergoing preoperative CVC placement. A US protocol aimed at direct
visualization of the CVC was completed by a single operator and two consecutive “bubble tests” were performed
independently by different physicians. Two parameters were considered: complete right atrium (RA) opacization
versus visualization of “no or few bubbles” and time from agitated saline injection to visualization of micro-bubbles
in the RA (“push-to-bubbles” time).
Results. CXR identified 14 (13%) CVC malpositions. Vascular US showed a sensitivity of 64% and a specificity
of 100% while visualization of “no or few bubbles” at bubble test yielded a sensitivity of 50% and a specificity of
100%. “Push-to-bubbles” times were ≈9 times longer in patients with compared to those without CVC malposition
(1400 [702-2160] ms versus 167 [123-228] ms, P<0.001). A cut off value of 500 ms had a sensitivity of 100% and a
specificity of 99% for CVC malposition with an inter-observer agreement of 99% (kappa 0.96, P<0.001).
Conclusion. CVC malposition was observed in a sizeable proportion of patients undergoing preoperative central
venous cannulation. Measurement of “push-to-bubbles” time is a fast, accurate and highly reproducible tool for
verifying the correct CVC position. (Minerva Anestesiol 2015;81:989-1000)
Key Words: Central venous catheters – Echocardiography - Ultrasonography - Complications.

C entral venous catheterization is common

practice in several clinical settings but rep-
resents a source of possible complications, such
as hematoma in the site of puncture, pneumo-
thorax and malposition.1 The latter is of partic-
ular concern because it is often concealed and
can cause life threatening complications such as
malfunction, vessel perforation, cardiac tampon-
Comment on p. 943.
ade and, in case of arterial cannulation, systemic
embolism.2-4
Chest x-ray (CXR) is usually performed af-
ter central venous catheters (CVC) placement
to rule out pneumothorax and malposition.1, 2
However, when a CVC is inserted in the pre-
operative setting for hemodynamic monitoring
and drug administration, immediate CXR may
not be feasible because of time and logistic is-
sues, with the risk of complications related to a

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 989

MEGGIOLARO CONFIRMATION OF CVC POSITION BY BUBBLE TEST
possible malposition. Several non-radiological
methods to verify the CVC position have been
proposed as possible alternatives to CXR, in-
cluding vascular ultrasound (US),9-11 echocardi-
ography,12 intracardiac electrocardiology 13 and,
more recently, the so-called “bubble test”.14-16
While vascular US and echocardiography rely on
direct visualization of the catheter, the rationale
of the bubble test is that the right atrium (RA)
should be immediately and completely filled
with micro-bubbles when a bolus of agitated sa-
line is administered through the distal lumen of
a normally positioned CVC. The micro-bubbles
can be easily recognized by echocardiography
because they appear brighter than the blood.
However, only few studies have assessed the
value of bubble test for confirmation of correct
CVC position and none of the above mentioned
techniques has been tested in the preoperative
setting.
The primary end-point of our study was to
assess the rate of CVC malpositions in the pre-
operative setting and to verify the diagnostic ac-
curacy and reproducibility of US and bubble test
to detect CVC malpositions compared to stand-
ard CXR. As a secondary end-point, we assessed
whether standard or bubble test-enhanced echo-
cardiography can reliably identify an intra-atrial
(“too-low”) CVC tip position.
Materials and methods
Study population, setting and central venous can-
nulation
The study was conducted from January to
September 2013 in the surgical theaters of
the University Hospital of Padoa, Italy, with
a prospective observational design. It was ap-
proved by the local Ethical Committee and
any informed consent from human subjects
was obtained as required.������������������
The study popula-
tion included a cohort of consecutive patients
older than 18 years-old that underwent preop-
erative central venous cannulation before gen-
eral, vascular, plastic and urologic surgery and
consented to participate in the study. Clini-
cally unstable patients were excluded, as well
as those with known unrepaired intra-cardiac
990 MINERVA ANESTESIOLOGICA September 2015
shunts (atrial or ventricular septal defects) or
suspicious arterial cannulation to minimize the
risk of systemic air embolism during bubble
test. The CVC (non-tunneled, dual-lumen 7F,
20-cm long catheters) were inserted percutane-
ously, either in a preparation room before oro-
tracheal intubation or in the operating room
after intubation. Four vascular approaches were
used: right internal jugular vein (IJV), left IJV,
right subclavian vein (SV) and left SV. The
vascular approach as well as depth of CVC in-
sertion was chosen by the attending anesthesi-
ologist (either specialist or in training) based
on patient characteristics and clinical demand.
The CVC were inserted with the patient in a
supine position using the Seldinger technique.
US guide was always used for IJV cannulation
while a traditional approach based on recogni-
tion of anatomical landmarks was used for SV
cannulation. After CVC placement, the US
protocol and two bubble tests were performed.
CXR to confirm the correct CVC position was
normally performed in the early postoperative
period but was anticipated if US or bubble test
findings suggested a possible CVC malposi-
tion.14-16 Images were analyzed by a single radi-
ologist (A. Lauro), who was blinded to the study
results. Time from CXR request to final report
availability was recorded. The CVC course was
considered normal when the body of the cath-
eter was straight and located inside the superior
vena cava (SVC). CVC malposition was diag-
nosed if the catheter had an abnormal vascular
course, made a loop or was impinged upon the
lateral wall of the SVC.
For the purpose of this study (short-term in-
tra-operative use), an intra-atrial CVC tip posi-
tion was not considered a malposition. However,
as a secondary end-point, we also evaluated the
diagnostic ability of US and bubble test to dis-
tinguish between intra-atrial and SVC tip posi-
tion.
Ultrasound protocol
After CVC insertion, an US study was per-
formed by a single experienced operator (M.
Meggiolaro) using an iE33 system (Philips Ul-
trasound, Bothell, WA, USA) equipped with
CONFIRMATION OF CVC POSITION BY BUBBLE TEST MEGGIOLARO

a 4-8 MHz linear probe and a 2.5-3.5 MHz
sector probe. The exam was aimed at direct
visualization of the CVC tip and exclusion of
pneumothorax. The IJV and SV were scanned
bilaterally to directly visualize the CVC. The
presence of the catheter distally to the site of
puncture (suggesting a catheter loop) or in ves-
sels not used for catheter insertion (suggest-
ing an abnormal CVC course) was considered
a malposition. The RA was visualized from
the apical 4-chambers and subcostal long-axis
views to identify the presence of the CVC tip
inside the heart (intra-atrial CVC tip position).
The protocol to rule out pneumothorax has
been described in details elsewhere17. In brief,
longitudinal and trasversal scanning of both
hemithoraces was performed and pneumotho-
rax was diagnosed if sliding sign and “B-lines”
were absent and “lung point” was present. The
time needed to perform the complete US pro-
tocol was recorded.
Bubble test protocol
The “bubble test” consisted of two separate in-
jections into the distal CVC lumen of 5 mL of a
contrast agent made of 90% saline and 10% air
mixed with a three-way stopcock by exchange of
saline/air mixture between the syringes.18 Due to
their peculiar acoustic properties, the microbub-
bles contained in the solution appear brighter
than the blood and can be easily distinguished by
echocardiography. For the purpose of the study,
only two parameters were evaluated: complete
opacization of the RA by microbubbles versus
visualization of no or few bubbles (qualitative
parameter) and time from injection to visualiza-
tion of the first bubbles (“push-to-bubbles”) in
the RA (quantitative parameter). The subcostal
long-axis view was considered the first choice
for RA visualization because of its simplicity. If
technical issues (such as drain tubes, obesity or
pneumoperitoneum) limited its use, the 4-cham-

A C
Figure 1.—Bubble test performed with electrical switch.
A) Electrical switch attached to the plunger of the syringe and wired to an auxiliary ultrasound machine’s plug. When the injec-
tion is started, the thumb of the operator presses the switch, and a typical squared artifact is recorded on the ECG tracing; B) at
frame #45 of a 5 second recording, the marker (white arrow) corresponds to the beginning of the squared artifact which marks
the agitated saline injection start. At this time, there are no bubbles in the right atrium, which appears black (white asterix); C) 9
frames after, sudden and complete opacization of the right atrium by microbubbles is noted. As the total number of frames in this
5 second clip was 183 (27 ms/frame), “push-to-bubbles” time was 243 ms. Chest X-ray showed a correct CVC course with the tip
in the superior vena cava.

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 991

MEGGIOLARO CONFIRMATION OF CVC POSITION BY BUBBLE TEST

ber apical view as second choice and the paraster-
nal short-axis view as third choice were utilized.
Other infusions were stopped before the bubble
test. A 5 seconds clip was recorded during the
injection. To accurately determine the exact mo-
ment of infusion we utilized a customized electric
switch attached to the plunger of the syringe and
wired to an auxiliary ultrasound machine’s plug
that produces a distinct electrical artifact on the
ECG trace of the echo system when the thumb
of the operator presses the plunger (Figure 1).
The time interval of a single frame was calculated
dividing the 5 seconds of recording by the total
number of frames in the clip and the “push-to-
bubbles” time was derived multiplying the num-
ber of frames from the beginning of the electrical
artifact to microbubbles visualization by the time
interval of a single frame. In each patient, two
consecutive tests were performed and reviewed
independently by different operators (M. Meg-
giolaro, A. Scatto) in order to evaluate the test re-
producibility and inter-observer agreement. The
two operators had both a ≈3 year-experience in
focused vascular and cardiac ultrasound for CVC
placement and hemodynamic monitoring. The
“push-to-bubbles” time assigned to each patient
corresponded to the mean of the two values. The
time needed to complete the study protocol (i.e.
the two consecutive bubble tests) was recorded.
The electric switch allows a more precise
measurement but it is not commercially avail-
able; hence, we also performed a preliminary
study and tested a second “clinically available”
method consisting of an ECG lead attached to
the syringe to produce an artifact during con-
trast injection in a second group of 20 different
patients (Figure 2). Particular attention was paid
to wait for the ECG tracking to be flat and clear
before starting clip recording and injection.
Statistical analysis
Results are summarized as n. (%) for di-
chotomous variables or median with 25-75%

A C
Figure 2.—Bubble test performed with ECG lead.
A) An ECG lead of the ultrasound machine is attached to the syringe. When the injection is started, flexion of the operator’s fin-
gers produces a vibration of the ECG lead which results in an artifact recorded on the ECG tracing; B) the marker (white arrow)
reaches the beginning of the ECG artifact marking the agitated saline injection at frame #38. At this time, there are no bubbles
in the right atrium, which appears black (white asterix); C) at frame #47, sudden and complete opacization of the right atrium by
microbubbles is noted. As the total number of frames in this 5 second clip was 201 (24 ms/frame), “push-to-bubbles” time was
216 ms. Chest X-ray showed a CVC-tip position in the atrio-cava junction.

992 MINERVA ANESTESIOLOGICA September 2015

CONFIRMATION OF CVC POSITION BY BUBBLE TEST MEGGIOLARO

iles for continuous variables. Categorical dif-
ferences between groups were evaluated by the
χ2 test or the Fisher exact test, as appropriate,
while the Rank Sum test was used to compare
continuous variables. The receiving operator
curve (ROC) method was used to estimate the
best cut-off value of “push-to-bubbles” time to
discriminate between patients with and without
CVC malposition in the entire study group and
between patients with and without intra-atrial
CVC tip position in the subgroup with normal
CVC course. The two tests were considered
concordant when both “push-to-bubbles” times
obtained in the same patient by the two differ-
ent operators were above or below the cut-off
value. The inter-observer agreement was evalu-
ated by the Cohen kappa statistic test. To assess
whether the diagnostic accuracy is improved by
averaging repeated measures we compared the
ROC curves of two single bubble tests to the
ROC curve of the mean values of the two tests
by the method described by DeLong et al.19
The sensitivity and specificity of different tech-
niques (US protocol, complete RA opacization
and cut-off values of “push-to-bubbles” time)
were assessed in relation to CXR. The lower
limits of the 95% confidence interval (C.I.) of
sensitivity and specificity of the “push-to-bub-
bles” time cut-off value for CVC malposition
were calculated by taking into account an ex-
pected sensitivity and specificity of 95%.20 All
P-values reported are 2 sided, and P<0.05 was
considered statistically significant. Data were
analyzed with SPSS 17 (SPSS Inc; Chicago, IL,
USA) and MedCalc 12.7 (MedCalc Software,
Ostend, Belgium).
Results
General characteristics, US protocol and bubble test
results
During the study period, 105 patients un-
dergoing elective surgery were included. The

Table I.—General characteristics, vascular access and bubble test findings in the overall study sample and according to the pres-
ence of malposition.
No malposition by Malposition
Overall CXR by CXR P
N.=105 N.=91 N.=14
General characteristics
Gender (male) 70 (67) 61 (67) 9 (64) 0.84
Age 74 (67-80) 74 (67-80) 74 (68-81) 0.80
Weight (kg) 70 (65-85) 70 (65-85) 72 (63-86) 0.98
Height (cm) 170 (165-175) 170 (165-175) 170 (160-175) 0.55
BMI (kg/m2) 25 (23-28) 25 (22-28) 25 (23-29) 0.75
Mechanical ventilation 52 (50) 48 (53) 4 (29) 0.15
CVP (mmHg) 8 (6-9) 8 (6-9) 8 (7-10) 0.73
Mean AP (mmHg) 79 (72-90) 78 (70-90) 86 (70-89) 0.23
CVC characteristics
Vascular approach
Right IJV 69 (66) 63 (69) 6 (44) 0.37
Left IJV 10 (9) 7 (8) 3 (21)
Right SV 19 (18) 16 (18) 3 (21)
Left SV 7 (67) 5 (5) 2 (14)
Complete insertion (20 cm) 77 (73) 67 (74) 10 (71) 1.0
Positive vascular US and standard echocardiogram 9 (9) 0 9 (64) <0.001
Bubble test
RA visualization
Subcostal view 98 (93) 85 (93) 13 (93) 1.0
Apical 4-chamber 7 (7) 6 (7) 1 (7)
Few or no bubbles 7 (7) 0 7 (50) <0.001
Push-to-bubbles frames 8 (6-13) 7 (6-10) 68 (37-87) <0.001
Push-to-bubbles time (ms) 180 (131-316) 167 (123-228) 1400 (702-2160) <0.001
Data are expressed as median (25%-75%iles) or N. (%).
AP: aortic pressure; BMI: body mass index; CVC: central venous catheter; CVP: central venous pressure; CXR: chest x-ray; IJV: internal jugular
vein; RA: right atrium; SV: subclavian vein.

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 993

MEGGIOLARO CONFIRMATION OF CVC POSITION BY BUBBLE TEST

characteristics of the study sample are shown
in Table I. There were 70 males and 35 females.
No one was morbidly obese (BMI>40). Fifty-
two (49%) patients were mechanically venti-
lated during CVC insertion. In 69 (66%) pa-
tients the catheter was inserted in the right IJV
and in 77 (73%) it was advanced completely
(20 cm) in the vessel. Vascular US and stand-
ard echocardiography (i.e. without bubble test)
identified 9 patients with CVC malposition
and 19 patients with an intra-atrial CVC tip
position. No pneumothorax was observed. The
median duration of the US protocol was 10 (7-
20) minutes. The bubble tests were performed
using either the sub-costal view (93%) or the
apical 4-chamber (7%) view for RA visualiza-
tion. No patient was excluded because the RA
could not be visualized. In 98 (94%) patients,
complete filling of the RA with microbubbles
was demonstrated while in the remaining 7
(6%) no or few bubbles were observed. Median
“push-to-bubbles” time was 180 (131-316) ms,
corresponding to 8 (6-13) frames. “Push-to-
bubbles” times were similar in spontaneously
breathing (189 [132-319] ms) and in mechan-
ically ventilated patients (172 [129-285] ms,
P=0.22). The difference between the longest
and the shortest “push-to-bubbles” time in
two consecutive tests was 25 (0-55) ms, cor-
responding to 1 (0-2) frame. The median dura-
tion of the bubble test protocol was 5 (5-10)
minutes.
Chest radiography findings
In 10 patients the CXR was acquired preop-
eratively, because of suspicious findings at US or
long (>2 sec) “push-to-bubbles” times, while in
the remaining it was performed postoperatively.
The median time between CVC insertion and
availability of the CXR report was 67 (42-84)
minutes. The exam revealed CVC malposition in
14 (13%) patients (Table II). Among the 91 pa-
tients with a correct CVC course, an intra-atrial
CVC tip position was reported in 33 (31%). No
pneumothorax was observed.
Identification of CVC malposition
General characteristics, vascular approach and
bubble test results in the overall study sample
and according to the presence of malposition
are shown in Table II. All 9 malpositions identi-
fied by US were confirmed by CXR (specificity
100%). There were 5 false negatives (sensitiv-
ity 64%): in 1 case the CVC distal tip was in
the proximal right SV, in 1 in the left internal
thoracic vein and in the remaining 3 the cath-
eter was impinged upon the lateral wall of the
SVC. All 7 patients with visualization of no or

Table II.—Characteristics of patients with CVC malposition.

Push-to-bubbles
Age/ Vascular Tip Vascular
Patient # Time (ms)
gender access position US (1st/2nd) mean
1 85/F RIJV RJV loop Positive (909/969) 697
3 75/F LSV LITV Negative (3077/3214) 3146
7 64/F RSV RJV Positive (1674/1502) 1588
12 81/M RSV RJV Positive (2564/2564) 2564
34 82/F RIJV RJV loop Positive (1717/1803) 1760
41 58/M RIJV RJV loop Positive (1077/1051) 1064
48 79/M RIJV RSV Positive (1225/1186) 1206
50 70/M RIJV RJV loop Positive (2026/1974) 2000
52 75/M LIJV Impinged upon SVC Negative (622/484) 553
66 69/M LSV RSV Negative (2500/2540) 2520
76 66/M RIJV Impinged upon SVC Negative (579/579) 579
87 70/M LIJV LJV loop Positive (814/644) 729
92 70/M RSV RJV Positive (1974/2103) 2039
98 87/F LIJV Impinged upon SVC Negative (494/751) 623
F: female; M: male; LITV: left internal thoracic vein; LIJV: left internal jugular vein; LSV: left subclavian vein; RIJV: right internal jugular vein;
RSV: right subclavian vein; SVC: superior vena cava.

994 MINERVA ANESTESIOLOGICA September 2015

CONFIRMATION OF CVC POSITION BY BUBBLE TEST MEGGIOLARO

few bubbles at bubble test had CVC malposition
confirmed by CXR (sensitivity 50%, specificity
100%). “Push-to-bubbles” times were signifi-
cantly longer in patients with malposition com-
pared to those without (1400 [702-2160] ver-
sus 167 [123-228] ms, P<0.001). In only 5/14
(36%) patients with CVC malposition “push-to-
bubbles” times were >2 sec. The difference be-
tween “push-to-bubbles” times measured during
the two consecutive tests performed by different
operators was 24 (0-48) ms in patients without
malposition and 94 (32-170) ms in patients
with malposition (P<0.001). In all patients with
a normal CVC course the difference between the
two “push-to-bubbles” times was <50 ms. Ac-
cording to the ROC curve (AUC 0.99) analy-
sis, a cut off value of 500 ms had the highest
diagnostic accuracy with a sensitivity of 100%
(95% C.I. lower limit=60%) and a specificity of
99% (95% C.I. lower limit=84%) for CVC mal-
position. The inter-observer agreement on a cut
off of 500 ms was 99% (kappa 0.96, P<0.001).
There were no differences between the ROC
curves of the “push-to-bubbles” times obtained
by the two single operators (AUC=0.99 and
0.98, respectively) compared to the ROC curve
of the mean values (AUC=0.99, P=0.78 and 1.0,
respectively). Representative examples of bubble
test and CXR findings in patients with and with-
out CVC malposition are shown in Figures 1-3.
Identification of intra-atrial CVC tip position
The 14 patients with CVC malposition were
not included in this analysis. General character-
istics, vascular access and bubble test findings
in the group of 91 patients with a normal CVC
course according to the position of the catheter
distal tip (SVC versus intra-atrial) are shown in
Table III. An intra-atrial CVC position was sig-
nificantly associated with insertion of the entire
20 cm-long CVC and use of a right-sided vascu-
lar approach. The sensitivity and specificity for
intra-atrial position of direct CVC tip visualiza-
tion by echocardiography were 48% and 95%,
respectively. The “push-to-bubbles” times were

A C
Figure 3.—Bubble test performed in a patient with malposition.
A) Chest X-ray showing a CVC inserted through the right subclavian vein which is abnormally positioned in the right jugular
vein. B-C) bubble test showing complete opacization of the right atrium (white asterix) but “push-to-bubbles” time (white arrows)
of 71 frames (1.8 second).

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 995

MEGGIOLARO CONFIRMATION OF CVC POSITION BY BUBBLE TEST

Table III.—General characteristics, vascular access and bubble test findings in the subpopulation of 91 patients with a normal
CVC course according to site of the catheter distal tip.
SVC tip position RA tip position
by CXR by CXR P
N.=58 N.=33
General characteristics
Gender (male) 37 (64) 24 (73) 0.38
Age 73 (69-80) 71 (65-79) 0.27
Weight (kg) 72 (62-85) 70 (65-88) 0.55
Height (cm) 170 (165-175) 170 (165-175) 0.53
BMI (kg/m2) 25 (22-28) 24 (23-28) 0.72
Mechanical ventilation 28 (48) 20 (61) 0.26
CVP (mmHg) 8 (6-10) 8 (6-9) 0.62
Mean AP (mmHg) 78 (69-89) 78 (72-90) 0.59
CVC characteristics
Vascular approach
Right IJV 38 (66) 25 (76) 0.04
Left IJV 7 (12) 0(
Right SV 8 (14) 8 (24)
Left SV 5 (9) 0(
Complete insertion (20 cm) 38 (66) 29 (88) 0.02
Positive standard echocardiogram 3 (5) 16 (48)
Bubble test
RA visualization
Subcostal view 55 (95) 30 (91) 0.68
Apical 4-chamber view 3 (5) 3 (9)
Few or no bubbles 0( 0( 1.0
Push-to-bubbles frames 9 (6-13) 5 (4-8) 0.002
Push-to-bubbles time (ms) 198 (141-293) 102 (86-162) 0.002
Data are expressed as median (25-75% quartiles) or N. (%).
AP: aortic pressure: BMI: body mass index; CVC: central venous catheter; CVP: central venous pressure; CXR: chest x-ray; IJV: internal jugular
vein; RA: right atrium; SV: subclavian vein.

significantly longer in patients with SVC posi-
tion compared to those with intra-atrial posi-
tion (198 [141-293] ms versus 102 [86-162] ms,
P=0.002). According to the ROC curve analy-
sis (AUC 0.71); a cut off value of 150 ms had
the highest diagnostic accuracy with a sensitiv-
ity of 67% and a specificity of 70% for intra-
atrial position. The inter-observer agreement
on a cut-off of 150 ms was 87% (kappa=0.73,
P<0.001). There were no differences between the
ROC curves of the “push-to-bubbles” times ob-
tained by the two single operators (AUC=0.72
and 0.68, respectively) compared to the ROC
curve of the mean values (AUC=0.71, P=0.84
and P=0.57, respectively).
“ECG lead” measure method
An alternative measure method (Figure 2)
was tested in a group of 20 different patients (12
males, mean age 71 years) with similar baseline
characteristics to the study cohort. In this group,
CXR revealed one malposition (CVC inserted
through the left SV with tip in the left IJV). All
19 patients with a normal CVC course showed
complete opacization of the RA and “push-to-
bubbles” time <500 ms while the patient with
CVC malposition showed a delayed (≈1.5 sec)
and incomplete RA opacization.
Discussion
The aim of this study was to assess the rate
of CVC malpositions in the preoperative setting
and to test the ability of vascular US and bubble
test to confirm a normal CVC course. Our main
findings were that: 1) CVC malpositions were
not uncommon (13%) in this clinical setting;
2) the sensitivity of US (aimed at direct visuali-
zation of the catheter) and “qualitative” bubble
test (complete RA opacization versus “few or no
bubbles”) for identification of CVC malposi-
tions was limited; 3) “quantitative” bubble test”
with “push-to-bubbles” times measurement was

996 MINERVA ANESTESIOLOGICA September 2015

CONFIRMATION OF CVC POSITION BY BUBBLE TEST MEGGIOLARO
a fast, accurate and highly reproducible method
to identify CVC malpositions.
Malposition of CVC is a potentially severe
complication and can cause catheter dysfunc-
tion, vascular perforation or systemic embolism.
The food and drug administration as well as
other guidelines recommend performing CXR
after each CVC insertion to rule out mechanical
complications such as pneumothorax and con-
firm the correct catheter course.7, 21, 22 However,
the CXR has some clinical limitations related to
x-ray exposure and time needed for the exam to
be performed and reported. In the preoperative
setting, where CVC are often inserted for hemo-
dynamic monitoring and drug administration,
time and logistic issues usually prevent the CXR
from being performed immediately, with the
inherent risks of using a potentially misplaced
catheter.7 As such, an alternative technique to
confirm the correct CVC position should be
implemented. Several different non-radiological
methods to identify CVC malpositions have
been recently proposed. Among them, US is par-
ticularly promising because it can be performed
bedside with portable machines and has virtually
no contraindications. The protocol for verifying
the CVC course consists in direct visualization
of the catheter inside the vessel used for inser-
tion and exploration of the other extra-thoracic
veins. The presence of the catheter in vessels not
used for catheter insertion or distally to the site
of puncture suggests a malposition. In addition,
US can be used to guide CVC placement and
rule out pneumothorax. The accuracy and fea-
sibility of this method have been demonstrated
both in the intensive care unit and in the emer-
gency department.9-11 The main limitations of
the technique are the need for a specific train-
ing, the operator dependent variability and the
time required to perform the entire protocol.
These drawbacks can be of particular concern
in the preoperative setting because of limited
time availability and different staff experience.
Furthermore, vascular US cannot identify intra-
thoracic CVC malpositions (e.g. concealed can-
nulation of the internal thoracic vein or CVC
impinged upon the SVC.)23, 24
Recently, a new protocol combining US with
bubble test-enhanced echocardiography has
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 997
been proposed.14-16 The bubble test consists of
administration of agitated saline as a contrast
agent, because the microbubbles which are dis-
persed in the solution appear brighter than the
blood at echocardiography. In current practice,
the test is performed to rule out an intra-cardiac
shunt because the microbubbles are cleared in
the lungs and should not be visible in the left
cardiac chambers if the agitated saline is admin-
istered through a venous line. The rationale for
its use to verify the CVC position is that rapid
and complete filling of the RA by micro-bubbles
should be observed at echocardiography if the
catheter tip is inside or very close to the atrium
(i.e. in the SVC). On opposite, infusion through
a misplaced CVC should result in delayed and
incomplete RA opacization. A previous study
found that combining US and bubble test-en-
hanced echocardiography yielded a very high di-
agnostic accuracy for CVC malposition.15 How-
ever, the protocol of this investigation consisted
in evaluating several US and bubble test find-
ings, which may require a significant amount of
time and experience. In the preoperative setting,
when time is usually a major issue and the ex-
perience of the anesthesiology staff diversified, a
single simple and fast test would be preferred.
More recently, a retrospective analysis of patients
undergoing echo-guided CVC placement into
the IJV showed a very high accuracy of “quali-
tative” bubble-test (i.e. complete RA opaciza-
tion by microbubbles) for the diagnosis of CVC
misplacement.14 However, the statistical analysis
was limited by the presence of only 2 malposi-
tions among the 219 CVCs and lack of quantita-
tive data.
In our study, we tested separately the diagnos-
tic accuracy of vascular US, qualitative evalua-
tion of complete filling of the RA by microbub-
bles and time from agitated saline infusion to
visualization of micro-bubbles in the RA. In con-
trast to previous reports, we found that neither
“qualitative-only” bubble test nor vascular US is
sensitive enough to rule out CVC malposition.
In fact, we observed that complete opacization
of the RA occurred in half of our patients with
CVC malposition (Figure 3) and that vascular
US could not identify intra-thoracic abnormal
course. Instead, we found that “push-to-bub-
MEGGIOLARO CONFIRMATION OF CVC POSITION BY BUBBLE TEST
bles” times followed a binomial distribution, be-
ing ≈9 times longer in patients with malposition
compared to those with a normal CVC course.
As a result, an intermediate cut-off (500 ms)
yielded a diagnostic accuracy >99%. Bubble test
with “push-to-bubbles” time measurement was
not only accurate, but also fast and simple: in
>90% of cases it could be performed using only
the echocardiographic sub-costal view, which is
very simple and easy to learn, and required only
5 minutes. Furthermore, the test showed a very
high inter-observer agreement; however, this
study is limited by lack of intra-observer agree-
ment analysis. In accordance with previous stud-
ies,14-16 the bubble test demonstrated a very high
safety profile with no complications.
In the present study, we measured “push-to-
bubbles” times precisely with a customized elec-
trical switch which is not clinically available.
Hence, we conducted a preliminary study in a
second group of 20 patients and observed that
the test may be performed in the daily practice
using a simple and clinically available method
consisting of an ECG lead of the US machine
attached to the syringe plunger (Figure 2). In all
these patients, a cut-off value of 500 ms correctly
classified the CVC position. We acknowledge
that this is a preliminary observation that needs
to be validated in a larger patient group.
Whether an intra-atrial CVC tip position
should be regarded as a malposition is a mat-
ter of debate. Traditionally, the intra-atrial CVC
position is considered sub-optimal because early
studies reported an increased risk of perfora-
tion and thrombosis. More recent investigations
showed that intra-atrial CVC tip position is
not associated with a higher rate of complica-
tions, probably thanks to the use of newer ma-
terials.11, 22 Accordingly, we considered an intra-
atrial tip position acceptable for a short-term
(perioperative) CVC use. However, as a second-
ary study end-point, we assessed whether direct
visualization of the catheter tip by echocardiog-
raphy or “push-to-bubbles” time can diagnose
a “too low” CVC tip position and found that
neither test is accurate enough. Interestingly, all
33 CVCs with an intra-atrial tip position were
inserted through a right-sided vascular approach
and the majority was advanced completely. This
998 MINERVA ANESTESIOLOGICA September 2015
finding underlines that systematic complete (20
cm) CVC insertion should be avoided, especially
when using right-sided vascular approaches.
The incidence of CVC malposition in the
literature is variable, depending on the setting,
experience of operators and definition of malpo-
sition. Excluding intra-atrial CVC tip position,
Gladwin et al. reported a 5.6% rate of CVC mal-
position.8 A similar incidence (6.7%) was found
in a prospective study on nearly 1800 CVC
placed by experienced operators, with the high-
est rate for the left IJV.1 Ruesch et al performed
a meta-analysis of six trials and found a 7.5%
rate of malposition, which was higher for subcla-
vian (9.3%) than for jugular (5,3%) approach.25
Finally, Solozano reported a higher (14%) in-
cidence of unusual CVC course in surgical pa-
tients.26 In our study, we also observed a high
(13%) incidence of CVC malposition. Although
we did not perform or supervise the CVC place-
ment and cannot retrieve precise information on
the level of experience of each operator, we sus-
pect that the reason for this finding may be two-
fold: first, our institution is a teaching hospital
where several young anesthesiologists in training
work; second, many anesthesiologists experi-
enced in the landmark technique were learning
how to use the US guide for CVC insertion dur-
ing the study period. This high rate of malposi-
tion prompted us to retrain all anesthesiologists
and fellows of our institution on choosing the
correct CVC insertion depth and limiting the
risk of malposition with an appropriate insertion
technique.
Key Messages
—— In the preoperative setting, verification
of the central venous catheter (CVC) course
by chest X-ray is usually performed only in
the postoperative period with the risks relat-
ed to possible malposition. A simple and fast
tool for confirming the position of the CVC
would be useful.
—— If the CVC course is normal, the tip
of the catheter lies inside or very close to
CONFIRMATION OF CVC POSITION BY BUBBLE TEST MEGGIOLARO
the right atrium (RA), and when agitated
saline is injected through the catheter rapid
and complete filling of the atrium by micro-
bubbles is observed at echocardiography. On
opposite, infusion through a misplaced CVC
should result in delayed and incomplete RA
opacization.
—— This study showed that a prolonged
interval between agitated saline infusion and
microbubbles visualization in the RA is high-
ly predictive for CVC malposition. Bubble
test with “push-to-bubbles” time measure-
ment is not only accurate, but also fast and
simple to perform.
Conclusions
In conclusion, we demonstrated that mea-
surement of time from agitated saline infusion
to echocardiographic opacization of the RA is
a highly accurate and reproducible tool to rule
out CVC malposition in the preoperative set-
ting. According to our findings, preoperative
CXR to verify the CVC position or reposition-
ing should be considered in patients with abnor-
mally long “push-to-bubbles” times. Given the
very high predictive value of the 500 ms cut-off,
even slight prolongation of “push-to-bubbles”
time above this cut-off should raise the doubt
of a malposition and prompt preoperative CXR
confirmation of CVC course. Prospective inves-
tigations performed on larger study sample are
required to confirm our study results.
Finally, it must be underlined that the “bub-
ble test” is generally safe, but may be potentially
dangerous in case of inadvertent intra-arterial
micro-bubbles injection. Hence, it should not
be used to verify the CVC position in case of
suspicious arterial cannulation.
References
  1. Schummer W, Schummer C, Rose N, Niesen WD, Sakka
SG. Mechanical complications and malpositions of central
venous cannulations by experienced operators. A prospec-
tive study of 1794 catheterizations in critically ill patients.
Intensive Care Med 2007;33:1055-9.
 2. Polderman KH, Girbes AJ. Central venous catheter use.
Part 1: mechanical complications. Intensive Care Med
2002; 28:1-17.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 999
  3. McGee DC, Gould MK. Preventing complications of cen-
tral venous catheterization. N Engl J Med 2003;348:1123-
33.
  4. Sznajder JI, Zveibil FR, Bitterman H, Weiner P, Bursztein
S. Central vein catheterization. Failure and complication
rates by three percutaneous approaches. Arch Intern Med
1986;146:259-61.
  5. Scott WL. Central venous catheters. An overview of Food
and Drug Administration activities. Surg Oncol Clin
NorthAm 1995;4:377-93.
 6. Lanza C, Russo M, Fabrizzi G. Central venous cannula-
tion: are routine chest radiographs necessary after B-mode
and colour Doppler sonography check? Pediatr Radiol
2006;36:1252-6.
  7. Food and Drug administration. Precautions necessary with
central venous catheter. Washington DC; US government
printing office: 1989. p. 15-6.
 8. Gladwin MT, Slonim A, Landucci DL, Gutierrez DC,
Cunnion RE, Cannulation of the internal jugular vein: is
postprocedural chest radiography always necessary? Crit
Care Med 1999;27:1819-23.
  9. Matsushima K, Frankel HL. Bedside ultrasound can safely
eliminate the need for chest radiographs after central ve-
nous catheter placement: CVC sono in the surgical ICU
(SICU). J Surg Res 2010;163:155-61.
10. Maury E, Guglielminotti J, Alzieu M, Guidet B, Offenstadt
G. Ultrasonic examination: an alternative to chest radiog-
raphy after central venous catheter insertion? Am J Respir
Crit Care Med 2001;164:403-5.
11. Zanobetti M, Coppa A, Bulletti F, Piazza S, Nazerian P,
Conti A et al. Verification of correct central venous cath-
eter placement in the emergency department: comparison
between ultrasonography and chest radiography. Intern
Emerg Med 2013;8:173-80.
12. Kim SC, Heinze I, Schmiedel A, Baumgarten G, Knuefer-
mann P, Hoeft A et al. Ultrasound confirmation of central
venous catheter position via a right supraclavicular fossa
view using a microconvex probe: an observational pilot
study. Eur J Anaesthesiol 2015;32:29-36.
13. Pittiruti M, Bertollo D, Briglia E, Buononato M, Capozzoli
G, De Simone L et al. The intracavitary ECG method for
positioning the tip of central venous catheters: results of an
Italian multicenter study. J Vasc Access 2012;13:357-65.
14. Wen M, Stock K, Heemann U, Aussieker M, Küchle C.
Agitated Saline Bubble-Enhanced Transthoracic Echocardi-
ography: A Novel Method to Visualize the Position of Cen-
tral Venous Catheter. Crit Care Med 2014; e231-3.
15. Vezzani A, Brusasco C, Palermo S, Launo C, Mergoni M,
Corradi F. Ultrasound localization of central vein catheter
and detection of postprocedural pneumothorax: an alterna-
tive to chest radiography. Crit Care Med 2010; 38:533-8.
16. Rath GP, Bithal PK, Toshniwal GR, Prabhakar H, Dash
HH. Saline flush test for bedside detection of misplaced
subclavian vein catheter into ipsilateral internal jugular
vein. Br J Anaesth 2009;102:499-502.
17. Lichtenstein DA, Menu Y. A bedside ultrasound sign ruling
out pneumothorax in the critically ill. Lung sliding. Chest
1995;108:1345-8.
18. Jauss M, Zanette E. Detection of right-to-left shunt with
ultrasound contrast agent and transcranial Doppler sonog-
raphy. Cerebrovasc Dis 2000;10:490-6.
19. De Long ER, De Long DM, Clarke-Pearson DL. Compar- �������
ing the Areas under two or more correlated receiver oper-
ating characteristics curves: a nonparametric approach. Bi-
ometrics 1998;44:837-845.
20. Flahault A, Cadilhac M, Thomas G.Sample size calculation
should be performed for design accuracy in diagnostic test
studies. J Clin Epidemiol 2005;58:859-62.
21. American Society of Anesthesiologists Task Force on Cen-
tral Venous Access, �����������������������������������
Rupp SM, Apfelbaum JL, Blitt C, Ca-
plan RA, Connis RT et al. Practice guidelines for central
MEGGIOLARO CONFIRMATION OF CVC POSITION BY BUBBLE TEST

venous access: a report by the American Society of Anesthe-
siologists Task Force on Central Venous Access. Anesthesi-
ology 2012;116:539-73.
22. Fletcher SJ, Bodenham AR. Safe placement of central ve-
nous catheters: where should the tip of the catheter lie? Br J
Anaesth 2000;85:188-91.
23. Shapiro MJ, Allen HM, Talpos GB. Internal thoracic vein
cannulation as a complication of central venous catheteriza-
tion. Am Surg 1982;48:408-11.
24. Venugopal AN, Koshy RC, Koshy SM. Role of chest X-ray
in citing central venous catheter tip: a few case reports with
a brief review of the literature. J Anaesthesiol Clin Pharma-
col 2013;29:397-400.
25. Ruesch S, Walder B, Tramèr MR. Complications of central
venous catheters: internal jugular versus subclavian access –
A systematic review. Crit Care Med 2002;454-60.
26. Solorzano P, Salazar A, Brogly N, Schiraldi R., Guasch E,
Gilsanz F. Central venous catheter malposition in surgical
patients: a retrospective observational study. Eur J Anaes-
thesiol 2013;30:241.

Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on May 17, 2014 - Accepted for publication on November 4, 2014. - Epub ahead of print on November 6, 2014.
Corresponding author: M. Meggiolaro, Division of Anesthesiology and Intensive Care, University Hospital, Via Giustiniani, 2, 31100
Padova, Italy. Tel: +390498212745. E-mail: meggiomarco@libero.it

1000 MINERVA ANESTESIOLOGICA September 2015



E X PE RT O P I N I O N

Continuous regional anesthesia and

inflammation: a new target
I. GROSU, P. LAVAND’HOMME

Department of Anesthesiology, St Luc Hospital UCL Medical School, Brussels, Belgium

ABSTRACT
Inflammation can be defined as the host response when confronted with an aggression. The purpose of the inflam-
matory reaction is the defense of the host for re-establishing the baseline homeostasis of the organism. Compared
to the neuroendocrine changes associated to the stress response to injury, the inflammatory reaction is the major
determinant of patient’s recovery in the perioperative period. Perioperative inflammation is involved in the occur-
rence of various postoperative adverse outcomes other than only acute pain. By consequence, perioperative strategies
which limit or control the inflammatory response might have beneficial effects on patient’s recovery. The present
review summarizes the current knowledges on the interactions between some of these strategies, specifically regional
anesthesia (RA) techniques, and inflammation in the context of perioperative medicine. Regional anesthesia through
its components i.e. local anesthetics and analgesic adjuvants like alpha-2 adrenergic agonists (clonidine, dexmedeto-
midine) modulates the inflammatory response consecutive to tissue injury by various mechanisms, at different levels.
While experimental studies have shown that RA techniques modulate both local and systemic inflammatory reac-
tions, in contrast, clinical findings are inconsistent as actual RA techniques fail to impact major patients’ outcomes
beyond immediate postoperative analgesia. The discrepancy between experimental findings and clinical observations
asks questions and argues for a different view of perioperative inflammatory processes, in other words for an indi-
vidualized management of the patients. Future developments of tools to quantify inflammatory and immune profile
of patients might certainly lead to exciting findings and to major improvements in perioperative medicine.
(Minerva Anestesiol 2015;81:1001-9)
Key words: Inflammation - Perioperative care - Anesthesia.

Perioperative inflammation: physiology help to destroy infective agents, to prevent fur-

and clinical consequences
Inflammation can be defined as the host re-
sponse when confronted with an aggression. The
purpose of the inflammatory reaction is the de-
fense of the host for re-establishing the baseline
homeostasis of the organism. Deficiencies or ex-
cesses of the inflammatory response cause mor-
bidity and shorten lifespan.1, 2
Inflammation is a complex phenomenon
which involves many cells and mediators, both
locally and systemically. Local mediators like
cytokines synthetized at the site of tissue injury
ther tissue damage and to activate wound heal-
ing. Inflammation and pain are interrelated as
the local release of proinflammatory mediators
also leads to peripheral sensitization, hence, pain
augmentation (i.e. hyperalgesia). Some media-
tors are released into the circulation where they
will induce a systemic inflammatory response
(SIR) and act on distant organs e.g. interleukin
(IL)-6 which will induce the synthesis of acute-
phase response proteins in the liver. Some oth-
ers like interleukin (IL)-1β crosses the blood
brain barrier and will induce a state of inflam-
mation, “neuroinflammation”, into the central

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1001

GROSU CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION

nervous system (CNS). Microglia (the immune
cells of CNS) is first activated then astrocytes
(the support cells of neurons), those later show-
ing a durable activation.3 Neuroinflammation
is one of the mechanisms underlying central
sensitization process.3 Whether acute inflamma-
tion is essential for recovery from tissue injury,
when it persists, inflammatory processes underly
both the development and the maintenance of
chronic diseases like atherosclerosis, cancer, de-
mentia, obesity, asthma, chronic pain.2 Because
of the deleterious consequences of uncontrolled
inflammation, the inflammatory response is a
highly regulated process. Termination of the in-
flammatory reaction is no more seen as a passive
process by elimination of local pro-inflammato-
ry mediators but is considered as an active pro-
cess which involves the synthesis of specific pro-
resolving mediators (e.g. resolvins, protectins,
maresins…) stimulating the resolution without
immune suppression.2
Tissue lesions combined to anesthesia drugs
and techniques are responsible for the impor-
tant inflammatory reaction and the suppres-
sion of cell-mediated immunity observed dur-
ing the perioperative period.1 Compared to the
neuroendocrine changes associated to the stress
response to injury, the inflammatory reaction is
the major determinant of patient’s recovery in
the perioperative period. Perioperative inflam-
mation is involved in the occurrence of various
postoperative adverse outcomes other than acute
pain:4 persistent pain, fatigue and delirium, car-
diovascular events, cancer recurrence. By conse-
quence, perioperative strategies which limit or
control the inflammatory response might have
beneficial effects on patient’s recovery. The pur-
pose of the present review is to summarize the
current knowledges on the interactions between
some of these strategies, specifically regional an-
esthesia (RA) techniques, and inflammation in
the context of perioperative medicine (Figure 1).

Preoperative proinflammatory profile:

low-grade systemic inflammation, ageing, etc.

Peroperative inflammation caused by

surgery (tissue injury)
Anesthesia drugs and techniques

Postoperative resolution of inflammation

Early and delayed effects on recovery

Acute pain Poor outcome:

fatigue, chronic postsurgical pain,
cancer recurrence

Major adverse events:

cardiovascular events
RiskCancer
of pre-existing disease
recurrence
(MI, stroke), delirium, etc. progression ?
(Alzheimer disease, osteoarthritis…)

Figure 1.—Consequences of perioperative inflammation.

1002 MINERVA ANESTESIOLOGICA September 2015

CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION GROSU
Mechanisms by which RA modulates the
perioperative inflammatory response
RA is a very popular technique, used alone or
more frequently, in combination with general
anesthesia to provide satisfactory postoperative
analgesia in major thoracoabdominal and ortho-
pedic procedures. To date, however, the benefits
of RA with regards to perioperative morbidity
and mortality reduction continue to be debated.5
Further, the benefits on later outcomes which are
far different than postoperative pain control and
which will condition patient’s rehabilitation and
return to preoperative situation are not clear.5 As
previously mentioned, individual’s inflamma-
tory response plays a major role in most if not all
of these outcomes. Whether the impact of RA
on the inflammatory response has been studied
in various experimental models, well-designed
clinical studies are scarce bringing unconclusive
results. Nevertheless, without doubt, RA can
modulate the inflammatory response consecu-
tive to tissue injury by various mechanisms, at
different levels.
Local anesthetics (LAs) are the major com-
ponent of RA techniques. The most cited anti-
inflammatory mechanism of LAs is the interrup-
tion of nociceptive transmission at the injury site
and the reduction of “neurogenic inflammation”.
LAs block transmembrane proteins voltage-gated
sodium channels on nociceptors sensitized by lo-
cally released pro-inflammatory mediators (pros-
taglandins E2, interleukins, etc.). Small-diame-
ter primary afferent fibers when sensitized also
release pronociceptive neuropeptides (substance
P, CGRP, neurokinin) within their immediate
vicinity through an axon reflex mechanism. By
consequence, reduction of neurogenic inflam-
mation by LAs will modulate both peripheral
and central sensitizations processes as local neu-
Table I.—Mechanisms involved in modulation of inflammation by local anesthetics and analgesic adjuvants.
Effects on PMN
↓ leukocyte adherence
↓ migration through the endothelium
↓ phagocytosis
↓ ICAM-1expression
↓ the release of chemo-attracting agents
↓ superoxide anion production of PAF-primed PMN
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1003
rogenic inflammation participates to the general
inflammatory response.6
Besides, LAs also possess intrinsic anti-inflam-
matory properties of which all the mechanisms
are not completely elucidated yet but differ from
that of sodium channels blockade.7 LAs modu-
late various steps of the inflammatory cascade
as summarized in Table I. In the context of RA
techniques, LAs may act on regular immune
cells e.g. macrophages recruited on injury site as
well as they may interact with resident cells like
keratinocytes and fibroblasts which are critically
involved in local inflammatory and healing proc-
esses.8, 9 Furthermore, when used for RA, LAs
may also be absorbed into the systemic circula-
tion. LAs attenuate the excessive inflammatory
response without impairing the physiological
host defense in contrast with other immunosup-
pressant drugs.10 Leucocytes have no sodium
voltage-dependent channels yet their immune
function is influenced by LAs at more than one
level. Priming of leucocytes is currently consid-
ered as playing a major role in the “overstimula-
tion” of inflammatory pathways. The majority of
inflammatory mediators (tumor necrosis factor
–TNF-α, lipopolysaccharide [LPS], platelet ac-
tivating factor [PAF], interleukin [IL-8]) signal
through G-protein coupled receptors to enhance
the response of leukocytes. LAs significantly at-
tenuate G-protein mediated human neutrophil
priming,11 the most potent inhibition being
achieved by ester compounds and being inverse-
ly correlated with the octanol-buffer coefficient
of the LA. Among other mechanisms involved in
the anti-inflammatory effect of LAs, protection
of the endothelial barrier reduces neutrophils ad-
hesion and endothelial hyperpermeability caus-
ing major organ dysfunction. Both ropivacaine
and lidocaine effectively block pro-inflammato-
ry TNF-α signalling in endothelial cells.12 This
Effects on the production of proinflammatory cytokines
↓ NF-kB signaling
↓ the axonal transport of TNF α
↓ the TNF α receptor expression in DRG
↓ PG biosynthesis and the release of TX B2 and LT B4
↓ synthesis of IL1 α, IL 1 β, IL 2, IL 8
↓ synthesis of TNF α, IFN ϒ
GROSU CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION
mechanism may provide therapeutic benefit in
acute inflammatory conditions including in the
context of carcinologic surgery.10 It is also worth
noting that LAs, specifically lidocaine, modu-
lates central neuroinflammation by decreasing
the production of proinflammatory cytokines by
microglial cells.13
Analgesic adjuvants used in RA techniques
may also interfere with the inflammatory reac-
tion at different levels. Among these adjuvants,
α2-adrenergic agonists clonidine and more re-
cently dexmedetomidine have been the most
studied. Their capacity to modulate the inflam-
matory response caused by tissue injury mainly
relies on α2-adrenergic receptors binding on
stimulated proinflammatory cells like macro-
phages and leucocytes,14 yielding to reduction
of the expression of Toll-like receptors (TLR) 4,
hence to reduced activation and translocation of
nuclear transcription factor NF-κB and subse-
quent decrease of pro-inflammatory cytokines
production (Table I).15, 16
How different RA techniques affect the
perioperative inflammatory response in
animal models and clinical conditions
Two important concerns need to be raised
preliminary. First, the relevance of the animal
models used to the mimic clinical conditions.
Carrageenan injection, i.e. a chemical irritant, is
a pure inflammatory model not really represent-
ative of incisional pain. More, it is mandatory
to take into account major immune differences
between rodents and humans, yielding to con-
sider with caution extrapolation of animal data
to patients. Second, inflammatory parameters
in the majority of clinical studies are biased by
the concommittant use of drugs which interfere
with the immune and inflammatory responses.
The most examplative is the administration of
high doses of opioids in control groups without
RA. Opioids display imuno-supressive effects
and proinflammatory responses in the central
nervous system likely via activation of micro-
glia.3 By consequence, better outcomes for the
patients who benefit of RA could actually be
translated as better outcomes for patients with
no or less opioid intake (i.e. balanced anesthesia
1004 MINERVA ANESTESIOLOGICA September 2015
using opioid sparing regimens). So more studies
should be performed, comparing RA with either
opioid-sparing or opioid-free general anesthesia
in order to state a clear anti-inflammatory effect
of RA techniques. A recent review on preven-
tive analgesia by local anesthetics administered
by either peripheral nerve blocks or intravenous
route suggests that the significant analgesic and
antihyperalgesic effect occurs for their use dur-
ing the postoperative period.17 It is worth noting
that other outcomes different from only acute
pain control have not be considered here and
that, the intraoperative use of RA is mandatory
to maximize the opioids sparing effect, hence the
anti-inflammatory effect, of the technique.18
With RA, LAs are partly absorbed in the sys-
temic circulation. For anti-inflammatory effects
of systemic LAs, special attention has been at-
tributed to intravenous lidocaine administra-
tion, particularly in major abdominal surgery.
Postoperative ileus and systemic inflammation
caused by surgery increase intestinal epithelial
permeability which are associated with bacterial
translocation and contribute to multiple organ
failure. Meta-analysis have confirmed the early
benefits of intravenous lidocaine in patients un-
dergoing abdominal procedures but surprisingly
not in those undergoing other types of surger-
ies like cardiac or orthopaedic procedures (Table
II).19
As intravenous lidocaine is also called “the
poor man’s epidural”, it is interesting to deter-
mine the benefit related to the use of neuraxial
blocks. In experimental conditions, epidural
bupivacaine attenuates mesenteric ischemia-
reperfusion related inflammatory response and
intestinal damage.20 Epidural lidocaine anesthe-
sia also reduces endotoxin-induced gut epithelial
injury by decreasing monocytic extravasation
and intestinal nitrosative stress.21 In patients un-
dergoing colonic surgery, thoracic epidural anal-
gesia with lidocaine provides better postopera-
tive pain relief and opioid sparing effect, earlier
return of bowel function and lesser production
of inflammatory cytokines than intravenous li-
docaine during 72 hours.22 The greater benefit
of epidural lidocaine over systemic lidocaine cer-
tainly questions the role of the central sympa-
thetic nervous system in controlling the systemic
CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION GROSU

Table II.—Modulation of perioperative inflammatory reaction by local anesthetics in experimental and clinical conditions:
comparison of RA and systemic administration.
Clinical parameters Biological parameters
Edema, Pain, Local Systemic Spinal PGE2
redness hyperalgesia cytokines cytokines
Abdominal visceral surgery
Laparotomy model – rat 34 CWI ropivacaine NA + 0 0 NA
Systemic NA + + ++ NA
ropivacaine
Laparotomy – human 22 Epidural lidocaine NA ++ NA ++ NA
Intravenous lidocaine NA + NA + NA
Orthopedic limb surgery
Carrageenan model – rat 29, 30 PNB bupivacaine + ++ + ++ +
Systemic bupivacaine 0 0 0 ++ 0
Knee arthroplasty – human 31, 32 PNB bupivacaine / + ++ 0 0 NA
ropivacaine
RA: regional anesthesia; CWI: continuous wound infusion; PNB: peripheral nerve block.
(+) positive modulatory effect of inflammatory parameters i.e. reduction of inflammation; (0) no modulatory effect of inflammatory parameters.
NA: data not available.

inflammatory response.23 Regional sympathetic
block increases intestinal perfusion. Further, the
sympathetic nervous system shows overactivity
under stress conditions i.e. surgical trauma and
RA may help to modulate this overeactivity and
to re-establish the immune balance in the perio-
perative period. Most of the studies assessing the
anti-inflammatory effect of neuraxial anesthesia
and analgesia have compared the combination of
epidural with general anesthesia to general anes-
thesia alone. The results have been disappointing,
showing only partial and temporary effect on the
acute inflammatory stress response in term of cy-
tokines expression.24 Spinal anesthesia exhibits
perioperative anti-inflammatory properties that
seem to be moderate and transient when com-
pared with epidural anesthesia. Spinal anesthesia
is often used for relatively short lasting proce-
dures. By example, in patients undergoing knee
arthroplasty, no difference was found regarding
systemic modulation of the inflammatory re-
sponse between spinal anesthesia followed by
systemic morphine analgesia and epidural anes-
thesia followed by postoperative epidural analge-
sia.25 Regarding the analgesic adjuvants, it is well
known that α2-adrenergic agonists blunt central
sympathetic outflow. Compared with epidural
bupivacaine, epidural clonidine demonstrates
greater anti-inflammatory effects (reduction of
systemic proinflammatory cytokines expres-
sion) after colorectal surgery, what lets suppose
that the modulation of the sympathetic nervous
system effectively plays a role in controlling the
SIR as aforementioned.26, 27 More, the reduc-
tion of the sympathetic outflow may increase the
role of the parasympathetic system. The cholin-
ergic anti-inflammatory pathway through the
stimulation of the vagus nerve is a well-known
mechanism of neural inhibition of inflamma-
tory response. Acetylcholine, the principal neu-
rotransmitter of the vagus, dose-dependently in-
hibits the release of pro-inflammatory cytokines
by macrophages.28
A series of experimental studies using the
carrageenan model of peripheral inflammation
in rats have questioned the anti-inflammatory
properties of peripheral nerve blocks.29 The
authors have demonstrated that LAs act at dif-
ferent levels when administered systemically or
via a nerve block. To summarize, LAs peripheral
nerve blocks decrease both paw edema and hy-
peralgesia caused by local tissue inflammation,
even when the nerve block is performed on the
contralateral limb. Systemic absorption of LAs
partly may explain the anti-inflammatory effect
of a contralateral nerve block. Both systemic ad-
ministration of LAs and perineural one reduce
the systemic inflammatory response (SIR) while
systemic LAs have no effect on paw edema and
local hyperalgesia.30 Spinal increase of prosta-
glandin (PG) E2 correlates with pain and hyper-
algesia after tissue injury in animal models and

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1005

GROSU CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION
clinical situations.4 After hindpaw carrageenan
injection, bupivacaine sciatic block inhibits
COX-2 expression in both dorsal root ganglion
and spinal cord, leading to a reduction of spinal
PGE2, while systemic bupivacaine administra-
tion does not modify COX-2 expression.30 In
humans, the anti-inflammatory effects of pe-
ripheral nerve blocks have been explored after
major orthopaedic procedures like knee arthro-
plasty.31, 32 Despite the use of continuous nerve
blocks with bupivacaine or ropivacaine, the
inflammatory parameters i.e. systemic and lo-
cal cytokines expression were not really affected
while knee edema, early postoperative pain and
mobilization were improved in the block groups.
Peri-neural administration of clonidine in the
context of nerve injury reduces the development
of hypersensitivity by altering the balance of pro-
and anti-inflammatory cytokines locally released
during the process of Wallerian degeneration.14
Drug effect is mediated through binding to α2-
adrenergic receptors expressed on local mac-
rophages. Similar findings have been recently
reported for dexmedetomidine, which is more
selective for α2-adrenergic receptors than clo-
nidine.15 Here, it is interesting to mention that
systemic administration of clonidine improves
posthypoxic endothelial function and modulates
inflammation during limb reperfusion in both
animal models and healthy volunteers.33
Wound infiltration is a non-invasive, easy
technique often considered as useful adjunct
to multimodal analgesia for postoperative pain
management as it reduces pain and shows an
opioid sparing effect. The catheter location is
important because of the role played by deep
structures into postoperative pain e.g. deep mus-
cular layers, peritoneum. Today, pre-peritoneal
infusion of LAs is considered as an alternative
to epidural analgesia for abdominal procedures.
Using an animal model of laparotomy, Kfoury et
al.34 have tried to explain the mechanisms of ac-
tion involved in intrawound LAs analgesic effect.
They have demonstrated that both high doses of
preperitoneal infusion of ropivacaine and sys-
temic ropivacaine similarly reduced mechanical
and visceral sensitivity while systemic adminis-
tration alone was associated with an anti-inflam-
matory effect (reduced stimulated production of
1006 MINERVA ANESTESIOLOGICA September 2015
TNF-α and IL-1β in whole blood and perito-
neal macrophages cultures). In humans, there
is a lack of correlation between systemic i.e. se-
rum levels and local expression of intra wound
cytokines and PGs after caesarean section.8 In
the same clinical model, continuous intrawound
infusion of bupivacaine seemed to enhance the
local proinflammatory response while showing
clinical analgesic effects 9. Recent experimental
data have highlighted the proinflammatory ef-
fect of bupivacaine on peripheral nerve and the
protective effect of the α2-adrenergic agonist
dexmedetomidine in that setting.35 Moreover,
local injection of dexmedetomidine counter-
acts the inflammatory reaction caused by in-
traplantar carrageenan in rats as it reduces paw
edema, leucocytes activation and decreases the
local production of TNF-α and the local expres-
sion of COX-2.36
RA, inflammatory response and patient’s
outcomes in clinical practice
While it is clear from experimental data that
RA display anti-inflammatory and immunologic
modulatory properties, in clinical practice the
impact on perioperative inflammation is not so
evident. First, there is clearly a dissociation be-
tween the clinical effects reported (i.e. reduction
of local edema and acute pain) and the modu-
lation of the biologic markers of inflammation
(i.e. local cytokines expression), particularly in
orthopaedic setting. Second, RA benefits in term
of patients’ major outcomes, other than acute
pain control (e.g. persistent pain and cancer re-
currence) are unconclusive. Epidural anesthesia
and paravertebral block, respectively, may pre-
vent persistent postsurgical pain after thoracot-
omy and breast surgery in about one out every
four to five patients.37 In orthopedic procedures,
where long-lasting preoperative pain and inflam-
mation are very common, the benefits are even
less evident as 4-days continuous peripheral
nerve blocks only improve rehabilitation at 6
weeks after hip or knee arthroplasty but not later
and do not prevent the development of persist-
ent pain.38 Inflammation not only plays a central
role at different stages of tumor development but
also contributes to the initiation of antitumoral
CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION GROSU
immune response.39 In regard to that complex
situation, it is not surprising that systematic re-
views based on retrospective studies have found
little evidence to support major benefit of RA on
survival after oncologic surgery.40, 41
Unsolved issues and questions deserving
future research in the field
The poor clinical results certainly incite us to
question the reasons for RA failure to influence
longterm outcomes. First, the severity and the
duration of ongoing perioperative inflammatory
processes may differ from one individual to the
other. Experimental data have clearly shown that
a prolonged nerve block is necessary to decrease
the systemic consequences of the local inflam-
matory reaction 6 as well as only repeated peri-
neural injections of clonidine modulate local
cytokines expression yielding to longlasting de-
crease of hypersentisivity 14. Under clinical con-
ditions, prolonged peripheral nerve blocks i.e.
lasting longer than the usual postoperative dura-
tion (2 to 4 days) might be necessary to control
perioperative inflammation and to demonstrate
beneficial effect on patient’s outome as demon-
strated in the context of limb amputation (medi-
an duration of sciatic block of 30 days; extreme
values 3 to 80 days).42 Second, studies have not
taken into account the preoperative basal status
of the patients while several preoperative condi-
tions are “inflammatory prone” like asthma, ob-
sesity, rheumatoid arthritis, inflammatory bowel
diseases, what could have blunted the impact of
some RA techniques. The widespread use of ul-
trasounds has contributed to RA success in both
acute and chronic pain settings. In the periop-
erative period, the anti-inflammatory impact
of techniques like TAP block has not yet been
assessed. Furthermore, when used for RA, LAs
may also be absorbed into the systemic circula-
tion. The use of ultrasound-guided RA has led
to a decrease in the volume of LAs administered,
reducing not only the risk of systemic toxicity
but perhaps also blunting expected benefits of
LAs systemic concentrations. Finally, clinical
observations also question the value of clinical
markers currently used to quantify RA effect on
the perioperative inflammation as major discrep-
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1007
ancies between local and systemic expression of
common cytokines are usually noted, and clini-
cal reduction of pain and inflammation does
not really correlates with decrease in biological
makers.31, 32 Such inflammation-based scores
are already used to assess the prognostic of can-
cer patients 43 but prospective studies aimed to
validate them in perioperative situations are cur-
rently missing. It is also interesting to note that
RA impact on the mediators involved in the ac-
tive resolution of inflammation has not yet been
assessed.2
Conclusions
Compared to the neuroendocrine changes
associated to the stress response to injury, the
inflammatory reaction is the major determi-
nant of patient’s recovery in the perioperative
period. Uncontrolled inflammation is involved
in the majority of adverse outcomes after sur-
gery. Experimental studies have shown that RA
techniques using LAs and analgesic adjuvants
modulate both local and systemic inflamma-
tory reaction caused by surgical injury. In con-
trast, clinical findings are inconsistent as actual
RA techniques fail to impact major patients’
outcomes beyond immediate postoperative an-
algesia. The discrepancy between experimental
findings and clinical observations ask questions
and argues for a different view of perioperative
inflammatory processes, in other words for indi-
vidualized management of the patients. Future
developments of tools to quantify inflammatory
and immune profile of patients might certainly
lead to exciting findings and to major improve-
ments in perioperative medicine.
Key messages
—— Local anesthetics (LAs) are the major
component of RA technique. LAs-related
anti-inflammatory mechanisms involve both
the interruption of nociceptive transmission
(reduction of neurogenic inflammation) and
intrinsic anti-inflammatory properties un-
related to sodium channels blockade (direct
action on immune cells). By contrast with
GROSU CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION
immunosuppressant drugs, LAs attenuate
the excessive inflammatory response without
impairing the physiological host defences.
—— RA-related modulation of periop-
erative inflammation also relies on a well-
known perioperative opioid-sparing effect
as opioids display immuno-suppressive and
pro-inflammatory properties. To state a clear
anti-inflammatory effect of RA techniques,
well-designed studies should compare RA
with opioid-sparing general anesthesia.
—— Analgesic adjuvants used in combina-
tion with LAs in RA also possess intrinsic
anti-inflammatory and immune-modulatory
properties (e.g. alpha-2 adrenergic agonists
clonidine and dexmedetomidine).
—— Clinical observations mainly dem-
onstrate the anti-inflammatory modulatory
effects of RA as a reduction of acute post-
operative pain and hyperalgesia although the
relationship between clinical markers of the
inflammatory reaction and biological ones
remains unclear. Anti-inflammatory impact
on major delayed outcomes like persistent
pain, cancer recurrence, progression of in-
flammatory diseases… is still unknown and
mostly unexplored.
References
  1. Nicholson G, Hall GM. Effects of anaesthesia on the in-
flammatory response to injury. Curr Opin Anaesthesiol
2011;24:370-4.
  2. Serhan C, Chiang N. Resolution phase lipid mediators of
inflammation: agonists of resolution. Curr Opin Pharma-
cology 2013;18:632-40.
 3. Watkins LR, Hutchinson MR, Milligan ED, Maier SF.
“Listening” and “talking” to neurons: implications of im-
mune activation for pain control and increasing the efficacy
of opioids. Brain Res Rev 2007;56:148-69.
  4. Buvanendran A, Kroin JS, Berger RA, Hallab NJ, Saha C,
Negrescu C et al. Upregulation of prostaglandin E2 and
interleukins in the central nervous system and peripheral
tissue during and after surgery in humans. Anesthesiology
2006;104:403-10.
  5. Curatolo M. Adding regional analgesia to general anaesthe-
sia: increase of risk or improved outcome? Eur J Anaesthe-
siol 2010;27:586-91.
  6. Pham-Marcou TA, Gentili M, Asehnoune K, Fletcher D,
Mazoit JX. Effect of neurolytic nerve block on systemic
carrageenan-induced inflammatory response in mice. Br J
Anaesth 2005;95:243-6-
 7. Hollmann MW, Durieux ME.L ocal anesthetics and the
inflammatory response:a new therapeutic indication? An-
esthesiology 2000;93:858-75.
  8. Carvalho B, Clark DJ, Angst MS. Local and systemic re-
1008 MINERVA ANESTESIOLOGICA September 2015
lease of cytokines, nerve growth factor, prostaglandin E2,
and substance P in incisional wounds and serum following
cesarean delivery. J Pain 2008;9:650-7.
  9. Carvalho B, Clark DJ, Yeomans DC, Angst MS. Continu-
ous subcutaneous instillation of bupivacaine compared
to saline reduces interleukin 10 and increases substance p
in surgical wounds after cesarean delivery. Anesth Analg
2010;111:1452-9.
10. Votta-Velis EG, Piegeler T, Minshall RD, Aguirre J, Beck-
Schimmer B, Schwartz DE et al. Regional anaesthesia and
cancer metastases: the implication of local anaesthetics.
Acta Anaesthesiol Scand 2013;57:1211-29.
11. Picardi S, Cartellieri S, Groves D, Hahnenkamp K, Gerner
P, Durieux ME et al. Local anesthetic-induced inhibition of
human neutrophil priming: the influence of structure, li-
pophilicity, and charge. Reg Anesth Pain Med 2013;38:9-15.
12. Piegeler T, Votta-Velis EG, Bakhshi FR, Mao M, Carnegie
G, Bonini MG et al. Endothelial barrier protection by local
anesthetics: ropivacaine and lidocaine block tumor necrosis
factor-alpha-induced endothelial cell src activation. An-
esthesiology 2014;120:1414-28.
13. Jeong HJ, Lin D, Li L, Zuo Z.D elayed treatment with
lidocaine reduces mouse microglial cell injury and cytokine
production after stimulation with lipopolysaccharide and
interferon gamma. Anesth Analg 2012;114:856-61.
14. Lavand’homme PM, Eisenach JC.Perioperative administra-
tion of the alpha2-adrenoceptor agonist clonidine at the site
of nerve injury reduces the development of mechanical hy-
persensitivity and modulates local cytokine expression. Pain
2003;105:247-54.
15. Huang Y, Lu Y, Zhang L, Yan J, Jiang J, Jiang H.
Perineural dexmedetomidine attenuates inflammation in
rat sciatic nerve via the NF-kappaB pathway. Int J Mol Sci
2014;15:4049-59
16. Wu Y, Liu Y, Huang H, Zhu Y, Zhang Y, Lu F et al. Dexme-
detomidine inhibits inflammatory reaction in lung tissues
of septic rats by suppressing TLR4/NF-kappaB pathway.
Mediators Inflamm 2013:562154.
17. Barreveld A, Witte J, Chahal H, Durieux ME, Strichartz G.
Preventive analgesia by local anesthetics: the reduction of
postoperative pain by peripheral nerve blocks and intrave-
nous drugs. Anesth Analg 2013;116:1141-61.
18. Richebe P, Rivat C, Liu SS. Perioperative or postopera-
tive nerve block for preventive analgesia: should we care
about the timing of our regional anesthesia? Anesth Analg
2013;116:969-70.
19. McCarthy GC, Megalla SA, Habib AS.I mpact of intra-
venous lidocaine infusion on postoperative analgesia and
recovery from surgery: a systematic review of randomized
controlled trials. Drugs 2010;70:1149-63.
20. Bedirli N, Akyurek N, Kurtipek O, Kavutcu M, Kartal S,
Bayraktar AC. Thoracic epidural bupivacaine attenuates
inflammatory response, intestinal lipid peroxidation, oxida-
tive injury, and mucosal apoptosis induced by mesenteric
ischemia/reperfusion. Anesth Analg 2011;113:1226-32.
21. Schaper J, Wagner A, Enigk F, Brell B, Mousa SA, Habazet-
tl H, Schafer M. Regional sympathetic blockade attenuates
activation of intestinal macrophages and reduces gut barrier
failure. Anesthesiology 2013;118:134-42.
22. Kuo CP, Jao SW, Chen KM, Wong CS, Yeh CC, Sheen MJ,
Wu CT:Comparison of the effects of thoracic epidural anal-
gesia and i.v. infusion with lidocaine on cytokine response,
postoperative pain and bowel function in patients undergo-
ing colonic surgery. Br J Anaesth 2006;97:640-6
23. Ganapathy S, Murkin JM, Dobkowski W, Boyd D. Stress
and inflammatory response after beating heart surgery ver-
sus conventional bypass surgery: the role of thoracic epi-
dural anesthesia. Heart Surg Forum 2001;4:323-7.
24. Fant F, Tina E, Sandblom D, Andersson SO, Magnuson A,
Hultgren-Hornkvist E et al. Thoracic epidural analgesia in-
hibits the neuro-hormonal but not the acute inflammatory
CONTINUOUS REGIONAL ANESTHESIA AND INFLAMMATION GROSU
stress response after radical retropubic prostatectomy. Br J
Anaesth 2013;110:747-57.
25. Chloropoulou P, Iatrou C, Vogiatzaki T, Kotsianidis I,
Trypsianis G, Tsigalou C et al. Epidural anesthesia followed
by epidural analgesia produces less inflammatory response
than spinal anesthesia followed by intravenous morphine
analgesia in patients with total knee arthroplasty. Med Sci
Monit 2013;19:73-80.
26. Persec J, Persec Z, Husedzinovic I. Postoperative pain and
systemic inflammatory stress response after preoperative
analgesia with clonidine or levobupivacaine:a randomized
controlled trial. Wien Klin Wochenschr 2009;121:558-63.
27. Wu CT, Jao SW, Borel CO, Yeh CC, Li CY, Lu CH et al.
The effect of epidural clonidine on perioperative cytokine
response, postoperative pain, and bowel function in patients
undergoing colorectal surgery. Anesth Analg 2004;99:502-
9, table of contents.
28. Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina
GI, Watkins LR et al. Vagus nerve stimulation attenuates
the systemic inflammatory response to endotoxin. Nature
2000;405:458-62.
29. Beloeil H, Mazoit JX. [Effect of local anesthetics on the
postoperative inflammatory response]. Ann Fr Anesth Re-
anim 2009;28:231-7.
30. Beloeil H, Gentili M, Benhamou D, Mazoit JX. The effect
of a peripheral block on inflammation-induced prostaglan-
din E2 and cyclooxygenase expression in rats. Anesth Analg
2009;109:943-50.
31. Bagry H, de la Cuadra Fontaine JC, Asenjo JF, Bracco D,
Carli F. Effect of a continuous peripheral nerve block on
the inflammatory response in knee arthroplasty. Reg Anesth
Pain Med 2008;33:17-23.
32. Martin F, Martinez V, Mazoit JX, Bouhassira D, Cherif
K, Gentili ME et al. Antiinflammatory effect of peripheral
nerve blocks after knee surgery: clinical and biologic evalu-
ation. Anesthesiology 2008;109:484-90.
33. Gourdin M, Dubois P, Mullier F, Chatelain B, Dogne JM,
Marchandise B et al. The effect of clonidine, an alpha-2
adrenergic receptor agonist, on inflammatory response
and postischemic endothelium function during early
reperfusion in healthy volunteers. J Cardiovasc Pharmacol
2012;60:553-60.
Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on April 22, 2014.- Accepted for publication on October 13, 2014. - Epub ahead of print on October 15, 2014.
Corresponding author: P. Lavand’homme, Department of Anesthesiology, St Luc Hospital UCL Medical School, Avenue Hippocrate 10-
1821, B-1200 Brussels, Belgium. E-mail: patricia.lavandhomme@uclouvain.be
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1009
34. Kfoury T, Mazoit JX, Schumacher M, Benhamou D, Be-
loeil H. A comparison of different dosages of a continu-
ous preperitoneal infusion and systemic administration
of ropivacaine after laparotomy in rats. Anesth Analg
2011;113:617-25.
35. Tufek A, Kaya S, Tokgoz O, Firat U, Evliyaoglu O, Celik
F et al. The protective effect of dexmedetomidine on bupi-
vacaine-induced sciatic nerve inflammation is mediated by
mast cells. Clin Invest Med 2013;36:E95-102.
36. Sukegawa S, Higuchi H, Inoue M, Nagatsuka H, Maeda
S, Miyawaki T. Locally injected dexmedetomidine inhibits
carrageenin-induced inflammatory responses in the injected
region. Anesth Analg 2014;118:473-80.
37. Andreae MH, Andreae DA. Local anaesthetics and re-
gional anaesthesia for preventing chronic pain after surgery.
Cochrane Database Syst Rev 2012;10:CD007105.
38. Ilfeld BM, Shuster JJ, Theriaque DW, Mariano ER, Girard
PJ, Loland VJ et al. Long-term pain, stiffness, and function-
al disability after total knee arthroplasty with and without
an extended ambulatory continuous femoral nerve block:a
prospective, 1-year follow-up of a multicenter, randomized,
triple-masked, placebo-controlled trial. Reg Anesth Pain
Med 2011;36:116-20.
39. Forget P, De Kock M. Perspectives in anaesthesia for cancer
surgery. J Cancer Res Clin Oncol 2013;140:353-9.
40. Cata JP, Hernandez M, Lewis VO, Kurz A. Can regional
anesthesia and analgesia prolong cancer survival after or-
thopaedic oncologic surgery? Clin Orthop Relat Res
2014;472:1434-41.
41. Binczak M, Tournay E, Billard V, Rey A, Jayr C. Major
abdominal surgery for cancer: does epidural analgesia have
a long-term effect on recurrence-free and overall survival?
Ann Fr Anesth Reanim 2013;32:e81-8.
42. Borghi B, D’Addabbo M, White PF, Gallerani P, Toccaceli
L, Raffaeli W et al. The use of prolonged peripheral neural
blockade after lower extremity amputation: the effect on
symptoms associated with phantom limb syndrome. Anesth
Analg 2010;111:1308-15.
43. Bugada D, Allegri M, Lavand’homme P, De Kock M, Fanel-
li G. Inflammation-based scores:a new method for patient-
targeted strategies and improved perioperative outcome in
cancer patients. Biomed Res Int 2014;2014:142425.


E X PE RT O P I N I O N

Blood glucose amplitude variability

in critically ill patients
G. MEYFROIDT

Intensive Care Medicine, University Hospitals Leuven, Leuven, Belgium

ABSTRACT
While the discussion on the optimal blood glucose (BG) level target in critically ill patients is on-going, attention
shifts towards other aspects of the BG signal, such as hypoglycemia and blood glucose amplitude variability (BGAV).
A large number of observational and mostly retrospective studies have demonstrated an association between in-
creased BGAV and worse outcomes. This observed association could partially be explained by endogenous factors
such as changes in the status of the patient that cannot be externally influenced. On the other hand, exogenous fac-
tors such as insulin and caloric infusions could play a role in increasing or decreasing BGAV. In this review article,
intuitive concept of “variability” will be clarified, and possible metrics to quantify BGAV are discussed. Whether
it is feasible to actively minimize BGAV in order to improve the outcome of critically ill patients, is questionable.
(Minerva Anestesiol 2015;81:1010-18)
Key words: Blood glucose - Hyperglycemia - Hypoglycemia - Intensive care - Critical Illness - Insulin.

I n critically ill patients, stress-induced hyper-

glycaemia is associated with worse outcomes.
Observational studies have suggested a J-shaped
or U-shaped relationship between blood glucose
(BG) and worse outcome, with a nadir approxi-
mately between 5 and 8 mmol/L, in patients
after myocardial infarction 1, 2 and in critical ill-
ness.3, 4 Single-center interventional trials, per-
formed in Leuven, Belgium, have demonstrated
that targeting age-adjusted normoglycaemia (in
comparison with not treating BG below the re-
nal threshold), by using continuous infusions of
insulin, was able to reduce mortality and mor-
bidity, in adult surgical,5 medical,6 as well as in
paediatric 7 intensive care unit (ICU) patients.
The beneficial effect of such tight glycemic con-
trol (TGC) protocol was initially confirmed in
implementation trials, and small RCT’s using
intermediate endpoints, in selected subpopula-
tions.8-11 Multicenter trials on BG control 12-15
have not been able to replicate the results of the
Leuven trials. However, these studies were un-
derpowered to detect a significant benefit be-
tween 2 strategies. The large NICE-Sugar study,
set up with adequate statistical power,16 showed
worse outcomes in the TGC group. When look-
ing into detail at the methodology of these tri-
als, it is clear that none of them were actual
replications of the original complex Leuven in-
tervention. Different ranges for BG in control
and intervention groups have been targeted or
obtained, different routes for insulin adminis-
tration and types of infusion-pumps were used,
glucose was sampled at different sampling sites,
glucometers have been used that are inaccurate
in the critically ill,17 and there were huge differ-
ences in nutritional strategies.18 These important
methodological differences are not trivial, and
can possibly or partially explain the observed dif-
ferences in outcomes. Also, it is possible that the
optimal BG level might be different in different
patient groups, as has been suggested in a recent

1010 MINERVA ANESTESIOLOGICA September 2015

BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS MEYFROIDT
large retrospective study where non-diabetic
patients had a clear decreased risk of mortality
with TGC (4.4-6.1 mmol/L), as compared to an
intermediate target (5.0-7.8 mmol/L), whereas
in patients with pre-existing diabetes, the inter-
mediate target was associated with a lower risk
of mortality.19
Additional dimensions of the BG signal, not
taken into account in the different trials, might
serve as alternative explanations for the observed
discordant results. Not only an elevated mean or
median BG level is associated with worse out-
come in critically ill patients. Indeed, hypogly-
cemia, and BG amplitude variability (BGAV),
have both been associated with worse outcomes,
independent of BG level and severity of ill-
ness.20-22
Although hypoglycaemia has been associated
with worse outcome independent of the BG
level in many studies, Mackenzie 21 has demon-
strated that it is not completely independent of
other metrics of glucose control. Thus, it is dis-
putable whether hypoglycaemia is a true sepa-
rate dimension of BG control: hypoglycemia is
merely the lower side of the BG level spectrum,
and fits perfectly in the abovementioned J- of U-
shaped relationship between BG and outcome.
Obviously, an interventional trial were patients
are randomized to hypoglycaemia independent
of the average BG level will never be performed.
Hypoglycemia is a marker for severity of illness
and will occur more frequently in sicker patients,
but at the same time can be influenced by BG
management protocols. By aiming at levels of BG
control closer to normal fasting levels, the risk of
inducing hypoglycemic events is increased. Re-
gardless of the desired BG level target, it is ob-
vious that any BG control algorithm should be
designed to minimize the risk of hypoglycemia,
and should have special considerations for those
patients with a higher inherent probability of hy-
poglycemia. Particularly the use of well-designed
computerized TGC protocols might reduce the
incidence of hypoglycaemia while still being able
to adequately control BG levels.23
The present review article will focus on the
third dimension of BG control, namely BGAV.
Growing evidence from experimental mod-
els,24-26 and studies of patients with diabetes
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1011
mellitus,27 suggest that fluctuations in BG levels
can increase the risk of hyperglycaemia-induced
oxidative stress. First, BGAV will be defined as
a concept. Second, the different measures of
BGAV and their association with outcome in
critically ill patients will be reviewed. Third, im-
plications for clinical practice, and suggestions
for future research will be made.
BGAV: variability versus complexity
In the literature, the terms “glycemic vari-
ability”, “BG variability”, “glucose variability”,
“glucose variation” are often used, to describe
properties of the BG signal. These terms refer
to an intuitive concept, elegantly defined by
Braithwaite in a recent review:28 “the propensity
of a single patient to develop repeated excursions of
BG over a relatively short period of time that exceed
the amplitude expected in normal physiology”. Be-
cause the variability takes place in the amplitude
domain, BGAV is the most unambiguous col-
lective term for these metrics, and will be con-
sistently used in this review. In order to quan-
tify these excursions mathematically, a measure
should ideally describe the amplitude of the ex-
cursion away from the central tendency of the
BG signal, and should at the same time take into
account the frequency of these excursions within
a timeframe. A BGAV metric should make it
possible to identify or quantify pathological or
abnormal excursions and to assess the difference
with normal BG homeostasis, where smaller os-
cillations from the fasting BG range are observed
following a disturbance such as food intake.
Glucose complexity refers to another compo-
nent of the BG signal: short-term glucose oscil-
lations. Glucose complexity is considered to be a
description of the patient’s endogenous glucose
regulation system, independent from exogenous
factors.29 It is well known that loss of complex-
ity in the output signal of a biological system
can be one of the earliest symptoms of disease.30
Examples are loss of heart rate variability,31 or
reduced complexity in the pulsatility of endo-
crine systems.32 Usually, a signal becomes more
complex when more regulators are involved in
its control (and vice versa). In general, loss of
complexity can be considered as the loss of a
MEYFROIDT BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS
“regulator”, which is indicative of a pathologi-
cal state, and thus a worse outcome. In diabetic
patients, as compared to a non-diabetic or pre-
diabetic status, loss of complexity of the BG
signal is observed. In critically ill patients, three
studies have examined the relationship between
complexity of the BG signal and mortality. In
the study by Meyfroidt et al.,33 increased com-
plexity, calculated with jack-knifed approximate
entropy, was associated with increased hospi-
tal mortality. Lundelin et al.34 have calculated
complexity with detrended fluctuation analysis
(DFA), and found an opposite result: decreased
complexity was associated with increased ICU
mortality. In the most recent study, by Brunner
et al.,35 the association between complexity, cal-
culated with DFA, and increased mortality was
not linear, and increased as well as decreased
complexity, was associated with increased mor-
tality. Complexity might be considered a fourth
dimension of the BG signal. With regards to BG
regulation, the actual physiological meaning of
these complexity metrics is unknown, and their
clinical significance unclear. The apparent diver-
gent results of the three studies in critically ill
patients could indicate that the current sampling
resolution of the BG signal is not ideal for these
complexity metrics. In the future, continuous or
near-continuous BG sensors might allow for a
more precise calculation of BG complexity, and
its potential therapeutic or diagnostic implica-
tions. Based on the existing data, BG complexity
cannot be considered as therapeutic target. It is
not known whether and how an additional ex-
ternal regulator could be designed that can influ-
ence complexity, and whether adding complex-
ity to BG regulation, could affect the outcome of
a critically ill patient.
Therefore, this review will focus on BGAV,
and not complexity.
Measures to quantify BGAV, and
their association with outcome
BGAV is an intuitive concept, for which
several metrics have been described. There is
no consensus definition of BGAV, nor is there
consensus on which metrics should be used to
report on BGAV. Two excellent review articles
1012 MINERVA ANESTESIOLOGICA September 2015
have summarized in total 16 different BGAV
metrics that have been described in in critically
ill patients, for which an association with out-
come has been investigated.28, 36 There is much
heterogeneity between the different published
studies; most of them were observational and
retrospective, all have their methodological limi-
tations. Additionally, the diversity and the high
number of reported metrics, bears a high risk of
reporting bias, because possibly only indicators
that demonstrate an association with mortality
might have been reported.36 Moreover, in only
a minority of these studies the interdependence
between the different metrics, or whether the
observed association with outcome is independ-
ent of severity of illness, the BG control strategy
or the BG level, has been studied. Only 7 of the
16 metrics have been described in more than one
study in critically ill patients: the standard de-
viation (SD) of the BG signal; the coefficient of
variation (CV); the occurrence of hypoglycemia
and hyperglycaemia in the same admission; the
mean daily difference between minimum and
maximum BG (mean daily δ BG); the mean
amplitude of glucose excursions (MAGE); the
mean absolute glucose change per hour (MAG);
and the Glycemic Lability Index (GLI). For the
sake of the present review, we will only discuss
these 7 metrics. They are summarized in Table
I. Although from a theoretical and conceptual
point of view, some of the other metrics might
be of value to describe the characteristics of the
BG signal, they cannot be widely recommended
as reference BGAV measures because of the lim-
ited data.
SD
The first trials that described the association
between BGAV and outcome have used SD as a
metric.37-41 Meanwhile, it has become the most
commonly used indicator, and shows a uniform
association with worse outcome,21, 33, 42 except
in one study were none of the BGAV indicators
differed between survivors and non-survivors.35
The SD is measure of dispersion around a cen-
tral tendency. It has the advantage of being easy
to calculate, that most clinicians will be familiar
with its meaning, and that it is expressed in the
BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS MEYFROIDT

Table I.—Summary of the most frequently used BGAV measures.

BGAV metric What? Advantages Drawbacks
SD Dispersion around a central –– Easy to calculate –– Requires normally distributed
tendency –– Familiarity of clinicians data
–– Expressed in the same units as –– Does not differentiate between
BG multiple or single excursions
–– Refers to excursions with
reference to mean BG
CV SD/mean BG –– Easy to calculate –– Requires normally distributed
–– Expressed in the same units as data
BG –– Does not differentiate between
–– Can be used to compare between multiple or single excursions
BG signals with different means
Hypo/hyper Presence of both hypo- and –– Easy –– Binary (yes/no)
hyperglycaemia in an interval –– Does not require high frequency –– Inconsistent association with
data outcome
Mean daily δ BG Average of difference between –– Easy –– Inconsistent association with
highest and lowest BG value of –– Does not require high frequency outcome
each day data –– Different behaviour than SD in
TGC setting
–– Does not differentiate between
multiple or single excursions
MAGE Mean of the absolute values of any δ –– Designed and validated for –– Dependent on frequency of
BG from consecutive measurements diabetes patients sampling
that were higher than the SD
MAG Sum of all absolute glucose changes, –– Takes order and time interval –– Blunts out single major excursions
divided by the total time into account of BG
–– Dependent on frequency of
sampling
GLI Squared difference of consecutive –– Adapted from a validated index –– Dependent on frequency of
glucose measures per unit of actual for diabetics sampling
time between the samples, per day
SD: standard deviation; CV: coefficient of variation; Hypo-hyper: hypoglycemia and hyperglycemia in the same admission; Mean daily δ BG:
the mean daily difference between minimum and maximum blood glucose; MAGE: mean amplitude of glucose excursions; MAG: mean absolute
glucose change per hour; GLI: Glycemic Lability Index.

same units as BG. However, the SD does not
cover every aspect of BGAV. For instance, a high
SD could represent multiple high amplitude os-
cillations of BG, or one single large BG excur-
sion. Additional drawbacks of SD are the fact
that it requires a “normal” distribution of data in
order to make sense (which is not always the case
with BG data), and that SD refers to excursions
with reference to the mean BG.
CV
The CV is the SD, divided by the mean BG
(CV=SD/mean), and is associated with worse
outcome.38, 43 It represents the relative value of
BG excursions in proportion to the mean. CV
has been proposed as an alternative metric for
SD when comparing excursions for differing
means. Nevertheless, it is unknown, whether it
is the absolute magnitude of BG excursions, or
the relative magnitude as compared to the mean
BG, that would be most associated with worse
outcomes. CV has been used in recent trials
that have demonstrated an association between
BGAV and outcome in diabetic patients.44, 45
Hypoglycemia and hyperglycemia in the same ad-
mission (hypo-hyper criterion)
Early studies have used the presence of both
hypoglycaemia and hyperglycaemia in a pre-
defined interval as a measure of variability.46-48
However, the different thresholds used to define
hyper- and hypoglycemia, as well as the differ-
ent timeframes used, makes it hard to compare
between the different studies. Moreover, the fact

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1013

MEYFROIDT BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS
that this criterion is not a quantitative but a cat-
egorical (binary) variable (either present or not
present), and the inconsistent association of this
measure with outcome, indicate that this “hypo-
hyper criterion” is probably not a good descrip-
tor of the BGAV phenomenon. In addition, it
can be argued whether the hypo-hyper criterion
is a true marker of BGAV, or rather a metric to
assess the quality of BG level control.
The mean daily difference between minimum and
maximum BG (mean Daily δ BG)
The average of the difference of the highest
and the lowest BG value of each day in the ICU
has been used as a BGAV metric in 2 studies.33, 35
In one study, mean Daily δ BG was increased by
TGC, and was independently associated with
worse outcome.33 Of note, SD was not increased
by TGC in this study. Another study could not
demonstrate an effect on outcome of this meas-
ure.35 Moreover, mean daily δ BG is a very rough
indicator of glycemic excursions: first, only the
highest and the lowest value on each day are used
for its calculation, losing a lot of the granular-
ity of the BG signal; second, even more than is
the case with SD, a single glycaemic excursion
cannot be differentiated from several glycaemic
peaks. For these reasons, and in addition because
of its inconsistent association with outcome, and
the different behaviour to TGC as compared
with SD, the mean daily δ BG is probably not
an optimal measure of BGAV.
MAGE
The MAGE is a metric that was originally de-
signed for ambulatory use in the setting of con-
tinuous glucose monitoring, and has been modi-
fied for the use in hospitalized patients that are
monitored intermittently. MAGE is calculated
as the mean of the absolute values of any delta
BG from consecutive measurements that were
higher than the SD of the entire set of glucose
values. MAGE is dependent of the frequency of
sampling. By lowering the lower sampling rate,
peaks and lows in the BG signal can be missed,
which will influence the MAGE.21, 38, 42 Never-
theless, there is an independent correlation of
1014 MINERVA ANESTESIOLOGICA September 2015
MAGE with outcomes in critically ill patients,
in most 21, 38, 42 studies.
MAG
The MAG is calculated as the sum of all ab-
solute glucose changes divided by the total time
in hours, and is expressed in the unit of glucose
concentration per hour.49 MAG takes order and
time interval into account, and will blunt out
single major excursions of BG. MAG is associ-
ated with outcome in most studies,42, 49 and in
only one of these studies,50 this association was
stronger than with SD.
GLI
The GLI was also first developed and used
in diabetic patients whose glucose values were
continuously measured, and was defined as the
squared difference between consecutive glucose
measures per unit of actual time between the
samples per week.50 Meynaar has adapted the
GLI to be calculated over the samples per day.42
GLI is associated with outcome in critically ill
patients,21, 42, 51, 52 and was superior to SD and
MAGE in one study.38
BGAV: a possible therapeutic target?
BGAV is independently associated with worse
outcomes in most, but not all 35, 53 observational
studies in critically ill patients. This association
is partially attributed to endogenous patient re-
lated factors, which cannot be influenced, such
as the inflammatory status or the severity of ill-
ness of the patient. The specific effect of excessive
administration of nutrients on BGAV has never
been investigated. The multicentre EPaNIC trial
has demonstrated that avoiding early parenteral
nutrition (PN) up to full caloric target improved
recovery and reduced complications in critically
ill patients.54 Early PN led to a small but sig-
nificant increase in mean BG levels, a lower inci-
dence of hypoglycaemia, and almost doubled the
amount of insulin that had to be administered,
but did not report the effect on BGAV.
Although research to delineate the exact un-
derlying pathophysiological mechanism of the
BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS MEYFROIDT
observed increased BGAV is lacking, it is be-
ing suggested that a modern BG control policy
should aim at controlling the central tendency in
the normoglycaemic range (controlling median
or average BG), while at the same time minimiz-
ing hypoglycemia and BGAV.22 Obviously, the
association of a measure with a certain outcome
does not prove a causal relationship. Causality
can only be demonstrated by randomized con-
trolled trials (RCTs), comparing two clearly
distinct levels in one or more domains. With
regards to BGAV, it is questionable whether it
will be feasible to conduct such a trial. First, a
uniform indicator reference set for BGAV, to be
used in future trials, is lacking.36 No actual rec-
ommendation exists on which measure or meas-
ures should be included in such set, and this ex-
plains the wide variety in published indicators.
The literature does not allow to recommend any
BGAV metric above the other, and therefore it
appears wise to report on the effect on multiple
metrics when possible: a proposal could be to use
at least SD or CV, in combination with MAGE
(provided a regular sampling interval with suffi-
cient BG samples per day). MAG or GLI can be
used as additional metrics. Second, it is impor-
tant to realize that BGAV metrics are used to as-
sess variability in retrospect, but are not suitable
as a therapeutic target. A metric that uses pre-
vious values to determine the acceptable range
for next values, that can be calculated online and
presented to the bedside clinician, would have
to be developed and validated. Until that time,
BGAV can at the best be an additional evalu-
ation criterion to compare blood glucose regu-
lation protocols afterwards. Third, and funda-
mentally: can BGAV be reduced in settings were
TGC to the normoglycemic levels is performed
adequately? Several studies were TGC was inten-
sified and improved by clinical decision support
or computerized protocols, have observed a re-
duction in BGAV at the same time. In a small
study of 9 medical-surgical critically ill patients,
the implementation of the STAR-Liege 2 BG
control protocol, targeting 5.5-7.8 mmol/L with
continuous insulin and glucose infusions, and
a BG measurement interval of 2-3 hours, lead
to an improved BG level control and a reduced
BGAV at the same time.55 In a larger study of
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1015
197 surgical ICU patients, the implementation
of a software-guided intensive insulin protocol
(target range 5.3-7.5 mmol/L) lead to a greater
time in the desired BG range, a reduction in
the number of hypoglycaemic events, and a re-
duction of BGAV per patient.56 BGAV was an
evaluation criterion, but not a clinical target in
these protocols, which implies that TGC in it-
self reduced BGAV. In a small before/after study
in 105 mixed medical/surgical ICU patients,
standardizing the management of hypoglycae-
mia was able to reduce BGAV when compared
to treating hypoglycemia at the discretion of the
prescriber.57 Others have questioned whether
BGAV can be influenced at all.29 In the Leuven
studies, TGC did not decrease or increase BGAV
(using SD as metric).33
Continuous or near-continuous BG moni-
tors are being developed by the industry, and
in recent years, several small studies have been
published where they have been tested for ac-
curacy and safety, with varying results.58-60 Nev-
ertheless, it can be expected that, when proven
reliable in larger prospective clinical trials, some
of these systems might actually be marketed in
the upcoming years. It has been hypothesized
that differences between the native endogenous
glucoregulatory system, and the non-physiolog-
ical way by which nutrients and insulin are in-
fused in critically ill patients, might play a role
in magnifying or inducing BGAV.61 Indeed,
the response time of the islets of Langerhans to
changes in BG level is fast, with response times
of 5-10 minutes. Insulin is secreted in a pulsatile
way, and the interval as well as the amount of
insulin secreted will vary in response to the rate
with which glucose changes. In response to hy-
poglycemia, the liver will immediately increase
its glucose output to 2-3 times the baseline with-
in a period of 1 hour, first by glycogenolysis, and
later by increased rates of glyconeogenesis. In or-
der to adequately characterize the BG dynamics,
a sampling frequency of at least every 10 minutes
would be needed. In contrast, most of the cur-
rent algorithms to control BG use continuous
infusions of insulin and nutrients, and BG is
sampled at a much lower frequency of once every
1-4 hours. Simulations have shown that BGAV,
and the rates of both hyper- and hypoglycemia,
MEYFROIDT BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS
will reduce when the BG measurements are tak-
en more frequently, and will increase when the
BG measurement method has a higher impre-
cision.62 This simulation demonstrates that, in
theory, increasing the BG sampling frequency,
with a continuous or near-continuous sensor,
will inherently reduce BGAV, provided that the
measurement accuracy is sufficient. The MAG
change is the only marker for BGAV that has
been formally evaluated in the context of con-
tinuous BG monitoring, and correlates well with
indices describing intraday BGAV.63 In a recent
study by Brunner, strict TGC using nearly con-
tinuous BG monitoring did not further decrease
BGAV.35
Another technological advancement in BG
control in the ICU is the development of com-
puterized BG management systems. Currently,
some of these systems are being prospectively
tested in clinical trials. It is almost certain that in
the very near future, with the help of these sys-
tems, it will be possible to target any desired BG
range with minimal side effects, in particular a
minimal risk of hypoglycemia and BGAV. When
producing replicable results, widespread use of
this technology could help in the design of large
clinical trials to define the optimal BG range in
subgroups of patients. In addition, these BG
controllers could be used to detect endogenous
changes in insulin sensitivity, variability or com-
plexity, that maybe indicative for a change in the
clinical state of the patient.
Conclusions
BGAV is an intuitive concept, and describes
repeated excursions of the BG signal, exceeding
the amplitude expected in normal physiology. In
most observational clinical studies, an increase
in BGAV is associated with worse outcomes in
critically ill patients. Several measures to assess
BGAV in retrospect have been proposed, and
recommendations on an optimal reference set of
indices are currently lacking. It is advisable to re-
port on multiple measures when possible: at least
SD or CV, in combination with MAGE (pro-
vided a regular sampling interval with sufficient
BG samples per day) appears to be a reasonable
indicator set. Whether BGAV should be a target
1016 MINERVA ANESTESIOLOGICA September 2015
for therapy, in addition to controlling BG level,
has never been prospectively investigated, and
therefore cannot be recommended. Moreover, it
is questionable whether it is feasible to specifical-
ly target BGAV. In most studies, a tighter blood
glucose control in a certain range automatically
limits BG excursions in the extremes, thus re-
ducing BGAV. In the future, computerized pro-
tocols in combination with (near)-continuous
BG sensors might allow for a more precise BG
regulation without increasing BGAV.
Key messages
—— Observational studies have demon-
strated an independent association between
measures of increased BGAV and worse out-
comes in critically ill patients.
—— An optimal reference set of BGAV
measures is currently lacking, but with regu-
lar sampling intervals, the SD or CV, in com-
bination with the MAGE, are proposed.
—— BGAV is probably independent from
the BG level.
—— At this stage, BGAV can at the best be
an additional evaluation criterion to com-
pare blood glucose regulation protocols af-
terwards.
—— In the near future, computerized pro-
tocols in combination with (near)-continu-
ous BG sensors might allow for a more pre-
cise BG regulation without increasing BGAV.
References
 1. Pinto DS, Skolnick AH, Kirtane AJ, Murpy SA, Barron
HV, Giugliano RP et al. U-shaped relationship of blood
glucose with adverse outcomes amongst patients with ST-
segment elevation myocardial infarction. J Am Coll Cardiol
2005;46:178-80.
  2. Kosiborod M, Inzucchi SE, Krumholz HM, Xiao L, Jones
PG, Fiske S et al. Glucometrics in patients hospitalized
with acute myocardial infarction: defining the optimal out-
comes-based measure of risk. Circulation 2008;117:1018-
27.
  3. Bagshaw SM, Egi M, George C, Bellomo R. Early blood
glucose control and mortality in critically ill patients in
Australia. Crit Care Med 2009;37:463-70.
  4. Siegelaar SE, Hermanides J, Oudemans-Van Straaten HM,
van der Voort PH, Bosman RJ, Zandstra DF et al. Mean
glucose during ICU admission is related to mortality by a
U-shaped curve in surgical and medical patients: a retro-
spective cohort study. Crit Care 2010;14:R224.
BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS MEYFROIDT
  5. Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bru-
yninckx F, Schetz M et al. Intensive insulin therapy in the
critically ill patients. N Engl J Med 2001;345:1359-67.
  6. Van den Berghe G, Wilmer A, Hermans G, Meersseman W,
Wouters PJ, Milants I et al. Intensive insulin therapy in the
medical ICU. N Engl J Med 2006;354:449-61.
  7. Vlasselaers D, Milants I, Desmet L, Wouters PJ, Vanhore-
beek I, van den Heuvel I et al. Intensive insulin therapy
for patients in paediatric intensive care: a prospective, ran-
domised controlled study. Lancet 2009;373:547-56.
  8. Krinsley JS. Effect of an intensive glucose management pro-
tocol on the mortality of critically ill adult patients. Mayo
Clin Proc 2004;79:992-1000.
  9. Lecomte P, Foubert L, Nobels F, Coddens J, Nollet G, Cas-
selman F et al. Dynamic tight glycemic control during and
after cardiac surgery is effective, feasible, and safe. Anesth
Analg 2008;107:51-8.
10. Lecomte P, VanVlem B, Coddens J, Cammu G, Nollet G,
Nobels F et al. Tight perioperative glucose control is as-
sociated with a reduction in renal impairment and renal
failure in non-diabetic cardiac surgical patients. Crit Care
2008;12:R154.
11. Furnary AP, Cheek DB, Holmes SC, Howell WL, Kelly SP.
Achieving tight glycemic control in the operating room: les-
sons learned from 12 years in the trenches of a paradigm
shift in anesthetic care. Semin Thorac Cardiovasc Surg
2006;18:339-45.
12. Brunkhorst FM, Engel C, Bloos F, Meier-Hellmann A, Ra-
galler M, Weiler N et al. Intensive insulin therapy and pen-
tastarch resuscitation in severe sepsis. N Engl J Med 2008;
358:125–139
13. Preiser JC, Devos P, Ruiz-Santana S, Mélot C, Annane D,
Groeneveld J et al. A prospective randomised multi-centre
controlled trial on tight glucose control by intensive insulin
therapy in adult intensive care units: the Glucontrol study.
Intensive Care Med 2009;35:1738-48.
14. Arabi YM, Dabbagh OC, Tamim HM, Al-Shimemeri
AA, Memish ZA, Haddad SH et al. Intensive versus con-
ventional insulin therapy: a randomized controlled trial in
medical and surgical critically ill patients. Crit Care Med
2008;36:3190-7.
15. De La Rosa G del C, Donado JH, Restrepo AH, Quintero
AM, Gonzalez LG, Saldarriaga NE et al. Strict glycaemic
control in patients hospitalised in a mixed medical and sur-
gical intensive care unit: a randomised clinical trial. Crit
Care 2008;12:R120.
16. Finfer S, Chittock DR, Su SY, Blair D, Foster D, Dhingra V
et al. Intensive versus conventional glucose control in criti-
cally ill patients. N Engl J Med 2009;360:1283-97.
17. Scott MG, Bruns DE, Boyd JC, Sacks DB. Tight glucose
control in the intensive care unit: are glucose meters up to
the task? Clin Chem 2009;55:18-20.
18. Van den Berghe G, Schetz M, Vlasselaers D, Hermans G,
Wilmer A, Bouillon R et al. Insulin therapy in critically ill
patients. J Clin Endocrinol Metab 2009;94:3163-70.
19. Lanspa MJ, Hirshberg EL, Phillips GD, Holmen J, Stod-
dard G, Orne J. Moderate glucose control is associated
with increased mortality compared with tight glucose
control in critically ill patients without diabetes. Chest
2013;143:1226-34.
20. Bagshaw SM, Bellomo R, Jacka MJ, Egi M, Hart GK,
George C. The impact of early hypoglycaemia and blood
glucose variability on outcome in critical illness. Critical
Care 2009;13:R91.
21. Mackenzie IMJ, Whitehouse T, Nightingale PG. The met-
rics of glycaemic control in critical care. Intensive Care Med
2011;37:435-43.
22. Krinsley JS. Understanding glycemic control in the criti-
cally ill: three domains are better than one. Intensive Care
Med 2011;37:382-4.
23. Vriesendorp TM, DeVries JH, Hoekstra JB. Hypoglycemia
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1017
and strict glycemic control in critically ill patients. Curr
Opin Crit Care 2008;14:397-402.
24. Risso A, Mercuri F, Quagliaro L, Damante G, Ceriello A.
Intermittent high glucose enhances apoptosis in human
umbilical vein endothelial cells in culture. Am J Physiol En-
docrinol Metab 2001;281:E924-E930.
25. Piconi L, Quagliaro L, Da Ros R, Assaloni R, Giugliano
D, Esposito K et al. Intermittent high glucose enhances
ICAM-1, VCAM-1, E-selectin and interleukin-6 expres-
sion in human umbilical endothelial cells in culture: the
role of poly (ADP-ribose) polymerase. J Thromb Haemost
2004;2:1453-9.
26. Quagliaro L, Piconi L, Assaloni R, Martinelli L, Motz E,
Ceriello A. Intermittent
������������������������������������������������
high glucose enhances apoptosis re-
lated to oxidative stress in human umbilical vein endothelial
cells. The role of protein kinase C and NAD(P)H-oxidase
activation. Diabetes 2003;52:2795-804.
27. Monnier L, Mas E, Ginet C, Michel F, Villon L, Cristol
JP, et al:Activation of oxidative stress by acute glucose fluc-
tuations compared with sustained chronic hyperglycemia in
patients with type 2 diabetes. JAMA 2006;295:1681-7.
28. Braithwaite SS. Glycemic variability in hospitalized pa-
tients: choosing metrics while awaiting evidence. Curr Diab
Rep 2013;13:138-54.
29. van Hooijdonk RTM, Abu-Hanna A, Schultz MJ. Glyc-
emic variability is complex – is glucose complexity variable?
Crit Care 2012;16:178.
30. Goldberger AL, Amaral LA, Hausdorff JM, Ivanov P, Peng
CK, Stanley HE. Fractal dynamics in physiology: altera-
tions with disease and ageing. Proc Natl Acad Sci USA
2002;99(Suppl 1):2466-72.
31. Goldstein B, Fiser DH, Kelly MM, Mickelsen D, Ruthman
U, Pollack MM. Decomplexification in critical illness and
injury: relationship between heart rate variability, severity of
illness, and outcome. Crit Care Med 1998;26:352-7.
32. Boonen E, Meersseman P, Vervenne H, Meyfroidt G, Guïza
F, Wouters PJ et al. Reduced nocturnal ACTH-driven corti-
sol secretion during critical illness. Am J Physiol Endocrinol
Metab 2014;306:E883-92.
33. Meyfroidt G, Keenan DM, Wang X, Wouters PJ, Veldhuis
JD, Van den Berghe G. Dynamic characteristics of blood
glucose time series during the course of critical illness: ef-
fects of intensive insulin therapy and relative association
with mortality. Crit Care Med 2010;38:1021-9.
34. Lundelin K, Vigil L, Bua S, Gomez-Mestre I, Honrubia T,
Varela M. Differences in complexity of glycemic profile in
survivors and nonsurvivors in an intensive care unit:a pilot
study. Crit Care Med 2010;38:849-54.
35. Brunner R, Adelsmayr G, Herkner H, Madl C, Holzinger
U. Glycemic variability and glucose complexity in critically
ill patients:a retrospective analysis of continuous glucose
monitoring data. Crit Care 2012;16:R175.
36. Eslami S, Taherzadeh Z, Schultz MJ, Abu-Hanna A. Glu-
cose variability measures and their effect on mortality:a sys-
tematic review. Intensive Care Med 2011;37:583-593
37. Egi M, Bellomo R, Stachowski E, French CJ, Hart G. Vari-
ability of blood glucose concentration and short-term mor-
tality in critically ill patients. Anesthesiology 2006;105:244-
52.
38. Ali NA, O’Brien Jr JM, Dungan K, Philips G, Marsh CB,
Lemeshow S et al. Glucose variability and mortality in pa-
tients with sepsis. Crit Care Med 2008;36:2316-21.
39. Krinsley JS. Glycemic variability:a strong independent pre-
dictor of mortality in critically ill patients. Crit Care Med
2008;36:3008-13.
40. Dossett LA, Cao H, Mowery NT, Dortch MJ, Morris Jr
JM, May AK. Blood glucose variability is associated with
mortality in the surgical intensive care unit. Am Surg
2008;74:679-85.
41. Waeschle RM, Moerer O, Hilgers R, Herrmann P, Neu-
mann P, Quintel M. The impact of the severity of sepsis
MEYFROIDT BLOOD GLUCOSE AMPLITUDE VARIABILITY IN CRITICALLY ILL PATIENTS

on the risk of hypoglycaemia and glycaemic variability. Crit
Care 2008;12:R129.
42. Meynaar IA, Eslami S, Abu-Hanna A, van der Voort P, de
Lange DW, de Keizer N. Blood glucose amplitude vari-
ability as predictor for mortality in surgical and medical in-
tensive care unit patients:a multicenter cohort study. J Crit
Care 2012;27:119-24.
43. Krinsley JS. Glycemic variability in critical illness and the
end of chapter 1. Crit Care Med 2010;38:1206-8.
44. Krinsley JS, Egi M, Kiss A, Devendra AN, Schuetz P, Mau-
rer PM et al. Diabetic status and the relation of the three
domains of glycemic control to mortality in critically ill pa-
tients: an international multicenter cohort study. Crit Care
2013;17:R37.
45. Subramaniam B, Lerner A, Novack V, Khabbaz K, Pary-
ente-Wiesman M, Hess P. Increased glycemic variability in
patients with elevated preoperative HbA1C predicts adverse
outcomes following coronary artery bypass grafting surgery.
Anesth Analg 2014;118:277-87.
46. Hirshberg E, Larsen G, Van Duker H. Alterations in glu-
cose homeostasis in the pediatric intensive care unit: hy-
perglycemia and glucose variability are associated with
increased mortality and morbidity. Pediatr Crit Care Med
2008;9:361-6.
47. Bagshaw SM, Bellomo R, Jacka MJ, Egi M, Hart GK,
George C. The impact of early hypoglycemia and blood
glucose variability on outcome in critical illness. Crit Care
2009;13:R91.
48. Jacka MJ, Torok-Both CJ, Bagshaw SM. Blood glucose con-
trol among critically ill patients with brain injury. Can J
Neurol Sci. 2009;36:436-42.
49. Hermanides J, Vriesendorp TM, Bosman RJ, Zandstra
DF, Hoekstra JB, DeVries JH. Glucose variability is asso-
ciated with intensive care unit mortality. Crit Care Med
2010;38:838-42.
50. Hirsch IB, Brownlee M. Should minimal blood glucose
variability become the gold standard of glycemic control? J
Diabetes Complications 2005;19:178-81.
51. Dungan KM, Binkley P, Nagaraja HN, Schuster D, Osei
K. The effect of glycaemic control and glycaemic variability
on mortality in patients hospitalized with congestive heart
failure. Diabetes Metab Res Rev 2011;27:85-93.
52. Zuo YY, Kang Y, Yin WH, Wang B, Chen Y. The asso-
ciation of mean glucose level and glucose variability with
intensive care unit mortality in patients with severe acute
pancreatitis. J Crit Care 2012;27:146-52.
53. Lipska KJ, Venkitachalam L, Gosch K, Kovatchev B, Van
den Berghe G, Meyfroidt G et al. Glucose variability and
mortality in patients hospitalized with acute myocardial in-
farction. Circ Cardiovasc Qual Outcomes 2012;5:550-557
54. Casaer MP, Mesotten D, Hermans G, Wouters PJ, Schetz
M, Meyfroidt G et al. Early versus late parenteral nutrition
in critically ill adults. N Engl J Med 2011;11:506-17.
55. Penning S, Le Compte AJ, Massion P, Moorhead KT, Pretty
CG, Preiser JC et al. Second pilot trials of the STAR-Liege
protocol for tight glycemic control in critically ill patients.
Biomedical Engineering Online 2012;11:58.
56. Saur NM, Kongable GL, Holewinski S, O’Brien K, Nasra-
way SA Jr. Software-guided insulin dosing: tight glycemic
control and decreased glycemic derangements in critically
ill patients. Mayo Clin Proc 2013;88:920-9.
57. Arnold P, Paxton RA, McNorton K, Szpunar S, Edwin SB.
The effect of a hypoglycemia treatment protocol on glyc-
emic variability in critically ill patients. J Intensive Care
Med 2015;30:156-60.
58. Rabiee A, Andreasik V, Abu-Hamdah R, Galiatsatos P,
Khouri Z, Gibson BR et al. Numerical and clinical accuracy
of a continuous glucose monitoring system during intrave-
nous insulin therapy in the surgical and burn intensive care
units. J Diabetes Sci Technol 2009;3:951-9.
59. Logtenberg SJ, Kleefstra N, Snellen F, Groenier KH, Slin-
gerland RJ, Nierich AP et al. Pre- and postoperative accura-
cy and safety of a real-time continuous glucose monitoring
system in cardiac surgical patients:a randomized pilot study.
Diabetes Technol Ther 2009;11:31-7.
60. Brunner R, Kitzberger R, Miehsler W, Herkner H, Madl
C, Holzinger U. Accuracy and reliability of a subcutane-
ous continuous glucose-monitoring system in critically ill
patients. Crit Care Med 2011;39:659-64.
61. DeJournett L. Essential elements of the native glucoregula-
tory system, which, if appreciated, may help improve the
function of glucose controllers in the intensive care unit set-
ting. J Diabetes Sci Technol 2010;4:190-8.
62. Boyd JC, Bruns DE. Effects of measurement frequency on
analytical quality required for glucose measurements in in-
tensive care units: assessment by simulation models. Clini-
cal Chemistry 2014;60:644-50.
63. Kohnert K-D, Heinke P, Fritsche G, Vogt L, Augstein P,
Salzsieder E. Evaluation of the Mean Absolute Glucose
Change as a measure of glycemic variability using con-
tinuous glucose monitoring data. Diabetes Technol Ther
2013;15:448-54.

Funding.—G. Meyfroidt is supported by the research foundation, Flanders (FWO) as senior clinical investigator (1846113N).
Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on June 10, 2014. - Accepted for publication on December 15, 2014. - Epub ahead of print on December 17, 2014.
Corresponding author: G. Meyfroidt, Intensive Care Medicine, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium.
E-mail: geert.meyfroidt@uzleuven.be

1018 MINERVA ANESTESIOLOGICA September 2015



REVIEW

Colloids versus crystalloids in the prevention of

hypotension induced by spinal anesthesia in elective
cesarean section. A systematic review and meta-analysis
J. RIPOLLÉS MELCHOR 1, 2, Á. ESPINOSA 2, 3, E. MARTÍNEZ HURTADO 1, 2,
R. CASANS FRANCÉS 2, 4, R. NAVARRO PÉREZ 1, 2, A. ABAD GURUMETA 2, 5,
J. M. CALVO VECINO 1, 2

1Department of Anesthesia, Computense University of Madrid, Hospital Universitario Infanta Leonor, Madrid, Spain;
2E.A.R. (Evidence Anesthesia Review) Group; 3Thorax Intensive Care Centre, Örebro County Council Hospital, Örebro
University, Örebro, Sweden; 4Department of Anesthesia, Hospital Clínico Universitario Lozano Blesa, Zaragoza, Spain;
5Department of Anesthesia, Hospital Universitario la Paz, Madrid, Spain

ABSTRACT
The incidence of hypotension associated to spinal anesthesia in elective cesarean section is high. To determine the
effects of colloids and crystalloids in the incidence of hypotension induced by spinal anesthesia in elective cesarean
section, an attempt was made to define which type of fluid and what total volume should be administered. Fol-
lowing the PRISMA methodology a systematic review and meta-analysis were carried out. A systematic Medline/
PubMed, EMBASE and Cochrane Library search was made to identify trials where women were scheduled for elec-
tive cesarean section with spinal anesthesia and volume loading (preload or co-load). The primary outcome was the
incidence of hypotension. Stratification into subgroups was made for the primary outcome according to the type
of colloid administered, differentiating those studies employing new generation colloids (HES 6% 130/0.4) from
those not using such colloids, based on the volume of colloid administered and the combination of a vasopressor.
The secondary outcome was the incidence of intraoperative nausea and vomiting. Two-hundred and twenty-seven
controlled clinical trials were analyzed; eleven randomized clinical trials including 990 patients were included. A
significative decrease of incidence of hypotension associated to spinal anesthesia was observed with the use of col-
loids compared to crystalloids (RR [95% CI] 0.70 [0.53-0.92], P=0.01). However, there was no difference between
crystalloid and colloid in the risk of intraoperative nausea and vomiting (RR [95% CI] 0.75 [0.41-1.38]; P=0.33).
This meta-analysis shows colloid administration to significantly reduce the incidence of hypotension associated to
spinal anesthesia in elective cesarean section compared with of crystalloid use.
(Minerva Anestesiol 2015;81:1019-30)
Key Words: Fluid therapy - Cesarean section - Colloid - Anesthesia, Spinal.

S pinal anesthesia (SA) is the anesthetic tech-

nique of choice in healthy pregnant women
undergoing elective cesarean section.1 The inci-
dence of hypotension secondary to spinal block
is high (about 70%). The risk factors associated
to hypotension in pregnant women include pre-
operative hypertension, age, the type of SA used,
and the weight of the newborn infant.2 On the
other hand, pregnant women experience an in-
crease in sympathetic versus parasympathetic ac-
tivity,3 together with increased susceptibility to
the effects of sympathetic block – a fact that can
give rise to increased vasodilatation.
Prolonged severe hypotension can have
harmful effects including organ ischemia, loss
of consciousness, cardiovascular collapse and
uteroplacental hypoperfusion.4 The physiologi-
cal aim during SA is to maintain cardiac output
(CO), and specifically uteroplacental perfusion.
CO is presently regarded as a better predictor of

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1019

RIPOLLÉS MELCHOR COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION
uteroplacental perfusion than arterial pressure,5
though the latter is commonly used as a surro-
gate marker of CO.
The recent literature underscores that the pre-
vention of hypotension during SA for cesarean
section is fundamentally based on the prophy-
lactic use of vasopressors. Phenylephrine is the
current vasopressor of choice for the prevention
of maternal hypotension and nausea.6 However,
the administration of fluids remains useful for
further reducing the incidence and severity of
hypotension and/or the need for vasopressor
drugs. The best way to administer these fluids is
still subject to controversy. Despite the evidence
supporting associated vasopressor drug use,2 flu-
id therapy alone is often employed for avoiding
hypotension secondary to SA.7
Anesthetists can choose four different fluid
loading or expansion regimens: there are two
types of fluids (colloids or crystalloids), and each
can be administered either before SA (preload)
or with SA (co-load).
The present systematic review and meta-anal-
ysis compares colloids versus crystalloids for the
prevention of hypotension in elective cesarean
section, in an attempt to define which type of
fluid and what total volume should be admin-
istered.
Material and methods
Screening criteria
The Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) method-
ology8 was used to identify the studies, based on
the following inclusion criteria:
—— participants: women scheduled for elec-
tive cesarean section with SA and volume load-
ing (preload or co-load), with or without associ-
ated vasopressor treatment;
—— types of intervention: administration at
any time (preload or co-load) of any type or
volume of colloid), with or without associated
vasopressor treatment for the prevention of hy-
potension secondary to SA in patients subjected
to elective cesarean section under SA;
—— types of comparator: administration at
any time (preload or co-load) of any type or vol-
1020 MINERVA ANESTESIOLOGICA September 2015
ume of crystalloid), with or without associated
vasopressor treatment for the prevention of hy-
potension secondary to SA in patients subjected
to elective cesarean section under SA;
—— outcomes: the primary outcome or end-
point was the incidence of hypotension (using
the definition of hypotension used in differ-
ent studies) The secondary outcome was the
incidence of intraoperative nausea and vomit-
ing (IONV), defined as the unpleasant subjec-
tive urge to vomit. In contrast, vomiting is the
propulsive abdominal muscular spasms associ-
ated with the expulsion of gastric contents; and
which are produced during the intraoperative
period from the completion of SA until comple-
tion of cesarean section;
—— types of studies: selection was made
of those published randomized clinical trials
(RCTs) meeting the mentioned inclusion crite-
ria and with a Jadad score of ≥3.
Sources of information
Following the PRISMA protocol8, different
strategies were used to identify those relevant
studies that complied with the aforementioned
inclusion criteria, based on a Medline/PubMed,
EMBASE and Cochrane Library search. There
were no restrictions regarding date of publica-
tion. The search was limited to articles in English
(last search update: September 2014).
Search terms
The search was carried out using the following
key words: hydroxyethyl starch derivatives, col-
loids, crystalloids, fluid therapy, cesarean deliv-
ery, cesarean section, SA. The search was limited
to controlled, randomized clinical trials (RCTs).
Selection of articles and data extraction
Two independent investigators reviewed each
title and abstract in order to discard all irrel-
evant clinical trials and identify the potentially
relevant publications. The latter in turn were
thoroughly analyzed, selecting those which met
the abovementioned inclusion criteria. Data ex-
traction from the selected studies was carried out
COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION RIPOLLÉS MELCHOR
by two different investigators. All discrepancies
were again analyzed and validated by a third in-
vestigator. The authors reviewed the data analy-
sis in order to avoid any possible transcription
errors. Two independent researchers conducted
the quality assessment of the included RCTs us-
ing the Jadad scale.9
Outcomes
The primary outcome or endpoint was the
incidence of hypotension (using the definition
of hypotension used in different studies). Pre-
defined stratification into subgroups was made
for the primary outcome according to the type
of colloid administered, differentiating those
studies employing new generation colloids (HES
6% 130/0.4) from those not using such colloids,
based on the volume of colloid administered and
the combination of a vasopressor. ��������������
Since the com-
parison of preload vs coload was analyzed in a
previous meta-analysis,10 we decided not carry
out this comparison. The secondary outcome
was the incidence of IONV.
Figure 1.—PRISMA Flow Diagram.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1021
Statistical analysis
Review Manager (RevMan) 11 was used for the
statistical analysis. The meta-analysis was carried
out using the random binary effects model, re-
porting the data as relative risk (RR) with the
respective 95% confidence interval (95% CI).
Forest plots were used, considering a statisti-
cally significant effect when P<0.05. Statistical
heterogeneity was evaluated using the I2 statis-
tic. I2 values of <25%, 25-50% and >75% were
defined as indicating low, moderate and high
heterogeneity, respectively. Sensitivity analyses
were conducted by restricting the analysis to tri-
als that had a Jadad Score of ≥4. Publication bias
was assessed by funnel plot using the relative risk
of hypotension as an end-point.
Results
Study screening
Finally, 11 RCTs, which included 990 pa-
tients, were included in the meta-analysis.12-22
RIPOLLÉS MELCHOR COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION

Figure 1 shows the flowchart used for the selec-
tion of items.
Characteristics of included studies
The selected articles describe the results of
RCTs in which colloid load is compared with
crystalloid load. The included studies were
conducted in healthy patients, ASA I-II; in all
studies the groups were comparable in weight
and height, generally obese patients were ex-
cluded. After SA, every pregnant woman re-
mained in supine and with left-sideways in or-
der to avoid the aorto-cava compression. The
characteristics of the included RCTs are shown
in Tables I, II.
Of the 11 RCTs included in the meta-analy-
sis,12-22 10 used preloading 1-19, 21, 22 and one used
co-loading.20 One study compared preloading
with crystalloid versus co-loading with colloid.20
The characteristics of the colloids and crystal-
loids are shown in Table II. Nine of the included
RCTs did not use vasopressor medication asso-
ciated to fluids for the prevention of hypoten-
sion,12-19, 22 while two RCTs used phenylephrine
associated to fluids.20, 21
Assessment of risk of bias in individual studies
Into the included studies, two RCTs had a
Jadad score of 3,12, 13 four RCTs a Jadad score
of 4,14-16, 19 and five RCTs a Jadad score of
5.17, 18, 20-22 Futhermore, all studies were double-
blind and randomized correctly. Table I shows
the Jadad score of included studies.
Primary outcome
Incidence of hypotension
The 11 included studies comprised a total of
990 patients, and were analyzed in relation to

Table I.—PICOS characteristics of included studies.

Study Year Patients Intervention
Karien et al.12 1995 Healthy parturients scheduled for elective cesarean Preload with0.5l HES 6% N. 13
section. Spinal anesthesia
Riley et al.13 1995 Healthy parturients scheduled for elective cesarean Preload with 0.5l HES 6% + lactated
section. Spinal anesthesia Ringer’s 1l N. 20
French et al.14 1999 Healthy parturients scheduled for elective cesarean Preload with 15ml/kg of 10%Pentastarch
section. Spinal anesthesia N. 20

Siddik et al.15 2000 Healthy parturients scheduled for elective cesarean Preload with 0.5l HES 10% N. 20
section. Spinal anesthesia
Cardoso et al.16 2004 Healthy parturients scheduled for elective cesarean Preload with 1l HES 6% N. 20
section. Spinal anesthesia
Dahlgren et al.17 2005 Healthy parturients scheduled for elective cesarean Preload with 1l 3% Dextran 60 N. 56
section. Spinal anesthesia
Dahlgren et al.18 2007 Healthy parturients scheduled for elective cesarean Preload with 1l 3% Dextran 60 in patients
section. with positive or negative stress test. Spinal positive stress test N. 9, or negative
anesthesia stress test N. 19

Tamilselvan et al.19 2009 Healthy parturients scheduled for elective cesarean Preload with 0.5l HES 6% N. 20 or 1l
section. Spinal anesthesia HES 6% N. 20
McDonald et al.20 2011 Healthy parturients scheduled for elective cesarean Coload with 11 HES 6%+ phenylephrine
section. Spinal anesthesia infusion N. 30

Mercier et al.21 2014 Healthy parturients scheduled for elective cesarean Preload with 0.5l HES 6% + lactated
section. Spinal anesthesia Ringer’s 0.5l N. 82
Tawfik et al.22 2014 Healthy parturients scheduled for elective cesarean Preload with 0.5l HES 10% N. 103
section. Spinal anesthesia
HES: hydroxyethyl starches; RCT: randomized controlled trial; IONV: Intraoperative nause and vomiting.

1022 MINERVA ANESTESIOLOGICA September 2015

COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION RIPOLLÉS MELCHOR

the incidence of hypotension.12-22 A significa- Hypotension according to type of colloid

tive decrease of hypotension associated to SA A subgroup comparative analysis was
was observed with the use of colloids compared made of the studies that used HES 6%
to crystalloids (RR [95% CI] 0.70 [0.53-0.92], 130/0 20, 21 versus those that used some other type
P=0.01) (Figure 2). of colloid.12-19, 22 Both, HES 6% 130/0.4 sub-

Figure 2.—Effects of colloid versus crystalloid on incidence of hypotension.

Comparator Outcomes Study design Jadad score Funding Country

Preload with lactated Effects of volume preload in RCT Single center 3 Not declared Finland
Ringer’s 1l N. 13 uteroplacentar hemodynamics
Preload with lactated Incidence of hypotension; ephedrine RCT Single center 3 Not declared USA
Ringer’s 2l N. 20 consumption
Preload with lactated Incidence of hypotension; incidence of RCT Single center 4 Not declared Uk
Ringer’s 15 ml/kg IONV
N. 30
Preload with lactated Incidence of hypotension; ephedrine RCT Single center 4 Not declared Lebanon
Ringer’s 0.5l N. 20 consumption, incidence of IONV
Preload with lactated Incidence of hypotension; incidence of RCT Single center 4 Not declared Brazil
Ringer’s 1l N. 85 IONV
Preload with acetated Incidence of hypotension and severe RCT Single center 5 Not declared Sweeden
Ringer’s 1l N. 53 hypotension
Preload with acetated Incidence of hypotension. Ephedrine RCT Single center 5 Not declared Sweeden
Ringer’s 1l in consumption
patients positive
stress test N. 10, or
negative stress test
N. 15
Preload with lactated Effects of volume preload on blood RCT Single center 4 Deltex Medical UK
Ringer’s 1.5l N. 20 volume
Coload with lactated Effects of volume coload on cardiac RCT Single center 5 Smiths Medical UK
Ringer’s 1l + output. Incidence of hypotension
phenylephrine
infusion N. 30
Preload with lactated Incidence of hypotension, maternal RCT Multicentre 5 Fresenius Kabi France
Ringer’s 1l N. 85 outcomes
Coload with acetated Incidence of hypotension RCT Monocentric 5 Not declared Egypt
Ringer’s 1l N. 102

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1023

RIPOLLÉS MELCHOR COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION

Table II.—Characteristics of included studies.

Associated
Study Year Load Colloid Crystalloid Anesthesia
Vasopressor
Karien et al.12 1995 Preload Saline HES 6% Lactated No Spinal Anesthesia: L3-L4. 2.6 mg
(450/0.7/5) Ringer’s 1L Hyperbaric Bupivacaine
0.5 L
Riley et al.13 1995 Preload Saline HES 6% Lactated No Spinal Anesthesia: L2-L3 or L3-L4. 12
450/0.7/5 0.5 l Ringer´s mg Bupivacaine 0.75% + morphine
1-2L 0.2 mg, fentanyl 10 mcg
French et al.14 1999 Preload Saline HES 10% Lactated No Spinal Anesthesia: L2-L3.10-15mg
240/0.45 15 Ringer´s Bupivacaine 0.5
mL/kg 15 mL/kg
Siddik et al.15 2000 Preload Saline HES 10% Lactated No Spinal Anesthesia: L2-L3 or L3-L4. 13
200/0.5/6 0.5 L Ringer´s mg Bupivacaine 0.75%
0.5 L
Cardoso et al.16 2004 Preload Gelafundin 1 Ll Lactated No Spinal Anesthesia: L2-L3 or L3-L4.
Ringer´s 0.5% Hyperbaric Bupivacaine +
1L morphine 0.4 mg
Dahlgren et al.17 2005 Preload 3% Dextran 60 1 L Acetated No Spinal Anesthesia: L3-L4. 2.5
Ringer’s 1L Hyperbaric Bupivacaine 0.5%+
fentanyl 10 mcg.
Dahlgren et al.18 2007 Preload 3% Dextran 60 1 L Acetated No Spinal Anesthesia: L3-L4. 2.5
Ringer’s Hyperbaric Bupivacaine 0.5%+
1L fentanyl 10 mcg.
Tamilselvan et al. 19 2009 Preload Saline HES 6% Lactated No Spinal Anesthesia: L3-L4 Tetracaine
70/0.55/3 1 L Ringer´s hydrochloride 8mg + hydrochloride
1L morphine 100mcg
McDonald et al.20 2011 Coload Saline HES 6% Lactated phenylephrine Spinal Anesthesia: L3-L4. 12.5 mg
70/0.55/31 L Ringer´s Bupivacaine 0.5%+ fentanyl 15 mcg.
1L
Mercier et al.21 2014 Preload Saline HES 6% Lactated phenylephrine Spinal Anesthesia: L2-L3, L3-L4.or L4-
130/0.4/9 0.5 Ringer´s L5. 11 mg Hyperbaric Bupivacaine
1L 1L 0.5% + morphine 100umg +
suphentanyl 3mcg
Tawfik et al.22 2014 Preload/ Saline HES 6% Acetated No Spinal Anesthesia: L2-L3 or L3-L4.
Coload 130/0.4/9 0.5 L Ringer’s 12.5 mg Hyperbaric Bupivacaine
1L 0.5%+ fentanyl 10 mcg.
HES: hydroxyethyl starches. The first number appearing after HES refers to the molecular weight of the product in kilodaltons, the second number
is its molar substitution ratio and the third one is the C2/C6 ratio; SBP: systolic blood pressure; IONV: intraoperative nausea and vomiting.

group (RR [95% CI] 0.66 [0.50-0.88]; P=0.005
and the subgroup using other colloids showed a
reduction in the RR of hypotension). (RR [95%
CI] 0.70 [0.51-0.96]; P=0.03)(Figure 3).
Hypotension according to volume of col-
loid
A subgroup comparative analysis was
made of the studies that used 0.5 liters of col-
loid 13, 15, 17, 18, 21, 22 versus those that used one
liter.12, 16, 19, 20 The 0.5 liter subgroup showed a
significant reduction in hypotension (RR [95%
CI] 0.72 [0.58-0.91]; P<0.01), while the use of
larger volumes resulted in nonsignificant reduc-
tions in hypotension (RR [95% CI] 0.87 [0.53-
1.43]; P=0.59)(Figure 4).
Hypotension and vasopressor use
A subgroup comparative analysis was made
of the studies that used a vasopressor drug (phe-
nylephrine) associated to fluid therapy 20, 21 versus
those that did not.12-19, 22 Both showed a reduc-
tion in the RR of hypotension (Figure 5).
Secondary outcome. Intraoperative nausea and
vomiting
Of the 11 included studies, 5 analyzed
IONV.13, 16, 18, 21, 22 There was no difference

1024 MINERVA ANESTESIOLOGICA September 2015

COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION RIPOLLÉS MELCHOR

Complications
Hypotension defined as Profilaxis PONV Hypotension Treatment colloid associed
Decrease in SBP to less than 90 mmHg and less than 80% No Ephedrine 5-10 mg No
of the baseline value.

Decrease in SBP to less than 100 mmHg and less than 80% Metoclopramide 10 mg Ephedrine 5 mg No
of the baseline value.

Decrease in SBP to less than 90 mmHg and less than 70% No Ephedrine 3-6 mg No
of the baseline value.

Decrease in SBP to less than 100 mmHg and less than 80% No Ephedrine 5 mg No
of the baseline value.

Decrease in SBP to less than 80-90% of the baseline value. No Metaraminol 0.2mg No

Decrease in SBP to less than 100 mmHg and less than 80% No Ephedrine 5 mg No
of the baseline value.

Decrease in SBP to less than 100 mmHg and less than 80% No Ephedrine 5 mg No
of the baseline value.

Decrease in SBP to less than 100 mmHg and less than 80% No Ephedrine 6 mg No
of the baseline value.

Decrease in SBP to less than 80% of the baseline value. No phenylephrine 100 mcg No

Decrease in SBP to less than 80% of the baseline value. No phenylephrine 50-150 mcg No

Decrease in SBP to less than 90 mmHg and less than 80% No Ephedrine 5 mg No
of the baseline value.

between crystalloid and colloid in the risk of [95% CI] 0.76 [0.67-0.85]; P=0.01). Using
���������
hy-
IONV (RR [95% CI] 0.75 [0.41-1.38]; P=0.35) potension as an end-point, the funnel plot sug-
(Figure 6). gests that there may be a publication bias favour-
ing small positive studies using colloid (Figure
Complications 7).

None of the studies included examined renal Discussion

function, and no complications associated to
colloid use were reported.
Sensitivity analysis, assessment risk of bias across
studies and publication bias
Restricting the analysis to higher-quality stud-
ies did not change the hypotension results (RR
The main results of this meta-analysis are that
colloid administration significantly reduces the
incidence of hypotension compared with the use
of crystalloids. Likewise, the administration of
volumes in excess of 0.5 liters affords no added
benefit, and the association of vasopressor drugs
contributes to reduce the risk of hypotension.

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1025

RIPOLLÉS MELCHOR COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION

Figure 3.—Effects of colloid versus crystalloid according to type of colloid on incidence of hypotension.
HES: hydroxyethyl starches; M-H: Mantel-Haentzel.

Figure 4.—Effects of colloid versus crystalloid gruped by volume on incidence of Hypotension.

M-H: Mantel-Haentzel.

1026 MINERVA ANESTESIOLOGICA September 2015

COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION RIPOLLÉS MELCHOR

Figure 5.—Effects of colloid versus crystalloid with associated vasopressor treatment on incidence of hypotension.
M-H: Mantel-Haentzel.

Figure 6.—Effects of colloid versus crystalloid on incidence of IONV.

IONV: intraoperative nausea and vomiting; M-H: Mantel-Haentzel.

Figure 7.—Funnel Plot Hypotension compared between colloid group vs. crystalloid group.
SE (log RR): Standard error log Relative risk.

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1027

RIPOLLÉS MELCHOR COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION
Lastly, colloids do not offer a greater decrease
in the risk of IONV compared with the use of
crystalloids.
In our meta-analysis, most of the included
studies involved preloading, with confirmation
that colloids reduce the risk of hypotension
in SA for elective cesarean section. A previous
meta-analysis of the moment or time of fluid
administration and preload or co-load revealed
that the timing of fluid expansion does not influ-
ence the results obtained.10 On the other hand,
studies comparing preloading with colloid versus
co-loading with colloid have recorded no signifi-
cant differences.23
Recently, Tawfik et al.22 compared preloading
with colloid versus co-loading with crystalloid.
The authors concluded that the administration
of one liter of crystalloid co-load gives rise to a
decrease in the incidence of hypotension similar
to that afforded by a 0.5-liter colloid preload.
The fact that two distinct regimes of fluids were
compared, can be added heterogeneity to the re-
sults.
There is some controversy regarding the safety
of colloids, as a result of clinical trials in septic
patients that have recorded an increase in re-
nal failure and mortality.24None of the studies
included in our meta-analysis examined renal
function, and no complications associated to
colloid use were reported. However, all included
studies are underpowered to assess for these ef-
fects, including adverse effects on hemostasis,
possibility of allergic reaction, and cost. In view
of the results obtained by the subgroup analy-
sis, the use of HES 6% (130/0.4) can be recom-
mended for loading, since the observed safety
profile is better.
The administration of volumes in excess of 0.5
liters of colloid offers no added benefits. Larger
volumes not only result in no added benefits but
can lead to an unnecessary increase in the risk of
anaphylactic reactions,25 as well as to an increase
in healthcare costs.
Mercier et al.21 compared colloid with crys-
talloid versus crystalloid with phenylephrine un-
der algorithm guidance, with good results. Such
prophylaxis is therefore regarded as strongly
recommendable. In contrast, McDonald et al.20
recorded no differences in the incidence of hypo-
1028 MINERVA ANESTESIOLOGICA September 2015
tension – although this was not the primary out-
come or endpoint of their study. In the present
meta-analysis, the subgroup analysis involving
associated vasopressor treatment yielded good
results. Despite the decrease in the incidence of
hypotension with colloid administration, the in-
cidence remains high. Thus, on the basis of the
results of this meta-analysis, the combination
of a vasopressor (preferably phenylephrine, if
available) should be considered for the preven-
tion of hypotension. The most adequate way to
administer phenylephrine is subject to debate,26
though there is agreement that it should be re-
garded as the treatment of choice in hypotension
induced by SA.6 The administration of colloids
has not been shown to be superior to crystalloids
in the context of co-loading.20 Lastly, pregnant
women with a positive preoperative supine stress
test, i.e. patients at an increased risk of suffer-
ing hypotension, appear to derive greater benefit
from the administration of colloids,18 and the
latter reduce the incidence of severe or clinically
significant hypotension,17, 18, 21 or avoid it al-
most entirely,19 and moreover lessen the need for
vasopressor treatment.
Our meta-analysis was unable to demonstrate
that colloid use reduces IONV as the principal
manifestation of hypotension. A possible expla-
nation is the rapid treatment of hypotension
after the drop in blood pressure defined by the
authors.
Research implications
The preventive treatment of hypotension sec-
ondary to SA in elective cesarean section using
fluid therapy and phenylephrine appears to be
well established.27 Only two trials to date have
compared the fluids in the presence of vasopres-
sor. Further trials are necessary.
It remains to be seen whether the combination
of different prophylactic fluid and vasopressor
regimens offers advantages for healthy pregnant
women in reference to long-term complications
such as wound infection and sepsis, wound heal-
ing, postoperative bleeding, or the duration of
hospital stay. Evaluation of the safety of HES
administration over the long term and its effects
upon renal function is also needed.
COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION RIPOLLÉS MELCHOR
Limitations
The principal limitation is the heterogeneity
between studies. We acknowledge that the asym-
metry in the funnel plot could be explained by
factors such as methodological heterogeneity be-
tween studies; the possibility of publication bias
given the asymmetry of the plot and the small
number of studies cannot be totally excluded.
The different definitions of hypotension used in
the literature make it difficult to establish com-
parisons among the results of different studies,
simply because the definition of hypotension
conditions its incidence. On the other hand,
only one of the studies was a multicenter ran-
domized clinical trial. 21 The fact that most of
the publications corresponded to single-center
studies does not allow us to rule out the possibil-
ity of registry bias – though we only included
RCTs with a Jadad score of ≥3. Of the 11 in-
cluded studies only one have investigated more
than 100 patients per group.22 Differences in
methodological quality may cause heterogeneity.
Smaller studies tend to be conducted and ana-
lysed with less methodological rigour than larger
studies, and trials of lower quality also tend to
show larger intervention effects. Regarding pub-
lication bias may need to consider potential con-
flicts of interest, if the results and analyzes have
been standardized, and degree of trial registra-
tion. However, no such heterogeneity was found
in the predefined HES 6% 130/0.4 subgroup or
in the subgroup analysis with associated vaso-
pressor medication.
Only two 20, 21 RCTs used phenylephrine with
intravenous fluids suggesting poor external va-
lidity with respect to the study findings.
Conclusions
The results of this meta-analysis, within the
limitations of existing data, ������������������
show colloid load-
ing therapy for the prevention and treatment of
hypotension associated to SA in elective cesarean
section to be superior to the use of crystalloids.
New generation colloids (HES 6% 130/0.4) are
to be preferred, in view of their superior safety
profile and the results obtained in this study.
Colloids should be administered at the total
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1029
minimum effective dose (0.5 liters), since larger
volumes offer no added benefits. Likewise, their
use should be combined with phenylephrine
treatment algorithms, for although the incidence
of hypotension is reduced, the problem of hypo-
tension persists.
Key Messages
—— Colloid loading therapy for the pre-
vention and treatment of hypotension as-
sociated to SA in elective cesarean section is
superior to the use of crystalloids.
—— New generation colloids (HES 6%
130/0.4) administered at the total minimum
effective dose (0.5 liters), are to be preferred,
in view of their superior safety profile.
—— Phenylephrine decreases the incidence
of hypotension.
—— Incidence of intraoperative nausea and
vomiting remains unaffected.
References
  1. Cyna AM, AndrewM, Emmett RS, Middleton P, Simmons
SW. Techniques for preventing hypotension during spinal
anaesthesia for caesarean section. Cochrane Database Sys
tRev2006; CD002251
  2. NganKee WD. Prevention of maternal hypotension after
regional anaesthesia for caesarean section. Curr Opin An-
aesthesiol 2010;23:304-9.
 3. Lewinsky RM, Riskin-Mashiah S. Autonomic imbal-
ance in preeclampsia: evidence for increasedsympathetic
tone in response to the supine-pressor test. Obstet Gyne-
col1998;91:935-9
  4. Macarthur A, Riley ET. Obstetric anesthesia controversies:
vasopressor choice for postspinal hypotension during cae-
sarean delivery. Int Anesthesiol Clin 2007;45:115-32.
 5. Robson SC, Boys RJ, Rodeck C, Morgan B. Maternal
and fetal haemodynamic effects of spinal and extradural
anaesthesia for elective caesarean section. Br J Anaesth
1992;68:54-9.
 6. Habib AS. A review of the impact of phenylephrine ad-
ministration on maternal hemodynamics and maternal and
neonatal outcomes in women undergoing cesarean delivery
under spinal anesthesia. Anesth Analg 2012;114:377-90.
  7. Allen TK, Muir HA, George RB, Habib AS. A survey of the
management of spinal-induced hypotension for scheduled
cesarean delivery. Int J Obstet Anesth 2009;18:356-61.
 8. Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA
Group: Preferred Reporting Items for Systematic Re-
views and Meta-Analyses: the PRISMA statement. BMJ
2009,339:b2535
  9. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds
DJ, Gavaghan DJ et al. Assessing the quality of reports of
randomized clinical trials: is blinding necessary? Control
Clin Trials 1996;17:1-12.
10. Banerjee A, Stocche RM, Angle P, Halpern SH. Preload or
RIPOLLÉS MELCHOR COLLOIDS VERSUS CRYSTALLOIDS AGAINST HYPOTENSION

coload for spinal anesthesia for elective cesarean delivery: A
meta-analysis. Can J Anaesth 2010;57:24-31.
11. Higgins JPT, Green S (editors). Cochrane Handbook for
Systematic Reviews of Interventions Version 5.0.2 [updated
September 2009]. The Cochrane Collaboration, 2009.
12. Karinen J, Räsänen J, Alahuhta S, Jouppila R, Jouppila P.
Effect of crystalloid and colloid preloading on uteroplacen-
tal and maternal haemodynamic state during spinal anaes-
thesia for Caesarean section. Br J Anaesth 1995;75:531-5.
13. Riley ET, Cohen SE, Rubenstein AJ, Flanagan B. Preven-
tion of hypotension after spinal anesthesia for cesarean sec-
tion: Six percent hetastarch versus lactated Ringer’s solu-
tion. Anesth Analg 1995;81:838-42.
14. French GW, White JB, Howell SJ, Popat M. Comparison of
pentastarch and Hartmann’s solution for volume preload-
ing in spinal anaesthesia for elective caesarean section. Br J
Anaesth1999;83:475-7.
15. Siddik SM, Aouad MT, Kai GE, Sfeir MM, Baraka AS. ��� Hy-
droxyethylstarch 10% is superior to Ringer’s solution for
preloading before spinal anesthesia for Cesarean section.
Can J Anesth 2000;47:616-2.
16. Cardoso MMSC, Bliacheriene S, Freitas CRC, Cesar DS,
Torres MLA. Preload during spinal anesthesia for caesarean
section: comparison between crystalloid and colloid solu-
tions. Rev Bras Anestesiol 2004;54:781-7.
17. Dahlgren G, Granath F, Pregner K, Rosblad PG, Wessel
H, Irestedt L. Colloid vs. crystalloid preloading to prevent
maternal hypotension during spinal anesthesia for elective
cesarean section. Acta Anaesthesiol Scand 2005;49:1200-6.
18. Dahlgren G, Granath F, Wessel H, Irestedt L. Prediction of
hypotension during spinal anesthesia for cesarean section
and its relation to the effect of crystalloid or colloid preload.
Int J Obstet Anesth 2007;16:128-34.
19. Tamilselvan P, Fernando R, Bray J, Sodhi M, Columb M.
The effects of crystalloid and colloid preload on cardiac out-
put in the parturient undergoing planned cesarean delivery
under spinal anesthesia: A randomized trial. Anesth Analg
2009;109:1916-21.
20. McDonald S, Fernando R, Ashpole K, Columb M. Ma-
ternal cardiac output changes after crystalloid or colloid
coloadfol- lowing spinal anesthesia for elective cesarean
delivery: A randomized controlled trial. Anesth Analg
2011;113:803-10.
21. F. J. Mercier, P. Diemunsch, A.-S. Ducloy-Bouthors, A.
Mignon, M. Fischler, J.-M. Malinovsky et al. 6% Hydrox-
yethylstarch (130/0.4) vs Ringer’slactate preloading before
spinal anaesthesia for Caesarean delivery: the randomized,
double-blind, multicentre CAESAR trial. Br J Anaesth
2014;113:459-67.
22. Tawfik MM et al. Comparison between colloid preload
and crystalloid co-load in cesarean section under spinal an-
esthesia: a randomized controlled trial. Int J Obstet Anesth
2014;23:317-23.
23. Nishikawa K, Yokoyama N, Saito S, Goto F. Comparison
of effects of rapid colloid loading before and after spinal an-
esthesia on maternal hemodynamics and neonatal outcomes
in cesarean section. J Clin Monit Comput 2007;21:125-9.
24. Myburgh JA, Finfer S, Bellomo R, Billot L, Cass A, Gattas
D et-al. CHEST Investigators; Australian and New Zealand
Intensive Care Society Clinical Trials Group. Hydroxyethil
starch or saline for fluid resuscitation in intensive care. N
Engl J Med 2012;367:1901-11.
25. Barron ME, Wilkes MM, Navickis RJ. A system-
atic review of the comparative safety of colloids. Arch
Surg2004;139:552-63.
26. Allen TK, George RB, White WD, Muir HA, Habib AS.
A double-blind, placebo-controlled trial of four fixed rate
infusion regimens of phenylephrine for hemodynamic sup-
port during spinal anesthesia for cesarean delivery. Anesth
Analg 2010;111:1221-9.
27. Mercier FJ. Cesarean delivery fluid management. Curr
Opin Anaesthesiol 2012;25:286-91.

Conflicts of interest.—Ripollés Melchor received honoraria as a lecturer from Fresenius Kabi. The other authors certify that there is no
conflict of interest with any financial organization regarding the material discussed in the manuscript.
Received on September 19, 2014. - Accepted for publication on November 27, 2014. - Epub ahead of print on December 11, 2014.
Corresponding author: Javier Ripollés Melchor, Goya 111, 2, D, 28009, Madrid, Spain. E-mail: ripo542@gmail.com

1030 MINERVA ANESTESIOLOGICA September 2015



REVIEW

Safe mechanical ventilation in patients without

acute respiratory distress syndrome (ARDS)
S. R. PANNU 1, R. D. HUBMAYR 2

1Departmentof Pulmonary, Critical Care, Sleep and Allergy, Ohio State University, Wexner Medical Center, Columbus,
OH, USA; 2Pulmonary and Critical Care Medicine, Mayo Clinic College of Medicine, Mayo Clinic, Rochester, MN,
USA

ABSTRACT
Insights into the pathogenesis of lung deformation injury inspired a benchmark clinical trial, which demonstrated
that reducing tidal volumes compared to previous norms was associated with improved patient survival in acute respi-
ratory distress syndrome (ARDS). Since many critically ill patients without ARDS possess ventilator associated lung
injury (VALI) risk factors, there is no need to expose them to tidal volumes that are larger than would be needed to
achieve acceptable blood gas tensions. In the following perspective we will argue that lung protection from deforma-
tion injury should guide ventilator management in all patients, irrespective of the presence of ARDS. That is not to
say that all lung diseases share the same VALI risk, but we contend that adopting a low tidal ventilation strategy is a
simple and safe starting point in most instances. We will review studies in the medical and surgical literature that have
addressed “lung protective ventilation” in patients without ARDS and summarize them with a focus on tidal volume,
positive end expiratory pressure and oxygen supplementation settings. In addition, we will briefly discuss under what
circumstance one might consider deviating from a conventional approach. (Minerva Anestesiol 2015;81:1031-40)
Key words: Respiratory distress syndrome, adult - Ventilation - Lung injury.

M echanistic insights into the pathogen-

esis of ventilator associated lung injury
(VALI) in patients with acute respiratory distress
syndrome (ARDS), dominate the debate about
optimal ventilator care of patients with all cause
respiratory failure. To date, low tidal volume
ventilation has been the only therapy shown to
diminish VALI manifestations and was associ-
ated with improved ARDS survival.1, 2 There
remain nagging questions about optimum ven-
tilation mode and positive end-expiratory pres-
sure settings (PEEP), but the general approach
to “lung protective” ventilation in ARDS has not
changed during the past decade. In the following
perspective we will argue that lung protection
from deformation injury should guide ventila-
tor management in all patients, irrespective of
the presence of ARDS. That is not to say that
all lung diseases share the same VALI risk, but
we contend that adopting a low tidal ventila-
tion strategy is a simple and safe starting point
in most instances.
Biophysical mechanisms of VALI in patients
with ARDS include the radial stress (stretch) as-
sociated with “so-called” over distension of aer-
ated alveoli, and the interfacial stress associated
with “so-called” cyclic opening and collapse of
small lung units.3 Both injury mechanisms are
amplified by small scale heterogeneity in re-
gional mechanical properties causing local tissue
distortions (shear) on account of interdepend-
ence and by surfactant inactivation leading to
atelectasis, local flooding, airspace foam forma-
tion and/or unit collapse. ARDS lungs are par-
ticularly susceptible to VALI because they are
functionally small (the number and volume of

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1031

PANNU Safe mechanical ventilation in patients without ARDS
aerated alveoli typically equal those of a child) 4
and because they are prone to inflammation and
flooding. Thus, an “adult sized” tidal volume
may be forced into a “baby lung”, causing injury
from over distension.5 As the alveolus expands,
and is repeatedly stretched beyond its structural
limit (presumably at a volume near total lung
capacity, TLC) epithelial and endothelial cells,
which trim the blood-gas barrier, lose adherence
to the basement membrane, form blebs and
experience tight junction stress failure.6, 7 As a
consequence the overstretched alveolus floods
and becomes the site of local inflammation, of-
ten referred to as “biotrauma”.8 Contemporary
changes in structure and hydration state of the
connective tissue matrix contribute to activation
of inflammasome pathways and shape tissue re-
modeling.9 Cyst formation in non-dependent
portions of ARDS lungs have been attributed
to this mechanism.10 Flooding exposes the epi-
thelia of small conducting airways and alveoli
to injury by interfacial stress.11 Interfacial stress
injury is often referred to as low volume injury,
as atelectrauma, as opening and collapse injury
and as recruitment/derecruitment injury.12 It is
associated with cell wounding as surface pres-
sure gradients generated by air-liquid interfaces
(foam) sweep across the apical surfaces of epi-
thelia.13 While it is convenient to discuss over
distension and atelectrauma as if they were
distinct and separate lung injury mechanisms,
it is important to recall that commonly both
are consequences of the same biophysical in-
put, namely recurrent large parenchymal de-
formations (strains). Large deformations are
associated with increased alveolar wall stress,
impaired alveolar-capillary barrier properties,
local flooding, surfactant inactivation, as well
as agitation and thus foam formation of the al-
veolar exudate. The stresses associated with said
deformation responses are sufficient to produce
a proinflammatory response, which is typically
associated with cell and tissue microstructural
lesions. Consistent with these mechanisms Prot-
ti et al. showed that in experimental animals the
use of tidal volumes (TV) twice the volume at
relaxed end-expiration, thereby straining the pa-
renchyma by a factor of 2, was associated with
the development of ARDS.44
1032 MINERVA ANESTESIOLOGICA September 2015
Our brief review of biophysical lung injury
mechanisms makes it clear why normal lungs
tolerate tidal volume settings that would be
injurious in ARDS.14 For one, over disten-
sion to volumes exceeding total lung capacity
is unlikely to occur in a lung with a preserved
inspiratory capacity. Moreover, the greater the
number of alveoli that are capable of accepting
inspired gas, the smaller is the strain (measure
of deformation) of individual alveolar walls.
For that reason surfactant inactivation and lo-
cal flooding are much less likely to occur. The
same cannot be said of atelectrauma, because re-
cumbent, mechanically ventilated patients with
normal lungs are at risk for atelectasis, particu-
larly in the context of obesity and abdominal
distension.15 Moreover, the immune system of
critically ill patients is frequently “primed” for
biotrauma. Stressors such as sepsis, trauma or
brain injury render the lungs susceptible to a
“second hit”, manifest as an exaggerated inflam-
matory response to mechanical ventilation.16
Since many critically ill patients without ARDS
possess VALI risk factors, there is no need to ex-
pose them to tidal volumes that are larger than
would be needed to achieve acceptable blood
gas tensions. Moreover, in the absence of a com-
pelling reason to do otherwise, uniformity in
tidal volume guidelines may be appealing. The
incidence of human related medical errors in
the ICU is significant leading to increased pa-
tient morbidity.17, 18 Standardization of practice
patterns, such as uniformity in tidal volumes,
wherever possible may serve in reducing medi-
cal errors. Many studies in the medical and sur-
gical literature have addressed “lung protective
ventilation” in patients without ARDS. In this
paper, we will review and summarize them with
a focus on tidal volume, PEEP and FiO2 set-
tings. In addition, we will briefly discuss under
what circumstance one might consider deviat-
ing from a conventional approach. Given con-
straints in space and time we have not addressed
patient-ventilator interactions or the merits of
preserving diaphragm contractions during as-
sisted ventilation. These would warrant an in-
depth discussion of regional ventilation and its
determinants in health and disease, which is be-
yond the scope of this review.
Safe mechanical ventilation in patients without ARDS PANNU
Tidal volume
Five decades ago Greenfield et al. demonstrat-
ed that repeated inflations to high volumes and
pressures increase the surface tension of healthy
canine lungs and cause de novo injury.9 At the
time the clinical relevance of these observations
was not appreciated. The same was true for a sub-
sequent report by Webb and Tierney, who con-
firmed Greenfield’s observations in a rat model
and showed that the use of PEEP along with a
reduction in tidal volume was lung protective.19
It was not until Dreyfuss et al. confirmed Webb
and Tierney’s findings and provided detailed
data on lung microstructure and biophysical in-
jury mechanisms that the critical care commu-
nity began to pay attention.5, 6 By coining the
term “biotrauma” Tremblay and Slutsky under-
scored the importance innate immune responses
to deforming stress, thereby setting the stage for
numerous experimental and clinical intervention
trials, in which proinflammatory mechanisms
were targeted.20 However to date, the only prov-
en strategy associated with improved survival in
humans with ARDS is low tidal volume ventila-
tion.1
In normal resting humans tidal volume rang-
es between 6% and 8% of total lung capacity,
which like ideal body weight, is largely correlated
with height and gender.21 For that reason, the
recommendation to initiate mechanical ventila-
tion of all patients with tidal volumes of 6 mL/
kg predicted body weight amounts to choosing
a tidal volume in the “physiologic range”. In fact
Tenny et al. had argued that 6.3 mL/kg was an
appropriate tidal volume for all mammalian spe-
cies. Yet, anesthesiologists have historically cho-
sen much larger tidal volumes to support patients
during surgery. This is because in the absence of
PEEP, anesthesia is invariably associated with gas
absorption atelectasis, which may be transient-
ly “opened” with sighs and large breaths.22, 23
Given the focus on oxygenation, and concerns
about PEEP induced hypotension in this setting,
raising tidal volumes is certainly one means of
combatting hypoxemia. However, it is unclear,
if correcting hypoxemia with the aid of high
tidal ventilation in anesthetized individuals with
more or less normal lungs is safer than doing so
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1033
by raising FiO2 or by increasing PEEP. The lat-
ter requires the judicious use of fluids and ino-
tropes and is therefore associated with potential
cardiovascular risks. In general, the management
algorithms used in clinical trials and case series
focused on lung protective ventilation in the op-
erating room have not been explicit enough to
address this concern.
Can a lung without ARDS withstand
tidal volumes between 10-12 mL/kg ideal
body weight without adverse effects?
The incidence of ARDS acquired after MV
is between 6-25%. Various risk factors have
been identified for development of ARDS in
MV patients who did not have ARDS at onset.
In a look back study of all patients, who were
mechanically ventilated at Mayo Clinic during
2001 Gajic et al.24 identified tidal volume as sig-
nificant risk factor in hospital acquired ARDS
(OR 1.3 for each mL above 6 mL/kg pbw) along
with blood transfusions, female gender and his-
tory of restrictive lung disease. These findings
are consistent with a subsequent analysis of an
international cohort of mechanically ventilated
patients, where injurious MV settings such as
high VT (OR 2.6 for VT >700 mL) along with
other settings were associated with ARDS de-
velopment.25 High VT was also identified as an
independent predictor of ARDS in severe brain
patients injury 26 resulting in higher morbidity
and mortality.27-29
In the last decade, various studies were con-
ducted both in the medical and surgical arena
evaluating effects of low Vt <6 mL or less) ven-
tilation in patients without ARDS at onset.
A prospective randomized controlled clinical
trial comparing high vs. low tidal volumes (10
vs. 6 IBW) in MV patients without ARDS was
stopped prematurely because of the development
of lung injury in the high tidal volume group.30
They also found that sustained cytokine produc-
tion was associated with high tidal volumes.
A recent report by Severgnini et al. showed im-
provement in respiratory function when patients
undergoing elective abdominal surgery were ven-
tilated with relatively high levels of PEEP and VT
of 7 mL/kg body weight.31 Lung protective venti-
PANNU Safe mechanical ventilation in patients without ARDS
lation in patients undergoing esophageal surgery
was associated with a reduced systemic inflam-
matory response.32, 33 Several randomized clini-
cal trials in patients undergoing lung resection
have suggested that lung protective ventilation,
i.e., the application of recruitment maneuvers,
VT and modest amounts of PEEP, can lower the
incidence of postoperative pulmonary compli-
cations and thereby reduce cost.34,35 Substantial
reductions in tidal volumes delivered during one
lung ventilation in the operating room have re-
duced the incidence of “post-pneumonectomy
pulmonary edema” a syndrome, which closely
resembles “Ventilator Induced Lung Injury”, as
can be generated experimentally in animals with
normal lungs.36 Consistent with this interpreta-
tion, in a retrospective review pneumonectomy
patients, who developed respiratory failure in the
immediate postoperative period, had received
significantly larger tidal volumes in the operat-
ing room than patients, who recovered without
complications (median VT is 8.3 vs. 6.7 mL/kg
predicted body weight).37 In patients undergo-
ing cardiac surgery there is added concern about
hemodynamics, many surgeons favor high tidal
ventilation over PEEP as a recruitment strategy.
Also, cardiopulmonary bypass itself increases the
risk of ARDS.38 Yet, Zupanich et al. reported that
early open lung approach (small tidal volumes
and increased PEEP) attenuated the inflamma-
tory response associated with cardiopulmonary
bypass.39 In a retrospective analysis of cardiac sur-
gical patients, VT >10 mL/kg was identified as a
risk factor for organ failure and prolonged ICU
stay. In turn, patients, who had been given small-
er VT in the operating room, had a shorter wean-
ing duration and required fewer re-intubations.40
In contrast, Wrigge et al. failed to observe a cy-
tokine surge in patients with healthy lungs sub-
jected to high VT and concluded that, in the ab-
sence of “priming factors”, i.e., VALI risk factors,
MV with low PEEP and VT as high as 15 mL/kg
PBW was safe.41 A recent secondary analysis of
MV data failed to reveal an association of initial
VT with the development of ARDS.42 Moreo-
ver, a review of about thirty thousand patients,
suggested that the intra-operative use of low tidal
volume and minimal PEEP settings were associ-
ated with an increase in 30 day mortality.43
1034 MINERVA ANESTESIOLOGICA September 2015
Finally, a recent meta-analysis showed that
the use of low tidal volumes in patients with-
out ARDS was associated with a lower risk of
developing ARDS, lower mortality and other
improved clinical outcomes.45 Based on our un-
derstanding of the VALI pathogenesis and our
review of the clinical literature, we conclude that
tidal volumes between 6 and 8 mL/kg predicted
body weight are both safe and ought to meet gas
exchange targets in the vast majority of patients
without ARDS. We propose that standardization
of practice patterns, such as uniformity in tidal
volumes, wherever possible may serve in reduc-
ing medical errors.
Low tidal volumes for all. What are the barriers?
Some authors are of the opinion that low
tidal volume settings may not be necessary in
all patients and raise concerns about increased
risk of excessive sedation, paralysis and atelecta-
sis.46 Yet many of these undesired effects can be
avoided by means other than high tidal ventila-
tion, such as the judicious use of PEEP, atten-
tion to flow settings and adjustments in sedative
and narcotic medications. This has been shown
in a recent meta analysis where initiation of low
VT at onset of MV was not associated with in-
creased sedation requirements.47 It is true that
many patients, particularly after deep and pro-
longed sedation, demand large tidal volumes
during recovery. They are often transitioned
to unassisted breathing using a bi-level pres-
sure mode and in that mode are able to gener-
ate large tidal breaths. The risk for exacerbating
biotrauma at that point is presumably less than
during the acute stage of the syndrome, but it
may not be zero. Therefore, risks and benefits of
enforcing lung protective tidal volumes through
the use of extracorporeal and transtracheal CO2
removal options, neuromuscular blockade or
further adjustments in sedative and narcotic
dosing must be carefully considered. Unfortu-
nately, we are not aware of data, which would
truly guide this judgment. There are barriers to
the implementation of low tidal volume venti-
lation. A review of the ARDS-net database re-
vealed old age, short stature, high bicarbonate
levels, and a low lung injury score as reasons
Safe mechanical ventilation in patients without ARDS PANNU
why providers deviated from low tidal ventila-
tion guidelines.48 Nevertheless, these barriers
can be overcome using a multifaceted interdis-
ciplinary intervention strategy including web
based teaching, computerized order entry and
decision support tools.49 Prediction tools such
as the acute lung injury prediction score (LIPS),
surgical lung injury prevention (SLIP-II) model
and early acute lung injury (EALI) score may
direct providers towards low tidal volume man-
agement for patients at risk for ARDS early in
the course of their ED presentation.50-52
Positive end expiratory pressure
In their original description of patients with
ARDS, Ashbaugh and Petty already emphasized
the value of PEEP as a means to enhance pul-
monary oxygen transfer in hypoxemic patients
with edematous lungs. At least in that context,
the use of PEEP has become the cornerstone of
the “open lung approach” and it’s mechanisms of
action are well understood.53, 54 Yet the debate
about “best PEEP” is far from settled, because
there is no agreement on: 1) the relative VALI
risks attributed to over inflation vs. under recruit-
ment; 2) the confounding effects of tidal volume
on PEEP mediated lung protection; and 3) an
uncertainty if PEEP induced changes in arterial
oxygen tension are valid surrogates of efficacy.
While these issues dominate PEEP management
decisions in ARDS, they are only peripheral to
the topic at hand, namely, the risk/benefit ratio
of PEEP in patients without ARDS.
Is PEEP necessary to prevent injury to a normal
lung?
It is well known that airway intubation is as-
sociated with a decrease in end-expired lung
volume and that considerable amounts of PEEP
are required to prevent gas absorption atelecta-
sis during inhalational anesthesia.55 However,
the use of PEEP in the operating room is only
now emerging as standard of care in many North
American hospitals. When considered in a mech-
anistic context, the use of PEEP in anesthetized
individuals with relatively normal lungs hinges
on one’s concern for adverse hemodynamic ef-
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1035
fects relative to one’s belief that preventing atel-
ectasis is associated with a meaningful reduction
in postoperative pulmonary complications. Both
concerns are valid and supported by experimen-
tal evidence. While there are differences in the
micromechanics of ARDS associated alveolar
edema and anesthesia associated gas absorption
atelectasis, both conditions share the risk for
interfacial stresses injury.56 Even though numer-
ous experimental studies in anesthetized animals
have demonstrated PEEP associated lung pro-
tection, confounding effects of covariates such
as saline lavage, pneumothorax, negative end-
expiratory pressure or high tidal volume ventila-
tion cannot be discounted. On the other hand
one may reasonably infer from lung mechanics
measurements, that low tidal ventilation without
PEEP is associated with epithelial injury.12, 57-59
Therefore, the evidence to date favors the use of
PEEP even in patients with normal lungs.
It is common practice to initiate mechanical
ventilation with a PEEP setting of 5 cm H2O.
This “default setting” is insufficient and inappro-
priate for patients with ARDS, but seems rea-
sonable in most other circumstances, including
patients with fibrotic non-recruitable lungs.60, 61
The only exception applies to patients with in-
creased abdominal pressure on account of obes-
ity, ascites or ileus.62 Since the passive tension of
the relaxed diaphragm is quite small, abdominal
pressure is transmitted to the pleural space and
“squeezes” the lung until expiratory flow limita-
tion and airway closure prevent further empty-
ing.63, 64 Needless to say breathing near residual
volume is associated with an increased right
ventricular afterload, and predisposes the lung
to atelectasis, hypoxemia and interfacial stress
injury. Therefore, PEEP should be increased in
proportion to either body mass index or some
estimate of abdominal load as might be inferred
from body habitus and/or esophageal, gastric
and bladder pressure measurements.65, 66
Mergoni et al. measured the pressure-volume
characteristics of the chest-wall in non-obese
recumbent patients with ARDS and concluded
that only 5 cm H2O of PEEP are necessary to
counterbalance its elastic load, which was as-
sumed to equal the chest wall elastance (dPes/
dV).67 This estimate is much lower than the
PANNU Safe mechanical ventilation in patients without ARDS
“best PEEP” estimate derived by Talmor et al.
in a similar population.68 Importantly, the differ-
ence in the two load estimates lies not in the ac-
tual Pes measurements, but in the way the pres-
sure tracings were analyzed. Mergoni considered
the actual Pes to be biased by the weight of the
heart and mediastinal contents and therefore set
the end-expiratory Pes value to zero cm H2O. In
contrast, Talmor et al. noted that the end-expira-
tory Pes was quite high (between +5 and +20 cm
H2O) and argued that the high Pes reflected the
weight of the chest wall/abdomen on the ede-
matous lungs and needed to be counterbalanced
with PEEP. Although physiologists have long
subscribed to the notion of a “mediastinal arti-
fact, i.e., a positive Pes bias, Loring convincingly
argued that the very high end-expiratory Pes
values seen in many patients with ARDS can-
not be explained solely on that basis.66, 69 Since
a high end-expiratory Pes is the result of a large
superimposed load on lungs, which resist further
emptying, the same mechanisms and inferences
for PEEP management apply to any condition
in which the weight of the passive chest wall is
large enough to cause either airway closure or
dynamic airway collapse and expiratory flow
limitation. It follows that the terms elastic load
(the pressure at a specific volume) and elastance
(the change in pressure with volume) must not
be used interchangeably. In fact, obese patients
tend to have a greatly increased end-expiratory
Pes, i.e., load, yet normal or near normal chest
wall elastances.70 In keeping with this reasoning
investigators, who managed PEEP in recumbent
patients guided by its effect on gas exchange, ar-
rived at PEEP settings between 10 and 15 cm
H2O, a value significantly larger than one might
have inferred from chest wall elastance measure-
ments.67, 71-75
Arguments about “best PEEP” beg a discus-
sion about risks and benefits of recruitment
maneuvers. Alveolar recruitment maneuvers can
help transiently reverse the derecruitment associ-
ated with low tidal volume ventilation.76 How-
ever, while lung mechanics may suggest benefit,
a detrimental impact of repeated large inflations
on microvascular integrity and right ventricular
afterload cannot be discounted.77 Recruitment
maneuvers may be used to identify “best PEEP”
1036 MINERVA ANESTESIOLOGICA September 2015
but the best way to do this remains controver-
sial.78-80 Such maneuvers were thought to be
helpful in the management of obese patients
during surgery.81 However, the PROVHILO
trial did not favor the use of high PEEP (~12
cm) and recruitment maneuvers during elective
abdominal surgery, and suggested that low peep
without recruitment maneuvers was associated
with fewer postoperative complications.
Concerns about hemodynamic compromise
and vasopressor use persist and need to be bal-
anced in respective clinical contexts. Adverse ef-
fects on intracranial pressure and cerebral blood
flow have limited the use of PEEP in patients
with traumatic brain injury.82 Yet several re-
ports suggest that PEEP does not uniformly
raise intracranial pressure or reduce cerebral per-
fusion.82, 83 For that reason the principles of lung
protection and lung protective ventilator settings
should not be a priori ignored in that popula-
tion.84
In conclusion, a PEEP setting of 5 cm seems
reasonable in most circumstances, unless there is
impending or overt ARDS. In patients with in-
creased abdominal pressure on account of obes-
ity, ascites or ileus, PEEP may be raised to coun-
teract the elevated chest wall recoil pressure. It is
yet to be established if such PEEP adjustments
should be guided by esophageal manometry.
Moreover, is it not known if the benefit of PEEP
mediated lung protection justifies the risk as-
sociated with hemodynamic compromise when
there is need to support blood pressure with flu-
ids and inotropes.
Should different end-inspiratory plateau airway
pressures targets be considered in patients with-
out ARDS?
In mechanically ventilated patients, whose
respiratory muscles are relaxed at end-inflation,
the plateau airway pressure (Pplat) is the passive
recoil pressure of the respiratory system, i.e. that
of lungs plus chest wall. As inferred from lung
and chest wall mechanics recordings in seated
normal volunteers and embedded in virtually all
ARDS ventilator management trials, values in
excess of 30 to 35 cm H2O are considered inju-
rious to the lungs. As evident from the pioneer-
Safe mechanical ventilation in patients without ARDS PANNU
ing studies of Greenfield, Webb, and Dreyfuss,
there is no biologic reason to consider intrinsi-
cally normal lungs less susceptible to injury by
large parenchymal stresses and strains than are
lungs with ARDS.5, 19, 32 Therefore, as long as
the contribution of the chest wall to Pplat is ≤10
cm H2O, said thresholds imply that lung paren-
chymal stress (trans pulmonary pressure at end
inspiration) is high and potentially injurious.
As was already discussed in the context of PEEP
management, in recumbent patients with large
abdominal loads the chest wall recoil pressure
can substantially exceed lung recoil. In such pa-
tients plateau pressures in excess of 30 cm H2O
should in theory prove safe. However, because
Pes and therefore lung and chest wall mechanics
are rarely measured in clinical practice, exceed-
ing conventional Pplat limits should be contem-
plated with extreme caution.
Oxygen (FiO2)
High oxygen concentrations cause cell dam-
age through production of reactive intermediates
called “free radicals”.85, 86 In the experimental
setting O2 toxicity manifests as tracheobronchi-
tis, pulmonary edema, diffuse alveolar damage
and fibrosis.85, 87, 88 Since these manifestation
are indistinguishable from those seen in all-cause
ARDS, it is generally impossible in the clinical
setting to attribute changes in lung structure and
function to O2 toxicity. Yet, in a large retrospec-
tive study of all ICU patients in the Netherlands,
the delivery of high FiO2 to hypoxemic as well
as hyperoxic patients was associated with excess
mortality.89 Similarly, therapeutic hyperoxia fol-
lowing cardiac arrest was associated with poor
outcomes.90
Even though practice norms of O2 dosing in
critically ill hypoxemic patients are largely based
on empiric evidence, the biologic rationale for
targeting O2 saturations >90% is reasonably
strong. The same cannot be said for O2 supple-
mentation to patients with near normal arterial
O2 tensions. However, compared to the preven-
tion and treatment of hypoxemia, very little ef-
fort seems directed at the prevention of hyper-
oxia or the avoidance of an unnecessarily high
FIO2 exposure.
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1037
Conclusions
All mechanically ventilated patients deserve
concern for lung protection. While patients with
ARDS are particularly vulnerable to injury by
deforming stress, the biophysical determinants
of VALI are generic and apply across the board.
Physical injury can occur in normal lungs particu-
larly during general anesthesia and in an intensive
care setting. While patients without ARDS offer
many more degrees of freedom in the choice of
ventilator settings we see little benefit from apply-
ing different rules and guidelines to their care. We
have outlined our approach by focusing on tidal
volume and PEEP. In general, lower tidal volumes
and moderate PEEP’s may be useful in minimiz-
ing lung injury. We suggest that using a simpli-
fied uniform approach to apply low tidal volumes
(6-8 mL/kg) pbw, wherever possible, may serve
in reducing medical errors. Moderate PEEP with
an understanding of chest wall mechanics is ben-
eficial. Frequent Recruitment maneuvers in this
setting have not shown to be beneficial. Evidence
for intraoperative ventilator settings is conflict-
ing. Further studies in this domain are warranted
and could help establish directions for clinicians.
While there are numerous choices in ventilation
mode, their effects on the lung can always be
traced back to absolute lung volume and volume
change, i.e., to stress and strain.
Key messages
—— Tidal volumes between 6 and 8 mL/
kg predicted body weight are both safe and
ought to meet gas exchange targets in the
vast majority of patients without ARDS.
—— We propose that standardization of
practice patterns, such as uniformity in tidal
volumes, wherever possible may serve in re-
ducing medical errors.
—— PEEP setting of 5 cm seems reason-
able in most circumstances, unless there is
impending or overt ARDS.
—— Frequent recruitment maneuvers in
this setting have not shown to be beneficial.
—— Attempts should be made to prevent
hyperoxia.
PANNU Safe mechanical ventilation in patients without ARDS
References
  1. Ventilation with lower tidal volumes as compared with tra-
ditional tidal volumes for acute lung injury and the acute
respiratory distress syndrome. The Acute Respiratory Dis-
tress Syndrome Network. N Engl J Med 2000;342:1301-8.
  2. Amato MB, Barbas CS, Medeiros DM, Magaldi RB, Schet-
tino GP, Lorenzi-Filho G et al. Effect of a protective-venti-
lation strategy on mortality in the acute respiratory distress
syndrome. N Engl J Med 1998;338:347-54.
  3. Plataki M, Hubmayr RD. The physical basis of ventilator-
induced lung injury. Expert Rev Respir Med 2010;4:373-
85.
 4. Gattinoni L, Pesenti A, Avalli L, Rossi F, Bombino M.
Pressure-volume curve of total respiratory system in acute
respiratory failure. Computed tomographic scan study. Am
Rev Respir Dis 1987;136:730-6.
  5. Dreyfuss D, Soler P, Basset G, Saumon G. High inflation
pressure pulmonary edema. Respective effects of high air-
way pressure, high tidal volume, and positive end-expirato-
ry pressure. Am Rev Respir Dis 1988;137:1159-64.
 6. Dreyfuss D, Basset G, Soler P, Saumon G. Intermittent
positive-pressure hyperventilation with high inflation pres-
sures produces pulmonary microvascular injury in rats. Am
Rev Respir Dis 1985;132:880-4.
 7. John E, McDevitt M, Wilborn W, Cassady G. Ul-
trastructure of the lung after ventilation. Br J Exp Pathol
1982;63:401-7.
  8. Tremblay L, Valenza F, Ribeiro SP, Li J, Slutsky AS. Inju-
rious ventilatory strategies increase cytokines and c-fos m-
RNA expression in an isolated rat lung model. J Clin Invest
1997;99:944-52.
  9. Dolinay T, Kim YS, Howrylak J, Hunninghake GM, An
CH, Fredenburgh L et al. Inflammasome-regulated cy-
tokines are critical mediators of acute lung injury. Ameri-
can Journal of Respiratory and Critical Care Medicine
2012;185:1225-34.
10. Terragni PP, Rosboch G, Tealdi A, Corno E, Menaldo E,
Davini O et al. Tidal hyperinflation during low tidal vol-
ume ventilation in acute respiratory distress syndrome. Am
J Respir Crit Care Med 2007;175:160-6.
11. Bilek AM, Dee KC, Gaver DP, 3rd. Mechanisms of surface-
tension-induced epithelial cell damage in a model of pul-
monary airway reopening. J Appl Physiol 2003;94:770-83.
12. Muscedere JG, Mullen JB, Gan K, Slutsky AS. Tidal venti-
lation at low airway pressures can augment lung injury. Am
J Respir Crit Care Med 1994;149:1327-34.
13. Oeckler RA, Lee WY, Park MG, Kofler O, Rasmussen DL,
Lee HB et al. Determinants of plasma membrane wounding
by deforming stress. Am J Physiol Lung Cell Mol Physiol
2010;299:L826-33.
14. Protti A, Cressoni M, Santini A, Langer T, Mietto C, Fe-
bres D et al. Lung stress and strain during mechanical ven-
tilation: any safe threshold? Am J Respir Crit Care Med
2011;183:1354-62.
15. Hubmayr RD. Straining to Make Mechanical Venti-
lation Safe and Simple. Am J Respir Crit Care Med
2011;183:1289-90.
16. Pelosi P, Croci M, Ravagnan I, Cerisara M, Vicardi P, Lissoni
A et al. Respiratory system mechanics in sedated, paralyzed,
morbidly obese patients. J Appl Physiol 1997;82:811-8.
17. Wrigge H, Uhlig U, Zinserling J, Behrends-Callsen E, Ot-
tersbach G, Fischer M et al. The effects of different ven-
tilatory settings on pulmonary and systemic inflammatory
responses during major surgery. Anesth Analg 2004;98:775-
81, table of contents.
18. Giraud T, Dhainaut JF, Vaxelaire JF, Joseph T, Journois D,
Bleichner G et al. Iatrogenic complications in adult inten-
sive care units: a prospective two-center study. Crit Care
Med 1993;21:40-51.
19. Bracco D, Favre JB, Bissonnette B, Wasserfallen JB, Revelly
1038 MINERVA ANESTESIOLOGICA September 2015
JP, Ravussin P et al. Human errors in a multidisciplinary in-
tensive care unit: a 1-year prospective study. Intensive Care
Med 2001;27:137-45.
20. Webb HH, Tierney DF. Experimental pulmonary edema
due to intermittent positive pressure ventilation with high
inflation pressures. Protection by positive end-expiratory
pressure. Am Rev Respir Dis 1974;110:556-65.
21. Tremblay LN, Slutsky AS. Ventilator-induced injury: from
barotraumas to biotrauma. Proc Assoc Am Physicians
1998;110:482-8.
22. Goldman HI, Becklake MR. Respiratory function tests;
normal values at median altitudes and the prediction of
normal results. Am Rev Tuberc 1959;79:457-67.
23. Westbrook PR, Stubbs SE, Sessler AD, Rehder K, Hyatt
RE. Effects of anesthesia and muscle paralysis on respiratory
mechanics in normal man. J Appl Physiol 1973;34:81-6.
24. Hedenstierna G, Brismar B, Strandberg A, Lundquist H,
Tokics L. New aspects on atelectasis during anaesthesia.
Clin Physiol 1985;5(Suppl 3):127-31.
25. Gajic O, Dara S, Mendez J, Adesanya A, Festic E, Caples S
et al. Ventilator-associated lung injury in patients without
acute lung injury at the onset of mechanical ventilation.
Crit Care Med 2004;32:1817-24.
26. Gajic O, Frutos-Vivar F, Esteban A, Hubmayr RD, An-
zueto A. Ventilator settings as a risk factor for acute respira-
tory distress syndrome in mechanically ventilated patients.
Intensive Care Med 2005;31:922-6.
27. Mascia L, Zavala E, Bosma K, Pasero D, Decaroli D, An-
drews P et al. High tidal volume is associated with the
development of acute lung injury after severe brain in-
jury: an international observational study. Crit Care Med
2007;35:1815-20.
28. Bratton SL, Davis RL. Acute lung injury in isolated trau-
matic brain injury. Neurosurgery 1997;40:707-12; discus-
sion 12.
29. Holland MC, Mackersie RC, Morabito D, Campbell AR,
Kivett VA, Patel R et al. The development of acute lung in-
jury is associated with worse neurologic outcome in patients
with severe traumatic brain injury. J Trauma 2003;55:106-
11.
30. Fumeaux T, Rothmeier C, Jolliet P. Outcome of mechani-
cal ventilation for acute respiratory failure in patients with
pulmonary fibrosis. Intensive Care Med 2001;27:1868-74.
31. Determann RM, Royakkers A, Wolthuis EK, Vlaar AP,
Choi G, Paulus F et al. Ventilation with lower tidal volumes
as compared with conventional tidal volumes for patients
without acute lung injury: a preventive randomized con-
trolled trial. Crit Care 2010;14:R1.
32. Severgnini P, Selmo G, Lanza C, Chiesa A, Frigerio A,
Bacuzzi A et al. Protective mechanical ventilation dur-
ing general anesthesia for open abdominal surgery im-
proves postoperative pulmonary function. Anesthesiology
2013;118:1307-21.
33. Greenfield LJ, Ebert PA, Benson DW. Effect of Positive
Pressure Ventilation on Surface Tension Properties of Lung
Extracts. Anesthesiology 1964;25:312-6.
34. Michelet P, D’Journo XB, Roch A, Doddoli C, Marin V,
Papazian L et al. Protective ventilation influences systemic
inflammation after esophagectomy: a randomized control-
led study. Anesthesiology 2006;105:911-9.
35. Licker M, Diaper J, Villiger Y, Spiliopoulos A, Licker V,
Robert J et al. Impact of intraoperative lung-protective in-
terventions in patients undergoing lung cancer surgery. Crit
Care 2009;13:R41.
36. Schilling T, Kozian A, Huth C, Buhling F, Kretzschmar M,
Welte T et al. The pulmonary immune effects of mechanical
ventilation in patients undergoing thoracic surgery. Anesth
Analg 2005;101:957-65, table of contents.
37. Bigatello LM, Allain R, Gaissert HA. Acute lung injury af-
ter pulmonary resection. Minerva Anestesiol 2004;70:159-
66.
Safe mechanical ventilation in patients without ARDS PANNU
38. Yang M, Ahn HJ, Kim K, Kim JA, Yi CA, Kim MJ et al.
Does a protective ventilation strategy reduce the risk of pul-
monary complications after lung cancer surgery?: a rand-
omized controlled trial. Chest 2011;139:530-7.
39. Asimakopoulos G, Smith PL, Ratnatunga CP, Taylor KM.
Lung injury and acute respiratory distress syndrome after
cardiopulmonary bypass. Ann Thorac Surg 1999;68:1107-
15.
40. Reis Miranda D, Gommers D, Struijs A, Dekker R, Mekel
J, Feelders R et al. Ventilation according to the open lung
concept attenuates pulmonary inflammatory response in
cardiac surgery. Eur J Cardiothorac Surg 2005;28:889-95.
41. Sundar S, Novack V, Jervis K, Bender SP, Lerner A, Panzica
P et al. Influence of low tidal volume ventilation on time
to extubation in cardiac surgical patients. Anesthesiology
2011;114:1102-10.
42. Wrigge H, Zinserling J, Stuber F, von Spiegel T, Hering
R, Wetegrove S et al. Effects of mechanical ventilation
on release of cytokines into systemic circulation in pa-
tients with normal pulmonary function. Anesthesiology
2000;93:1413-7.
43. Chang SY, Dabbagh O, Gajic O, Patrawalla A, Elie MC,
Talmor DS et al. Contemporary ventilator management
in patients with and at risk of ALI/ARDS. Respir Care
2013;58:578-88.
44. Levin MA, McCormick PJ, Lin HM, Hosseinian L, Fischer
GW. Low intraoperative tidal volume ventilation with
minimal PEEP is associated with increased mortality. Br J
Anaesth 2014;113:97-108.
45. Serpa Neto A, Cardoso SO, Manetta JA, Pereira VG,
Esposito DC, Pasqualucci Mde O et al. Association be-
tween use of lung-protective ventilation with lower tidal
volumes and clinical outcomes among patients without
acute respiratory distress syndrome: a meta-analysis. JAMA
2012;308:1651-9.
46. Gattinoni L. Counterpoint: Is low tidal volume mechani-
cal ventilation preferred for all patients on ventilation? No.
Chest 2011;140:11-3; discussion 4-5.
47. Serpa Neto A, Simonis FD, Barbas CS, Biehl M, Deter-
mann RM, Elmer J et al. Association between tidal vol-
ume size, duration of ventilation, and sedation needs in
patients without acute respiratory distress syndrome: an
individual patient data meta-analysis. Intensive Care Med
2014;40:950-7.
48. Walkey AJ, Wiener RS. Risk factors for underuse of lung-
protective ventilation in acute lung injury. J Crit Care
2012;27:323 e1-9.
49. Yilmaz M, Keegan MT, Iscimen R, Afessa B, Buck CF,
Hubmayr RD et al. Toward the prevention of acute lung
injury: protocol-guided limitation of large tidal volume
ventilation and inappropriate transfusion. Crit Care Med
2007;35:1660-6.
50. Elie-Turenne MC, Hou PC, Mitani A, Barry JM, Kao EY,
Cohen JE et al. Lung injury prediction score for the emer-
gency department: first step towards prevention in patients
at risk. Int J Emerg Med 2012;5:33.
51. Kor DJ, Warner DO, Alsara A, Fernandez-Perez ER, Mal-
inchoc M, Kashyap R et al. Derivation and diagnostic accu-
racy of the surgical lung injury prediction model. Anesthe-
siology 2011;115:117-28.
52. Levitt JE, Calfee CS, Goldstein BA, Vojnik R, Matthay
MA. Early acute lung injury: criteria for identifying lung
injury prior to the need for positive pressure ventilation*.
Crit Care Med 2013;41:1929-37.
53. Ashbaugh DG, Bigelow DB, Petty TL, Levine BE. Acute
respiratory distress in adults. Lancet1967. p. 319-23.
54. Haitsma JJ, Lachmann B. Lung protective ventilation
in ARDS: the open lung maneuver. Minerva Anestesiol
2006;72:117-32.
55. Hachenberg T, Lundquist H, Tokics L, Brismar B, Heden-
stierna G. Analysis of lung density by computed tomogra-
Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1039
phy before and during general anaesthesia. Acta Anaesthe-
siol Scand 1993;37:549-55.
56. Hubmayr RD. Perspective on lung injury and recruitment:
a skeptical look at the opening and collapse story. Ameri-
can Journal of Respiratory and Critical Care Medicine
2002;165:1647-53.
57. Sinclair SE, Chi E, Lin HI, Altemeier WA. Positive end-ex-
piratory pressure alters the severity and spatial heterogeneity
of ventilator-induced lung injury: an argument for cyclical
airway collapse. J Crit Care 2009;24:206-11.
58. D’Angelo E, Calderini E, Wolfler A, Pecchiari M. Factors
influencing the shape of the inspiratory flow. Respir Physiol
2001;126:211-9.
59. Pecchiari M. Effects of Heliox in Stable COPD Patients at
Rest and during Exercise. Pulm Med 2012;2012:593985.
60. Nava S, Rubini F. Lung and chest wall mechanics in venti-
lated patients with end stage idiopathic pulmonary fibrosis.
Thorax 1999;54:390-5.
61. Fernandez-Perez ER, Yilmaz M, Jenad H, Daniels CE, Ryu
JH, Hubmayr RD et al. Ventilator settings and outcome of
respiratory failure in chronic interstitial lung disease. Chest
2008;133:1113-9.
62. Aldenkortt M, Lysakowski C, Elia N, Brochard L, Tramer
MR. Ventilation strategies in obese patients undergoing
surgery: a quantitative systematic review and meta-analysis.
Br J Anaesth 2012;109:493-502.
63. Jones RL, Nzekwu MM. The effects of body mass index on
lung volumes. Chest 2006;130:827-33.
64. Rubinstein I, Zamel N, DuBarry L, Hoffstein V. Airflow
limitation in morbidly obese, nonsmoking men. Ann In-
tern Med 1990;112:828-32.
65. Mattingley JS, Holets SR, Oeckler RA, Stroetz RW, Buck
CF, Hubmayr RD. Sizing the lung of mechanically venti-
lated patients. Crit Care 2011;15:R60.
66. Loring SH, O’Donnell CR, Behazin N, Malhotra A, Sarge
T, Ritz R et al. Esophageal pressures in acute lung injury: do
they represent artifact or useful information about transpul-
monary pressure, chest wall mechanics, and lung stress? J
Appl Physiol 2010;108:515-22.
67. Mergoni M, Martelli A, Volpi A, Primavera S, Zuccoli P,
Rossi A. Impact of positive end-expiratory pressure on chest
wall and lung pressure-volume curve in acute respiratory
failure. Am J Respir Crit Care Med 1997;156(3 Pt 1):846-
54.
68. Talmor D, Sarge T, O’Donnell CR, Ritz R, Malhotra A,
Lisbon A et al. Esophageal and transpulmonary pressures in
acute respiratory failure. Crit Care Med 2006;34:1389-94.
69. Gillespie DJ, Lai YL, Hyatt RE. Comparison of esophageal
and pleural pressures in the anesthetized dog. J Appl Physiol
1973;35:709-13.
70. Suratt PM, Wilhoit SC, Hsiao HS, Atkinson RL, Roches-
ter DF. Compliance of chest wall in obese subjects. J Appl
Physiol 1984;57:403-7.
71. Pelosi P, Ravagnan I, Giurati G, Panigada M, Bottino N,
Tredici S et al. Positive end-expiratory pressure improves res-
piratory function in obese but not in normal subjects dur-
ing anesthesia and paralysis. Anesthesiology 1999;91:1221-
31.
72. Reinius H, Jonsson L, Gustafsson S, Sundbom M, Du-
vernoy O, Pelosi P et al. Prevention of atelectasis in mor-
bidly obese patients during general anesthesia and pa-
ralysis: a computerized tomography study. Anesthesiology
2009;111:979-87.
73. Talab HF, Zabani IA, Abdelrahman HS, Bukhari WL,
Mamoun I, Ashour MA et al. Intraoperative ventilatory
strategies for prevention of pulmonary atelectasis in obese
patients undergoing laparoscopic bariatric surgery. Anesth
Analg 2009;109:1511-6.
74. Valenza F, Vagginelli F, Tiby A, Francesconi S, Ronzoni
G, Guglielmi M et al. Effects of the beach chair position,
positive end-expiratory pressure, and pneumoperitoneum
PANNU Safe mechanical ventilation in patients without ARDS

on respiratory function in morbidly obese patients during
anesthesia and paralysis. Anesthesiology 2007;107:725-32.
75. Erlandsson K, Odenstedt H, Lundin S, Stenqvist O. Posi-
tive end-expiratory pressure optimization using electric
impedance tomography in morbidly obese patients during
laparoscopic gastric bypass surgery. Acta Anaesthesiol Scand
2006;50:833-9.
76. Richard JC, Maggiore SM, Jonson B, Mancebo J, Lemaire
F, Brochard L. Influence of tidal volume on alveolar recruit-
ment. Respective role of PEEP and a recruitment maneuver.
Am J Respir Crit Care Med 2001;163:1609-13.
77. Mekontso Dessap A, Voiriot G, Zhou T, Marcos E, Dudek
SM, Jacobson JR et al. Conflicting physiological and ge-
nomic cardiopulmonary effects of recruitment maneuvers
in murine acute lung injury. Am J Respir Cell Mol Biol
2012;46:541-50.
78. Badet M, Bayle F, Richard JC, Guerin C. Comparison of
optimal positive end-expiratory pressure and recruitment
maneuvers during lung-protective mechanical ventilation
in patients with acute lung injury/acute respiratory distress
syndrome. Respir Care 2009;54:847-54.
79. Jardin F. Recruitment maneuvers during lung protective
ventilation in acute respiratory distress syndrome. Am J
Respir Crit Care Med 2003;167:92; author reply
80. Villagra A, Ochagavia A, Vatua S, Murias G, Del Mar Fern-
andez M, Lopez Aguilar J et al. Recruitment maneuvers
during lung protective ventilation in acute respiratory dis-
tress syndrome. Am J Respir Crit Care Med 2002;165:165-
70.
81. Forgiarini Junior LA, Rezende JC, Forgiarini SG. Alveolar
recruitment maneuver and perioperative ventilatory sup-
port in obese patients undergoing abdominal surgery. Rev
Bras Ter Intensiva 2013;25:312-8.
82. Huynh T, Messer M, Sing RF, Miles W, Jacobs DG, Tho-
mason MH. Positive end-expiratory pressure alters intrac-
ranial and cerebral perfusion pressure in severe traumatic
brain injury. J Trauma 2002;53:488-92; discussion 92-
3.
83. Frost EA. Effects of positive end-expiratory pressure on
intracranial pressure and compliance in brain-injured pa-
tients. J Neurosurg 1977;47:195-200.
84. Mascia L, Grasso S, Fiore T, Bruno F, Berardino M, Ducati
A. Cerebro-pulmonary interactions during the application
of low levels of positive end-expiratory pressure. Intensive
Care Med 2005;31:373-9.
85. Bryan CL, Jenkinson SG. Oxygen toxicity. Clin Chest Med
1988;9:141-52.
86. Kazzaz JA, Xu J, Palaia TA, Mantell L, Fein AM, Horowitz
S. Cellular oxygen toxicity. Oxidant injury without apopto-
sis. J Biol Chem 1996;271:15182-6.
87. Comroe J. OXYGEN TOXICITY, The effect of breath-
ing high concentration of oxygen for 24 hours in nor-
mal men at sea level and simulated at 18,000 feet. JAMA
1945;128:710-7.
88. Nash G, Blennerhassett JB, Pontoppidan H. Pulmonary
lesions associated with oxygen therapy and artifical ventila-
tion. N Engl J Med 1967;276:368-74.
89. de Jonge E, Peelen L, Keijzers PJ, Joore H, de Lange D, van
der Voort PH et al. Association between administered oxy-
gen, arterial partial oxygen pressure and mortality in me-
chanically ventilated intensive care unit patients. Crit Care
2008;12:R156.
90. Kilgannon JH, Jones AE, Shapiro NI, Angelos MG, Mil-
carek B, Hunter K et al. Association between arterial hyper-
oxia following resuscitation from cardiac arrest and in-hos-
pital mortality. JAMA 2010;303:2165-71.

Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on May 6, 2013. - Accepted for publication on January 16, 2015. - Epub ahead of print on January 19, 2015.
Corresponding author: R. D. Hubmayr, MD, 200 First street SW, Rochester, MN, USA. E-mail: rhubmayr@mayo.edu

1040 MINERVA ANESTESIOLOGICA September 2015



LETTER TO THE EDITOR

Non-adherence to CPAP and prevention of

postoperative complications in OSA: what
are the limits of anesthesia consultation?
A. M. ESQUINAS 1, G. INSALACO 2

1IntensiveCare  and Non Invasive Ventilatory Unit,Hospital Morales Meseguer, Murcia, Spain; 2National Research
Council of Italy, Institute of Biomedicine and Molecular Immunology, Sleep Laboratory, Palermo, Italy

Dear Editor, Third, the authors used an interview to detect this high risk
population during anesthesiologist consults. Although this pro-
cedure is simple and appropriate, information derived from this
O bstructive sleep apnea syndrome (OSA) represents a high
risk condition related to perioperative cardiopulmonary
complications.1 Continuous positive airway pressure (CPAP) is
interview and potential mechanical psychological/educational
interventions are still controversial. We know that impact of
educational program is low with low quality of evidence. Two
the keystone in OSA. However, in some cases, non-adherence
main questions: could a brief short-term interview and educa-
to CPAP is a mainly adverse factor that requires early knowl-
tional intervention result in a modest increase in CPAP usage
edge by the anesthesiologist. Currently, there are several specific
before surgery? Which is the definition of optimal timing of
protocols to improve adherence to CPAP that are not always
interventions?
well-encoded.
We think that a proper response to these main questions
We read with great interest the original study by Deflandre
still remains uncertain. Further prospective long term clinical
et al.2, which described how non-adherent CPAP users show
trials need to define interventions and times in non-adherent
“feeling of breathlessness” which is a key determinant symptom
CPAP- OSA population that require surgery interventions.5
of non-adherence to treatment and three criteria of adherence
(awareness of the risk of complications, awareness of treatment
efficacy and feeling to be less tired with CPAP therapy) that References
should preoperatively be sought, in order to detect non-adher-
ent patients more efficiently during anesthesia consultation. We   1. Hai F, Porhomayon J, Vermont L, Frydrych L, Jaoude P,
agree that before surgery, detect non-adherent CPAP patients is El-Solh AA. Postoperative complications in patients with
a rational and practical strategy to prevent postoperative compli- obstructive sleep apnea: a meta-analysis. J Clin Anesth
2014;26:591-600.
cations, however, in this study some others non-adherent CPAP   2. Deflandre E, Degey S, Bonhomme V, Donneau AF, Poirrier
causes should be taken into account. R, Brichant JF et al. Preoperative adherence to continuous
First, the “feeling of breathlessness” reported during CPAP positive airway pressure among obstructive sleep apnea pa-
therapy could be improved by other factors associated with OSA tients. Minerva Anestesiol 2015;81:960-7.
that could be measured easily such as: 1) evaluation of psycho-   3. Guralnick AS, Pant M, Minhaj M, Sweitzer BJ, Mokhlesi
B. CPAP adherence in patients with newly diagnosed ob-
logical factors, especially prevalence of depressive symptoms that structive sleep apnea prior to elective surgery. J Clin Sleep
are often found during CPAP therapy in non-adherent patients;3 Med 2012;8:501-6.
2) influence of pharmacology drugs as antidepressants and anxio-   4. Sopkova Z, Dorkova Z, Tkacova R. Predictors of compli-
lytics which induce sleep disturbances and are predictors of poor ance with continuous positive airway pressure treatment in
CPAP compliance;4 and 3) evaluation of cognitive function. patients with obstructive sleep apnea and metabolic syn-
Second, information regarding some key physiological drome. Wien Klin Wochenschr. 2009;121:398-404.
 5. Haniffa M, Lasserson TJ, Smith I. Interventions to im-
measurements, such as: 1) mean sleep SpO2; 2) mask leaks as prove compliance with continuous positive airway pressure
predictor of CPAP compliance; and 3) information regarding for obstructive sleep apnoea. Cochrane Database Syst Rev
polysomnography (OPS) during visits.4 2004;CD003531.

Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on February 13, 2015. - Accepted for publication on February 20, 2015. - Epub ahead of print on February 25, 2015.
Corresponding author: A. M. Esquinas, Avenida Marques de Los Velez s/n, Murcia, 30.008; Spain. E-mail: antmesquinas@gmail.com

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1041


LETTER TO THE EDITOR
Obstructive sleep apnea and detection
of non-adherence to CPAP in OSA:
limits of the preanesthesia visit
E. DEFLANDRE 1, J. F. BRICHANT 2, V. BONHOMME 3
1Department of Anesthesia, Clinique Saint-Luc, Bouge, Belgium & Cabinet Medical ASTES, Jambes,
Belgium;2Department of Anesthesia and ICM, CHU Liege, Liege, Belgium;3University Department of Anesthesia and
ICM, CHR Citadelle and CHU Liege, Liege, Belgium
Dear Editor,
W e read with great interest the letter by Esquinas and
Insalaco and thank them for their pertinent comments
on our paper.1 The aim of our study was to isolate factors of
poor-adherence in order to detect these patients and give them
appropriate postoperative care such as opioid use limitation,
close monitoring in the postanesthesia care unit.2
First, we agree with them in the fact that feeling of breath-
lessness could be complemented by several other elements.
They mention the most frequently cited.1 To date, no research
has specifically addressed the impact of an educational program
on these factors for non-adherence.
Second, we also agree with them that several measurements
should be made during the preoperative visit. In our Institu-
tion, we evaluate the mask leaks and collect the SpO2 (mean
and minimum) during polysomnography (PSG). We consider
the analysis of Oxygen Desaturation Index (ODI) derived
from the PSG ‑ representing the number of arterial desatura-
tion per hour ‑ very interesting. We believe that these param-
eters could be more accurate at predicting postoperative com-
plications than the Apnea Hypopnea Index (AHI, i.e. number
of apneas and hypopneas per hour). Nevertheless, to date, this
has never been clearly demonstrated by a clinical study. The
PSG can also give us very important information to differenti-
ate apneic patients. For example, some patients present much
more apneas during REM sleep. Knowing that patients have
a rebound of REM sleep around the third postoperative day,3
those REM sleep apneic patients should deserve special atten-
tion during that period. This has never been looked over ei-
ther. Finally, the type of surgery should also help us applying
the right care to the right patient. All apneic patients do not
need the same precautions. In our practical expertise, we take
1042 MINERVA ANESTESIOLOGICA September 2015
account of the results of the OSA Scoring System (proposed
by the American Society of Anesthesiologists) when managing
these patients during our daily practice. These six-point scale
gives a theoretical probability of postoperative complications
among OSA patients.4
Third, we agree with the authors on the fact that an edu-
cational program is still controversial in ameliorating general
adherence to CPAP, and particularly adherence outside the
preoperative period. As stated by Shi and Warner, the preop-
erative period is certainly a “teachable moment”.5 But, actually,
there is no clear answer to knowing whether a brief short-term
interview and educational intervention will result in a modest
increase in CPAP usage before surgery. The optimal timing of
interventions is also not well defined. Specific studies should
be designed to address these points, with foreseeable limita-
tions: indeed, it will be uneasy to blindly compare a group
with educational intervention and a group without interven-
tion. Comparing long-term and brief intervention should be
easier to perform.
As one can see, there is still a lot of research to be done in
the field of OSA syndrome and perioperative care.
References
 1. Esquinas AM, Insalaco G. Non-adherence to CPAP and
prevention of postoperative complications in OSA. What
are the limits of anesthesia consultation? Minerva Anest-
esiol 2015;81:1041.
  2. Deflandre E, Degey S, Bonhomme V, Donneau AF, Poirrier
R, Brichant JF et al. Preoperative adherence to continuous
positive airway pressure among obstructive sleep apnea pa-
tients. Minerva Anestesiol 2014 [Epub ahead of print].
  3. Krenk L, Jennum P, Kehlet H. Sleep disturbances after fast-
track hip and knee arthroplasty. Br J Anaesth 2012;109:769-
75.
 LETTER TO THE EDITOR

 4. Practice Guidelines for the Perioperative Management of Sleep Apnea. Anesthesiology 2014;120:268-86 10.1097/
Patients with Obstructive Sleep Apnea: An Updated Report ALN.0000000000000053.
by the American Society of Anesthesiologists Task Force on   5. Shi Y, Warner D. Surgery as a teachable moment for smok-
Perioperative Management of Patients with Obstructive ing cessation. Anesthesiology 2010;112:102-7.

Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on March 15, 2015. - Accepted for publication on March 23, 2015. - Epub ahead of print on March 24, 2015.
Corresponding author: E. Deflandre, Chaussee de Tongres 29, B 4000 Liege, Belgium. E-mail: eric.deflandre@gmail.co

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1043



LETTER TO THE EDITOR

An additional tip to facilitate glidescope intubation

T. TURKSTRA, M. RACHINSKY, P. BATOHI

Department of Anesthesia and Peri-operative Medicine, London, Ontario, Canada

Dear Editor, caution is warranted, similar to the portion of ETT insertion

W e read with interest the recent excellent review on the
GlideScope® videolaryngoscopes.1 They, and video-
laryngoscopes in general, have proved to be a valuable aid to
intubation, demonstrating improved laryngoscopic view and
rescuing many airway emergencies.1 Their use is strongly sup-
ported in the recent ASA and CAFG guidelines 2,3 and some
have proposed their default use.4
While we would hesitate to accept expert designation,1 we
would share an additional “tip” for to Table II.1 As noted, there
have been several cases where, despite a reasonable view of the
glottic aperture via the GlideScope® videolaryngoscope, it has
been difficult to impossible to advance the endotracheal tube
(ETT) into the trachea.1, 5, 6 Intubation has not proved possible
despite a good view of the glottic aperture.
When we encountered this phenomenon of a reasonable
glottic view but difficulty in advancing the ETT into the tra-
chea, there was the perception that the GlideScope® blade was
impeding the advance of the ETT at some point in the phar-
ynx. The situation was not improved by manipulation of the
cricoid structures or mouth/lip manipulation by an assistant.
We then employed a novel technique in changing the inser-
tion sequence. The GlideScope® was removed and the ETT
inserted gently on its own into the pharynx, in the general
vicinity of the epiglottis. When the GlideScope® was subse-
quently inserted, the tip of the ETT was noted to be in close
proximity to the glottis and the ETT was easily advanced into
the trachea to complete the intubation. The ETT was utilized
with a malleable stylet, with a 90 degree bend, 8 cm from the
distal tip.7
Subsequent to this, we have been called several times to as-
sist colleagues using the GlideScope®, and have successfully
used this technique of primarily inserting the ETT to easily
complete intubation.
Trauma has been reported with GlideScope® use.8, 9 Con-
sidering that the ETT is inserted without direct visualisation,
between direct-view and videolaryngoscope-view with the
standard GlideScope® technique.
Thus, we would offer this technique to other clinicians as a
potential rescue technique if they find difficulty in advancing
the ETT despite a good view of the glottic aperture. We would
emphasize that we have used this technique only with the Gli-
deScope®, and it may not necessarily be extrapolated to other
videolaryngoscopes or devices.
References
 1. Agrò FE, Doyle DJ, Vennari M. Use of GlideScope® in
adults: an overview. Minerva Anestesiol 2015;81:342-51.
 2. Enterlein G, Byhahn C, American Society of Anesthesi-
ologists Task Force. [Practice guidelines for management
of the difficult airway: update by the American Society of
Anesthesiologists task force]. Anaesthesist 2013;62:832-5.
  3. Law JA, Broemling N, Cooper RM, Drolet P, Duggan LV,
Griesdale DE et al. The difficult airway with recommen-
dations for management--part 1--difficult tracheal intuba-
tion encountered in an unconscious/induced patient. Can J
Anaesth 2013;60:1089-118.
  4. Zaouter C, Calderon J, Hemmerling TM. Videolaryngosco-
py as a new standard of care. Br J Anaesth 2015;114:181-3.
  5. Doyle DJ. The GlideScope video laryngoscope. Anaesthesia
2005;60:414-5.
  6. Niforopoulou P, Pantazopoulos I, Demestiha T, Koudouna
E, Xanthos T. Video-laryngoscopes in the adult airway
management: a topical review of the literature. Acta Anaes-
thesiol Scand 2010;54:1050-61.
 7. Jones PM, Turkstra TP, Armstrong KP, Armstrong PM,
Cherry RA, Hoogstra J, Harle CC. Effect of stylet angula-
tion and endotracheal tube camber on time to intubation
with the GlideScope. Can J Anaesth;2007;54:21-7.
  8. Chin KJ, Arango MF, Paez AF, Turkstra TP. Palatal injury
associated with the GlideScope. Anaesthesia and intensive
care 2007;35:449-50.
 9. Cooper RM. Complications associated with the use of the
GlideScope videolaryngoscope. Can J Anaesth 2007;54:54-7.

Conflicts of interest.—The authors certify that there is no conflict of interest with any financial organization regarding the material discussed
in the manuscript.
Received on April 15, 2015. - Accepted for publication on April 29, 2015. - Epub ahead of print on April 30, 2015.
Corresponding author: T. P. Turkstra, Department of Anesthesia and Peri-operative Medicine, University of Western Ontario, London
Health Sciences Center, 339 Winderemere Road, London, Ontario, Canada N6A 5A5. E-mail: timothy.turkstra@lhsc.on.ca

1044 MINERVA ANESTESIOLOGICA September 2015



T O P 5 0 M I N E RVA A N E S T E S I O L O G I C A R E V I E W E R S

Most active reviewers between February 2015-July 2015

Number
Surname Name of revisions
Verheijde Joseph L. 7
Peris Adriano 7
Caruselli Marco 7
Brogly Nicolas 7
Faraoni David 6
Brazzi Luca 6
Cata Juan P. 6
Cattano Davide 6
Borghi Battista 6
Savoia Gennaro 6
Leone Marc 5
Della Rocca Giorgio 5
Gómez-Ríos Manuel Ángel 5
Chelazzi Cosimo 5
Bertini Pietro 5
Dalfino Lidia 5
Lirk Philipp 5
Dauri Mario 5
Deflandre Eric P. 5
Torgersen Christian 5
Alston Theodore 4
Abad Gurumeta Alfredo 4
Del Sorbo Lorenzo 4
Agrò Eugenio Felice 4
Barbas Silvia Valente Carmen 4
Langer Thomas 4
Vincent Jean Louis 4
Aceto Paola 4
Pabelick Christina Maria 4
Sbaraglia Fabio 4
Cereda Maurizio 4
Tritapepe Luigi 3
Masclans Joan R 3
Oropello John Mark 3
Diaz-Gomez Jose L. 3
Li Bassi Gianluigi 3
Loop Torsten 3
Czarnik Tomasz 3
Ranucci Marco 3
Aly Essam 3
Staals Lonneke 3
Page Valerie 3
Guarracino Fabio 3
Bruder Nicolas 3
De Gasperi Andrea 3
Caldiroli Dario 2
Mariano Filippo 2
Mauri Tommaso 2
Carmona Paula 2
Cressoni Massimo 2

Vol. 81 - No. 9 MINERVA ANESTESIOLOGICA 1045