World J Surg

https://doi.org/10.1007/s00268-019-05357-5

SCIENTIFIC REVIEW

Meta-analysis of Enhanced Recovery After Surgery (ERAS)

Protocols in Emergency Abdominal Surgery
Shahab Hajibandeh1 • Shahin Hajibandeh2 • Victor Bill3 • Thomas Satyadas4

 Société Internationale de Chirurgie 2020

Abstract
Objectives To evaluate enhanced recovery after surgery (ERAS) protocols in emergency abdominal surgery.
Methods The electronic data sources were explored to capture all studies that evaluated the impact of ERAS
protocols in patients who underwent emergency abdominal surgery. The quality of randomised and non-randomised
studies was evaluated by the Cochrane tool and the Newcastle–Ottawa scale, respectively. Random or fixed effects
modelling were utilised as indicated.
Results Six comparative studies, enrolling 1334 patients, were eligible. ERAS protocols resulted in shorter post-
operative time to first flatus (mean difference: -1.40, P \ 0.00001), time to first defecation (mean difference: -1.21,
P = 0.02), time to first oral liquid diet (mean difference: -2.30, P \ 0.00001), time to first oral solid diet (mean
difference: -2.40, P \ 0.00001) and length of hospital stay (mean difference: -3.09, -2.80, P \ 0.00001). ERAS
protocols also resulted in lower risks of total complications (odds ratio: 0.50, P \ 0.00001), major complications
(odds ratio: 0.60, P = 0.0008), pulmonary complications (odds ratio: 0.38, P = 0.0003), paralytic ileus (odds ratio:
0.53, 0.88, P = 0.01) and surgical site infection (odds ratio: 0.39, P = 0.0001). Both ERAS and non-ERAS protocols
resulted in similar risk of 30-day mortality (risk difference: -0.00, P = 0.94), need for re-admission (risk difference:
-0.01, P = 0.50) and need for re-operation (odds ratio: 0.83, P = 0.50).
Conclusions Although ERAS protocols are commonly used in elective settings, they are associated with favourable
outcomes in emergency settings as indicated by reduced post-operative complications, accelerated recovery of bowel
function and shorter post-operative hospital stay without increasing need for re-admission or re-operation. There
should be an effort to incorporate ERAS protocols into emergency abdominal surgery settings.

Introduction

Shahab Hajibandeh and Shahin Hajibandeh had equal contributions in

Enhanced recovery after surgery (ERAS) protocols via a
this study proposing joined first authorship. multidisciplinary team approach aim to decrease periop-
erative surgical stress, sustain post-operative physiological
& Shahab Hajibandeh function and facilitate recovery in surgical patients [1].
shahab_hajibandeh@yahoo.com
ERAS protocols comprise preoperative (consultations
1
Glan Clwyd Hospital, Rhyl, Denbighshire, UK before procedure, use of carbohydrate drinks before the
2 procedure, preparation of bowel when indicated), intraop-
Sandwell and West, Birmingham Hospitals NHS Trust,
Birmingham, UK erative (prevention of venous thromboembolism, prophy-
3 lactic use of antibiotics, maintaining temperature during
North Manchester General Hospital, Manchester, UK
4
procedure, objective fluid replacement, use of the best
Manchester Royal Infirmary Hospital, Manchester, UK

123
 World J Surg

possible surgical approach and incision) and post-operative • Major complications

(appropriate pain control following procedure, early • Pulmonary complications
removal of tubes including nasogastric tube, urinary • Paralytic ileus
catheter and drains, early mobilisation and commencement • Surgical site infection
of enteral feeding) components [2]. • 30-Day mortality
ERAS protocols were originally applied in colorectal • Need for re-admission
surgical setting and were found to be successful in • Need for re-operation.
decreasing duration of hospital stay and improving post-
operative outcomes [3]. ERAS approach is regarded as the
Search methods
standard of care in various elective surgical settings [4–9].
Although the significance of ERAS protocols in terms of
Two separate authors explored the following sources:
post-operative outcomes has been investigated extensively
in elective settings, the effectiveness of these protocols in • CINAHL
emergency abdominal surgical settings is not established. • EMBASE
Hence, our main purpose was to complete a systematic • MEDLINE
review and meta-analysis of outcomes to evaluate the • The Cochrane Central Register of Controlled Trials
impact of ERAS protocols in emergency abdominal sur- • ClinicalTrials.gov
gical settings. • WHO International Clinical Trials Registry
• SRCTN Register.
The date for the last literature search was 15 April 2019.
Methods
Moreover, in order to identify more eligible studies, the
references mentioned within full-text of relevant papers
This study was designed according to an agreed protocol
were searched. There were no language limitations in the
which was in compliance with PRISMA statement stan-
search strategies. ‘‘Appendix’’ demonstrates the main lit-
dards [10].
erature search strategy which was modified according to
each database.
Eligibility criteria
Study selection and data extraction
All comparative studies (randomised or non-randomised)
comparing ERAS protocols with standard protocols in
The title and abstract of the articles found after running the
patients undergoing emergency abdominal surgery were
search strategy in the aforementioned databases were
considered qualified for inclusion. All adult patients who
reviewed. This was followed by obtaining the full texts of
underwent emergency abdominal procedures (emergency
the identified studies and selecting the appropriate studies
resection of small bowel, segmental colectomy, total
meeting the eligibility criteria. A data collection proforma
colectomy, operation for perforated viscus, adhesiolysis or
was then designed and tested in randomly chosen studies.
laparotomy) were regarded as the participants of interest.
This was in compliance with Cochrane’s guidelines. As
ERAS protocols, comprising components recommended by
outlined in the tables and figures, the data collection pro-
ERAS group [11], were regarded as intervention of interest,
forma contained information on:
and non-ERAS (standard or conventional) perioperative
protocols were regarded as comparison of interest. • Bibliometric parameters of the eligible studies
• Characteristics of the eligible patients at baseline
Primary outcomes • Outcome measures.
The study selection and data extraction steps were done
• Length of hospital stay
by two separate reviewers. Any disagreements during
• Total post-operative complications.
selection of the included studies or data extraction process
were discussed between the first two authors, and a sepa-
Secondary outcomes rate author was involved when required.

• Post-operative time to first flatus Risk of bias assessment

• Post-operative time to first defecation
• Post-operative time to first oral liquid diet The methodological quality of RCTs was evaluated using
• Post-operative time to first oral solid diet the Cochrane tool that examines the quality of the study

123
World J Surg

design in terms of selection, performance, detection, attri- Results

tion, reporting and other sources of bias. The method-
ological quality of non-randomised comparative studies Results of the search
was evaluated using the Newcastle–Ottawa scale [12] that
examines the quality of the study design in terms of Overall, 300 articles were identified after applying the
selection of the eligible patients, comparability of the search strategy in the aforementioned databases. Among
intervention and comparison groups within the study and these, 6 studies [13–18] were qualified for inclusion
outcome assessments. This process was done by two sep- (Fig. 1). These were 2 RCTs [14, 17] and 4 observational
arate reviewers. Any disagreements were discussed studies [13, 15, 16, 18], including 1334 patients. Among
between the first two authors, and a separate author was the eligible patients, 690 patients were treated according to
involved when required. the ERAS protocols and 644 patients were treated
according to standard non-ERAS protocols. Both ERAS
Data analysis and non-ERAS groups were similar in terms of mean age
(57.9 versus 57.4, P = 0.95) and gender (male: 59% versus
In terms of summary measures, odds ratio (OR) was 62%, P = 0.25; female: 41% versus 38%, P = 0.25). Three
computed for dichotomous outcomes and mean difference studies [13, 15, 18] included patients with bowel
(MD) for continuous outcomes; risk difference (RD) was
computed in the case of no occurrence of the outcome of
interest (zero event) in either groups in more than a third of
the studies. The unit of analysis was an individual patient,
and the analyses were according to intention to treat
information. The Review Manager (RevMan, version 5.3.
Copenhagen, 2014) software was used for data synthesis.
Random or fixed effects modelling was utilised as indi-
cated; when between-study heterogeneity was conspicuous,
random effects models were applied. The heterogeneity
was quantified and reported as I2 using Cochran Q test (v2).
We interpreted I2 as below:
• 0–50% means low heterogeneity
• 50–75% means moderate heterogeneity
• 75–100% means high heterogeneity.
For the outcomes reported by at least 10 studies, we
aimed to generate funnel plots and planned to examine the
possibility of publication bias by assessing the symmetry of
the funnel plots.

Additional analyses

The robustness of the analyses was examined by per-

forming sensitivity analyses for the outcomes reported by a
minimum number of four comparative studies. The first set
of analyses included separate calculation of risk ratio, OR
and RD for dichotomous outcomes. The second set of
analyses aimed to examine the impact of each study on the
total effect size and heterogeneity. This involved elimina-
tion of one study at a time followed by repeating the
analyses. The third set of analyses involved independent
analyses using random effects and fixed effect models.
Furthermore, where applicable, we did subgroup analysis
according to design of the studies (RCTs and observational
studies) and reason for emergency operation (visceral
Fig. 1 Study flow diagram
perforation and bowel obstruction).

123
 World J Surg

obstruction; two studies [14, 17] included patients with P \ 0.00001). The reported between-study heterogeneity
perforated duodenal ulcer; one study [16] included patients was judged to be moderate (I2 = 69%, P = 0.01).
with abdominal pathology requiring major abdominal
procedure (Tables 1, 2). Table 3 highlights the ERAS Time to first flatus
components considered as part of ERAS protocols in each
of the eligible studies. Analysis of 795 patients from three studies [13, 14, 18]
showed that ERAS protocols resulted in shorter time to first
Risk of bias in included studies flatus compared with non-ERAS protocols (mean differ-
ence: -1.40, 95% CI -1.52, -1.27, P \ 0.00001). The
Figure 2 highlights the outcomes of methodological quality reported between-study heterogeneity was judged to be low
assessment based on the Cochrane tool and Newcastle– (I2 = 0%, P = 0.69).
Ottawa scale.
Time to first defecation
Outcome synthesis (Fig. 3)
Analysis of 795 patients from three studies [13, 14, 18]
Length of hospital stay showed that ERAS protocols resulted in shorter time to first
defecation compared with non-ERAS protocols (mean
Analysis of 964 patients from five studies [13–15, 17, 18] difference: -1.21, 95% CI -2.19, - 0.23, P = 0.02). The
showed that ERAS protocols resulted in shorter length of reported between-study heterogeneity was judged to be
hospital stay compared with non-ERAS protocols (mean high (I2 = 94%, P \ 0.00001).
difference: -3.09, 95% CI -3.37, -2.80, P \ 0.00001).
The reported between-study heterogeneity was judged to Time to first oral liquid diet
be low (I2 = 14%, P = 0.32).
Analysis of 735 patients from two studies [13, 14] showed
Total post-operative complications that ERAS protocols resulted in shorter time to first oral
liquid diet compared with non-ERAS protocols (mean
Analysis of 964 patients from five studies [13–15, 17, 18] difference: -2.30, 95% CI -2.37, -2.23, P \ 0.00001).
showed that ERAS protocols resulted in lower risk of total The reported between-study heterogeneity was judged to
post-operative complications compared with non-ERAS be low (I2 = 12%, P = 0.29).
protocols (odds ratio: 0.50, 95% CI 0.38, 0.66,

Table 1 Baseline characteristics of the included studies

References Country Journal Design Included populations Sample size
Total ERAS Non-
ERAS

Shang et al. China Medicine Retrospective Patients with obstructive colorectal cancer undergoing 636 318 318
[13] observational the unplanned emergency operation
Mohsina India J Gastrointest Randomised Patients undergoing emergency operation for 99 50 49
et al. [14] Surg controlled perforated duodenal ulcer
trial
Shida et al. Japan BMC Surgery Retrospective Patients undergoing emergency operation for 122 80 42
[15] observational obstructive colorectal cancer
Wisely et al. Australia ANZ J Surg Retrospective Patients undergoing major emergency abdominal 370 201 169
[16] observational surgery
Gonenc Turkey Am J Surg Randomised Patients undergoing emergency operation for 47 21 26
et al. [17] controlled perforated duodenal ulcer
trial
Lohsiriwat Thailand World J Retrospective Patients undergoing emergency operation for 60 20 40
et al. [18] Gastroenterol observational obstructing colorectal adenocarcinoma
ERAS enhanced recovery after surgery

123
World J Surg

Table 2 Baseline characteristics of the included population

Age Male Female BMI (kg/m2) ASA1–2 ASA3–4
ERAS versus ERAS versus ERAS versus ERAS versus non- ERAS versus ERAS versus
non-ERAS non-ERAS non-ERAS ERAS non-ERAS non-ERAS

Shang et al. Median 192/318 versus 126/318 versus 25.1 ± 2.6 versus 216/318 versus 102/318 versus
[13] 66 versus 65 201/318 117/318 24.9 ± 2.5 220/318 98/318
Mohsina et al. Mean: 44/50 versus 6/50 versus 5/49 24.82 ± 0.35 versus 50/50 versus 0/50 versus 0/49
[14] 46 versus 44 44/49 24.6 ± 0.34 49/49
Shida et al. Mean: 52/80 versus 28/80 versus NR NR NR
[15] 69 versus 68 25/42 17/42
Wisely et al. Median 87/201 versus 114/201 versus NR 71/201 versus 130/201 versus
[16] 68 versus 68 85/169 84/169 74/169 95/169
Gonenc et al. Mean 16/21 versus 5/21 versus 6/26 NR NR NR
[17] 35 versus 38 20/26
Lohsiriwat Mean 14/20 versus 6/20 versus 16/40 21.7 ± 3.3 versus 16/20 versus 4/20 versus 4/20
et al. [18] 58 versus 62 24/40 22.8 ± 3.4 36/40

NR not reported, ERAS enhanced recovery after surgery, BMI body mass index, ASA American Society of Anaesthesiologists

Table 3 Enhanced recovery after surgery components considered in the included studies
ERAS component Shang et al. Mohsina et al. Shida et al. Wisely et al. Gonenc et al. Lohsiriwat et al.
[13] [14] [15] [16] [17] [18]

Preoperative counselling Yes Yes Yes Yes Yes Yes

Preoperative fasting and carbohydrate Yes NR Yes Yes Yes Yes
loading
Mechanical bowel preparation Yes NR NR Yes Yes NR
Thromboprophylaxis Yes NR NR NR Yes NR
Antibiotic prophylaxis NR Yes NR NR Yes NR
Maintenance of intraoperative Yes NR Yes NR Yes Yes
normothermia
Goal-directed fluid therapy Yes NR Yes Yes Yes Yes
Surgical incision and approach Yes Yes NR Yes Yes Yes
Avoidance of nasogastric tubes Yes Yes Yes Yes Yes Yes
Post-operative analgesia Yes Yes Yes Yes Yes Yes
Prevention of post-operative ileus Yes Yes Yes Yes Yes Yes
Avoidance of peritoneal drains Yes Yes NR Yes Yes Yes
Early removal of urinary catheter Yes Yes Yes Yes Yes Yes
Early post-operative enteral nutrition Yes Yes Yes Yes Yes Yes
Early mobilisation patients should Yes Yes Yes Yes Yes Yes
ERAS enhanced recovery after surgery, NR not reported

Time to first oral solid diet Major complications

Analysis of 842 patients from four studies [13, 14, 17, 18] Analysis of 818 patients from three studies [13, 15, 18]
showed that ERAS protocols resulted in shorter time to first showed that ERAS protocols resulted in lower risk of
oral solid diet compared with non-ERAS protocols (mean major post-operative complications compared with non-
difference: -2.40, 95% CI -2.62, -2.19, P \ 0.00001). ERAS protocols (odds ratio: 0.60, 95% CI 0.45, 0.81,
The reported between-study heterogeneity was judged to P = 0.0008). The reported between-study heterogeneity
be moderate (I2 = 69%, P = 0.02). was judged to be low (I2 = 0%, P = 0.70).

123
 World J Surg

Fig. 2 Risk of bias summary and graph showing authors’ judgements about each risk of bias item for: a randomised trials, b observational
studies

Pulmonary complications reported between-study heterogeneity was judged to be low

(I2 = 0%, P = 0.61).
Analysis of 638 patients from four studies [14–17] showed
that ERAS protocols resulted in lower risk of pulmonary 30-Day mortality
complications compared with non-ERAS protocols (odds
ratio: 0.38, 95% CI 0.22, 0.64, P = 0.0003). The reported Analysis of 1334 patients from six studies [13–18] showed
between-study heterogeneity was judged to be low that ERAS protocols and non-ERAS protocols were similar
(I2 = 0%, P = 0.65). in terms of 30-day mortality risk (risk difference: -0.00,
95% CI - 0.02, 0.02, P = 0.94). The reported between-
Paralytic ileus study heterogeneity was judged to be low (I2 = 0%,
P = 0.92).
Analysis of 904 patients from four studies showed that
ERAS protocols resulted in lower risk of paralytic ileus Need for re-admission
compared with non-ERAS protocols (odds ratio: 0.53, 95%
CI 0.32, 0.88, P = 0.01). The reported between-study Analysis of 1334 patients from six studies [13–18] showed
heterogeneity was judged to be low (I2 = 0%, P = 0.97). that ERAS protocols and non-ERAS protocols were similar
in terms of need for re-admission (risk difference: -0.01,
Surgical site infection 95% CI - 0.04, 0.02, P = 0.50). The reported between-
study heterogeneity was judged to be low (I2 = 7%,
Analysis of 516 patients from three studies [14, 16, 17] P = 0.37).
showed that ERAS protocols resulted in lower risk of
surgical site infection compared with non-ERAS protocols
(odds ratio: 0.39, 95% CI 0.24, 0.64, P = 0.0001). The

123
World J Surg

Fig. 3 Forest plots of the comparisons of outcomes between the ERAS and non-ERAS groups: a length of hospital stay; b total post-operative
complications; c time to first flatus; d time to first defecation; e time to first liquid diet; f time to first solid diet; g major complications;
h pulmonary complications; i paralytic ileus; j surgical site infection; k 30-day mortality; l need for re-admission; m need for re-operation

123
 World J Surg

Fig. 3 continued

123
World J Surg

Fig. 3 continued

Need for re-operation unaffected when RRs, ORs and RDs were computed
independently. Eliminating one study at a time in the
Analysis of 1274 patients from five studies [13–17] showed second set of analyses did not affect the results. Applica-
that ERAS protocols and non-ERAS protocols were similar tion of fixed effect and random effects modelling in the
in terms of need for re-operation (odds ratio: 0.83, 95% CI third set of analyses did not affect the results.
0.49, 1.42, P = 0.50). The reported between-study hetero-
geneity was judged to be low (I2 = 0%, P = 0.71). Randomised controlled trials

Additional analyses Analysis of 146 patients from two RCTs [14, 17] showed
that ERAS protocols resulted in shorter post-operative time
Sensitivity analyses to first flatus (mean difference: -1.47, 95% CI -1.83,
-1.11, P \ 0.00001), time to first defecation (mean dif-
The outcomes presented by a minimum number of four ference: -2.26, 95% CI -2.68, -1.84, P \ 0.00001), time
comparative studies were qualified for sensitivity analyses. to first oral liquid diet (mean difference: -2.72, 95% CI
First set of analysis showed that the results stayed -3.49, -1.95, P \ 0.00001), time to first oral solid diet

123
 World J Surg

(mean difference: -2.51, 95% CI -3.05, -1.97, -3.71, 95% CI -4.57, -2.85, P \ 0.00001). ERAS pro-
P \ 0.00001) and length of hospital stay (mean difference: tocols resulted in lower risks of total complications (odds
-3.71, 95% CI -4.57, -2.85, P \ 0.00001). ERAS pro- ratio: 0.18, 95% CI 0.09, 0.37, P \ 0.00001), pulmonary
tocols resulted in lower risks of total complications (odds complications (odds ratio: 0.23, 95% CI 0.06, 0.86,
ratio: 0.18, 95% CI 0.09, 0.37, P \ 0.00001), pulmonary P = 0.03) and surgical site infection (odds ratio: 0.36, 95%
complications (odds ratio: 0.23, 95% CI 0.06, 0.86, CI 0.15, 0.88, P = 0.03). Both ERAS and non-ERAS pro-
P = 0.03) and surgical site infection (odds ratio: 0.36, 95% tocols resulted in similar risk of paralytic ileus (odds ratio:
CI 0.15, 0.88, P = 0.03). Both ERAS and non-ERAS pro- 0.52, 95% CI 0.15, 1.85, P = 0.32), risk of 30-day mor-
tocols resulted in similar risk of paralytic ileus (odds ratio: tality (risk difference: -0.01, 95% CI - 0.06, 0.03,
0.52, 95% CI 0.15, 1.85, P = 0.32), risk of 30-day mor- P = 0.59), need for re-admission (risk difference: 0.04,
tality (risk difference: -0.01, 95% CI - 0.06, 0.03, 95% CI - 0.03, 0.10, P = 0.30) and need for re-operation
P = 0.59), need for re-admission (risk difference: 0.04, (odds ratio: 0.61, 95% CI 0.13, 2.96, P = 0.54).
95% CI - 0.03, 0.10, P = 0.30) and need for re-operation
(odds ratio: 0.61, 95% CI 0.13, 2.96, P = 0.54). Emergency operation for bowel obstruction Overall, 818
patients underwent emergency operation for bowel
Observational studies obstruction. ERAS protocols resulted in shorter post-op-
erative time to first flatus (mean difference: -1.39, 95% CI
Analysis of 1188 patients from four observational studies -1.52, -1.25, P \ 0.00001), time to first defecation (mean
[13, 15, 16, 18] showed that ERAS protocols resulted in difference: -0.73, 95% CI -1.40, - 0.06, P = 0.03), time
shorter post-operative time to first flatus (mean difference: to first oral liquid diet (mean difference: -2.30, 95% CI
-1.39, 95% CI -1.52, -1.25, P \ 0.00001), time to first -2.37, -2.23, P \ 0.00001), time to first oral solid diet
defecation (mean difference: -0.73, 95% CI -1.40, (mean difference: 2.39, 95% CI -2.62, -2.15,
- 0.06, P = 0.03), time to first oral liquid diet (mean dif- P \ 0.00001) and length of hospital stay (mean difference:
ference: -2.30, 95% CI -2.37, -2.23, P \ 0.00001), time -3.01, 95% CI -3.31, -2.71, P \ 0.00001). ERAS pro-
to first oral solid diet (mean difference: 2.39, 95% CI tocols resulted in lower risks of total complications (odds
-2.62, -2.15, P \ 0.00001) and length of hospital stay ratio: 0.60, 95% CI 0.45, 0.81, P = 0.0008), major com-
(mean difference: -3.01, 95% CI -3.31, -2.71, plications (odds ratio: 0.60, 95% CI 0.45, 0.81,
P \ 0.00001). ERAS protocols resulted in lower risks of P = 0.0008) and paralytic ileus (odds ratio: 0.53, 95% CI
total complications (odds ratio: 0.60, 95% CI 0.45, 0.81, 0.31, 0.93, P = 0.03). Both ERAS and non-ERAS proto-
P = 0.0008), major complications (odds ratio: 0.60, 95% cols resulted in similar risk of pulmonary complications
CI 0.45, 0.81, P = 0.0008), pulmonary complications (odds (odds ratio: 0.10, 95% CI 0.00, 2.15, P = 0.14), risk of
ratio: 0.42, 95% CI 0.24, 0.76, P = 0.004), paralytic ileus 30-day mortality (risk difference: -0.00, 95% CI - 0.02,
(odds ratio: 0.53, 95% CI 0.31, 0.93, P = 0.03) and surgical 0.01, P = 0.92), need for re-admission (risk difference:
site infection (odds ratio: 0.41, 95% CI 0.23, 0.72, -0.03, 95% CI - 0.07, 0.01, P = 0.11) and need for re-
P = 0.002). Both ERAS and non-ERAS protocols resulted operation (odds ratio: 0.76, 95% CI 0.30, 1.88, P = 0.55).
in similar risk of 30-day mortality (risk difference: 0.00,
95% CI - 0.02, 0.02, P = 0.95), need for re-admission
(risk difference: -0.02, 95% CI - 0.05, 0.02, P = 0.33) Discussion
and need for re-operation (odds ratio: 0.87, 95% CI 0.49,
1.53, P = 0.62). In this study, analysis of the outcomes related to 1334
patients from six comparative studies showed that ERAS
Subgroup analyses (Table 4) protocols resulted in favourable outcomes in emergency
settings as indicated by reduced post-operative complica-
Emergency operation for visceral perforation Overall, tions, accelerated recovery of bowel function, and shorter
146 patients underwent emergency operation for visceral length of hospital stay without increasing the need for re-
perforation. ERAS protocols resulted in shorter post-oper- admission or re-operation. The sensitivity and subgroup
ative time to first flatus (mean difference: -1.47, 95% CI analyses confirmed the consistency of the results.
-1.83, -1.11, P \ 0.00001), time to first defecation (mean Although the impact of ERAS protocols on post-oper-
difference: -2.26, 95% CI -2.68, -1.84, P \ 0.00001), ative outcomes in elective settings has been investigated by
time to first oral liquid diet (mean difference: -2.72, 95% previous meta-analyses, this is the first meta-analysis in the
CI -3.49, -1.95, P \ 0.00001), time to first oral solid diet literature evaluating ERAS protocols in emergency set-
(mean difference: -2.51, 95% CI -3.05, -1.97, tings. Our results are comparable with the evidence pro-
P \ 0.00001) and length of hospital stay (mean difference: vided by previously published studies in elective settings.

123
World J Surg

Table 4 Results of subgroup analyses

Outcomes Randomised controlled Observational studies Visceral perforation Bowel obstruction
trials
Summary P value Summary P value Summary P value Summary P value
measure (95% measure (95% measure (95% measure (95%
CI) CI) CI) CI)

Length of hospital MD: -3.71 \0.00001 MD: -3.01 \0.00001 MD: -3.71 \0.00001 MD: -3.01 \0.00001
stay (4.57, (- 3.31, (4.57, (- 3.31,
- 2.85) - 2.71) - 2.85) - 2.71)
Total post-operative OR: 0.18 (0.09, \0.00001 OR: 0.60 (0.45, 0.0008 OR: 0.18 (0.09, \0.00001 OR: 0.60 (0.45, 0.0008
complications 0.37) 0.81) 0.37) 0.81)
Time to first flatus MD: -1.47 \0.00001 MD: -1.39 \0.00001 MD: -1.47 \0.00001 MD: -1.39 \0.00001
(- 1.83, (- 1.52, (- 1.83, (- 1.52,
- 1.11) - 1.25) - 1.11) - 1.25)
Time to first MD: -2.26 \0.00001 MD: -0.73 0.03 MD: -2.26 \0.00001 MD: -0.73 0.03
defecation (- 2.68, (- 1.40, (- 2.68, (- 1.40,
- 1.84) - 0.06) - 1.84) - 0.06)
Time to first oral MD: -2.72 \0.00001 MD: -2.30 \0.00001 MD: -2.72 \0.00001 MD: -2.30 \0.00001
liquid diet (- 3.49, (- 2.37, (- 3.49, (- 2.37,
- 1.95) - 2.23) - 1.95) - 2.23)
Time to first oral MD: -2.51 \0.00001 MD: 2.39 \0.00001 MD: -2.51 \0.00001 MD: 2.39 \0.00001
solid diet (- 3.05, (- 2.62, (- 3.05, (- 2.62,
- 1.97) - 2.15) - 1.97) - 2.15)
Major complications NR NR OR: 0.60 (0.45, 0.0008 NR NR OR: 0.60 (0.45, 0.0008
0.81) 0.81)
Pulmonary OR: 0.23 (0.06, 0.03 OR: 0.42 (0.24, 0.004 OR: 0.23 (0.06, 0.03 OR: 0.10 (0.00, 0.14
complications 0.86) 0.76) 0.86) 2.15)
Paralytic ileus OR: 0.52 (0.15, 0.32 OR: 0.53 (0.31, 0.03 OR: 0.52 (0.15, 0.32 OR: 0.53 (0.31, 0.03
1.85) 0.93) 1.85) 0.93)
Surgical site OR: 0.36 (0.15, 0.03 OR: 0.41 (0.23, 0.002 OR: 0.36 (0.15, 0.03 NR NR
infection 0.88) 0.72) 0.88)
30-Day mortality RD: -0.01 0.59 RD: 0.00 0.95 RD: -0.01 0.59 RD: -0.00 0.92
(- 0.06, (- 0.02, (- 0.06, (- 0.02,
0.03) 0.02) 0.03) 0.01)
Need for re- RD: 0.04 0.3 RD: -0.02 0.33 RD: 0.04 0.3 RD: -0.03 0.11
admission (- 0.03, (- 0.05, (- 0.03, (- 0.07,
0.10) 0.02) 0.10) 0.01)
Need for re- OR: 0.61 (0.13, 0.54 OR: 0.87 (0.49, 0.62 OR: 0.61 (0.13, 0.54 OR: 0.76 (0.30, 0.55
operation 2.96) 1.53) 2.96) 1.88)
MD mean difference, OR odds ratio, RD risk difference, CI confidence interval, NR not reported

Ji et al. [19] completed a meta-analysis of 3694 patients and appropriate in emergency setting. Full preoperative
from twenty comparative studies and concluded that ERAS counselling, which is known to reduce post-operative
protocols resulted in improved complication rates, better stress, pain and anxiety [24], may not be possible in
post-operative recovery and shorter hospital stay in elective emergency setting; nevertheless, information such as
pancreatic surgery. Wang et al., in another meta-analysis details of procedure, possible perioperative complications,
[20], reached similar conclusions in elective gastric cancer need for creation of stoma and length of hospitalisation can
surgery. Other studies have reported comparable findings be communicated with patients and their family before the
in patients undergoing hepatectomy [21], oesophageal procedure. On the other hand, although complete optimi-
cancer surgery [22] and colorectal surgery [23]. sation of medical conditions cannot be achieved in emer-
Although some of the preoperative components of gency setting, objective intravenous fluid and antibiotic
ERAS protocols (e.g. nutritional support and carbohydrate resuscitation in emergency major abdominal surgery are
loading) cannot be achieved in emergency setting, most crucial and feasible [25]. Apart from these, the rest of
components of ERAS protocols are considered applicable intraoperative and post-operative components of ERAS

123
 World J Surg

protocols [26–31] are achievable in emergency settings. Funding There were no funding resources for this study
Consistent with the above, our results also support the
Compliance with ethical standards
applicability of the above components of ERAS protocols
in emergency general surgery setting. Conflict of interest All authors confirmed that they had no conflict of
Use of ERAS protocols in emergency setting may face interest.
some barriers postoperatively. The critically unwell
Ethical approval Ethical approval was not required for this study.
patients undergoing emergency abdominal procedures may
require longer period of post-operative ventilation and
circulation supports in intensive care unit compared to
Appendix
those undergoing minor emergency abdominal procedures.
This would limit early mobilisation, early removal of
drains or catheters and early enteral nutrition. Although the
available data from the included studies do not support this Search Search description
argument as almost all studies demonstrated feasibility of
Search enhance* near2 recovery T, A, K
the above components of ERAS protocols, we could not
#1
perform subgroup analysis for patients who were critically
Search ERAS T, A, K
unwell postoperatively requiring high level of ventilation #2
and circulation supports. Therefore, applicability of ERAS Search Search #1 OR Search #2 Combined with OR
protocols in this subgroup needs further evaluation in #3
future studies. Search emergency near2 surgery T, A, K
In the current study, an objective approach was followed #4
to report the best available comparative evidence and to Search MeSH term: [laparotomy] explode all trees
examine the possibility of bias in the current evidence. Our #5
findings stayed unaffected during sensitivity analyses, Search emergency near2 laparotomy T, A, K
#6
separate analyses of randomised and non-randomised
Search Search #4 OR Search #5 OR Search Combined with OR
studies and subgroup analyses. These would strongly sup- #7 #6
port our conclusions of the published evidence. However,
Search Search #3 AND Search #7 Combined with
this study was associated with some limitations. Only six #8 AND
eligible studies were found after applying the search
T, A, K: titles, abstracts, keywords
strategy in the aforementioned databases of which four
were non-randomised observational studies, and their
results may be associated with inevitable selection bias.
Finally, because less than 10 comparative studies were
qualified for inclusion; publication bias could not be
examined.
References

1. Fearon KC, Ljungqvist O, Von Meyenfeldt M, Revhaug A,

Conclusions Dejong CH, Lassen K et al (2005) Enhanced recovery after
surgery: a consensus review of clinical care for patients under-
Although ERAS protocols are commonly used in elective going colonic resection. Clin Nutr 24(3):466–477
settings, they are associated with favourable outcomes in 2. Watt DG, McSorley ST, Horgan PG, McMillan DC (2015)
Enhanced recovery after surgery: which components, if any,
emergency settings as indicated by reduced post-operative impact on the systemic inflammatory response following col-
complications, accelerated recovery of bowel function and orectal surgery? A systematic review. Medicine (Baltimore)
shorter hospital stay without increasing need for re-ad- 94(36):e1286
mission or re-operation. There should be an effort to 3. Basse L, Raskov HH, Hjort Jakobsen D, Sonne E, Billesbølle P,
Hendel HW et al (2002) Accelerated postoperative recovery
incorporate ERAS protocols into emergency abdominal programme after colonic resection improves physical perfor-
surgery settings. mance, pulmonary function and body composition. Br J Surg
89:446–453
4. Barbieri A, Vanhaecht K, Van Herck P, Sermeus W, Faggiano F,
Author contributions SH was the lead of the project and designed Marchisio S et al (2009) Effects of clinical pathways in the joint
the study. SH and SH performed the statistical analyses required in replacement: a meta-analysis. BMC Med 7:32
this study and wrote the first draft of the article. All authors con- 5. Azhar RA, Bochner B, Catto J, Goh AC, Kelly J, Patel HD et al
tributed to collection of data, revision of the first draft and approving (2016) Enhanced recovery after urological surgery: a
the final version of the article

123
World J Surg

contemporary systematic review of outcomes, key elements, and 19. Ji HB, Zhu WT, Wei Q, Wang XX, Wang HB, Chen QP (2018)
research needs. Eur Urol 70:176–187 Impact of enhanced recovery after surgery programs on pancre-
6. Arsalani-Zadeh R, ElFadl D, Yassin N, MacFie J (2011) Evi- atic surgery: a meta-analysis. World J Gastroenterol
dence-based review of enhancing postoperative recovery after 24(15):1666–1678
breast surgery. Br J Surg 98:181–196 20. Wang LH, Zhu RF, Gao C, Wang SL, Shen LZ (2018) Appli-
7. Hughes MJ, McNally S, Wigmore SJ (2014) Enhanced recovery cation of enhanced recovery after gastric cancer surgery: an
following liver surgery: a systematic review and meta-analysis. updated meta-analysis. World J Gastroenterol 24(14):1562–1578
HPB (Oxford) 16:699–706 21. Li L, Chen J, Liu Z, Li Q, Shi Y (2017) Enhanced recovery
8. Pisarska M, Małczak P, Major P, Wysocki M, Budzyński A, program versus traditional care after hepatectomy: a meta-anal-
Pe˛dziwiatr M (2017) Enhanced recovery after surgery protocol in ysis. Medicine (Baltimore) 96(38):e8052
oesophageal cancer surgery: systematic review and meta-analy- 22. Pisarska M, Małczak P, Major P, Wysocki M, Budzyński A,
sis. PLoS ONE 12:e0174382 Pe˛dziwiatr M (2017) Enhanced recovery after surgery protocol in
9. de Groot JJ, Ament SM, Maessen JM, Dejong CH, Kleijnen JM, oesophageal cancer surgery: systematic review and meta-analy-
Slangen BF (2016) Enhanced recovery pathways in abdominal sis. PLoS ONE 12(3):e0174382
gynecologic surgery: a systematic review and meta-analysis. Acta 23. Spanjersberg WR, van Sambeeck JD, Bremers A, Rosman C, van
Obstet Gynecol Scand 95:382–395 Laarhoven CJ (2015) Systematic review and meta-analysis for
10. Liberati A, Altman DG, Tetzlaff J et al (2009) The PRISMA laparoscopic versus open colon surgery with or without an ERAS
statement for reporting systematic reviews and meta-analyses of programme. Surg Endosc 29(12):3443–3453
studies that evaluate healthcare interventions: explanation and 24. Shuldham C (1999) A review of the impact of pre-operative
elaboration. BMJ 339:b2700 education on recovery from surgery. Int J Nurs Stud 36:171–177
11. Lassen K, Soop M, Nygren J, Cox PB, Hendry PO, Spies C et al 25. Sethi A, Debbarma M, Narang N, Saxena A, Mahobia M, Tomar
(2009) Consensus review of optimal perioperative care in col- GS (2018) Impact of targeted preoperative optimization on
orectal surgery: enhanced Recovery After Surgery (ERAS) Group clinical outcome in emergency abdominal surgeries: a prospec-
recommendations. Arch Surg 144:961–969 tive randomized trial. Anesth Essays Res 12:149–154
12. Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos M. 26. Bilgin H (2017) Inadverdent perioperative hypothermia. Turk J
The Newcastle-Ottawa Scale (NOS) for assessing the quality of Anaesthesiol Reanim 45:124–126
nonrandomised studies in meta-analyses. http://www.ohri.ca/pro 27. Huddart S, Peden CJ, Swart M, McCormick B, Dickinson M,
grams/clinical_epidemiology/nos_manual.pdf. Accessed 19 Apr Mohammed MA et al (2015) Use of a pathway quality
2019 improvement care bundle to reduce mortality after emergency
13. Shang Y, Guo C, Zhang D (2018) Modified enhanced recovery laparotomy. Br J Surg 102:57–66
after surgery protocols are beneficial for postoperative recovery 28. Sapkota R, Bhandari RS (2013) Prophylactic nasogastric
for patients undergoing emergency surgery for obstructive col- decompression after emergency laparotomy. JNMA J Nepal Med
orectal cancer: a propensity score matching analysis. Medicine Assoc 52:437–442
(Baltimore) 97(39):e12348 29. Lohsiriwat V (2016) Pelvic drain after colorectal anastomosis:
14. Mohsina S, Shanmugam D, Sureshkumar S, Kundra P, Maha- useful or useless. Transl Cancer Res 5:S1404–S1407
lakshmy T, Kate V (2018) Adapted ERAS pathway vs. standard 30. Klappenbach RF, Yazyi FJ, Alonso Quintas F, Horna ME,
care in patients with perforated duodenal ulcer—a randomized Alvarez Rodrı́guez J, Orı́a A (2013) Early oral feeding versus
controlled trial. J Gastrointest Surg. J 22(1):107–116 traditional postoperative care after abdominal emergency surgery:
15. Shida D, Tagawa K, Inada K, Nasu K, Seyama Y, Maeshiro T a randomized controlled trial. World J Surg 37:2293–2299.
et al (2017) Modified enhanced recovery after surgery (ERAS) https://doi.org/10.1007/s00268-013-2143-1
protocols for patients with obstructive colorectal cancer. BMC 31. Boden I, Sullivan K, Hackett C, Winzer B, Lane R, McKinnon M
Surg 17(1):18 et al (2018) ICEAGE (Incidence of Complications following
16. Wisely JC, Barclay KL (2016) Effects of an Enhanced Recovery Emergency Abdominal surgery: Get Exercising): study protocol
After Surgery programme on emergency surgical patients. ANZ J of a pragmatic, multicentre, randomised controlled trial testing
Surg 86(11):883–888 physiotherapy for the prevention of complications and improved
17. Gonenc M, Dural AC, Celik F, Akarsu C, Kocatas A, Kalayci physical recovery after emergency abdominal surgery. World J
MU et al (2014) Enhanced postoperative recovery pathways in Emerg Surg 13:29
emergency surgery: a randomised controlled clinical trial. Am J
Surg 207(6):807–814 Publisher’s Note Springer Nature remains neutral with regard to
18. Lohsiriwat V (2014) Enhanced recovery after surgery vs con- jurisdictional claims in published maps and institutional affiliations.
ventional care in emergency colorectal surgery. World J Gas-
troenterol 20(38):13950–13955

123