Journal of Parenteral and Enteral

Nutrition http://pen.sagepub.com/

Nutrition Therapy for the Critically Ill Surgical Patient: We Need To Do Better!
John W. Drover, Naomi E. Cahill, Jim Kutsogiannis, Giuseppe Pagliarello, Paul Wischmeyer, Miao Wang, Andrew G. Day
and Daren K. Heyland
JPEN J Parenter Enteral Nutr 2010 34: 644
DOI: 10.1177/0148607110372391

The online version of this article can be found at:

http://pen.sagepub.com/content/34/6/644

Published by:

http://www.sagepublications.com

On behalf of:

The American Society for Parenteral & Enteral Nutrition

Additional services and information for Journal of Parenteral and Enteral Nutrition can be found at:

Email Alerts: http://pen.sagepub.com/cgi/alerts

Subscriptions: http://pen.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

>> Version of Record - Nov 19, 2010

What is This?

Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014

Knowledge in Translation
Nutrition Therapy for the Critically
Ill Surgical Patient: We Need
To Do Better!
John W. Drover, MD, FRCSC1; Naomi E. Cahill, RD, MSc2,4;
Jim Kutsogiannis, MD, MHS, FRCPC7; Giuseppe Pagliarello, MD, FRCSC6;
Paul Wischmeyer, MD5; Miao Wang, MD, MSc4;
Andrew G. Day, MSc4; and Daren K. Heyland, MD, FRCPC2,3,4
Financial disclosure: Naomi Cahill holds a Canadian Institute for Health Research fellowship in Knowledge Translation.
Background: To identify opportunities for quality improvement,
the nutrition adequacy of critically ill surgical patients, in con-
trast to medical patients, is described. Methods: International,
prospective, and observational studies conducted in 2007 and
2008 in 269 intensive care units (ICUs) were combined for
purposes of this analysis. Sites provided institutional and patient
characteristics and nutrition data from ICU admission to ICU
discharge for maximum of 12 days. Medical and surgical
patients staying in ICU at least 3 days were compared. Results:
A total of 5497 mechanically ventilated adult patients were
enrolled; 37.7% had surgical ICU admission diagnosis. Surgical
patients were less likely to receive enteral nutrition (EN) (54.6%
vs 77.8%) and more likely to receive parenteral nutrition (PN)
(13.9% vs 4.4%) (P < .0001). Among patients initiating EN in
ICU, surgical patients started EN 21.0 hours later on average
(57.8 vs 36.8 hours, P < .0001). Consequently, surgical patients
received less of their prescribed calories from EN (33.4% vs
T
he benefits of nutrition in critically ill patients
have been demonstrated in several studies and
reviews.1-3 In the literature, a variety of potential
From the 1Department of Surgery, 2Department of Community
Health and Epidemiology, 3Department of Medicine, Queen’s
University, Kingston, Ontario, Canada; 4Clinical Evaluation
Research Unit, Kingston General Hospital, Kingston, Ontario,
Canada; 5Department of Anesthesiology, University of Colorado
School of Medicine, Denver, Colorado; 6Departments of Surgery
and Critical Care, Ottawa Hospital and University of Ottawa,
Ottawa, Ontario, Canada; and 7Division of Critical Care
Medicine, University of Alberta, Edmonton, Alberta, Canada.
Submitted on October 14, 2009. Accepted for publication
February 2, 2010.
Address correspondence to: Daren K. Heyland, 76 Stuart St,
Angada 4, Room 5-416, Kingston, ON, K7L 2V7 Canada;
e-mail: dkh2@queensu.ca.
644
Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014
Journal of Parenteral and
Enteral Nutrition
Volume 34 Number 6
November 2010 644-652
© 2010 American Society for
Parenteral and Enteral Nutrition
10.1177/0148607110372391
http://jpen.sagepub.com
hosted at
http://online.sagepub.com
49.6%, P < .0001) or from all nutrition sources (45.8% vs
56.1%, P < .0001). These differences remained after adjustment
for patient and site characteristics. Patients undergoing cardio-
vascular and gastrointestinal surgery were more likely to use
PN, were less likely to use EN, started EN later, and had lower
total nutrition and EN adequacy rates compared with other
surgical patients. Use of feeding and/or glycemic control proto-
cols was associated with increased nutrition adequacy.
Conclusions: Surgical patients receive less nutrition than medi-
cal patients. Cardiovascular and gastrointestinal surgery patients
are at highest risk of iatrogenic malnutrition. Strategies to
improve nutrition performance, including use of protocols, are
needed. (JPEN J Parenter Enteral Nutr. 2010;34:644-652)
Keywords:   enteral nutrition; parenteral nutrition; nutrition
therapy; nutrition support; intensive care unit; critical care;
critically ill; surgery
benefits are attributed to nutrition in surgery-associated
critical illness. The physiologic improvements associated
with nutrition therapy lead to improved wound healing,
as demonstrated in a variety of animal and human stud-
ies.4,5 Haydock et al6 proposed a model of improved col-
lagen deposition and wound strength. The effects of
stress-induced hypercatabolism are blunted via enteral
nutrition (EN).7 EN has also been shown to reverse
enteric mucosal atrophy and enhance epithelial cell
function.8,9 The role of EN in the maintenance of gut
mucosa has been linked to improved immunologic func-
tion and a reduction in the systemic inflammatory
response to surgical trauma,10-13 although this concept
remains controversial.14 Improved patient outcome has
also been attributed to postoperative EN. These include
decreased complications, reduced hospital length of stay
and hospital costs, and improved safety compared with
parenteral nutrition (PN).7,15-17
 Nutrition for Critically Ill Surgical Patients / Drover et al   645
Patients with preoperative malnutrition have a sig-
nificantly higher risk of postoperative complication and
death.16,17 Perioperative malnutrition has also been asso-
ciated with increased hospital length of stay and overall
costs.17 Poor nutrition compromises the function of the
immune response and many organ systems17 and impairs
wound healing. In the postoperative period, delayed or
insufficient nutrition has been associated with higher
complication rates and mortality.17 These deleterious
effects of nutrition deficiency suggest that perioperative
nutrition support may positively affect outcomes in surgi-
cal patients.
As described above, a substantial body of evidence
shows that nutrition therapy and specifically EN are
associated with improved patient outcomes in surgical
patients. In the context of a large, multicenter observa-
tional study, we have collected key nutrition process and
outcome data on all intensive care unit (ICU) patients.
This allows us to compare how nutrition is delivered in
surgical patients compared with medical patients. This
analysis then allows us to look at gaps that may exist
between what would be considered nutrition best prac-
tice and nutrition therapy as it is actually practiced.
The purpose of this study is to describe the nutrition
practices for critically ill surgical patients and identify
opportunities for improving nutrition therapies in this
population.
Methods
Study Design and Participants
In 2007 and 2008, we conducted 2 international, pro-
spective, observational studies of nutrition practices in
critical care units around the world. Given that the same
inclusion criteria and data collection strategies were used
for both studies, the data sets were combined to make a
larger, more comprehensive data set for the purposes of
this analysis. The details of our methods have been previ-
ously reported, but in brief, sites that had participated in
previous audits were invited to take part again.18-20
New sites were invited to participate through mem-
bership lists of dietetic associations and critical care
research communities around the world. Clinicians
known to have a particular interest in critical care nutri-
tion were contacted individually. We advertised the
survey on various international websites, including our
own.21
To be eligible, ICUs had to have at least 8 beds and a
willing individual with knowledge of clinical nutrition to
complete data collection. We excluded ICUs with <8 beds
because these units do not typically care for ventilated
patients for >24 hours and may not have the resources for
completing data collection. However, smaller units that
expressed a keen interest in participating and had the
Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014
resources to do so were granted permission to participate
on a case by case basis.
Data Collection
Participating sites provided information that described the
hospital, the characteristics of the ICU, and the use of
nutrition-related protocols within their unit. Participating
researchers identified all eligible patients who were in the
ICU on the days that we initiated the survey (January 14,
2007, and May 14, 2008). Critically ill adult patients (ie,
≥18 years of age) who were mechanically ventilated within
the first 48 hours of admission to ICU and who remained
in ICU for >72 hours were consecutively enrolled in the
study. Patients were excluded if they were not intubated or
were receiving mask ventilation or if they were in the ICU
for <3 days prior to death or permanent progression to
exclusively oral feeding. Each participating ICU aimed to
recruit 20 patients. Data were subsequently collected from
the hospital records (with exception of head of the bed
elevation, which was optional and obtained through direct
observation at the point of enrollment). Data were entered
online using a secure web-based data collection tool. The
clinical management of study patients was left to the dis-
cretion of the treating clinical team; no attempt was made
to influence or change management. The information col-
lected on each enrolled patient included sex, age, admis-
sion category (surgical vs medical), Acute Physiology and
Chronic Health Evaluation (APACHE) II score22 and diag-
nosis category, presence of acute respiratory distress syn-
drome (ARDS), height, weight, and baseline nutrition
assessment (ie, energy and protein prescribed). The pres-
ence or absence of ARDS was based on chart abstraction
using standard definitions23; no verification of this determi-
nation was made centrally. Daily nutrition information was
collected on the type and amount of nutrition received,
morning blood glucose levels, total insulin dose, supple-
mental glutamine and selenium use, and the use of promo-
tility drugs. Reasons for which EN was contraindicated
were recorded and included mechanical bowel obstruction,
bowel ischemia, small bowel ileus, small bowel fistulae,
gastrointestinal perforation, and short gut syndrome. Daily
information was recorded for a maximum of 12 days unless
death or ICU discharge occurred sooner. ICU and hospital
outcomes were determined at 60 days.
Data Management
Range, data completeness, and logic checks were incor-
porated into the web-based data collection tool and data-
base. Participants were required to resolve queries
regarding accuracy of the data prior to being permitted to
finalize data entry of a study patient. The entered data
were checked by the project leader to identify errors,
inconsistencies, and omissions. Data queries were sent to
the participating ICU for clarification and resolution.
646   Journal of Parenteral and Enteral Nutrition / Vol. 34, No. 6, November 2010

Statistical Analysis Table 1.   Characteristics of Participating

ICUs (N = 269)a
Our primary objective was to compare the characteris-
tics, nutrition processes, and outcome variables between Region, n (%)
medical and surgical patients. Appropriate nutrition Canada 57 (21.2)
adequacy and adequacy from the enteral route alone Australia and New , Zealand 35 (13.0)
were our primary goals. We defined appropriate nutri- USA 77 (28.6)
tion adequacy as the total amount of calories or protein Europe and South Africa 46 (17.1)
received from EN and propofol plus PN when there is a China 26 (9.7)
contraindication for EN over the first 12 ICU days, Asia 14 (5.2)
divided by the amount prescribed as per the baseline Latin America 14 (5.2)
assessment and expressed as a percentage. Days without Hospital type, n (%)
Teaching 213 (79.2)
EN or PN were included and counted as 0% adequacy.
Nonteaching 56 (20.8)
Days following permanent progression to exclusive oral Size of hospital (beds), mean (range) 647.8 (108-4000)
intake were excluded from the calculation of nutrition ICU structure, n (%)
adequacy. EN adequacy was calculated similarly but Open 70 (26.0)
includes only the calories or protein received by the EN Closed 195 (72.5)
route in the numerator. Other 4 (1.5)
Statistical analyses include all eligible patients Presence of medical director, n (%)
contributed from both years at each site. Categorical Yes 250 (92.9)
variables were compared between groups by the Rao- No 19 (7.1)
Scott adjusted χ2 method to adjust for potential hetero- Case types, n (%)
Medical 233 (86.6)
geneity between ICUs due to the clustered 2-stage
Surgical 240 (89.2)
sampling design.24 Length of stay variables were com-
Trauma 164 (61.0)
pared between groups by the score test with robust Pediatric 33 (12.3)
standard errors. This method is similar to the log-rank Neurological 177 (65.8)
test in the absence of between-ICU heterogeneity. Neurosurgical 137 (50.9)
Continuous variables were compared between groups Cardiac surgery 87 (32.3)
by a linear mixed effects model that accounted for Burns 46 (17.1)
dependence between patients within the same ICU and Other 30 (11.5)
year (a 3-level multilevel model). We modeled appro- Size of ICU (beds), mean (range) 17.6 (4-75)
priate nutrition adequacy by first modeling the effect ICU has feeding protocol, n (%)
of each patient and site-level characteristic separately Yes 208 (77.3)
No 61 (22.7)
after adjusting for year and number of the first 12 days
Presence of dietitian(s), n (%)
in the ICU. Then backward selection with a retention
Yes 214 (79.6)
criterion of P < .1 was used to select a model of inde- No 55 (20.5)
pendent predictors of appropriate nutrition adequacy. Full-time equivalent dietitian
We assessed whether continuous predictors were linear (per 10 beds), mean (range) 0.4 (0-6.7)
by using the likelihood ratio test to measure the statis- Glycemic control protocol, n (%)
tical significance of the nonlinear terms modeled by a Yes 232 (86.3)
restricted cubic spline with 5 knots.25 Each predictor No 37 (13.8)
was tested for an interaction with the indicator variable a
102 ICUs participated in 2007 only; 102 ICUs participated in
for admission category. A prespecified subgroup analy- 2008 only; 65 ICUs that participated in both 2007 and 2008 are
sis compared the following 3 groups of surgical patients: reported using 2008 responses.
(1) cardiovascular/vascular, (2) gastrointestinal/pan-
creatitis, and (3) other surgical admission diagnoses.
All tests were 2-sided. Analyses were completed with
SAS (statistics analysis software) 9.1 (SAS Institute, Results
Cary, NC).
Institutional ethics approval was obtained from the In 2007, 167 ICUs from 21 countries participated in
Health Sciences Research Ethics Board at Queen’s the survey. In 2008, the same number from 18 countries
University, Kingston, Ontario, and additional centers if participated; 65 ICUs from 10 countries participated in
required for their participation. The need for informed both years. Thus, data from 269 unique ICUs from 29
patient consent was waived given the observational countries are included in this analysis. The characteris-
nature and deidentified data capture of this study. tics of participating ICUs are displayed in Table 1. In

Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014

 Nutrition for Critically Ill Surgical Patients / Drover et al   647

Table 2.   Patient Characteristics and Clinical Outcomes of Medical and Surgical Patients

Medical (n = 3425) Surgical (n =2072) P Valuea

Age, y, mean (range) 60.1 (13-99) 58.4 (12-94) .22

Sex, n (%) .001
Female 1403 (41.0) 1771 (36.1)
Male 2022 (59.0) 2645 (63.9)
Admission diagnosis, n (%) <.0001b
Cardiovascular/vascular 498 (14.5) 417 (20.1)
Respiratory 1331 (38.9) 130 (6.3)
Pancreatitis 61 (1.8) 32 (1.5)
Gastrointestinal 155 (4.5) 636 (30.7)
Neurologic 392 (11.5) 285 (13.8)
Sepsis 484 (14.1) 0
Trauma 172 (5.0) 389 (18.8)
Metabolic 123 (3.6) 0
Hematologic 36 (1.1) 0
Renal 58 (1.7) 22 (1.1)
Gynecologic 0 8 (0.4)
Orthopedic 0 24 (1.2)
Bariatric surgery 0 6 (0.3)
Burns 50 (1.5) 0
Other 65 (1.9) 123 (5.9)
APACHE II score, mean (range) 23.1 (1-54) 21.0 (1-72) <.0001
Presence of ARDS, n (%) <.0001
Yes 509 (14.9) 182 (8.8)
No 2916 (85.1) 1890 (91.2)
Length of mechanical ventilation, d, median [IQR] 9.2 [4.4-20.5] 7.4 [3.4-16.3] <.0001
Length of ICU stay, d, median [IQR]c 12.4 [7.1-24.7] 11.2 [6.7-21.2] .0004
Length of hospital stay, d, median [IQR]c 27.7 [14.7-60.0d] 28.2 [16.5-56.1] .7859
Patient died (within 60 d), n (%)c <.0001
Yes 1134 (33.1) 441 (21.3)
No 2291 (66.9) 1631 (78.7)
APACHE, Acute Physiology and Chronic Health Evaluation; ARDS, acute respiratory distress syndrome; ICU, intensive care unit;
IQR, interquartile range.
a
P values account for ICU-level clustering, by using random ICU and ICU by year effects for age and APACHE II score, Rao-Scott
χ2 method clustering by ICU for categorical outcomes, and the score test with robust standard errors for the length of stay variables.
b
P value calculation was based on cardiovascular/vascular, respiratory, gastrointestinal, neurologic, trauma, renal, and others (all
remaining groups).
c
Based on 60-day survivors only. Time before ICU admission is not counted.
d
Actual discharge rate by day 60 was 74%, so third quartile was not quite reached.

total, 5497 eligible patients (37.7% surgical) are included,
but 50 of these are excluded from the assessment of
nutrition adequacy because of missing prescriptions. The
patient characteristics and clinical outcomes of surgical
and medical patients are compared in Table 2.
Surgical patients were significantly less likely to
receive EN (54.6% vs 77.8%) and more likely to receive
PN (13.9% vs 4.4%, P < .0001) (see Table 3).
Furthermore, among patients who began EN in the
ICU, surgical patients started feeding an average of
21.0 hours later (57.8 vs 36.8 hrs, P < .0001).
Consequently, surgical patients received a substantially
lower proportion of their initial prescription from
EN (33.4 vs 49.6%, P < .0001) or from EN + propofol
+ appropriate PN (45.8 vs 56.1%, P < .0001). In
patients who experienced feeding interruptions sec-
ondary to high gastric residual volumes, surgical
patients were more likely to receive motility agents and
small bowel feeding tubes (66.9% vs 59.4%, P = .04,
and 18.8% vs 11.8%, P = .0007, respectively).
The patient characteristics and clinical outcomes of
cardiovascular, gastrointestinal, and other surgical
patients are compared in Table 4. Within surgical patients,
those undergoing cardiovascular and gastrointestinal sur-
gery were more likely to use PN, were less likely to use
EN, started EN later, and had lower total nutrition ade-
quacy and EN adequacy rates compared with other surgi-
cal patients (see Table 5).

Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014

648   Journal of Parenteral and Enteral Nutrition / Vol. 34, No. 6, November 2010

Table 3.   Nutrition Outcomes of Medical and Surgical Patients

Medical (n = 3425) Surgical (n = 2072) P Valuea

Adequacy of calories from total appropriate nutrition 56.1 ± 29.7 45.8 ± 31.9 <.0001
support,b %, mean ± SD
Adequacy of protein from total appropriate nutrition 53.2 ± 36.4 43.1 ± 34.2 <.0001
support,b %, mean ± SD
Type of nutrition support, n (%) <.0001
EN only 2662 (77.8) 1131 (54.6)
PN only 151 (4.4) 287 (13.9)
EN + PN 476 (13.9) 493 (23.8)
None 133 (3.9) 161 (7.8)
Adequacy of calories from EN, %, mean ± SD 49.6 ± 30.2 33.4 ± 29.5 <.0001
Adequacy of protein from EN, %, mean ± SD 50.4 ± 36.9 33.6 ± 32.0 <.0001
Mean hours from ICU admission to initiation of EN,c %, 36.8 ± 38.7 57.8 ± 52.1 <.0001
mean ± SD
EN, enteral nutrition; ICU, intensive care unit; PN, parenteral nutrition; SD, standard deviation.
a
P values account for ICU-level clustering, by using random ICU and ICU by year effects for continuous outcomes and Rao-Scott
χ2 method clustering by ICU for categorical outcomes.
b
Includes EN, propofol, and PN when there is a contraindication for EN.
c
Includes 2893 medical and 1542 surgical patients who started EN feeding in the ICU. These numbers are slightly lower than the
sum of the EN only and EN + PN rows above, because some of these patients began enteral feeding prior to ICU admission.

Table 4.   Patient Characteristics and Clinical Outcomes of Surgical Subgroups

Cardiovascular Gastrointestinal Other Surgical

(n = 417) (n = 668) (n = 987) P Valuea

Age, y, mean (range) 66.8 (19-94) 64.1 (18-94) 51.0 (12-92) <.0001
Sex, n (%) .007
Female 144 (34.5) 273 (40.9) 331 (33.5)
Male 273 (65.5) 395 (59.1) 656 (66.5)
APACHE II score, mean (range) 22.3 (5-48) 21.6 (1-72) 19.9 (1-57) <.0001
Presence of ARDS, n (%) <.0001
Yes 19 (4.6) 88 (13.2) 75 (7.6)
No 398 (95.4) 580 (86.8) 912 (92.4)
Length of mechanical ventilation,b d, median [IQR] 6.7 [3.1-13.8] 6.6 [2.7-16.6] 8.2 [3.9-17.2] .1091
Length of ICU stay,b d, median [IQR] 10.9 [6.7-19.9] 9.8 [5.9-22.1] 12.4 [7.1-21.0] .9345
Length of hospital stay,b d, median [IQR] 23.9 [14.6-47.1] 33.3 [18.4- 27.8 [16.6-54.5] .0009
undefinedc]
Patient died (within 60 days), n (%) .4
Yes 91 (21.8) 153 (22.9) 197 (20.0)
No 326 (78.2) 515 (77.1) 790 (80.0)
APACHE, Acute Physiology and Chronic Health Evaluation; ARDS, acute respiratory distress syndrome; ICU, intensive care unit;
IQR, interquartile range.
a
P values account for ICU-level clustering, by using random ICU and ICU by year effects for age and APACHE II score, Rao-Scott
χ2 method clustering by ICU for categorical outcomes, and the score test with robust standard errors for length of stay variables.
b
Based on 60-day survivors only. Time before ICU admission is not counted.
c
Undefined because 29 of patients remained in hospital at day 60.

The effect of site- and patient-level variables on ade-
quacy of appropriate nutrition therapy (EN + propofol +
appropriate PN received/baseline prescription) is presented
in Table 6. Appropriate nutrition adequacy over the first 12
ICU days was 10.5% less in surgical patients compared with
medical patients (−11.7% for elective surgery and −9.8% for
urgent surgery). After adjustment for significant patient- and
site-level characteristics, the expected total appropriate ade-
quacy remained 6.7% (95% confidence interval [CI],
4.8%−8.5%, P < .0001) less for surgical compared with medical
patients (8.3% less for elective surgery and 5.7% less for urgent
surgery). Admission category (surgical/medical) interacted sig-
nificantly with admission diagnosis (P < .0001) and APACHE
II score (P = .007). That is, the difference between medical

Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014

 Nutrition for Critically Ill Surgical Patients / Drover et al   649

Table 5.   Nutrition Outcomes of Surgical Subgroups

Cardiovascular Gastrointestinal Other Surgical

(n = 417) (n = 668) (n = 987) P valuea

Adequacy of calories from total appropriate 40.5 ± 29.5 42.2 ± 35.6 50.5 ± 29.5 <.0001
nutrition support,b %, mean ± SD
Adequacy of protein from total appropriate 37.2 ± 32.6 40.9 ± 37.4 47 ± 32 <.00001
nutrition support,b %, mean ± SD
Type of nutrition support, n (%) <.0001
EN only 248 (59.5) 147 (22.0) 736 (74.6)
PN only 19 (4.6) 225 (33.7) 43 (4.4)
EN + PN 93 (22.3) 249 (37.3) 151 (15.3)
None 57 (13.7) 47 (7.0) 57 (5.8)
Adequacy of calories from EN, %, mean ± SD 32.2 ± 28.2 18.2 ± 24.1 44.2 ± 28.6 <.0001
Adequacy of protein from EN, %, mean ± SD 32.7 ± 31.7 17.9 ± 24.8 44.5 ± 31.9 <.0001
Hours from ICU admission to initiation of ENc, 66.2 ± 51.2 80.0 ± 61.9 44.6 ± 42.8
mean ± SD
EN, enteral nutrition; ICU, intensive care unit; PN, parenteral nutrition; SD, standard deviation.
a
P values account for ICU-level clustering, by using random ICU and ICU by year effects for continuous outcomes and Rao-Scott
χ2 method clustering by ICU for categorical outcomes.
b
Includes EN, propofol and PN when there is a contraindication for EN.
c
Includes 320 cardiovascular, 380 gastrointestinal, and 842 other surgical patients who started EN feeding in the ICU. These num-
bers are slightly lower than the sum of the EN only and EN + PN rows above, because some of these patients began enteral feeding
prior to ICU admission.

and surgical patients decreased as APACHE II score
increased, and the medical patients had significantly greater
nutrition adequacy than surgical patients for cardiovascular/
vascular, respiratory, trauma, and renal diagnoses, but had
worse adequacy for pancreatitis (not statistically significant).
Patients with admission diagnoses of gastrointestinal or neu-
rologic disease had similar nutrition adequacy regardless of
whether their admission was medical or surgical (data not
shown). Other than admission diagnosis, the only other sig-
nificant variables associated with clinically important differ-
ences in nutrition adequacy had to do with the use of
feeding protocols (4.8% improved adequacy) and glycemic
control protocols (4.2% improved adequacy).
Discussion
Our study compares the nutrition adequacy of surgical
patients in ICUs around the world. In the context of the
international survey we were able to compare the perfor-
mance in surgical patients to patients with a medical admis-
sion category as well as compare different types of surgical
patients. Overall, surgical patients received less nutrition
during the course of the early phase of their critical illness,
with a delayed initiation of nutrition and less use of EN.
Surgeons and intensivists may choose to delay initia-
tion of nutrition for several different reasons. Clinicians
may anticipate that the patient may need to return to the
operating room or that the patient will soon be extubated
and able to have volitional intake. If there has been a
gastrointestinal anastomosis distally, there may be con-
cern that EN may compromise its integrity.
Mitigating factors may explain the differences seen in
the surgical population. To explore this, we performed sub-
group analysis of the 2 larger subgroups, which were those
with cardiovascular and gastrointestinal admission diagno-
sis. The other groups were collapsed together as a com-
parison group. We observed that patients within the
cardiovascular and gastrointestinal surgery groups received
less adequate nutrition overall and less EN. The cardiovas-
cular group received the least nutrition and the least EN of
the 3 groups. This is unlikely to be explained by concern
about a dysfunctional gastrointestinal tract. It could be
postulated that hemodynamic instability may play a role in
the delay to initiate EN in this patient population.
Hemodynamic instability continues to be a concern for
clinicians, leading to reluctance to prescribe EN. This may
be the predominant reason in the subgroup of patients
undergoing cardiovascular surgery but may be a factor in
many other critically ill surgical patients. This concern
persists despite data showing that early EN is associated
with improved outcomes in critically ill patients with
hemodynamic compromise.26 A strategy to lessen the con-
cern of delivering full-dose EN into the compromised gut
is to deliver a reduced dose of enteral feeding, so-called
trickle or trophic feeds for the first day, and reassess the
following day.27 This can provide luminal nutrition to the
gut without the anticipation that caloric goals will be met
in the first day or two after admission. It should be antici-
pated that hemodynamic stability will be achieved within

Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014

650   Journal of Parenteral and Enteral Nutrition / Vol. 34, No. 6, November 2010

Table 6.   Effect of Patient- and ICU-Level Variables Total Appropriate Adequacy
Multiple Predictor Adjusted
Single Predictor Modelsb Model
n/Na df β ± SE P Value β ± SE P Value

Patient-level variables
Number of ICU daysb 9 <.0001 <.0001
12 2673/5447 Referent Referent
11 178/5447 −5.7 ± 2.0 −5.1 ± 2.0
10 273/5447 −7.8 ± 1.7 −7.4 ± 1.6
9 290/5447 −7.9 ± .1.6 −8.0 ± 1.6
8 316/5447 −13.3 ± 1.6 −13.2 ± 1.5
7 339/5447 −16.3 ± 1.5 −16.3 ± 1.5
6 363/5447 −22.3 ± 1.5 −22.3 ± 1.4
5 396/5447 −30.9 ± 1.4 −31.0 ± 1.4
4 383/5447 −30.9 ± 1.4 −31.4 ± 1.4
3 236/5447 −36.7 ± 1.9 −36.6 ± 1.8
Age (per decade) 5447 1 −0.39 ± 0.2 .064
Female (vs male) 2124/5447 1 3.4 ± 0.7 <.0001 2.8 ± 0.7 .0001
Admission category 2 <.0001 <.0001
Medical 3396/5447 Referent Referent
Surgical: elective 704/5447 −11.7 ± 1.1 −8.3 ± 1.3
Surgical: emergency 1347/5447 −9.8 ± 0.9 −5.7 ± 1.1
Admission diagnosisc 5447 14 Not shownd <.0001 Not shownd <.0001
Days in hospital prior to ICU admission 2 .319 .025
<1 3353/5427 Referent Referent
1 to ≤7 1210/5427 −0.6 ± 0.9 0.4 ± 0.9
>7 862/5427 1.1 ± 1.0 2.9 ± 1.0
Ventilated prior to ICU admission 2185/5417 1 −1.0 ± 0.8 .226
ARDS in first 72 h 691/5447 1 0.2 ± 1.1 .844
APACHE II score (per point) 5395 1 −0.15 ± 0.05 .002 −0.2 ± 0.05 .0006
BMI 6 .0004 .036
<20 521/5447 3.4 ± 1.3 2.0 ± 1.3
20 to <25 1947/5447 Referent Referent
25 to <30 1607/5447 −2.4 ± 0.9 −2.0 ± 0.9
30 to <35 741/5447 −1.3 ± 1.1 −0.9 ± 1.1
35 to <40 317/5447 −1.8 ± 1.6 −1.5 ± 1.6
>40 314/5447 0.5 ± 1.6 0.4 ± 1.6
Site-level variables
Year of survey (2008 vs 2007)b 2722/5447 1 1.4 ± 1.1 .193
Country 5447 27 Not shownd .0002 Not shownd .0004
Teaching (vs nonteaching) 4300/5447 1 −1.2 ± 2.2 .602
Hospital size (per 1000 beds) 5447 1 −1.8 ± 1.9 .348
ICU beds (per 10 beds) 5447 1 −0.5 ± 0.9 .6
Closed ICU (vs open) 4239/5447 1 −3.5 ± 2.0 .072
ICU medical director 5165/5447 1 −2.6 ± 3.5 .458
ICU has feeding protocol 4371/5447 1 6.7 ± 1.8 .0002 4.8 ± 1.9 .011
Dietician FTE per 10 ICU beds 5447 1 −0.1 ± 1.9 .959
Use protocol to monitor BS control
or insulin administration 4670/5447 1 3.8 ± 2.3 .091 4.2 ± 2.3 .066

APACHE, Acute Physiology and Chronic Health Evaluation; ARDS, acute respiratory distress syndrome; BMI, body mass index; BS, blood
sugar, df, degrees of freedom; FTE, full-time equivalent; ICU, intensive care unit; SE, standard error.
Appropriate nutrition adequacy is the percentage of the prescription received by EN, propofol, and PN when there is a contraindication to
EN. The β coefficient (±SE) provides the mean difference in appropriate nutrition adequacy of calories between the given category and the
reference group.
a
n provides number in each row; N is the total number of nonmissing values. Denominators of some variables are slightly <5447 because of
sporadic missing values. The multivariable modeling is based on 5376 patients because of 52 patients with missing APACHE II scores.
b
Single predictor models all adjust for number of days in ICU (from 1 to 12) and year of survey
c
See Table 2 for a complete list of admission diagnoses.
d
Because admission diagnosis or country have many categories with no inherent referent category, we have not provided the specific β esti-
mates that contrast each category to a referent. Only the global P value that is unaffected by the arbitrary choice of referent is provided. It
may be seen that both admission category and country are significantly associated with appropriate nutrition adequacy.

Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014

 Nutrition for Critically Ill Surgical Patients / Drover et al   651
hours of admission and certainly within 2 days, and at that
point, feeds can be advanced to achieve goal calories.
There are data to support the hypothesis that EN in
scheduled surgery is safe and associated with improved
patient outcomes. A meta-analysis has compared postop-
erative EN with the common practice of providing no
nutrition by mouth. This included 11 randomized control-
led trials with 837 patients undergoing elective gastrointes-
tinal surgery.3 These patients were fed within 24 hours of
their surgery, and the analysis included both tube and oral
feeding protocols. The results demonstrated a reduction in
the risk of infection (Relative Risk [RR] 0.79; 95% CI,
0.54–0.98) and a reduction in length of stay by 0.84 days
(P < .001). Results also showed an increased risk of vomit-
ing (RR 1.27; 95% CI, 1.01–1.61). In another systematic
review, a statistically significant mortality benefit (RR 0.41;
95% CI, 0.18–0.93) was associated with EN for patients
undergoing scheduled colorectal surgery.28
Surgeons have generally adopted the idea that routine
nasogastric decompression is not necessary after colorec-
tal resection29 but have been slower to adopt the idea that
a period of fasting is unnecessary. A major barrier to initia-
tion of EN in the early postoperative period for large bowel
surgery is the concern of increasing the risk of anasto-
motic dehiscence. In the above-noted meta-analysis, 6 of
the studies involved colon surgery.5,30-34 Seven studies of
immune-modulating formulas have evaluated patients
undergoing colon surgery.35-41 The most recent study of
patients undergoing colorectal resection was by Feo et al.42
Consistent findings are that early (within 24 hours)
enteral feeding after colorectal surgery is safe and that
there is an association between early feeding and the pres-
ence of vomiting and the need for a nasogastric tube.
However, early feeding has been associated with decreased
septic complications and length of hospital stay.32
Investigators have also examined the need for fasting after
major gynecologic surgery.43,44 In this group of patients,
early oral feeding was well tolerated and was associated
with a decreased length of hospital stay.
We explored the data set looking for factors that may
be associated with improved nutrition adequacy in surgical
patients. The 2 factors that were associated with improved
delivery of EN were the use of protocols for feeding and
blood sugar control. It would be expected that units that
have a protocol for managing EN would also have a proto-
col for managing blood sugar, which could be a systematic
way of overcoming barriers to delivering nutrition in these
units. This observation supports the hypothesis that sys-
tematic barriers need to be addressed to narrow the gap
between best practice and current practice. We recom-
mend that feeding protocols be used in surgical patients as
a strategy to improve EN delivery. We also discourage the
strategy of feeding nothing by mouth.
Two strategies that have been promoted to optimize
the benefits and minimize the risks of EN are the use
of motility agents and small bowel feeding tubes.45 In this
Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014
data set, these strategies were used more often in surgical
patients than medical patients (Table 3) but they still were
underused in patients who demonstrated intolerance to
EN (eg, only 18% of such surgical patients received small
bowel feeding tubes). The use of these strategies was simi-
lar across the surgical subgroups that we analyzed (Table
5). There is an opportunity to improve EN delivery by
including motility agents in feeding protocols. The routine
use of small feeding tubes may be more difficult in some
surgical patients because technical considerations of place-
ment. Critically ill patients undergoing gastrointestinal
surgery should be routinely considered for placement of a
small bowel feeding tube at the time of surgery.
The strengths of this study are the prospectively iden-
tified large number of patients who were observed over 2
different data collection periods in ICUs in many coun-
tries. Data collection was thorough with few missing
variables, allowing for complete analysis. Limitations of
the study are the retrospective collection of the data from
chart review. Despite our statistical adjustments, biases
between groups cannot be accounted for completely. The
clinical outcome data are limited, which limits our ability
to completely explain how the surgical patients may have
been different from the medical patients.
Conclusion
Surgical patients in critical care units receive less
nutrition than comparable medical patients. Within
surgical patients, those undergoing cardiovascular and
gastrointestinal surgery are at highest risk of iatrogenic
malnutrition. Strategies to improve nutrition perfor-
mance, including the use of protocols, motility agents,
and small bowel feeding tubes, are needed.
Acknowledgments
We are grateful to the critical care practitioners from all
participating ICU sites for their dedication and commit-
ment to collecting data for this study.
References
1. Heyland DK, MacDonald S, Keefe L, Drover JW. Total parenteral
nutrition in the critically ill patient: a meta-analysis. JAMA.
1998;280:2013-2019.
2. Braunschweig CL, Levy P, Sheean PM, Wang X. Enteral compared
with parenteral nutrition: A meta-analysis. Am J Clin Nutr.
2001;74:534-542.
3. Lewis SJ, Egger M, Sylvester PA, Thomas S. Early enteral feeding ver-
sus “nil by mouth” after gastrointestinal surgery: systematic review
and meta-analysis of controlled trials. BMJ. 2001;323:773-776.
4. Heyland DK, Novak F, Drover JW, Jain M, Su X, Suchner U.
Should immunonutrition become routine in critically ill patients?
A systematic review of the evidence. JAMA. 2001;286:944-953.
652   Journal of Parenteral and Enteral Nutrition / Vol. 34, No. 6, November 2010
5. Schroeder D, Gillanders L, Mahr K, Hill GL. Effects of immediate
postoperative enteral nutrition on body composition, muscle function,
and wound healing. JPEN J Parenter Enteral Nutr. 1991;15:376-383.
6. Haydock DA, Hill GL, Haydock DA, Hill GL. Impaired wound
healing in surgical patients with varying degrees of malnutrition.
JPEN J Parenter Enteral Nutr. 1986;10:550-554.
7. Heyland DK, Heyland DK. Nutritional support in the critically ill
patients: a critical review of the evidence. Crit Care Clin 1998;
14:423-440.
8. Sigalet DL, Mackenzie SL, Hameed SM, Sigalet DL, Mackenzie
SL, Hameed SM. Enteral nutrition and mucosal immunity: impli-
cations for feeding strategies in surgery and trauma. Can J Surg.
2004;47:109-116.
9. Moss G, Greenstein A, Levy S, Bierenbaum A. Maintenance of GI
function after bowel surgery and immediate enteral full nutrition,
I: doubling of canine colorectal anastomotic bursting pressure and
intestinal wound mature collagen content. JPEN J Parenter Enteral
Nutr. 1980;4:535-538.
10. Moore FA, Feliciano DV, Andrassy RJ, et al. Early enteral feeding,
compared with parenteral, reduces postoperative septic complica-
tions: the results of a meta-analysis. Ann Surg. 1992;216:172-183.
11. Lipman TO. Bacterial translocation and enteral nutrition in
humans: an outsider looks in. JPEN J Parenter Enteral Nutr.
1995;19:156-165.
12. Jeejeebhoy KN. Enteral feeding. Curr Opin Clin Nutr Metab Care.
2002;5:695-698.
13. Scolapio JS. A review of the trends in the use of enteral and parenteral
nutrition support. J Clin Gastroenterol. 2004;38:403-407.
14. Correia MI, Waitzberg DL. The impact of malnutrition on morbidity,
mortality, length of hospital stay and costs evaluated through a multi-
variate model analysis. Clin Nutr. 2003;22:235-239.
15. Bozzetti F, Braga M, Gianotti L, Gavazzi C, Mariani L. Postoperative
enteral versus parenteral nutrition in malnourished patients with
gastrointestinal cancer: a randomised multicentre trial. Lancet.
2001;358:1487-1492.
16. Huckleberry Y. Nutritional support and the surgical patient. Am J
Health Syst Pharm. 2004;61:671-682.
17. Howard L, Ashley C. Nutrition in the perioperative patient. Annu
Rev Nutr. 2003;23:263-282.
18. Heyland DK, Schroter-Noppe D, Drover JW, et al. Nutrition sup-
port in the critical care setting: current practice in Canadian
ICUs—opportunities for improvement? JPEN J Parenter Enteral
Nutr. 2003;27:74-83.
19. Heyland DK, Dhaliwal R, Day A, Jain M, Drover J. Validation of the
Canadian clinical practice guidelines for nutrition support in mechan-
ically ventilated, critically ill adult patients: results of a prospective
observational study. Crit Care Med. 2004;32:2260-2266.
20. Jones NE, Dhaliwal R, Day A, Jiang X, Heyland DK. Nutrition
therapy in the critical care setting: what is “best achievable” prac-
tice? An international multicenter observational study. Crit Care
Med. 2010;38:395-401.
21. Critical Care Nutrition: What we offer. http://www.criticalcarenu-
trition.com. Accessed November 4, 2007.
22. Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II:
a severity of disease classification system. Crit Care Med.
1985;13:818-829.
23. Artigas A. Current definitions of acute lung injury and the acute
respiratory distress syndrome. Intensive Care Med. 2000;26:1019.
24. Rao JNK, Scott AJ. The analysis of categorical data from complex
surveys: chi-squared tests for goodness of fit and independence in
two-way tables. J Am Stat Assoc. 1981;76:221-230.
25. Harrell FE. Regression Modeling Strategies: With Application to
Linear Models, Logistic Regression and Survival Analysis. New York,
NY: Harrell; 2001.
26. Kazamias P, Kotzampassi K, Koufogiannis D, Eleftheriadis E.
Influence of enteral nutrition-induced splanchnic hyperemia on the
septic origin of splanchnic ischemia. World J Surg. 1998;22:6-11.
Downloaded from pen.sagepub.com at UNIV CALGARY LIBRARY on October 6, 2014
27. Marik PE, Karnack C, Varon J. The addition of trickle feeds
reduces the complications associated with parenteral nutrition.
Critical Care and Shock. 2002;5.
28. Andersen HK, Lewis SJ, Thomas S. Early enteral nutrition within
24h of colorectal surgery versus later commencement of feeding
for postoperative complications. Cochrane Database Syst Rev.
2009;(3).
29. Cheatham ML, Chapman WC, Key SP, Sawyers JL. A meta-analy-
sis of selective versus routine nasogastric decompression after
elective laparotomy. Ann Surg. 1995;221:469-476.
30. Binderow SR, Cohen SM, Wexner SD, Nogueras JJ. Must early
postoperative oral intake be limited to laparoscopy? Dis Colon
Rectum. 1994;37:584-589.
31. Reissman P, Teoh TA, Cohen SM, Weiss EG, Nogueras JJ,
Wexner SD. Is early oral feeding safe after elective colorectal
surgery? A prospective randomized trial. Ann Surg. 1995;222:
73-77.
32. Ortiz H, Armendariz P, Yarnoz C, Ortiz H, Armendariz P, Yarnoz C.
Is early postoperative feeding feasible in elective colon and rectal
surgery? Int J Colorectal Dis. 1996;11:119-121.
33. Hartsell PA, Frazee RC, Harrison JB, Smith RW. Early postopera-
tive feeding after elective colorectal surgery. Arch Surg.
1997;132:518-520.
34. Stewart BT, Woods RJ, Collopy BT, Fink RJ, Mackay JR, Keck JO.
Early feeding after elective open colorectal resections: a prospec-
tive randomized trial. Aust N Z J Surg. 1998;68:125-128.
35. Schilling J, Vranjes N, Fierz W, et al. Clinical outcome and immu-
nology of postoperative arginine, omega-3 fatty acids, and nucle-
otide-enriched enteral feeding: a randomized prospective
comparison with standard enteral and low calorie/low fat i.v. solu-
tions. Nutrition. 1996;12:423-429.
36. Braga M, Gianotti L, Radaelli G, et al. Perioperative immuno-
nutrition in patients undergoing cancer surgery: results of a
randomized double-blind phase 3 trial. Arch Surg. 1999;134:
428-433.
37. Braga M, Gianotti L, Nespoli L, Radaelli G, Di Carlo V. Nutritional
approach in malnourished surgical patients: a prospective rand-
omized study. Arch Surg. 2002;137:174-180.
38. Braga M, Gianotti L, Vignali A, Carlo VD. Preoperative oral
arginine and n-3 fatty acid supplementation improves the immu-
nometabolic host response and outcome after colorectal resection
for cancer. Surgery. 2002;132:805-814.
39. Gianotti L, Braga M, Nespoli L, Radaelli G, Beneduce A, Di Carlo
V. A randomized controlled trial of preoperative oral supplementa-
tion with a specialized diet in patients with gastrointestinal cancer.
Gastroenterology. 2002;122:1763-1770.
40. Xu J, Zhong Y, Jing D, Wu Z. Preoperative enteral immunonutrition
improves postoperative outcome in patients with gastrointestinal
cancer. World J Surg. 2006;30:1284-1289.
41. Finco C, Magnanini P, Sarzo G, et al. Prospective randomized
study on perioperative enteral immunonutrition in laparoscopic
colorectal surgery. Surg Endosc. 2007;21:1175-1179.
42. Feo CV, Romanini B, Sortini D, et al. Early oral feeding after color-
ectal resection: a randomized controlled study. ANZ J Surg.
2004;74:298-301.
43. Steed HL, Capstick V, Flood C, Schepansky A, Schulz J, Mayes
DC. A randomized controlled trial of early versus “traditional”
postoperative oral intake after major abdominal gynecologic sur-
gery. Am J Obst Gynecol. 2002;186:861-865.
44. Schilder JM, Hurteau JA, Look KY, et al. A prospective controlled
trial of early postoperative oral intake following major abdominal
gynecologic surgery. Gynecol Oncol. 1997;67:235-240.
45. Heyland DK, Dhaliwal R, Drover JW, Gramlich L, Dodek P;
Canadian Critical Care Clinical Practice Guidelines Committee.
Clinical practice guidelines for nutrition support in the adult
critically ill patient. JPEN J Parenter Enteral Nutr. 2003;27:
355-373.