British Journal of Anaesthesia, 121 (4): 706e721 (2018)

doi: 10.1016/j.bja.2018.04.036
Advance Access Publication Date: 20 June 2018
Review Article

CARDIOVASCULAR

Intraoperative hypotension and the risk of

postoperative adverse outcomes: a systematic review
E. M. Wesselink1,*, T. H. Kappen1, H. M. Torn1, A. J. C. Slooter2 and
W. A. van Klei1
1
Department of Anesthesiology, Utrecht, The Netherlands and 2Department of Intensive Care Medicine,
University Medical Center Utrecht, Utrecht University, Utrecht, The Netherlands

*Corresponding author. E-mail: [email protected]

This article is accompanied by an editorial: Making sense of the impact of intraoperative hypotension: from populations to the individual
patient by Ke et al., Br J Anesth 2018:121:689e691, doi: 10.1016/j.bja.2018.07.003.

Abstract
Background: Intraoperative hypotension is a common side effect of general anaesthesia and might lead to inadequate
organ perfusion. It is unclear to what extent hypotension during noncardiac surgery is associated with unfavourable
outcomes.
Methods: We conducted a systematic search in PubMed, Embase, Web of Science, and CINAHL, and classified the quality
of retrieved articles according to predefined adapted STROBE and CONSORT criteria. Reported strengths of associations
from high-quality studies were classified into end-organ specific injury risks, such as acute kidney injury, myocardial
injury, and stroke, and overall organ injury risks for various arterial blood pressure thresholds.
Results: We present an overview of 42 articles on reported associations between various absolute and relative intra-
operative hypotension definitions and their associations with postoperative adverse outcomes after noncardiac surgery.
Elevated risks of end-organ injury were reported for prolonged exposure (
10 min) to mean arterial pressures <80 mm Hg
and for shorter durations <70 mm Hg. Reported risks increase with increased durations for mean arterial pressures
<65e60 mm Hg or for any exposure <55e50 mm Hg.
Conclusions: The reported associations suggest that organ injury might occur when mean arterial pressure decreases
<80 mm Hg for
10 min, and that this risk increases with blood pressures becoming progressively lower. Given the
retrospective observational design of the studies reviewed, reflected by large variability in patient characteristics, hy-
potension definitions and outcomes, solid conclusions on which blood pressures under which circumstances are truly
too low cannot be drawn. We provide recommendations for the design of future studies.
Clinical registration number: (PROSPERO ID). CRD42013005171.

Keywords: acute kidney injury; hypotension; mortality; myocardial ischemia; stroke

Editorial decision: 2 May 2018; Accepted: 2 May 2018

© 2018 British Journal of Anaesthesia. Published by Elsevier Ltd. All rights reserved.
For Permissions, please email: [email protected]

706
 Hypotension and postoperative adverse outcomes - 707

Box 1
Editor’s key points Search string
 In a systematic review of the association between
intraoperative hypotension and adverse postoperative
Determinant: ((((hypotension[title and abstract] OR hypo-
outcomes in noncardiac surgery, 42 relevant studies
tensive[title and abstract]) AND (intraoperative[title and
were identified and analysed.
abstract] OR perioperative[title and abstract] OR intra-
 Elevated risks of end-organ injury were reported for
operatively[title and abstract] OR perioperatively[title
exposures to mean arterial pressures <80 mm Hg for
and abstract] OR peroperative[title and abstract] OR
>10 min, and for shorter durations <70 mm Hg.
peroperatively[title and abstract])))).
 Elevated risks were reported for increased durations for
Outcome: (mortal*[title and abstract] OR death[title and
mean arterial pressures <65e60 mm Hg or for any
abstract] OR “moribund”[title and abstract] OR die*[title
exposure <55e50 mm Hg.
and abstract] OR fatal[title and abstract]) OR ((kidney
 Future prospective studies are indicated with less
[title and abstract] OR renal[title and abstract]) AND
variability in patient characteristics and better defini-
(insuff*[title and abstract] OR failure[title and abstract]
tions of hypotension and adverse outcomes.
OR injury[title and abstract] OR “ATN”[title and ab-
stract]) OR (((heart[title and abstract] OR myocard*[title
and abstract] OR cardial[title and abstract] OR coronary
[title and abstract]) AND (ischem*[title and abstract] OR
Intraoperative hypotension is a common side-effect of general
ischaem*[title and abstract] OR infarct*[title and ab-
anaesthesia that has received much attention in recent years
stract]) OR (acute AND coronary AND syndrome[title
because of its frequent occurrence and presumed adverse
and abstract] OR ACS[title and abstract])) OR (((Brain
consequences. However, no widely accepted definition of
[title and abstract] OR cerebr*[title and abstract]) AND
intraoperative hypotension is available.1 Despite this lack of a
(Vascular[title and abstract] OR cerebrovascular[title
uniform definition, researchers have addressed the associa-
and abstract]) AND (embol*[title and abstract] OR acci-
tion between intraoperative hypotension and postoperative
dent*[title and abstract] OR complication*[title and ab-
mortality and organ dysfunction after general anaesthesia.
stract] OR ischaem*[title and abstract] OR ischem*[title
Monk and colleagues2 were one of the first groups to show a
and abstract] OR infarct*[title and abstract] OR incident*
significant association between duration of intraoperative
[title and abstract] OR stroke[title and abstract] OR
hypotension and mortality. More recent landmark studies
stroke*[title and abstract] OR apoplexy[title and ab-
have shown associations between hypotension and other
stract]) OR (((Delirium[title and abstract] OR Deliriou*
adverse outcomes such as acute kidney injury (AKI) and
[title and abstract])) OR ((admission[title and abstract]
myocardial injury (MI).3,4
OR stay[title and abstract]) AND (day*[title and abstract]
It remains a topic of debate if, and to what extent, hypo-
OR duration[title and abstract] OR LOS[title and ab-
tension disrupts organ perfusion resulting in organ damage.
stract] OR length[title and abstract]) OR (morbidity[title
Furthermore, such organ damage might depend on the degree
and abstract] OR complication[title and abstract] OR
and duration of the hypotensive episodes. A summary of what
“adverse event”[title and abstract] OR “adverse even-
is known about the effects of intraoperative hypotension on
ts”[title and abstract]))
postoperative organ dysfunction and mortality is essential for
anaesthesiologists to determine the range of blood pressures
acceptable during surgery. So far, no systematic search of the
literature has been conducted to summarise the available ev-
idence regarding the association between intraoperative hy- ischaemic stroke, delirium, and length of hospital stay
potension and adverse postoperative outcomes. As (LOS) as described in Box 1. The search filters were
hypotension has not clearly been defined, such a summary restricted to presence of the synonyms in titles and ab-
needs to include an analysis of which blood pressure stracts. No other limits were used. The articles obtained by
threshold an association with adverse outcomes becomes this search were independently screened by two reviewers
clinically relevant. (E.W. and H.M.T.). In case of inconsistency, consensus was
We studied the relationship between intraoperative hypo- achieved by a third independent reviewer (T.H.K.). The
tension and postoperative adverse outcomes after noncardiac reference lists of all selected and included articles were
surgery by performing a systematic search of the literature. checked to retrieve relevant publications that were not
We classified studies according to quality criteria and report found by the above described search strategy. The inclu-
strengths of associations for various blood pressure thresholds sion and exclusion criteria for publication type, study
and postoperative adverse outcomes. design, hypotension, and studied outcome definitions are
described in Box 2.

Methods
Search strategy and selection of articles Data extraction and quality assessment
We conducted a systematic search of literature in PubMed, Data on study design, hypotension definitions, studied out-
Embase, Web of Science, and CINAHL on March 8, 2017. comes, and (adjusted) strengths of association were extrac-
Synonyms and medical subject headings for intraoperative ted from all included studies (Tables 1 and 2).2e43 Commonly
hypotension were combined with synonyms and medical reported baseline characteristics were summarised by
subject headings for complication, mortality, AKI, MI, calculating weighted means of medians across study groups
 708 - Wesselink et al.

Box 2 normalisation was achieved by calculating the quality score,

Selection criteria defined as the number of positive items divided by the
maximum number of items for that study type and expressed
as a percentage.
Publication type

 Full reports published before 8 March 2017. Reporting and aggregation of results
 Written in English or Dutch.
For the reported strengths of association for absolute blood
Study design pressure thresholds and hypotension durations, multivari-
able associations from aetiological studies were presented
 Studies in which
50% of the adult patients underwent
when available; otherwise univariable strengths of associa-
general anaesthesia or general anaesthesia combined
tions were shown (Table 3). Strengths of associations were
with local or regional anaesthesia for noncardiac sur-
grouped according to (cumulative) durations of
1 min,
gery were included. Studies in animals or children and

5 min,
10 min, and
20 min exposure to intraoperative
emergency procedures for ruptured vessels were
hypotension below particular absolute thresholds. Studies
excluded.
that used a threshold that was relative to a baseline blood
 Studies were excluded when they selected a subgroup
pressure were grouped with the absolute threshold that
of patients with a specific comorbidity that was not
corresponded with the relative departure from the reported
part of the reason to perform the surgical procedure.
mean baseline blood pressure or a baseline of 140/90 mm Hg
 The study design had to be a randomised controlled
when no mean baseline was reported. For example,
clinical trial, a cohort study or a caseecontrol study
Hallqvist and colleagues14 used a relative threshold of a 50%
with >10 patients. Case series, case reports, meta-
decrease in systolic blood pressure (SBP). As no mean
analyses and (systematic) reviews were excluded.
baseline blood pressure was reported, the study was
 The association between intraoperative hypotension
grouped with absolute thresholds SBP <70 mm Hg (50% of
and at least one outcome (mortality, acute kidney
140 mm Hg).
injury, myocardial injury, stroke, delirium, length of
For studies that reported their strength of association per
stay) had to be reported. The definition of determinant
time-unit increase, the strengths of associations for other
or outcome did not belong to the inclusion or exclusion
durations were estimated from the reported strengths of
criteria. Studies focusing on intentional or induced
associations using the lower bound of the duration category.
intraoperative hypotension or on the effects of anti-
For example, Monk and colleagues2 reported a relative risk
hypertensive medication were excluded.
(RR) of 1.036 per min that SBP was <80 mm Hg. For the
Definition of intraoperative hypotension category SBP
5 min <80 mm Hg, the RR was then estimated
by 1.0365¼1.193.
 Intraoperative hypotension had to be defined in the
The reported strengths of associations were aggregated
article as an absolute or relative blood pressure
into single risk categories per blood pressure threshold and per
threshold. Blood pressure thresholds had to be clini-
hypotension duration for the five organ injury out-
cally relevant (i.e. not a mean blood pressure <100 mm
comesdmortality, AKI, MI, stroke, and deliriumdand com-
Hg or >5% decrease compared with baseline blood
bined into an overall organ injury risk per threshold and
pressure).
duration (Table 4). For each category, the highest association
[odds ratio (OR), RR, or hazard ratio (HR)] among the available
evidence of sufficient quality was classified into a risk of mild,
moderate, or high as defined below. Studies were considered
for each variable (Table 2). Two reviewers (E.M.W. and H.M.T.) qualitatively sufficient when they received a quality score
independently assessed the methodological quality of the
80% and defined an intraoperative hypotension analysis in
included articles. The Strengthening the Reporting of their primary of secondary objectives. The highest organ
Observational Studies in Epidemiology (STROBE)44 criteria injury risk for a specific blood pressure threshold and duration
and Consolidated Standards of Reporting Trials (CONSORT)45 category was then considered to be the overall organ injury
criteria were adapted and used for the composition of a risk.
checklist with predefined quality criteria. These criteria focus Several additional assumptions and conversions were
on the internal validity, external validity, bias, and precision made to compare studies and their strengths of associations.
(Supplementary Table S1). In short, all criteria on study First, strengths of associations that could not be converted to
design were scored as positive (þ), negative (e), unclear (?), or an OR, RR, or HR per blood pressure threshold and duration
not applicable (NA). The latter option was used when the category were not considered in the risk categorisation sum-
criterion was not appropriate for the specific study design, mary. Second, ORs, RRs, and HRs were deemed interchange-
such as loss-to-follow-up for a caseecontrol study. Depend- able in their magnitude, as the outcome incidences were
ing on the type of study design, a maximum of 13 ‘positive’ relatively low (the rare disease assumption). Third, the cut-off
items (caseecontrol studies), 14 items (randomised clinical to classify a strength of association as high-risk was chosen as
trials; RCTs), or 15 items (cohort studies) could be assigned a ‘doubled risk’ or more (OR/RR/HRhigh
2.0). The cut-off for
(Table 1, Supplementary Table S1). Disagreements between moderate risk was chosen at half the high-risk cut-off on an
both reviewers were discussed. In case of persistent exponential scale (√2¼1.4OR/RR/HRmoderate<2.0), with the
disagreement, the third reviewer made the final decision. mild-risk category starting at a minimal increased risk
Whereas the maximum score depended on study type, (1.0<OR/RR/HRmild<1.4). Fourth, all blood pressure thresholds
 Hypotension and postoperative adverse outcomes - 709

Table 1 Results of the methodological assessment of studies on intraoperative hypotension and postoperative adverse outcomes.
Scoring system to obtain a quality score for every included article based on 15 categories as described in Supplementary Table S1. þ, if
sufficient information is available and positive assessment; e, if insufficient information or negative assessment; ?, unknown; NA, not
applicable. Depending on the type of study design, a maximum of 13 points (caseecontrol studies), 14 points (randomised controlled
trials, RCTs) or 15 points (cohort studies) were assigned

First author (yr)ref Design A B C D E F G H I J K L M N O Quality score (%)

5
Hirsch (2015) Cohort þ þ þ þ þ þ þ þ þ þ þ þ þ þ þ 15 (100)
Monk (2015)6 Cohort þ þ þ þ þ þ þ þ þ þ ? þ þ þ þ 14 (93)
Willingham (2015)7 Cohort þ þ þ þ þ þ þ þ þ þ þ ? þ þ þ 14 (93)
Bijker (2012)8 Caseecontrol e þ þ þ þ NA NA þ þ þ þ þ þ þ þ 12 (92)
Mizota (2017)9 Cohort e þ þ þ þ þ þ þ þ þ þ e þ þ þ 13 (87)
Sun (2015)3 Cohort þ þ þ þ þ þ þ e þ þ þ ? þ þ þ 13 (87)
Schmid (2016)10 RCT þ e þ þ þ þ þ þ þ NA þ e þ þ þ 12 (86)
Roshanov (2017)11 Cohort þ þ þ þ þ þ þ e þ þ þ e þ þ þ 12 (80)
Salmasi (2017)12 Cohort e þ þ þ þ þ þ e þ þ þ e þ þ þ 12 (80)
Babazade (2016)13 Cohort e þ þ þ þ þ þ e þ þ þ ? þ þ þ 12 (80)
Hallqvist (2016)14 Cohort e þ þ þ þ þ þ e þ e þ þ þ þ þ 12 (80)
Van Waes (2016)15 Cohort e þ þ þ þ þ þ e þ þ þ ? þ þ þ 12 (80)
Mascha (2015)16 Cohort þ þ þ þ þ þ ? e þ þ þ ? þ þ þ 12 (80)
Pipanmekaporn (2014)17 Cohort þ e þ þ þ ? þ þ þ þ þ ? þ þ þ 12 (80)
Walsh (2013)4 Cohort e þ þ þ þ þ þ e þ þ þ ? þ þ þ 12 (80)
Bijker (2009)18 Cohort e þ þ e þ þ þ þ þ þ þ ? þ þ þ 12 (80)
Kheterpal (2009)19 Cohort þ þ þ þ þ þ þ e þ e þ ? þ þ þ 12 (80)
Monk (2005)2 Cohort þ e þ þ þ þ ? e þ þ þ þ þ þ þ 12 (80)
White (2016)20 Cohort e e þ þ þ þ þ þ þ þ þ ? þ e þ 11 (73)
Brinkman (2015)21 Cohort þ þ þ þ e þ ? þ þ þ þ þ þ e e 11 (73)
Petsiti (2015)22 Cohort þ þ þ þ þ þ e þ þ ? e þ e þ þ 11 (73)
Marcantonio (1998)23 Cohort þ þ þ þ þ þ ? þ þ 0 e ? þ e þ 11 (73)
Tallgren (2007)24 RCT þ e þ þ e þ ? þ þ NA þ þ þ þ e 10 (71)
House (2016)25 Cohort e e þ þ þ þ þ e þ e þ ? þ þ þ 10 (67)
Sessler (2012)26 Cohort e þ þ þ þ ? ? þ þ þ þ ? þ e þ 10 (67)
Sabate
(2011)27 Cohort þ e þ þ þ ? ? þ þ e þ ? þ þ þ 10 (67)
Taffe
(2009)28 Cohort þ þ þ þ þ ? ? e þ e þ ? þ þ þ 10 (67)
Sirivatanauksorn (2014)29 Cohort þ e þ þ þ þ ? e þ e þ ? þ e þ 9 (60)
Tassoudis (2011)30 Cohort þ þ þ þ e þ ? þ þ e e ? þ þ e 9 (60)
Stapelfeldt (2017)31 Cohort e þ þ e þ e ? e þ þ þ e þ e þ 8 (53)
Jiang (2016)32 Cohort e e þ þ þ þ ? þ e e e e þ þ þ 8 (53)
Yang (2016)33 Cohort e þ e þ þ þ e þ e e e e þ þ e 7 (47)
Yue (2013)34 Cohort e þ þ e þ ? ? e e e þ ? þ þ þ 7 (47)
Franck (2011)35 Cohort e þ þ e þ ? ? þ þ e e ? e þ þ 7 (47)
Patti (2011)36 Cohort þ e þ þ e ? ? e e e þ ? þ þ þ 7 (47)
Vasivej (2016)37 Caseecontrol e e þ þ þ NA NA e e e þ ? þ þ e 6 (46)
Thakar (2007)38 Cohort e e þ e e ? ? e þ e þ ? þ þ þ 6 (40)
Barone (2002)39 Caseecontrol e þ e e e NA NA e þ e þ ? þ þ e 5 (38)
Lima (2003)40 Cohort e e þ e e þ ? e e þ þ ? þ e e 5 (33)
Nakamura (2009)41 Caseecontrol e e þ e e NA NA e e e þ ? þ þ e 4 (31)
Davidovic (2017)42 Cohort þ e þ e e e ? e e e þ e þ e e 4 (27)
Sharma (2006)43 Caseecontrol e e e e e ? ? e e e þ ? þ þ e 3 (23)

were converted to mean arterial pressure (MAP) equivalents for acute kidney injury could not become a mild risk at
based on a pulse pressure of 40 mm Hg (i.e. an SBP threshold MAP<55 mm Hg, nor could it become ‘no risk’ because no
<90 mm Hg represents a blood pressure of 90/50 mm Hg, which appropriate study reporting an association for that threshold
is comparable with a MAP threshold of <65 mm Hg). The was available. This reflects the assumption that lower blood
40 mm Hg pulse pressure was chosen because it was consid- pressures or longer intraoperative hypotension episodes al-
ered the most plausible pulse pressure across all reported ways aggravate the risk of organ injury.
systolic blood pressure thresholds. MAP was calculated by
adding systolic blood pressure to two times diastolic blood
pressure divided by three. Fifth, once a specific MAP threshold Results
and hypotension duration reached a certain risk classification,
that classification carried over to all subsequent lower MAP
Included studies
thresholds or longer hypotension durations at the same MAP Our search strategy yielded 5224 articles in total. After
threshold. This means that a moderate risk at MAP<60 mm Hg removal of duplicates (n¼1955) and removing articles based
Table 2 Summary of the patient, surgery and hypotension characteristics of each study. Continuous variables are expressed as mean (SD) or x (x~x) ¼ mean (range) or x (x-x) ¼ median

710
(interquartile range). Categorical variables are expressed as xx% body mass index or weight (W) is expressed as kg m2 or kg. Any definition of history of arrhythmia or renal dysfunction
was included. If available, renal dysfunction was expressed as serum creatinine in mmol l1. e, not available; *, values are the weighted mean values of the study groups; y, median
(10the90th percentile). a-, absolute threshold expressed as mm Hg; AR, achievement rate; AUT, area under the threshold; BP blood pressure; CAD, coronary artery disease; Cr, creatinine;

-
CVD, cerebrovascular disease; DBP, diastolic blood pressure; Di, dichotomous; DM, diabetes mellitus; Du, duration under a blood pressure threshold; eGFR, estimated glomerular

Wesselink et al.
filtration rate; Gen. anaesth., general anaesthesia; GI, gastro-intestinal; HT, hypertension; LiverTX, liver transplantation; Low, lowest blood pressure; MAP, mean blood pressure; MI,
myocardial injury; r-, relative threshold expressed as percentage decrease from baseline BP; RCRI, Revised Cardiac Risk Index; SAS, blood pressure as part of Surgical Apgar Score; SBP,
systolic blood pressure; TIA, transient ischaemic attack; TL, triple low; TWA, time-weighted average; V, blood pressure variance; Var, variability or variance

Studies Patient Comorbidity (%) Procedure characteristics (%) Intraoperative hypotension

characteristics

First author Total Age Sex ASA Stroke HT DM CAD Renal Gen. Length of Emergency Type of BP threshold Threshold Analysis
disease anaesth. surgery surgery surgery type

(yr) n (yr) F, % 1 2 3 4 any min most

frequent

Hirsch (2015)5 594 74 (6) 51 48 e e 4 e e e e 71 300 (144) 0 Orthopaedic rSBP, rMAP, Y >10e40%, Du, Var
53 aMAP <50 mm Hg
6
Monk (2015) 18 756 60 (13) 7 3 26 58 13 e e e e 19 e 120 8 General 32 aSBP <80 mm Hg Du
(72e186)
Willingham 13 198 56 (44e66) 47 9 39 35 16 any: 3 48 17 20 e 100 178 e e aMAP, TL <75mm Hg Du
(2015)7 (115e259)
Bijker (2012)*,8 48 241 66 (57e76) 40 68 32 38 69 e e e e 163 e Vascular 48 aSBP, aMAP, <100e70 mm Hg, Du
(42/252) (130e232) rSBP, rMAP <70e40 mm Hg,
Y >10e40%
Mizota (2017)*,9 231 54 (44e60) 51 e e e e e e 19 e 22 100 838 e LiverTX 100 aMAP <40,<50 mm Hg Du, Di
(752e960)
,3
Sun (2015)* 5127 61 (14) 53 e e e e any: 2 48 15 11 16 e >120: 79 0 General 26 aMAP <65, <60, <55 Du
mm Hg
Schmid (2016)*,10 180 66 (12) 23 e e e e e e 56 20 6 100 e 100 Abdominal aMAP >70 mm Hg AR
100
Roshanov (2017)11 14 687 65 (12) 52 e e e e 7 47 19 12 eGFR: 79 e e 14 Orthopaedic aSBP <90 mm Hg Di
(23) 20
,12
Salmasi (2017)* 57 315 56 (15) 56 2 38 54 7 3 49 17 e 1 e 225(121) 4 Abdominal aMAP, rMAP <80e40 mm Hg, Du,
23 Y>10e60% TWA
13
Babazade (2016) 2521 56 (15) 45 45 50 5 e 42 14 e 4 100 199 1 Colorectal aSBP, aMAP <80mm Hg, Du
(142e265) 100 <55mm Hg
Hallqvist (2016)14 300 67 (57e74) 53 10 46 43 0.3 e 43 8 e e 39 e 0 Abdominal rSBP Y >50%, >5 min Di
40
van Waes (2016)*,15 890 74 (8) 31 1 14 36 49 21 e 10 e 8 100 191 (108) 30 (T)EVAR 24 aMAP <50, <60 mm Hg, Du,
Y>30%, Y >40% (AUT)
,16
Mascha (2015)* 104 401 57 (18) 53 5 40 47 8 1 48 17 14 6 e 174 5 e aMAP <80e50 mm Hg Du,
(114e252) TWA,
Var

Continued
Table 2 Continued

Studies Patient Comorbidity (%) Procedure characteristics (%) Intraoperative hypotension

characteristics

First author Total Age Sex ASA Stroke HT DM CAD Renal Gen. Length of Emergency Type of BP threshold Threshold Analysis
disease anaesth. surgery surgery surgery type

(yr) n (yr) F, % 1 2 3 4 any min most

frequent

Pipanmekaporn 719 49 (16) 29 14 58 28 e e 18 10 2 10 100 142 (65) 46 Thoracic 100 aSBP, aMAP <80 or <60 mm Hg, Di
(2014)*,17 >15 min
Walsh (2013)*,4 33 330 56 (16) 50 2 40 50 8 5 e 13 e eGFR: 93 e e 7 e aMAP <75e55 mm Hg Du
(27)
18
Bijker (2009) 1705 52 (16) 48 38 51 11 7 22 8 any: 15 e 88 112 0 General 88 aSBP, rSBP, <100e70 mm Hg, Du
(73e163) aMAP, rMAP Y >10e40%,
<70-40 mm Hg,
Y >10e40%
Kheterpal 7740
68: 23 49 e e e e any: 5 40 13 e 3 88 e 12 e aSBP, aMAP, <80e70 mm Hg, Di?
(2009)*,19 rSBP, rMAP <60e50 mm
Hg, Y >30e40%
Monk (2005)2 1064 51 (37e65) 64 13 52 35 4 33 4 6 e 100 186 e Orthopaedic aSBP <80 mm Hg Du
(138e258) 26
White (2016)20 11 085 83 (24e104) 72 3 30 55 12 11 55 7 e 14 54 e e Hip surgery aSBP, aMAP Lowest BP Low
100
Brinkman 40 69 (9) 35 e e e e e 68 13 e e 100 228 (84) 0 Aorta 100 aMAP <65 mm Hg AUT
(2015)*,21
Petsiti (2015)22 248 64 (11) 48 e e 32 1 47 13 8 3 100 232 (55) 0 Abdominal aMAP, rMAP <60 or <70 Di
100 mm Hg þ

Hypotension and postoperative adverse outcomes

Y >30%
Marcantonio 1341 67 (9) 55 e e e e e e e e e e e e Orthopaedic aSBP or rSBP <90 mm Hg Di
(1998)23 43 or Y >33%
Tallgren (2007)*,24 69 67 (60e74) 22 e e e e e 66 7 43 e 100 0 Aorta 100 aMAP <65 mm Hg, Di
>15 min
House (2016)*,25 46 799 54 (13) 47 41 3 43 16 e 4 e 162 (108) 4 e aMAP <40 mm Hg Di/SAS
Sessler (2012)26 24 120 e e e e e e e e e e e 100 e e e aMAP, TL <70 mm Hg Du
Sabate
(2011)27 3387 67 (47e81)y 52 8 55 33 4 e e e RCRI e 61 120 7 Orthopaedic aSBP or <100 mm Hg or Di

3: 7 (60e248)y 34 aMAP/rMAP Y >20 mm
Hg/20%, >60 min
Taffe
(2009)*,28 147 573 55 (18) 56 27 48 22 3 e e e e e 67 104 (?e?) 20 e rMAP Y >30%, >10 min Di
Sirivatanauksorn 81 53 (23e70) 31 e e e e e e e e Cr: 90 100 276 e LiverTX 100 aMAP <70 mm Hg, Di
(2014)*,29 (38e168) (168e438) >30 min
Tassoudis (2011)30 100 62 (14) 47 e 32 e e 1 46 12 8 3 100 195 (71) 0 Abdominal aMAP, rMAP <60 or Di
100 <70mm Hg þ
Y >30%

Continued

-
711
Table 2 Continued

712
Studies Patient Comorbidity (%) Procedure characteristics (%) Intraoperative hypotension
characteristics

-
First author Total Age Sex ASA Stroke HT DM CAD Renal Gen. Length of Emergency Type of BP threshold Threshold Analysis

Wesselink et al.
disease anaesth. surgery surgery surgery type

(yr) n (yr) F, % 1 2 3 4 any min most

frequent

Stapelfeldt (2017)31 152 445 e e e e e e e e e e e e 179 90 e aMAP <75e45 mm Hg Du

(118e259)
,32
Jiang (2016)* 451 65 (18) 50 e e e e e 14 6 e 1 100 164 (62) e Spine 100 aSBP <80 mm Hg Di
Yang (2016)*,33 480 81 (6) 51 e 71 29 e e 43 30 43 e 100 188 (32) 0 e r?BP Y >30% Di
Yue (2013)*,34 71 >70: 37 21 e e e e e e 11 e 23 100 e 38 Aorta 100 aSBP/aMAP <Y 30/<65 Di
mm Hg
Franck (2011)*,35 2350 53 (41e65) 50 50 50 e e e e e 100 98 (63e148) e e aSBP, rSBP <100 or Di
Y>30%, <92
mm Hg,
<80 mm Hg,
Y >20%
Patti (2011)*,36 100 70 (3) 60 9 41 29 21 e e e e e 100 121 (24) 0 Abdominal aMAP 60mm Hg Di
100
Vasivej (2016)*,37 55 648 58 (14) 52 e e e e 6 43 27 23 39 59 153 7 e aMAP <65 mm Hg Di
(42/168) (78e244)
Thakar (2007)*,38 504 43 (10) 83 e e e e e 57 26 e 4 100 e 0 Abdominal aMAP <60 mm Hg Di
100
,39
Barone (2002)* 25 501 74 (11) 45 Mean ASA: 2.9 e 40 19 any: 30 e e 124 (55) 27 e aSBP <100 mm Hg, Di
>10 min
,40
Lima (2003)* 92 44 (14) 48 e e e e e e e e Cr: 88 100 e e LiverTX 100 aMAP <60 mm Hg Di
(35)
Nakamura 72 71 (10) 29 e e e e any: 15 88 7 any: 19 11 e e 15 Aorta 100 aSBP <70 mm Hg Di
(2009)*,41
Davidovic (2017)42 450 66 (7) 12 e e e e e 70 e 26 11 100 e 0% Aorta 100% aSBP <100 mm Hg Di
Sharma (2006)*,43 1800 43 (9) 80 e e e e e 57 22 e 2 100 223 (63) 0% Abdominal aSBP <100 mm Hg, Di
100% >5 min
 Hypotension and postoperative adverse outcomes - 713

Table 3 Summary of reported and extrapolated strength of associations of mortality and organ injury in noncardiac patients. Grey cells
represent statistically not significant results. Bold cells represent statistically significant results. Italic cells represent extrapolated
results. *Adherence to dichotomous definition instead of an analysis of depth/duration of a certain threshold or continuous variable.
y
Relative threshold. zBased on combination relative and absolute threshold. ¶Based on both relative and absolute threshold(s) ana-
lysed. xException of duration or time definition; Bijker (2009), hazard risk 1 yr mortality translated to 30 day mortality. jjException of
hypotension definition or analysis; White (2016), odds ratio per mm Hg mean blood pressure decrease or per 5 mm Hg systolic blood
pressure increase; Sessler (2012), single low (low MAP/high bispectral index/high mean alveolar concentration); Sabate
(2011), com-
posite endpoint; Stapelfeldt (2017), odds ratio for percentage increase in the odds of the outcome per limit exceeded; Schmid (2016);
regression coefficient for the achievement rate time spend with mean blood pressure >70 mm Hg compared with total surgery time;
Monk (2015), odds ratio for systolic blood pressures 89e80 mm Hg, 79e70 mm Hg or 69e60 mm Hg for 2e4.9 min or >5 min, respec-
tively. AK, Acute Kidney Injury Network definition (AKIN); Clin, diagnosis based on clinical signs and symptoms; Cr, creatinine con-
centration; CK-MB, creatinine-kinase-Mb concentration; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, Fourth
Edition (DSM-IV); HR, hazard ratio; KD, Kidney Disease Improving Global Outcomes (KDIGO); NS, not significant; OR, odds ratio; RR,
relative risk; RF, Risk, Injury, Failure, Loss and End-stage kidney disease classification (RIFLE); TnT, troponin T concentration; TOAST,
Trial of ORG 10172 in Acute Stroke Treatment (TOAST).1
 714 - Wesselink et al.

Table 3 Continued
 Hypotension and postoperative adverse outcomes - 715

Table 3 Continued
 716 - Wesselink et al.

Table 3 Continued

on screening of title and abstract (n¼3128), 131 abstracts on goal-directed therapy during major abdominal surgery.10
adhered to the inclusion and exclusion criteria and the The other RCT investigated the efficacy of N-acetylcysteine
corresponding articles were retrieved (Fig. 1). After assess- in the prevention of acute kidney injury during elective
ment of the full publications, 89 articles were excluded with abdominal aorta repair.24 In these two trials, both intervention
the third reviewer adjudicating 10 of them. Eventually, 42 arms were analysed for the association between intra-
papers published between 2002 and 2017 were included for operative hypotension and AKI.
data extraction and quality assessment (Table 1). Eighteen
studies (43%) had a quality score
80%, whereas the median
quality score of the articles was 73% (interquartile range Patient characteristics
49e80%; Table 1). The number of included patients in the 42 included studies
Two observational substudies of randomised controlled ranged from 40 to 152 445 (Table 2, panel ‘studies’),21,31 with a
trials (RCTs) were included in which postoperative effects of median of 1523 patients (inter-quartile range 261e17 739). In
intraoperative hypotension were analysed. One RCT focused four studies (10%), the reported mean or median age was
<50 yr and in seven studies the reported mean or median age
was >70 yr (Table 2). Information about sex was reported in 40
studies. In 26 studies, the number of included females and
males was comparable (40e60%), while in the remainder
either males or females were overrepresented. Twenty-three
studies (55%) provided information on the ASA physical sta-
tus classification of the included patients (Table 2, panel ‘co-
morbidity’). In 36 studies (86%) any baseline information was
reported on the occurrence of stroke, hypertension, diabetes,
coronary artery disease or renal disease. In 15 (42%) of these 36
studies, information on at least four of these five conditions
was reported. Twenty-six studies (72%) provided information
on preoperative hypertension. In 22 of these 26 studies (85%),
hypertension was found in
40% of the included patients. In
22 studies (52%) all patients underwent general anaesthesia, in
one study <50% of the patients underwent general anaesthesia
and 12 studies (24%) did not report any information on type of
anaesthesia nor was it obvious from the included surgical
procedures that it always had to be general anaesthesia. In
several studies (n¼12; 29%), large groups of patients under-
went abdominal surgery, including liver transplantation (n¼3
studies; 7%). Other frequent types of surgery were orthopaedic
(n¼7 studies; 17%) and vascular surgery (n¼7 studies; 17%). In
11 studies (26%), no information was reported on type of sur-
gery (Table 2, panel ‘procedure characteristics’).

Intraoperative hypotension definitions

Fig 1. Flow chart of search strategy and article selection of studies on Types of blood pressure thresholds
intraoperative hypotension and postoperative adverse outcomes. In most studies, one or more hypotension definitions included
a threshold based on absolute blood pressures (Table 2, panel
 Hypotension and postoperative adverse outcomes - 717

Table 4 Summary of highest strength of associations of association of mortality and organ injury in noncardiac patients translated to
risk categories. *Not statistically significant; yHirsch (2015) performed a multivariable logistic regression model to analyse their data
but did not report odds ratios, only P-values (P¼0.409 for duration of mean blood pressure <50 mm Hg). HR, hazard ratio; MAP, mean
blood pressure; OR, odds ratio; RR, relative risk.2

‘intraoperative hypotension’). Of the 42 studies, 29 (69%) used associations for a minimum hypotension duration
1 min,
an intraoperative hypotension definition based on an absolute eight studies
5 min, 12 studies
10 min, and seven studies
MAP threshold and 17 studies (40%) used hypotension defini-
20 min.
tion based on an absolute SBP threshold. Seventeen (40%)
studies used a hypotension definition based on a relative blood
pressure threshold (a percentage-wise or absolute decrease Blood pressure threshold values
from baseline blood pressure). In nine studies, relative MAP After ranking the included studies according to blood pressure
thresholds were used and in seven studies relative SBP threshold, quality score, and studied outcome, strengths of
thresholds were used. One study did not report whether their associations per threshold were compared (Table 3, panel
relative threshold was based on a mean or systolic arterial ‘Intraoperative hypotension thresholds’). In addition, results
pressure.33 based on a blood pressure threshold including duration were
extrapolated to longer durations of hypotension. For each re-
ported MAP threshold between 50 mm Hg and 75 mm Hg
Intraoperative hypotension duration (5 mm Hg increments), from seven to twelve studies with MAP
Dichotomous analyses were performed in 20 studies (48%), of based thresholds were available. Six studies reported on MAP
which seven (17%) included minimum time duration in their thresholds 40 mm Hg, 45 mm Hg and 80 mm Hg. There
hypotension definition (Table 2, panel ‘intraoperative hypo- was no apparent relation between blood pressure threshold
tension’). Fifteen studies (36%) performed a comparative values and either quality score, intraoperative hypotension
analysis on whether the duration of hypotension was associ- duration, or studied outcome.
ated with any of the studied outcomes. Two studies (5%)
analysed time as the duration in minutes below a blood
Studied outcomes
pressure threshold. Two studies (5%) included an area under
the threshold, and three studies (7%) used a different type of Fourteen studies investigated mortality, with a follow-up
time-dependent analysis such as time-weighted average or duration between 1 day and 1 yr2,18,28 and an outcome inci-
percentage of the total procedure time. Four studies (10%) dence between 0.03% (follow-up: <1 day) and 5.6% (during
applied a different type of hypotension definition (e.g. lowest hospital admission)28,41 (Table 3, panel ‘Outcomes under
blood pressure, triple low conditions, or blood pressure as part study’). Twelve studies reported on associations between
of the Surgical Apgar Score). Fourteen studies reported intraoperative hypotension and AKI. Follow-up duration
 718 - Wesselink et al.

varied between 1 day and 30 days,21,40 and incidence of AKI studies and their observed associations. First, the included
between 2.8% (7 days) and 72% (7 days).20,29 Nine studies studies differed substantially in their selection of patient
investigated myocardial injury or infarction, with a follow-up groups or procedures. Table 2 demonstrates that few studies
duration between 1 and 30 days.11,14,17 The incidence of are comparable in terms of baseline characteristics of the
myocardial injury varied between 0.09% (in-hospital) and 30% patients included. Further, the selection of surgical procedures
(1 day).11,39 Stroke was reported in four studies, with in- ranged from very wide (e.g. noncardiac surgery) to very narrow
cidences varying from 0.004% (in-hospital) to 0.09% (10 (e.g. thoracic aortic aneurysm repair, gastric bypass surgery).
days).8,27 Five studies reported on delirium with incidences Finally, there was also large variability in what patient and
between 9% (2 days) and 33% (5 days).5,23 Another five studies procedure characteristics weredor were notdreported by the
reported on LOS, either prolonged LOS (three studies, in- various studies.
cidences between 29.7% and 37%) or duration of LOS (two Second, there was large variation in the way that intra-
studies, median LOS between 4 and 7 days). operative hypotension was defined and analysed. Definitions
of hypotension across the studies included a wide range of
degrees and durations for various types of blood pressure.
Summary of evidence for the most reported outcomes
Different thresholds were used for systolic, mean, or diastolic
Based on the methods described above, two studies with a blood pressure, or even multiple thresholds were combined
high-quality score were not used for the determination of or- into a single definition. In addition to a threshold definition,
gan injury risks. The study of Roshanov and colleagues11 was the variable for hypotension can also be modelled in different
excluded because studying hypotension was not part of the ways.46 The thresholds often introduce a cut-off: anything
primary or secondary research objectives. The strengths of above the threshold is considered to be the same (i.e. analysed
associations reported by Schmid and colleagues10 could not be as ‘zero’ or ‘no intraoperative hypotension’), even when the
converted into a comparable OR, RR, or HR as they reported a values are close to the threshold. However, everything below
regression coefficient for the decrease of creatinine clearance the threshold can be modelled in several ways: duration of
(e0.28 ml min1) per percent of total surgery time with MAP blood pressure below the threshold, the area under the

70 mm Hg. threshold, or simply a ‘one’ (i.e. ‘yes, the patient’s blood
The reported risks of any end-organ injury after noncardiac pressure was below the threshold’).
surgery started to increase with prolonged exposure (
10 min) Third, there was important variation in the way that post-
to MAPs <80 mm Hg, resulting in a mildly elevated risk, with operative adverse outcomes were defined, analysed, and re-
OR/RR/HRs between 1.0 and 1.4 (Table 4). For shorter durations ported. Six different groups of postoperative adverse
(<10 min), mildly elevated risks have been reported for outcomes were reported as outcomes in this review: mortality,
thresholds of 70 mm Hg and lower. The reported risks AKI, MI, ischaemic stroke, delirium, and LOS. Within each
increased to moderate (OR/RR/HRs between 1.4 and 2.0) with group different adverse outcomes with different definitions
exposures to MAPs <65e60 mm Hg for
5 min, or any exposure were studied. For example, the definition of myocardial injury
<55e50 mm Hg. High risks (OR/RR/HRs >2.0) were reported for ranged from only elevated biomarkers with or without ECG
MAPs <65 mm Hg for
20 min, MAPs <50 mm Hg for
5 min, or changes to cardiovascular complications. Residual confound-
any exposure <40 mm Hg. ing might have been present in studies that analysed post-
There were small differences between individual out- operative cardiac or renal biomarkers drawn by clinical
comes. For AKI and MI, the risks started at thresholds <65 mm indication compared with routine postoperative biomarker
Hg and increased gradually with degree and duration in a measurements.4,12,15 Furthermore, not all outcomes are
pattern that is largely similar between the two outcomes. For interchangeable in their severity and incidence rates. Delirium
mortality, associations were reported for higher thresholds might be an outcome more sensitive to find the low blood
than AKI and MI, starting when there was prolonged exposure pressure threshold, but regarding severity and incidence rates
to MAP <80 mm Hg. The reported risks were mild for thresh- it is not on par with mortality and AKI.
olds down to 55 mm Hg, at which the reported risks increased The fourth issue is a result of the three issues mentioned
with prolonged exposures to MAP 55 mm Hg. For ischaemic before. The large heterogeneity in baseline characteristics,
stroke, only non-significant, small strengths of associations hypotension definitions, and studied outcomes made it
were reported. For delirium, non-significant associations were challenging to come to a quantitative summary of the results.
found for a duration of MAP <50 mm Hg. For LOS, insufficient Hence, we made various conversions and assumptions on
data were available. how to merge definitions of intraoperative hypotension and
reported strengths of associations of these studies in a
qualitative way. Additionally, we only used high quality
Discussion studies (quality score
80%) with blood pressure thresholds
This systematic review summarised the current literature as converted to MAP thresholds for the organ-injury risk
of March 2017 on the relation between intraoperative hypo- classification.
tension and postoperative outcomes. It provides an overview Based on several assumptions and variations in patients,
of blood pressures that were reported to be associated with intraoperative hypotension definitions, outcome definitions,
inadequate organ perfusion. Prolonged exposure (
10 min) to and analyses, it is still difficult to reliably define a common
MAP <80 mm Hg and for shorter durations <70 mm Hg was ‘cut-off’ for which blood pressure is too low. Although the risk
associated with mildly elevated risks of any end-organ injury. of end-organ injury seems to increase rapidly with prolonged
Increased durations for MAP <65e60 mm Hg or for any expo- exposure to lower intraoperative blood pressures, based on
sure <55e50 mm Hg were associated with moderately or current evidence, we cannot prove a causal relation between
highly elevated risks. intraoperative blood pressures and outcomes. Current studies
The interpretation and clinical applicability of the results of on intraoperative hypotension aim to answer: ‘What blood
this review are hampered by the large differences between the pressure is too low?’, but their data can only be used to answer
 Hypotension and postoperative adverse outcomes - 719

‘What blood pressure is associated with adverse outcomes Authors’ contributions

given current treatment standards?’ In other words, this re-
Study concept and design: all authors.
view does not address the question on what blood pressure
Acquisition of data: E.M.W., T.H.K., H.M.T.
thresholds result in organ hypoperfusion, but whether there is
Interpretation of data: E.M.W., T.H.K., H.M.T., W.v.K.
remaining hypoperfusion despite present routines to manage
Drafting the manuscript: E.M.W., T.H.K., H.M.T.
patients’ blood pressures.
Critical revision of the manuscript for important intellectual
All contributing factors and interactions are difficult to un-
content: all authors.
ravel and discriminate and it seems unlikely that we will be able
Final approval of the version to be published: all authors.
to explore the contribution of separate factors using only
observational data. Therefore, intervention studiesdsuch as
pragmatic trialsdare required to understand the causal chain
Declaration of interest
of intraoperative low blood pressure and adverse outcomes.
After completion of the systematic search described in this None declared.
review, three trials concerning intraoperative blood pressure
manipulation have been published. In the first trial,27 elderly
patients with chronic hypertension who underwent major Funding
abdominal surgery were randomised to one of three target MAP Departmental.
groups. Vasopressor therapy and a fluid management protocol
based on stroke volume variation were used to adjust MAP. This
study showed that a target MAP of 80e95 mm Hg, compared Appendix A. Supplementary data
with lower (65e79 mm Hg) and higher targets (96e110 mm Hg)
Supplementary data related to this article can be found at
can decrease the incidence of AKI. Incidences of stroke and
https://doi.org/10.1016/j.bja.2018.04.036.
mortality did not significantly differ among groups. The lower
incidence of AKI in the midrange MAP group compared with the
lower MAP group is in accordance with results from observa-
References
tional studies (Table 4). Strict inclusion and exclusion criteria
regarding age, comorbidities, and preoperative medication use 1. Bijker JB, van Klei WA, Kappen TH, van Wolfswinkel L,
limit generalisability of this trial.47 In a second trial,48 two blood Moons KGM, Kalkman CJ. Incidence of intraoperative hy-
pressure management strategies and their effects on post- potension as a function of the chosen definition: literature
operative organ dysfunction in patients undergoing major definitions applied to a retrospective cohort using auto-
surgery were studied. This study showed that achieving a sys- mated data collection. Anesthesiology 2007; 107: 213e20
tolic blood pressure within 10% of the reference value by using 2. Monk TG, Saini V, Weldon BC, Sigl JC. Anesthetic man-
continuous vasopressor infusion may prevent postoperative agement and one-year mortality after noncardiac surgery.
organ dysfunction compared with a strategy of only treating Anesth Analg 2005; 100: 4e10
systolic blood pressures <80 mm Hg or <40% of the reference 3. Sun LY, Wijeysundera DN, Tait GA, Beattie WS. Associa-
value (standard care). However, anticipation of an expected tion of intraoperative hypotension with acute kidney
blood pressure decline was not allowed and the standard care injury after elective noncardiac surgery. Anesthesiology
group treatment thresholds might not really represent current 2015; 123: 515e23
clinical care. In a third trial,49 it was shown that avoidance of 4. Walsh M, Devereaux PJ, Garg AX, et al. Relationship be-
‘double low’ events defined as MAP<75 mm Hg and bispectral tween intraoperative mean arterial pressure and clinical
index <45 by automated alerts, did not significantly decrease outcomes after noncardiac surgery: toward an empirical
the 90-day mortality incidence in adults who underwent definition of hypotension. Anesthesiology 2013; 119: 507e15
noncardiac surgery. In this study, no standardised blood pres- 5. Hirsch J, DePalma G, Tsai TT, Sands LP, Leung JM. Impact
sure treatment protocols were used. Future studies on intra- of intraoperative hypotension and blood pressure fluctu-
operative hypotension should aim to explore blood pressure ations on early postoperative delirium after non-cardiac
thresholds within specific patient groups and for specific out- surgery. Br J Anaesth 2015; 115: 418e26
comes. These should include other variables that are indicative 6. Monk TG, Bronsert MR, Henderson WG, et al. Association
of underlying causes and mechanisms of hypotension, such as between intraoperative hypotension and hypertension
heart rate, pulse pressure variation, cardiac output estimated and 30-day postoperative mortality in noncardiac surgery.
by advanced techniques, and specific biomarkers. This will Anesthesiology 2015; 123: 307e19
allow us to study mechanistic hypotheses that are outcome- 7. Willingham MD, Karren E, Shanks AM, et al. Concurrence
specific and whether mechanism-specific interventions will of intraoperative hypotension, low minimum alveolar
improve outcomes. concentration, and low bispectral index is associated with
In conclusion, the reported associations suggest that organ postoperative death. Anesthesiology 2015; 123: 775e85
injury might occur when the mean arterial pressure decreases 8. Bijker JB, Persoon S, Peelen L, et al. Intraoperative hypo-
<80 mm Hg for
10 min, and that this risk increases with blood tension and perioperative ischemic stroke after general
pressures becoming progressively lower. Given the retro- surgery. Anesthesiology 2012; 116: 658e64
spective observational design of most studies, reflected by 9. Mizota T, Hamada M, Segawa H. Relationship between
large variability in patient characteristics, hypotension defi- intraoperative hypotension and acute kidney injury after
nitions, and outcomes, solid conclusions on which blood living donor liver transplantation: a retrospective anal-
pressures under which circumstances are truly too low cannot ysis. J Cardiothorac Vasc Anesth 2017; 31: 582e9
be drawn. We are in need of prospective interventional studies 10. Schmid S, Kapfer B, Heim M, et al. Algorithm-guided goal-
in specific patient groups and for specific outcomes to further directed haemodynamic therapy does not improve renal
unravel this topic. function after major abdominal surgery compared to good
 720 - Wesselink et al.

standard clinical care: a prospective randomised trial. Crit 27. Sabate
S, Mases A, Guilera N, et al. Incidence and pre-
Care 2016; 20: 1e11 dictors of major perioperative adverse cardiac and cere-
11. Roshanov PS, Rochwerg B, Patel A, et al. Enzyme inhibitors brovascular events in non-cardiac surgery. Br J Anaesth
or angiotensin II receptor blockers. Anesthesiology 2017; 2011; 107: 879e90
126: 16e27 28. Taffe
P, Sicard N, Pittet V, Pichard S, Burnand B. The
12. Salmasi V, Maheshwari K, Dongsheng Y, et al. Thresholds, occurrence of intra-operative hypotension varies between
and acute kidney and myocardial injury after noncardiac hospitals: observational analysis of more than 147,000
surgery a retrospective cohort analysis. Anesthesiology anaesthesia. Acta Anaesthesiol Scand 2009; 53: 995e1005
2017; 126: 47e65 29. Sirivatanauksorn Y, Parakonthun T, Premasathian N,
13. Babazade R, Yilmaz HO, Zimmerman NM, et al. Associa- et al. Renal dysfunction after orthotopic liver trans-
tion between intraoperative low blood pressure and plantation. Transplant Proc 2014; 46: 818e21
development of surgical site infection after colorectal 30. Tassoudis V, Vretzakis G, Petsiti A, et al. Impact of intra-
surgery: a retrospective cohort study. Ann Surg 2016; 6: operative hypotension on hospital stay in major abdom-
1058e64 inal surgery. J Anesth 2011; 25: 492e9
14. Hallqvist L, Martensson J, Granath F, Sahlen A, Bell M. 31. Stapelfeldt WH, Yuan H, Dryden JK, et al. The SLUScore: a
Intraoperative hypotension is associated with myocardial novel method for detecting hazardous hypotension in
damage in noncardiac surgery: an observational study. adult patients undergoing noncardiac surgical proced-
Eur J Anaesthesiol 2016; 33: 450e6 ures. Anesth Analg 2017; 124: 1135e52
15. van Waes JAR, van Klei WA, Wijeysundera DN, Van 32. Jiang X, Chen D, Lou Y, Li Z. Risk factors for postoperative
Wolfswinkel L, Lindsay TF, Beattie WS. Association be- delirium after spine surgery in middle- and old-aged pa-
tween intraoperative hypotension and myocardial injury tients. Aging Clin Exp Res 2017; 29: 1039e44
after vascular surgery. Anesthesiology 2016; 124: 35e44 33. Yang L, Sun DF, Han J, Liu R, Wang LJ, Zhang ZZ. Effects of
16. Mascha EJ, Yang D, Weiss S, Sessler DI. Intraoperative intraoperative hemodynamics on incidence of post-
mean arterial pressure variability and 30-day mortality in operative delirium in elderly patients: a retrospective
patients having noncardiac surgery. Anesthesiology 2015; study. Med Sci Monit 2016; 22: 1093e100
123: 79e91 34. Yue J, Luo Z, Guo D, et al. Evaluation of acute kidney
17. Pipanmekaporn T, Punjasawadwong Y, Charuluxananan S, injury as defined by the risk, injury, failure, loss, and end-
et al. Incidence of and risk factors for cardiovascular stage criteria in critically ill patients undergoing abdom-
complications after thoracic surgery for noncancerous le- inal aortic aneurysm repair. Chin Med J (Engl) 2013; 126:
sions. J Cardiothorac Vasc Anesth 2014; 28: 960e5 431e6
18. Bijker JB, van Klei WA, Vergouwe Y, et al. Intraoperative 35. Franck M, Radtke FM, Prahs C, et al. Documented intra-
hypotension and 1-year mortality after noncardiac sur- operative hypotension according to the three most com-
gery. Anesthesiology 2009; 111: 1217e26 mon definitions does not match the application of
19. Kheterpal S, O’Reilly M, Englesbe MJ, et al. Preoperative antihypotensive medication. J Int Med Res 2011; 39:
and intraoperative predictors of cardiac adverse events 846e56
after general, vascular, and urological surgery. Anesthesi- 36. Patti R, Saitta M, Cusumano G, Termine G, Di Vita G. Risk
ology 2009; 110: 58e66 factors for postoperative delirium after colorectal surgery
20. White SM, Moppett IK, Griffiths R, et al. Secondary anal- for carcinoma. Eur J Oncol Nurs 2011; 15: 519e23
ysis of outcomes after 11,085 hip fracture operations from 37. Vasivej T, Sathirapanya P, Kongkamol C. Incidence and
the prospective UK Anaesthesia Sprint Audit of Practice risk factors of perioperative stroke in noncardiac, and
(ASAP-2). Anaesthesia 2016; 71: 506e14 nonaortic and its major branches surgery. J Stroke Cere-
21. Brinkman R, HayGlass KT, Mutch WAC, Funk DJ. Acute brovasc Dis 2016; 25: 1172e6
kidney injury in patients undergoing open abdominal 38. Thakar CV, Kharat V, Blanck S, Leonard AC. Acute kidney
aortic aneurysm repair: a pilot observational trial. injury after gastric bypass surgery. Clin J Am Soc Nephrol
J Cardiothorac Vasc Anesth 2015; 29: 1212e9 2007; 2: 426e30
22. Petsiti A, Tassoudis V, Vretzakis G, et al. Depth of anes- 39. Barone JE, Bull MB, Cussatti EH, Miller KD, Tucker JB.
thesia as a risk factor for perioperative morbidity. Anes- Perioperative myocardial infarction in low-risk patients
thesiol Res Pract 2015; 2015: 829151 undergoing noncardiac surgery is associated with Intra-
23. Marcantonio ER, Goldman L, Orav EJ, Cook EF, Thomas HL, operative hypotension. J Intensive Care Med 2002; 17:
Lee THE. The association of intraoperative factors with 250e5
the development of postoperative delirium. Am J Med 40. Lima EQ, Zanetta DMT, Castro I, et al. Risk factors for
1998; 105: 380e4 development of acute renal failure after liver trans-
24. Tallgren M, Niemi T, Po € yhia
€ R, et al. Acute renal injury and plantation. Ren Fail 2003; 25: 553e60
dysfunction following elective abdominal aortic surgery. 41. Nakamura K, Matsuyama M, Yano M, et al. Open surgery
Eur J Vasc Endovasc Surg 2007; 33: 550e5 or stent repair for descending aortic diseases: results and
25. House LM, Marolen KN, St Jacques PJ, McEvoy MD, risk factor analysis. Scand Cardiovasc J 2009; 43: 201e7
Ehrenfeld JM. Surgical Apgar score is associated with 42. Davidovic LB, Maksic M, Koncar I, Ilic N. Open repair of
myocardial injury after noncardiac surgery. J Clin Anesth AAA in a high volume center. World J Surg 2017; 41:
2016; 34: 395e402 884e91
26. Sessler DI, Sigl JC, Kelley SD, et al. Hospital stay and 43. Sharma SK, McCauley R, Cottam D, et al. Acute changes in
mortality are increased in patients having a triple low of renal function after laparoscopic gastric surgery for
low blood pressure, low bispectral index, and low mini- morbid obesity. Surg Obes Relat Dis 2006; 2: 389e92
mum alveolar concentration of volatile anesthesia. Anes- 44. von Elm E, Altman DG, Egger M, et al. Strengthening the
thesiology 2012; 116: 1195e203 Reporting of Observational Studies in Epidemiology
 Hypotension and postoperative adverse outcomes - 721

(STROBE) statement: guidelines for reporting observa- gastrointestinal surgery in elderly hypertensive patients:
tional studies. BMJ 2007; 335: 806e8 a randomized study. J Clin Anesth 2017; 43: 77e83
45. Schulz KF, Altman DG, Moher D. Group C. CONSORT 2010 48. Futier E, Lefrant J-Y, Guinot P-G, et al. Effect of individu-
statement : updated guidelines for reporting parallel alized vs standard blood pressure management strategies
group randomised trials. BMC Med 2010; 8: 18 on postoperative organ dysfunction among high-risk pa-
46. Vernooij LM, van Klei WA, Machina M, Pasma W, tients undergoing major surgery a randomized clinical
Beattie WS, Peelen LM. Different methods of modelling trial. JAMA 2017; 318: 1346e57
intraoperative hypotension and their association with 49. McCormick PJ, Levin MA, Lin H, Sessler DI, Reich DL.
postoperative complications in patients undergoing non- Effectiveness of an electronic alert for hypotension and
cardiac surgery. Br J Anaesth 2018; 120: 1080e9 low bispectral index on 90-day postoperative mortality.
47. Wu X, Jiang Z, Ying J, Han Y, Chen Z. Optimal blood Anesthesiology 2017; 126: 1113e20
pressure decreases acute kidney injury after

Handling editor: H.C. Hemmings Jr