721639

research-article2017
VMJ0010.1177/1358863X17721639Vascular MedicineAitken et al.

Original Article

Vascular Medicine

Incidence, prognostic factors and impact of 2017, Vol. 22(5) 387­–397

© The Author(s) 2017
Reprints and permissions:
postoperative delirium after major vascular sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1358863X17721639
https://doi.org/10.1177/1358863X17721639
surgery: A meta-analysis and systematic review journals.sagepub.com/home/vmj

Sarah Joy Aitken1,2, Fiona M Blyth1,3,4 and Vasi Naganathan1,3,4

Abstract
Although postoperative delirium is a common complication and increases patient care needs, little is known about the
predictors and outcomes of delirium in patients having vascular surgery. This review aimed to determine the incidence,
prognostic factors and impact of postoperative delirium in vascular surgical patients. MEDLINE and EMBASE were
systematically searched for articles published between January 2000 and January 2016 on delirium after vascular surgery.
The primary outcome was the incidence of delirium. Secondary outcomes were contributing prognostic factors and
impact of delirium. Study quality and risk of bias was assessed using the QUIPS tool for systematic reviews of prognostic
studies, and MOOSE guidelines for reviews of observational studies. Quantitative analyses of extracted data were
conducted using meta-analysis where possible to determine incidence of delirium and prognostic factors. A qualitative
review of outcomes was performed. Fifteen articles were eligible for inclusion. Delirium incidence ranged between 5%
and 39%. Meta-analysis found that patients with delirium were older than those without delirium (OR 3.6, p<0.001).
Prognostic factors predicting delirium included increased age (OR 1.04, p<0.001), pre-existing cognitive impairment (OR
9.8, p=0.01), hypertension, pre-existing depression and open aortic surgery. Delirious patients remained in hospital 6
days longer (p<0.001) and had more complications than patients without delirium. Data were limited on the impact
of procedure complexity, endovascular compared to open surgery or type of anaesthetic. Postoperative delirium
occurs frequently, resulting in major morbidity for vascular patients. Improved quality of prognostic studies may identify
modifiable peri-operative factors to improve quality of care for vascular surgical patients.

Keywords
aging, delirium, epidemiology, postoperative complications, prognosis, risk factors, vascular surgical procedures

Introduction
Postoperative delirium is a frequent and often unrecognized
complication of surgery. Delirium is defined as an acute fluc-
tuation in cognitive status, with features of inattention and
altered levels of consciousness. Delirium is known to con-
tribute to increased peri-operative complications and length
of stay, functional decline and mortality.1,2 Delirium is both
preventable and treatable with environmental, supportive
and pharmacological interventions.2 For vascular surgery,
where patients are often older, frail and generally more at
risk of adverse outcomes, delirium represents a major poten-
tially modifiable risk factor. The incidence and impact of
postoperative delirium after major vascular surgery is
unknown. Although there are many clinically applicable
validated diagnostic tools for diagnosis of delirium, over
in multivariable prognostic models that predict disease end-
points, or can be used to stratify patients and identify those
most likely to benefit from a treatment or intervention. A
prognostic factor is one that is present in people with a par-
ticular disease at a given start point, and that influences the
clinical course and outcomes of a condition.6 In contrast,
risk factors are defined as those factors associated with the
1University of Sydney, Concord Clinical School, Sydney, NSW, Australia
2Concord Institute of Academic Surgery, Department of Vascular
Surgery, Concord Repatriation General Hospital, Sydney, NSW,
Australia
3Centre for Education and Research in Ageing, Concord, Sydney, NSW,
Australia
4Ageing and Alzheimers Institute, Concord, Sydney, NSW, Australia

50% of inpatient delirium is unrecognized.3 Research studies

have shown that there are multiple contributing factors for
the development of delirium, often with conflicting results.4,5
Prognostic factor research aims to identify and evaluate
those factors that may become targets of modifiable out-
come improvements, can be included as predictor variables
Corresponding author:
Sarah Joy Aitken, Department of Vascular Surgery, Concord Institute
of Academic Surgery, Concord Clinical School, University of Sydney,
Concord Repatriation General Hospital, Hospital Rd, Concord West,
Sydney, NSW 2139, Australia.
Email: sarah.aitken@sydney.edu.au
388
cause of a condition. Consideration of prognostic factors
for delirium in patients having major vascular surgery is
important so that specific interventions can be developed
for this cohort to improve outcomes. Understanding the
strength and consistency of the relationship of specific
prognostic factors for delirium can assist in understanding
the trajectories of illness and identify potentially modifia-
ble factors that may influence outcomes.
This study aims to systematically review the medical lit-
erature to establish the incidence of peri-operative delirium
after major vascular surgery, determine which vascular sur-
gical procedures have the highest risk of delirium, and
identify prognostic factors associated with developing
delirium. It will also assess the impact that delirium has
upon surgical outcomes such as length of stay and compli-
cation rates. This study will adhere to internationally recog-
nized standards for systematic reviews of prognostic
studies, incorporating the novel use of recently validated
risk of bias assessment methods for prognostic studies.
Methods
Search strategy and selection criteria
A systematic search for all articles that measured delirium
after major vascular surgery was conducted. MEDLINE
and EMBASE were systematically searched for articles
published between 1 January 2000 and 30 January 2016.
MESH terms included delirium, cognitive disorders and
vascular surgical procedures with a text word search
designed to include all major vascular surgical procedures
(see Online Supplemental Table 1). A hand search of con-
tents lists of major vascular surgery journals was also per-
formed. This review was registered with the Prospero
Register of Systematic Reviews in April 2015.7
Studies were eligible for inclusion if they reported origi-
nal data on delirium after major vascular surgery. Clinical
trials along with case–control, cohort studies and case-
series (with more than 10 participants) were included. Case
reports, review articles, editorials and conference proceed-
ings were excluded. Patients who had aortic, carotid,
peripheral or endovascular surgery formed the population
of interest.
The search strategy was developed in conjunction with a
medical librarian. Searches were limited to publications in
English. Search results were recorded according to the
PRISMA guidelines. After removing duplicates, titles and
abstracts were screened using the inclusion/exclusion crite-
ria. Remaining articles were assessed for eligibility by full
text review (SJA, FB, VN) and the decision on inclusion
decided by consensus. The references of included articles
were searched to identify any additional relevant studies.
The primary outcome was the incidence of postopera-
tive delirium, diagnosed by a validated diagnostic tool,
such as the Diagnostic and Statistical Manual of Mental
Disorders criteria (DSM-IV) or Delirium Observational
Scale (DOS). The differential risk associated with surgery
type (aortic vs other; open vs endovascular aortic) and
anaesthesia type (general vs loco-regional) was also
assessed. Secondary outcomes were prognostic factors
Vascular Medicine 22(5)
predicting delirium, the association of delirium with
advancing age, and outcomes of length of stay, mortality
and morbidity.
Prognostic factors were defined as those that were pri-
marily investigated as predictive factors for delirium inci-
dence after major vascular surgery using multivariable
analysis. Prognostic factors were categorized into patient,
procedural and peri-operative management factors.
Confounding factors were defined as other factors included
in multivariable analysis that were not defined as prognos-
tic factors in the study, but were included based on prior
evidence and clinical knowledge of their relationship with
the development of delirium.
Quality assessment (risk of bias)
Studies were assessed for risk of bias using the Quality in
Prognostic Studies (QUIPS) tool,8 a validated tool for
assessing risk of bias in prognostic factor studies.6,8,9
QUIPS provides a qualitative assessment of six key areas
– study participation, study attrition, prognostic factor
measurement, outcome measurement, study confounding
and statistical analysis and presentation.8 All three authors
(SJA, FB, VN) independently assessed each study, ranking
the risk of bias as high, moderate or low. If the authors disa-
greed on their risk of bias rating, a consensus agreement
was reached by joint review and discussion of the article.
This review has been conducted according to the quality
checklist developed by the Meta-analysis Of Observational
Studies in Epidemiology (MOOSE) guidelines.10 A full
description on how this study complies with the MOOSE
guidelines is provided in the Online Supplemental Table 2.
Statistical analysis
Study data were extracted using a pre-designed, standard-
ized proforma (primary extraction by SJA; verification of
accuracy assessed by FB and VN). This included informa-
tion on study characteristics (Table 1), patient characteris-
tics and quantification of results (odds ratio or hazard ratios
of prognostic factors identified on multivariable analysis,
mean and standard deviation of continuous variables).
Results were collated in Review Manager v5.3 (RevMan,
The Nordic Cochrane Centre, The Cochrane Collaboration,
Copenhagen). Meta-analysis of prognostic factors was con-
sidered where there were more than three studies reporting
an association between the prognostic factor and delirium.
Meta-analysis of prognostic factors was only deemed pos-
sible for those studies where the patient characteristics were
significantly similar at baseline in terms of age, sex, proce-
dural magnitude and study quality (design and confounders)
and where multivariable analysis included similar variables.
A random-effects inverse variance meta-analysis would be
calculated in RevMan using individual odds ratios trans-
formed to their logarithms to normalize their distribution,
and standard errors (SEs) calculated from the 95% confi-
dence intervals (CIs).8,11,12 Heterogeneity was anticipated to
be high (defined by I2 statistic as greater than 50%), as is
consistent with other meta-analyses of observational
studies13 due to the lack of randomization. The incidence of
Table 1. Studies on the incidence of delirium in vascular surgery patients.

Study Country Sample size Type of surgery Mean age, Sex (male: Delirium cases Diagnostic Incidence
Aitken et al.

years (range) female) tool

Visser et al.14 Netherlands 463 Major vascular 72 (66–77) 356:107 22 DOS 4.8%
(2015) surgery
Raats et al.15 Netherlands 206 General vascular 71.5 151:55 26 DOS 12.6%
(2015) surgery
Ellard et al.16 Canada 500 Major vascular 69 342:158 97 NEECHAM See Katznelson
(2014) (subset of Katzneslon surgery 2009
2009)
Sasajima Japan 299 Peripheral arterial 72 261:38 88 HDS-R 29%
et al.17 (2012) bypass surgery CAM
DRS
Salata et al.18 Canada 256 Open and 71.5 208:48 57 NEECHAM 22%
(2012) endovascular
aortic surgery
Pol et al.19 Netherlands 142 General vascular 68 (21–87) 100:42 10 DOS See Visser 201514
(2011) (subset of Visser14) surgery
Bryson et al.20 Canada 88 Open aortic 70.7 64:24 30 CAM 36%
(2011) surgery
Koebrugge Netherlands 107 Open and 70.6 82:25 25 DOS 23%
et al.21 (2010) endovascular
aortoiliac surgery
Katznelson Canada 582 Major vascular 69.5 414:168 128 NEECHAM 22%
et al.22 (2009) procedures
Rudolph Multinational 175 Major vascular 68.9 35 DSM-III 20%
et al.23 (2007) (vascular surgical subset procedures
only)
Minden et al.24 USA 35 Open aortic 67.5 (46–88) 31:4 8 CAM 23%
(2005) surgery MDAS
Benoit et al.25 Canada 102 Open aortic 70.8 (41–88) 79:23 34 DSM-IV 33%
(2005) surgery MMSE
Bohner et al.26 Germany 153 Major vascular 66 119:34 60 DRS 39%
(2003) surgery DSM-IV
Schneider Germany 47 Major vascular 66.8 (53–84) 38:9 17 DRS See Bohner 200326
et al.27 (2002) (subset of surgery DSM-IV
Bohner 200326)
Rosen et al.28 USA 188 Open aortic 66 (31–91) 133:55 53 TAMI 28%
(2002) surgery

DOS, Delirium Observation Scale; NEECHAM, Neelon and Champagne Confusion Score; HDS-R, Revised Hasegawa Dementia Scale; CAM, Confusion Assessment Method; DRS, Delirium Rating Scale; DSM-III, IV, Diagnostic
and Statistical Manual of Mental Disorders III, IV; MDAS, Memorial Delirium Assessment Scale; MMSE, Mini-Mental State Examination; TAMI, Transient Advanced Mental Impairment Score.
389
390
Figure 1. PRISMA flow diagram of search results.
postoperative delirium was presented as a percentage range
rather than an overall mean. Mean age, hospital length of
stay and duration of intensive care stay data were collated in
RevMan. Unadjusted mean difference was used for meta-
analysis of age and length of stay. A qualitative review was
performed of outcomes and prognostic factors that were not
eligible for meta-analysis.
Results
The results of the search strategy are reported in Figure 1.
Fifteen studies14–28 were eligible for inclusion and used to
establish the incidence of delirium: eight studies were
Vascular Medicine 22(5)
used to quantify hospital length of stay16,19–21,24–26,28, eight
studies for assessing the association of age with delir-
ium14,15,17,21,22,25–27 and 10 studies were used in the prog-
nostic factor analysis.14,15,17–19,21,22,26–28 Three studies
were included in the meta-analysis examining age as a
prognostic factor for delirium,15,18,22 and three studies in
the meta-analysis examining pre-existing cognitive
impairment.14,29 Three studies were earlier publications
which included the same patient cohort used in later pub-
lications.30–32 Where there were multiple publications
related to one study, the most recent publication was
selected unless different prognostic factors and out-
comes were reported in separate publications.14,16,19,22,27
Aitken et al.
Rudolph et al.23 was included only in the analysis of
delirium incidence as this study had no analysis of prog-
nostic factors. The major reasons for study exclusions
were that procedures were not major vascular surgery, or
due to study type (commentaries, editorials, or small
case series).
The design and characteristics of the included studies
are shown in Table 1. All were observational studies, with
the majority (n=9, 64%) being prospective cohort studies.
The median study size was 197 participants (range 35–582)
and the total number of participants was 2650 (excluding
duplicate studies).
Study quality and risk of bias
QUIPS ranking did not vary by more than one category
between raters for any criteria for each publication and con-
sensus was achieved for all rankings. Table 2 presents the
QUIPS analysis for risk of bias. Study quality was variable.
The most common reasons for studies having a moderate or
high risk of bias were poorly described study participation
(36%, n=5) and attrition rates (29%, n=4), inadequate
adjustment for confounders (79%, n=11) and inadequate
statistical analyses (43%, n=6).
Incidence of delirium
The incidence of postoperative delirium ranged between 5%
and 39%. The majority of studies found the incidence of
postoperative delirium to be greater than 20% (n=10). The
mean duration of delirium ranged between 1 and 30 days.
Based on the meta-analysis of eight studies,14,15,17,21,22,25–27
patients with delirium were more likely to be older than
patients without delirium (OR 3.6, 95% CI 1.6–5.5, p<0.001)
(Figure 2).
Prognostic factors for risk of developing
delirium
Ten14,15,17,18,19,21,22,26–28 of the 15 studies conducted multi-
variable analysis to investigate prognostic factors associ-
ated with the incidence of postoperative delirium. Table 3
summarizes these results , with further information included
in the Online Supplemental Table 3.
Increased age was the most frequently identified
prognostic factor for predicting delirium (seven of eight
studies).14,15,17,18,21,22,26,28 Three studies measured age as
a continuous variable15,18,22 and were suitable for meta-
analysis. On meta-analysis, increasing age was associ-
ated with the development of delirium (OR 1.04, 95% CI
1.02–1.06; I2=0%, p<0.001) (Figure 2). A further five
studies14,17,21,26,28 reported age as a dichotomous varia-
ble, with cut-points varying between >64 and >80 years
old. While these studies found an association with the
older age groups and delirium, they were unable to be
included in the meta-analysis.
Pre-existing cognitive impairment14,17,26,28 had a positive
association with delirium in three of four studies. On meta-
analysis of these three studies,14,17,26 pre-existing cognitive
impairment predicted the development of delirium (OR 9.79,
391
95% CI 1.7–57.2, p=0.01) (Figure 2). Six studies15,18,19,21,27,28
included the American Society of Anesthesiologists (ASA)
classification in multivariable analysis but were unable to be
combined in meta-analysis due to variation in ASA cut-
points. Only one of these studies showed ASA to be a prog-
nostic factor for delirium.29
Aortic surgery and delirium risk
Aortic surgery had an increased risk of delirium in a single
study15 (OR 1.77, 95% CI 1.04–3.02, p=0.004), while
patients having either aortic surgery or amputation were at
greater risk of delirium compared to other vascular surgical
procedures (OR 14.0, 95% CI 3.9–49.8, p<0.001) in another
study.14 None of the other 12 studies reported a comparison
of aortic surgery versus other non-aortic surgical proce-
dures. Open aortic surgery had a greater risk of postopera-
tive delirium compared to endovascular aortic surgery
(EVAR) in two of the three studies that compared these
procedures.18,21,22
Impact of anaesthetic type on incidence of
delirium
There were no studies that assessed anaesthetic type as a
prognostic factor in multivariable analysis. Three studies
examined the impact of anaesthetic type on rates of delir-
ium using crude ratios or univariate analysis.16,21,22 None of
these three studies found that general anaesthetic was prog-
nostic of delirium development; however, they did not
account for use of sedation, patient comorbidities or type of
anaesthesia such as general versus loco-regional anaesthe-
sia. The QUIPS risk of bias score demonstrated a moder-
ate–high risk of bias for confounding and analysis in two of
these three studies.16,22
Impact of postoperative delirium on peri-
operative outcomes
Postoperative delirium significantly increased the length of
hospital stay for vascular patients. Eight studies reported on
differences in length of stay between patients with and
without delirium (Figure 2). On meta-analysis, patients
with delirium remained over 6 days longer than those
patients without delirium (6.1 days, 95% CI 3.9–8.4,
p<0.001). Eight studies reported duration of intensive care
requirements for patients with delirium. Patients with delir-
ium did not have significantly longer intensive care unit
stays than those without delirium (1.3 days, 95% CI 0.7–
2.0, p<0.001).
Delirium increased the risk of discharge to nursing home
or institutional care facilities14,28 but was only examined in
two studies. In one study of patients with delirium after aor-
tic surgery, 38% were discharged to a nursing home after
previously living at home, compared to 11% of patients who
did not have delirium (p<0.001).28 Two studies investigated
longer-term patient outcomes. In one study, patients who
had postoperative delirium after endovascular aortic aneu-
rysm surgery had poorer physical function, social function
and energy at 1 and 6 months after surgery compared to
392 Vascular Medicine 22(5)

Table 2. Risk of bias assessment using the Quality in Prognostic Studies (QUIPS) tool8 for included studies.

Prognostic Statistical
Study Study Outcome Study
Study Design factor analysis and
participation attrition measurement confounding
measurement presentation

Visser et al14
Prospective cohort Low Low Low Low Low Low
2015
Raats et al15
Retrospective cohort High Low Low Low High Moderate
2015
Retrospective cohort
Ellard et al16 Low
(subset of Katznelson High Low Low High High
2014
2009)22
Sasajima et al17
Prospective cohort Low Low Low Low Moderate Low
2012
Salata et al18
Retrospective cohort Low Low Low Low Moderate Low
2012

Pol et al19 Prospective cohort

Low Low Low Low High Moderate
2011 (subset of Visser 2015)14

Bryson et al20
Prospective cohort High Moderate Low Low High Moderate
2011
Koebrugge et al21
Retrospective cohort Low Low Low Low Low Low
2010
Katznelson et al22
Prospective cohort Low Low Low Low Moderate Low
2009
Minden et al24
Prospective cohort Moderate Low Low Moderate High High
2005
Benoit et al25
Prospective cohort Low Moderate Low Low Moderate Low
2005
Bohner et al26
Prospective cohort Moderate Low Moderate Low Low Low
2003

Schneider et al27 Prospective cohort

Low Moderate Low Low Moderate Low
2002 (Subset of Bohner 2003)26

Rosen et al28
Retrospective cohort High Low Low Moderate Moderate Moderate
2002
White: Low risk of bias; Light shade: Moderate risk of bias; Dark shade: High risk of bias.

those without delirium.24 In the other study, residual postop- Discussion

erative cognitive impairment in patients who experienced
delirium in hospital was not significantly increased com-
pared to those without delirium at discharge or 3 months
post discharge.20
Complications of delirium
Overall rates of complications, including life-threatening
ones such as respiratory dysfunction, cardiac arrest, acute
renal failure and surgical revision, were higher in patients
with delirium than in those without.14,21,27,28 One study
found that patients with delirium were more likely to
remove venous or urinary catheters and have catheter site
infections than those without delirium.26
Mortality was not significantly increased in patients with
delirium in three of the four studies14,17,21,24 that reported
mortality data. Each of these studies measured mortality at
different time-points, from peri-operative, at discharge, or at
30 days, thus meta-analysis was not possible.
This systematic review demonstrates the high incidence of
delirium after major vascular surgery; over a third of patients
in some studies experienced postoperative delirium. With
one exception,14 the incidence of postoperative delirium
was between 13% and 39%. A number of key prognostic
factors were identified for development of delirium after
major vascular surgery. Increasing age was the most com-
monly identified prognostic factor in the studies included
and demonstrated a significant association with delirium on
meta-analysis. Cognitive impairment, pre-existing depres-
sion, hypertension and having open aortic surgery were all
prognostic factors for the development of delirium. Patients
who experienced delirium were more likely to have an
increased length of stay, increased risk of surgical complica-
tions, and were more likely to have institutional care needs
when discharged from hospital. Risk of bias analysis showed
moderate risk of bias in study populations and attrition rates,
as well as a lack of adjustment for confounders.
Aitken et al.
Figure 2. Results of meta-analysis on prognostic factors and outcomes for postoperative delirium.
Increasing age was associated with delirium in our
study. The association between increasing age and the
presence of delirium was seen on meta-analysis of mean
differences in age between patients with and without
delirium, and also on meta-analysis of age as a prognos-
tic factor for the development of delirium. Increasing
age has also been associated with delirium in other types
of surgery.1 Factors such as frailty may have a role in
the development of delirium independent of chronologi-
cal age and warrant further exploration; however, our
review did not find frailty was a prognostic factor for
393
delirium in the one study that examined this.19 Our sys-
tematic review found cognitive impairment to be a risk
factor for developing delirium, in concordance with
other studies of postoperative delirium.1,3,33 As many
studies excluded patients with underlying cognitive
impairment, a potentially vulnerable group may not
have been assessed for the risk and impact of delirium.
An elevated ASA score was not associated with delirium
but this is most likely due to differences in definitions
of significant cut-points for grading severe comorbidi-
ties with the ASA score.
 394

Table 3. Summary of prognostic factors related to post-operative delirium on multivariable analysis.

Study Patient-related prognostic Operative-related prognostic Peri-operative

factors factors management-related
prognostic factors

Increasing Elevated Pre-existing Pre-existing Diabetes Hypertension History Open Operative Blood Increased Emergency Low Elevated
age ASA cognitive depression of stroke vs EVAR duration transfusion fluid vs elective preoperative preoperative
score impairment or carotid aortic required infusion surgery haemoglobin urea /
disease surgery needs creatinine
Visser et al.14 + + +
(2015)
Raats et al.15 – – – –
(2015)
Sasajima et al.17 + + – – +
(2012)
Salata et al.18 + – + – – –
(2012)
Pol et al.19 (2011) – – –
Koebrugge et al.21 + – – – – – + – –
(2010)
Katznelson et al.22 + + + + – –
(2009)
Bohner et al.26 + + + +
(2003)
Schneider et al.27 – – +
(2002)
Rosen et al.28 + + – + + – –
(2002)
TOTAL number 8 6 4 2 2 2 3 3 3 4 2 2 4 5
of studies
TOTAL 7 1 3 2 1 2 1 2 0 1 1 1 0 1
POSITIVE
studies

Prognostic factors with a positive correlation to delirium examined in one study only included a history of previous myocardial infarction,28 smoking history,14 procedure type (aortic vs other)14 or amputation,26,29 preopera-
tive nursing care at home,15 preoperative beta-blocker,22 no preoperative statin22 or normal cholesterol,26,29 critical limb ischaemia,17 multi-segment arterial disease,17 history of supra-aortic arterial disease,26,29 body length
<170 cm,26 and low intraoperative potassium.26
A plus sign (+) indicates a significant predictive risk factor with p<0.05, a negative (–) sign indicates a non-significant or lack of an association with delirium, a blank cell indicates that this prognostic factor was not exam-
ined. Each of the prognostic factors shown was examined in two or more studies.
Vascular Medicine 22(5)
Aitken et al.
This review extends upon an earlier systematic review
conducted by Balasundaram and Holmes.34 Our review
includes 15 studies of postoperative delirium following
vascular surgery, and also includes meta-analysis and
bias assessment. By extending the scope of this review
to identify prognostic factors as well as incidence, this
study is able to report on factors of clinical relevance to
vascular surgical patients at risk of delirium. The emer-
gence of validated tools to assess prognostic study qual-
ity has led to advances in the methodology of systematic
reviews of prognostic factor studies. Risk of bias was
assessed using the QUIPS tool,8 which to our knowledge
has not previously been used in vascular surgery prog-
nostic studies. By using the QUIPS criteria8 to assess
risk of bias and by adhering the MOOSE guidelines,10
the reliability of our systematic review is enhanced and
informed by recent developments in the field of progno-
sis research methods.
Limitations
There are also limitations. Selective publication is an ongo-
ing issue for meta-analysis of prognostic studies. Kyzas
et al.35,36 highlight that the published prognostic literature
contains a significant bias towards factors with positive
associations, with endpoints associated with the highest
association preferentially reported. Furthermore, within
individual publications, reporting bias exists with a ten-
dency to report only those factors with positive associa-
tions, making meta-analysis difficult due to the lack of
non-predictive data. In this study, we have attempted to
ensure that all non-significant associations with delirium
are reported. We are conscious that by only including stud-
ies published in English, it is possible that important studies
that were published in languages other than English were
omitted; however, it was decided that the risk of selection
bias secondary to language was limited after a preliminary
review of the literature without language restrictions.
Variation in study quality and increased risk of bias from
confounders was evident and underscores the need for
standardization of methods and data reporting to enable
accurate characterization of the impact of postoperative
delirium. Dissimilarity in the presentation of results, with
different cut-points and measurement methods, reduced the
capacity for comparison with meta-analysis. Where meta-
analysis of unadjusted mean difference was performed, het-
erogeneity was high, as is common for meta-analysis of
observational studies,13 reflecting the lack of standardiza-
tion occurring when study groups are not randomized. On
sensitivity analysis, heterogeneity did not affect the asso-
ciation between increased age and the presence of delirium,
nor for the association between delirium and increased
length of hospital stay. I2 poorly reflects measures of het-
erogeneity when there are small numbers of studies
included, which limits the assessment of heterogeneity in
this review. For future studies, meta-analysis will be aided
by adherence to standardized reporting processes, includ-
ing measures such as registration of protocols, transparent
study methods37 and publication of results according to
395
guidelines such as the REMARK guidelines38 that recom-
mend utilizing continuous variables where possible and
including confidence intervals along with p-values. Another
potential drawback to our review is the assumption that all
methods for diagnosing delirium are equally robust; it is
beyond the scope of this review to assess the validity of
diagnostic tools for delirium. Future studies need to be
stringent in assessing and adjusting for potential confound-
ers to the development of delirium, with specific attention
paid to the reporting of measures affecting the onset and
outcomes of this condition.
Potential prognostic factors such as postoperative anal-
gesia requirements,39 anaesthetic choice,40 psychotropic
medication use41 and environmental factors2,5 affect postop-
erative delirium risk after general and orthopaedic surgery,
but our systematic review found that these potential prog-
nostic factors were not included in the studies of delirium
after vascular surgery. Early diagnosis and management of
delirium has been shown to decrease duration and severity
of delirium episodes, decrease mortality and reduce length
of stay in surgery other than vascular surgery.2,42 Treatment
protocols and diagnostic criteria for delirium already exist
but exploration into the impact of these criteria on clinical
decision-making is needed. Improvements in models of
care, with greater involvement of geriatricians and multidis-
ciplinary health teams into the postoperative management of
delirium patients, may lead to better outcomes.
Conclusion
This review has implications for clinical care of vascular
surgical patients. Our review suggests delirium is com-
mon and has an impact on length of stay, post discharge
care requirements and peri-operative complications.
Factors predicting an increased risk of delirium include
increased patient age, pre-existing cognitive impairment,
diabetes, hypertension and open aortic surgery. While the
prognostic factors identified in this systematic review,
such as age and comorbidities, are not modifiable, other
factors, such surgical or anaesthetic choice, models of
care, or pre-optimization of physiological parameters, are
potentially modifiable but have not been assessed in vas-
cular surgery. Arguably, these are the most valuable fac-
tors to study because of their potential for interventions.
Further study into the risks and impact of postoperative
delirium should focus upon potentially modifiable factors.
Well-designed clinical trials are needed to evaluate the
impact that modifying prognostic factors and using thera-
peutic interventions has on the development and outcomes
of delirium.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support
for the research, authorship, and/or publication of this article: the
primary author, SJ Aitken, is the recipient of the following grants
396
that have supported this research: Royal Australian College of
Surgeons Senior Lecturer Fellowship; Sydney Medical Foundation
Chapman Bequest for research in cardiovascular disease; Ageing
and Alzheimer’s Research Foundation, University of Sydney.
Supplementary material
The supplementary material is available at http://vmj.sagepub.
com/supplemental
References
1. Salluh JIF, Wang H, Schneider EB, et al. Outcome of delir-
ium in critically ill patients: Systematic review and meta-
analysis. BMJ 2015; 350: h2538.
2. Inouye SK, Bogardus ST, Charpentier PA, et al. A multi-
component intervention to prevent delirium in hospitalized
older patients. N Engl J Med 1999; 340: 669–676.
3. Milisen K, Foreman MD, Wouters B, et al. Documentation
of delirium in elderly patients with hip fracture. J Gerontol
Nurs 2002; 28: 23–29.
4. Dasgupta M, Dumbrell AC. Preoperative risk assessment for
delirium after noncardiac surgery: A systematic review. J Am
Geriatr Soc 2006; 54: 1578–1589.
5. Robinson T, Raeburn C, Tran Z. Postoperative delirium in
the elderly: Risk factors and outcomes. Ann Surg 2009; 249:
173–178.
6. Hayden J, Malats N, Abrams D, et al. Prognosis Research
Strategy (PROGRESS) 2: Prognostic factor research. PLoS
Med 2013; 10: e1001380.
7. University of York Centre for Reviews and Dissemination.
PROSPERO International prospective register of system-
atic reviews CRD42015019690. http://www.crd.york.ac.uk/
PROSPERO/ (2014, accessed 24 March 2015).
8. Hayden JA, Cote P, Bombardier C. Evaluation of the quality
of prognosis studies in systematic reviews. Ann Intern Med
2006; 144: 427–437.
9. Drubbel I, Numans ME, Kranenburg G, et al. Screening for
frailty in primary care: A systematic review of the psycho-
metric properties of the frailty index in community-dwelling
older people. BMC Geriatr 2014; 14: 27.
10. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of
observational studies in epidemiology: A proposal for report-
ing. Meta-analysis Of Observational Studies in Epidemiology
(MOOSE) group. JAMA 2000; 283: 2008–2012.
11. Higgins J, Green S (eds). Cochrane Handbook for Systematic
Reviews of Interventions. Version 5.1.0. London: The
Cochrane Collaboration, 2011.
12. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and
variance from the median, range, and the size of a sample.
BMC Med Res Methodol 2005; 5: 13.
13. Dwyer T, Couper D, Walter SD. Sources of heterogeneity in
the meta-analysis of observational studies: The example of
SIDS and sleeping position. J Clin Epi 2001; 54: 440–447.
14. Visser L, Prent A, van der Laan MJ, et al. Predicting post-
operative delirium after vascular surgical procedures. J Vasc
Surg 2015; 62: 183–189.
15. Raats JW, van Hoof-de Lepper CC, Feitsma MT, et al.
Current factors of fragility and delirium in vascular surgery.
Ann Vasc Surg 2015; 29: 968–976.
16. Ellard L, Katznelson R, Wasowicz M, et al. Type of anes-
thesia and postoperative delirium after vascular surgery.
J Cardiothorac Vasc Anesth 2014; 28: 458–461.
17. Sasajima Y, Sasajima T, Azuma N, et al. Factors related to
postoperative delirium in patients with lower limb ischaemia:
Vascular Medicine 22(5)
A prospective cohort study. Eur J Vasc Endovasc Surg 2012;
44: 411–415.
18. Salata K, Katznelson R, Beattie WS, et al. Endovascular
versus open approach to aortic aneurysm repair surgery:
Rates of postoperative delirium. Can J Anaesth 2012; 59:
556–561.
19. Pol RA, van Leeuwen BL, Visser L, et al. Standardised
frailty indicator as predictor for postoperative delirium after
vascular surgery: A prospective cohort study. Eur J Vasc
Endovasc Surg 2011; 42: 824–830.
20. Bryson GL, Wyand A, Wozny D, et al. A prospective
cohort study evaluating associations among delirium, post-
operative cognitive dysfunction, and apolipoprotein E gen-
otype following open aortic repair. Can J Anaesth 2011;
58: 246–255.
21. Koebrugge B, van Wensen RJ, Bosscha K, et al. Delirium
after emergency/elective open and endovascular aortoiliac
surgery at a surgical ward with a high-standard delirium care
protocol. Vascular 2010; 18: 279–287.
22. Katznelson R, Djaiani G, Mitsakakis N, et al. Delirium follow-
ing vascular surgery: Increased incidence with preoperative
beta-blocker administration. Can J Anaesth 2009; 56: 793–801.
23. Rudolph JL, Jones RN, Rasmussen LS, et al. Independent
vascular and cognitive risk factors for postoperative delir-
ium. Am J Med 2007; 120: 807–813.
24. Minden SL, Carbone LA, Barsky A, et al. Predictors and
outcomes of delirium. Gen Hosp Psychiatry 2005; 27:
209–214.
25. Benoit AG, Campbell BI, Tanner JR, et al. Risk factors and
prevalence of perioperative cognitive dysfunction in abdom-
inal aneurysm patients. J Vasc Surg 2005; 42: 884–890.
26. Bohner H, Hummel TC, Habel U, et al. Predicting delirium
after vascular surgery: A model based on pre- and intraopera-
tive data. Ann Surg 2003; 238: 149–156.
27. Schneider F, Bohner H, Habel U, et al. Risk factors for post-
operative delirium in vascular surgery. Gen Hosp Psychiatry
2002; 24: 28–34.
28. Rosen SF, Clagett GP, Valentine RJ, et al. Transient
advanced mental impairment: An underappreciated morbid-
ity after aortic surgery. J Vasc Surg 2002; 35: 376–381.
29. Bohner H, Friedrichs R, Habel U, et al. Delirium increases
morbidity and length of stay after vascular surgery. Results
of a prospective study. Chirurg 2003; 74: 931–936.
30. Pol RA. Frailty should determine type of anesthesia in reduc-
ing postoperative delirium after vascular surgery and not
vice versa. J Cardiothorac Vasc Anesth 2014; 28: e61–e2.
31. Sasajima Y, Sasajima T, Uchida H, et al. Postoperative delirium
in patients with chronic lower limb ischaemia: What are the spe-
cific markers? Eur J Vasc Endovasc Surg 2000; 20: 132–137.
32. Bohner H, Schneider F, Stierstorfer A, et al. Postoperative
delirium following vascular surgery. Anaesthesist 2000; 49:
427–433.
33. Morimoto Y, Yoshimura M, Utada K, et al. Prediction of
postoperative delirium after abdominal surgery in the elderly.
J Anesth 2009; 23: 51–56.
34. Balasundaram B, Holmes J. Delirium in vascular surgery.
Eur J Vasc Endovasc Surg 2007; 34: 131–134.
35. Kyzas PA, Denaxa-Kyza D, Ioannidis JPA. Quality of
reporting of cancer prognostic marker studies: Association
with reported prognostic effect. J Natl Cancer Inst 2007; 99:
236–243.
36. Kyzas PA, Loizou KT, Ioannidis JPA. Selective reporting
biases in cancer prognostic factor studies. J Natl Cancer Inst
2005; 97: 1043–1055.
Aitken et al.
37. Peat G, Riley R, Morley K, et al. Improving the transpar-
ency of prognosis research: The role of reporting, data
sharing, registration, and protocols. PLoS Med 2014; 11:
e1001671.
38. Altman DG, McShane LM, Sauerbrei W, et al. Reporting
Recommendations for Tumor Marker Prognostic Studies
(REMARK): Explanation and elaboration. PLoS Med 2012;
9: e1001216.
39. Lynch E, Lazor M, Gellis J. The impact of postoperative pain
on the development of postoperative delirium. Anesth Analg
1988; 86: 781–785.
397
40. American Geriatrics Society Expert Panel on Postoperative
Delirium in Older Adults. American Geriatrics Society
abstracted clinical practice guideline for posteropative delir-
ium in older adults. J Am Geriatr Soc 2015; 63: 142–150.
41. Marcantonio ER, Flacker J, Wright R. Reducing delirium
after hip fracture: A randomized trial. J Am Geriatr Soc
2001; 49: 516–522.
42. Vidán M, Sanchez E, Alonso M. An intervention integrated
into daily clinical practice reduces the incidence of delirium
during hospitalization in elderly patients. J Am Geriatr Soc
2009; 57: 2029–2036.