Journal of Gastrointestinal Surgery

, 13:2321

Enhanced Recovery after Surgery (ERAS) Programs for Patients Having Colorectal Surgery: A Meta-analysis of Randomized Trials

Authors

  • Cagla Eskicioglu
    • Department of SurgeryUniversity of Toronto
    • Department of Health Policy, Management, and EvaluationUniversity of Toronto
    • Zane Cohen Digestive Diseases Research CenterMount Sinai Hospital
  • Shawn S. Forbes
    • Department of SurgeryUniversity of Toronto
    • Department of Health Policy, Management, and EvaluationUniversity of Toronto
    • Zane Cohen Digestive Diseases Research CenterMount Sinai Hospital
  • Mary-Anne Aarts
    • Department of SurgeryUniversity of Toronto
    • Toronto East General Hospital
  • Allan Okrainec
    • Department of SurgeryUniversity of Toronto
    • Toronto Western HospitalUniversity Health Network
    • Department of SurgeryUniversity of Toronto
    • Department of Health Policy, Management, and EvaluationUniversity of Toronto
    • Zane Cohen Digestive Diseases Research CenterMount Sinai Hospital
    • Samuel Lunenfeld Research InstituteMount Sinai Hospital
Review Article

DOI: 10.1007/s11605-009-0927-2

Cite this article as:
Eskicioglu, C., Forbes, S.S., Aarts, M. et al. J Gastrointest Surg (2009) 13: 2321. doi:10.1007/s11605-009-0927-2

Abstract

Background

Enhanced recovery after surgery programs have been introduced with aims of improving patient care, reducing complication rates, and shortening hospital stay following colorectal surgery. The aim of this meta-analysis was to determine whether enhanced recovery after surgery programs, when compared to traditional perioperative care, are associated with reduced primary hospital length of stay in adult patients undergoing elective colorectal surgery.

Methods

MEDLINE, EMBASE, the Cochrane Central Registry of Controlled Trials, and the reference lists were searched for relevant articles. Only randomized controlled trials comparing an enhanced recovery program with traditional postoperative care were included.

Results

Three of four included studies showed significantly shorter primary lengths of stay for patients enrolled in enhanced recovery programs. There was no significant difference in postoperative mortality when the two groups were compared [relative risk (RR) = 0.53; 95% CI = 0.12–2.38; test for heterogeneity, p = 0.40 and I2 = 0], and patients in enhanced recovery programs were less likely to develop postoperative complications (RR = 0.61, 95% CI = 0.42–0.88; test for heterogeneity, p = 0.95 and I2 = 0).

Authors’ Conclusions

There is some evidence to suggest that enhanced recovery after surgery programs are better than traditional perioperative care, but evidence from a larger, better quality randomized controlled trial is necessary.

Keywords

Enhanced recovery after surgeryColorectal surgeryMeta-analysisPostoperative complications

Introduction

Colorectal surgery has been associated with complication rates ranging from 10% to 20% and mean postoperative hospital stays from 6 to 10 days.1 In an effort to improve postoperative outcomes in patients undergoing colorectal surgery, enhanced recovery after surgery (ERAS) programs have been designed and evaluated.2,3 ERAS programs, also known as fast-track surgery, are multimodal perioperative programs that aim to accelerate recovery, shorten hospital stay, and reduce complication rates following colorectal surgery. These programs address many factors thought to prolong hospital stay, including prolonged parenteral analgesia use, overuse of intravenous hydration, and delayed mobilization.4

ERAS programs described in the literature are incredibly diverse and include recommendations for a variety of interventions.4,5 These can be classified as preoperative, intraoperative, or postoperative interventions. The preoperative interventions include extensive preoperative counseling, avoidance of mechanical bowel preparation (MBP), avoidance of fasting, avoidance of premedication, administration of pre- and probiotics, and preoperative carbohydrate loading until 2 h prior to surgery. Intraoperative interventions are strict fluid management to avoid fluid overload, normothermia, hyperoxia, and tailored optimal analgesia. Finally, postoperative components include epidural anesthesia, early routine mobilization, early enteral nutrition, avoidance of nasogastric (NG) tubes, avoidance of peritoneal drains, and early removal of catheters.4,5 Many studies have shown that the avoidance of individual components, such as NG tubes or MBP, have not led to increased complications.4,6,7 Experience with other components, such as type of surgical incision or methods to address postoperative nausea and vomiting, has been limited, but these factors are still included in the comprehensive programs based mostly on consensus.4,5

ERAS programs are designed to minimize the stress response that is associated with surgery.4 Clinically, ERAS regimens result in better physical performance, measured by treadmill exercise, pulmonary function, and body composition as measured by lean body mass.8 It appears that the combination of many of these factors has a synergistic positive effect on postoperative outcomes following colorectal surgery as compared to each individual parameter alone.5

Designing a multimodal ERAS program and organizing a devoted surgical team to implement this program may be costly. Before embarking on an extensive intervention, it is important to clearly delineate the benefits of such programs. The individual studies evaluating ERAS programs have revealed both advantages and disadvantages of these programs. The purpose of this meta-analysis is to synthesize the evidence and determine the utility of ERAS programs for patients undergoing elective colorectal surgery.

The primary goal of this review was to determine whether ERAS programs, when compared to traditional perioperative care, are associated with reduced primary and total postoperative hospital stay in adult patients undergoing elective colorectal surgery. The impact of ERAS programs on other outcomes such as major and minor postoperative complications, mortality, rates of readmission, and reoperation is also reviewed.

Methods

Systematic Review

A systematic review of the medical literature was performed with the assistance of a medical librarian to identify all potential abstracts regardless of publication status or language that compared ERAS programs to traditional perioperative care in patients undergoing elective colorectal surgery. MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials were searched between 1950 and May 2008 using the following strategy: Explode “intestinal diseases,” “colectomy,” “laparotomy,” “laparoscopy,” “video-assisted surgery,” “cecal neoplasms,” or “colorectal neoplasms” combined with “multimodal,” “optimization,” “enhanced recovery,” or “fast track.” The associated text and title words were similarly searched. Standard limiters were used to identify randomized trials and review articles. Reference lists of randomized controlled trials (RCTs), meta-analyses, and systematic reviews were hand-searched for studies that were not captured by the initial electronic search. Content experts (MA, AO) were consulted to ensure no published or unpublished work had been missed. Two authors (CE, SSF) independently reviewed all citations generated by the literature search to include relevant studies. A consensus meeting was used to resolve any disagreements on selection.

Study Selection

All RCTs comparing ERAS programs to traditional perioperative care in patients undergoing elective colorectal surgery were included in this review. Studies evaluating adult patients (>18 years of age) undergoing elective, open or laparoscopic, colon or rectal resections regardless of indication for surgery (i.e., colon cancer, inflammatory bowel disease, diverticulosis, etc.) were included. Quasi-randomized trials, non-randomized trials, and uncontrolled studies were excluded. Trials investigating patients who had a colon resection for bowel obstruction, bowel perforation, or any emergent cause were excluded.

Data Extraction, Outcomes, and Study Quality

Two reviewers independently abstracted data and rated the methodological quality of all included studies, resolving discrepancies by consensus (CE, SSF). Methodological quality criteria were adapted from the US Preventative Task Force Criteria.9 The following criteria were assessed: concealment of the randomization allocation, description of the randomization method, blinding of the outcome assessor, similarity of baseline patient characteristics between the two study arms, definition of study outcomes, minimum of 80% patient follow-up, and use of intention-to-treat analysis.

One important study characteristic, which was abstracted in order to comment on clinical heterogeneity, was the nature of the ERAS intervention. This included the number and details of the parameters involved in each ERAS intervention. The primary outcomes were primary and total postoperative hospital stay. Primary hospital length of stay was defined as the number of days in hospital after surgery until discharge. Total hospital length of stay included primary hospital stay and any additional days during hospital readmission within 30 days of surgery. Secondary outcomes included postoperative mortality, postoperative complications (major and minor), and rate of readmission. Major complications included intra-abdominal infections, anastomotic leaks, cardiac complications, respiratory complications, venous thromboembolic disease. Minor complications included superficial surgical site infections, pneumonia, and urinary tract infections. For binary outcomes, the number of outcomes in the intervention and control groups were recorded. For continuous outcomes, the mean and standard deviation (SD) were recorded when available.

Statistical Analysis

Calculations of effect sizes for dichotomous variables are presented as relative risk (RR) with 95% confidence intervals (CI) and for continuous outcomes as weighted mean differences. Pooled analyses were performed using both the random effects and fixed effects models with the Mantel–Haenszel method when appropriate. Clinical heterogeneity was explored by examining study characteristics, specifically ERAS intervention modalities and the reported outcomes. Statistical heterogeneity was assessed using the Cochran’s Q test (p < 0.10) and the I2 statistic. Funnel plots were used to assess for publication bias. These analyses were performed using RevMan 5.0 (The Nordic Cochrane Center, The Cochrane Collaboration, Copenhagen, Denmark).

Results

Two hundred forty-three abstracts were retrieved from the search, of which 236 did not fulfill the inclusion criteria (Fig. 1). Of the seven identified for further review, three were excluded due to a non-randomized design.2,10,11 The four RCTs included in the review were published between 2003 and 2007 and enrolled a total of 198 patients.1215 There were no studies comparing one ERAS program to another. The number of ERAS interventions, which were addressed by the four studies, ranged from 8 to 14 (Table 1). Some of the components included in the ERAS intervention of the four studies were similar. However, there was some clinical heterogeneity regarding number of included interventions and some of the other intervention components. For example, one trial15 included the use of mechanical bowel preparation as part of the ERAS intervention, whereas two other trials included the omission of mechanical bowel preparation as part of the intervention.13,14
https://static-content.springer.com/image/art%3A10.1007%2Fs11605-009-0927-2/MediaObjects/11605_2009_927_Fig1_HTML.gif
Figure 1

Summary of literature search and study selection.

Table 1

Components of ERAS Interventions Used in the Four Included Studies

 

Delaney et al.12

Anderson et al.13

Gatt et al.14

Khoo et al.15

Preoperative

Preoperative counselling and written information

Verbal/written preoperative information

Verbal/written preoperative information

MBP—Fleet phospho-soda (2 bottles)

 

Pre-assessment by anesthesia/surgery

Pre-assessment by anesthesia/surgery

Oral fluids until 3 h prior to surgery

 

Synbiotics

Synbiotics

No intravenous (IV) fluids overnight

 

Omission of MBP

Omission of MBP

 
 

Oral carbohydrate loading

Oral carbohydrate loading

 
 

Preoperative fast 3 h prior to surgery

Preoperative fast 3 h prior to surgery

 

Intraoperative

NG tubes removed before extubation

80% FiO2

80% FiO2

Intra-op fluid restriction—1,500 cc unless blood loss >500 cc occurred

 

Transverse incision

Transverse incision

 

Postoperative

Offer clear fluids on the evening of surgery

No drains

No drains

Free oral fluids immediately postoperatively

Encouraged to ambulate on the evening of surgery

No NG tubes

No NG tubes

IV fluids discontinued once tolerating 200 cc of water over 30 min

Solid food POD#1 at dinner

Early fluid and diet reintroduction

Early fluid and diet reintroduction

NG tube removed in recovery room

Oral analgesia started on POD#2

Aggressive structured mobilization plan by physiotherapists

Aggressive structured mobilization plan by physiotherapists

Diet commenced immediately postoperatively

Wall chart to emphasize ambulation

Epidural analgesia with standing acetaminophen and ibuprofen, PRN morphine

Epidural analgesia until 24–36 h postoperatively

Regular domperidone, magnesium hydroxide 8% and liquid protein/calorie supplements from admission

PCA analgesia + ketorolac q6hr PRN

  

Thoracic epidural—infusion rate not adjusted unless there were features of narcotization

   

Epidurals discontinued at 48 h postoperatively

   

Standing paracetamol and ibuprofen starting immediately postoperatively

   

Urinary catheters removed 24 h post-colonic surgery and 72 h post-TME

   

Mobilization encouraged from night of OR and patients were encouraged to meet predefined mobility targets

All four studies reported randomized allocation to each study arm as well as concealment of allocation. Three studies1214 used sealed envelopes to allocate patients and one study15 used a central randomization system with a random number generator. None of the four studies reported the use of blinding of the surgeon, patient, or outcome assessors. Complete patient data (100% follow-up) were reported in all four studies. There was no evidence or reporting of protocol violations in three of the studies; as such, intention-to-treat (ITT) analysis was not necessary.1315 One study reported protocol violations and the subsequent use of ITT analysis Table 2.12
Table 2

Primary Hospital Length of Stay

Study

ERAS

Traditional perioperative care

p value

Delaney et al.12

Mean ± SD (days)

5.2 ± 2.5

5.8 ± 3

0.12

Anderson et al.13

Median (range, days)

3 (2–7)

7 (4–10)

0.002

Mean ± SD (hours)

95.1 ± 42.5

167.8 ± 49.6

0.002

Gatt et al.14

Median (range, days)

5 (4–9)

7.5 (6–10)

0.027

Khoo et al.15

Median (range, days)

5 (3–37)

7 (4–63)

<0.001

Hospital Stay

Pooling of the primary and overall hospital length of stay was not possible. One study12 reported the results as a mean ± SD. Two studies reported length of stay as a median and range.14,15 The fourth study13 reported both mean ± SD and median with range. Although it is possible to estimate mean and standard deviation from median and interquartile range, this was felt to be statistically inappropriate. In summary, three of these four studies showed significantly shorter primary length of stays for patients enrolled in ERAS programs (Table 2).1315 In the one study12 that did not find a statistically significant shorter primary hospital stay, the patients enrolled in the ERAS program were significantly older than those patients receiving traditional perioperative care despite randomization. Only two studies reported overall hospital stay and these results were also not pooled.12,15 Both studies,12,15 however, reported significantly shorter overall hospital length of stay in patients enrolled in ERAS programs (Table 3).
Table 3

Overall Hospital Length of Stay

Study

ERAS

Traditional perioperative care

p value

Delaney [12]

Mean ± SD (days)

5.4 ± 2.5

7.1 ± 4.8

0.022

Khoo [15]

Median (range)

5 (3–37)

7 (4–63)

<0.001

Postoperative Mortality and Complications

Three studies reported on postoperative mortality. Compared to patients receiving traditional perioperative care, there was no significant difference in postoperative mortality for patients enrolled in ERAS programs (RR = 0.53; 95% CI = 0.12–2.38; test for heterogeneity, p = 0.40 and I2 = 0; Fig. 2). All four studies reported on major and minor postoperative complications. Patients who were treated using ERAS programs were significantly less likely to develop postoperative complications than those patients receiving traditional perioperative care (RR = 0.61; 95% CI = 0.42–0.88; test for heterogeneity, p = 0.95 and I2 = 0; Fig. 3). There were no significant differences in the rates of major (RR = 0.40; 95% CI = 0.06–2.59; test for heterogeneity, p = 0.04 and I2 = 63%; Fig. 4) and minor complications (RR = 0.67; 95% CI = 0.37–1.23; test for heterogeneity, p = 0.17 and I2 = 41%; Fig. 5) in patients receiving ERAS care as compared to those patients receiving traditional perioperative care. Since significant statistical heterogeneity was observed for major complications, the random effect model was used to pool the data. Furthermore, there was no significant difference in readmission to hospital in ERAS patients compared to traditional perioperative care patients (RR = 0.67; 95% CI = 0.20–2.19; test for heterogeneity, p = 0.27 and I2 = 24%; Fig. 6). Only one trial reported on reoperation as an outcome, and this study showed a non-significant trend toward a reduced risk of reoperation in ERAS patients compared to traditional perioperative care patients (RR = 0.35; 95% CI = 0.04–3.23).12
https://static-content.springer.com/image/art%3A10.1007%2Fs11605-009-0927-2/MediaObjects/11605_2009_927_Fig2_HTML.gif
Figure 2

Pooled analysis of postoperative mortality, ERAS vs. traditional perioperative care (TPC).

https://static-content.springer.com/image/art%3A10.1007%2Fs11605-009-0927-2/MediaObjects/11605_2009_927_Fig3_HTML.gif
Figure 3

Pooled analysis of total complications (major and minor), ERAS vs. traditional perioperative care (TPC).

https://static-content.springer.com/image/art%3A10.1007%2Fs11605-009-0927-2/MediaObjects/11605_2009_927_Fig4_HTML.gif
Figure 4

Pooled analysis of major complications, ERAS vs. traditional perioperative care (TPC).

https://static-content.springer.com/image/art%3A10.1007%2Fs11605-009-0927-2/MediaObjects/11605_2009_927_Fig5_HTML.gif
Figure 5

Pooled analysis of minor complications, ERAS vs. traditional perioperative care (TPC).

https://static-content.springer.com/image/art%3A10.1007%2Fs11605-009-0927-2/MediaObjects/11605_2009_927_Fig6_HTML.gif
Figure 6

Pooled analysis of readmission to hospital, ERAS vs. traditional perioperative care (TPC).

Discussion

This meta-analysis evaluates postoperative outcomes which are relevant in the care of surgical patients by using data from four randomized controlled trials comparing ERAS with traditional perioperative care. The results for primary and total hospital length of stay could not be pooled; however, three of the four studies revealed a significantly shorter primary length of hospital stay in patients enrolled in ERAS programs as compared to traditional perioperative care. Furthermore, this meta-analysis revealed a significant decreased rate of overall complications with a trend towards decreased mortality and reoperation rate for patients enrolled in ERAS programs. However, there were only four deaths and 18 readmissions in total observed in the four trials, and this likely accounts for the failure to observe a statistically significant difference.

In addition to statistical heterogeneity, there was also some clinical heterogeneity that is important to note. In addition to differences in the particular components that were included in each ERAS program, the number of components also varied. For example, the trial by Delaney et al.12 included eight components in the ERAS intervention. Interestingly, this trial was the only one of the four not to detect a statistically significant difference in primary hospital length of stay between the control and intervention groups. The other three trials included 13 or 14 components, and these trials all showed a 2-day reduction of primary hospital length of stay in the ERAS intervention arm.1315 Also of note is that two of these three trials included very similar ERAS intervention components.13,14

This review includes evidence from randomized controlled trials only and therefore assesses the best available evidence. It is, however, limited by the small number of included studies and the low number of patients (total of 198) included in these trials. Furthermore, it is also limited by the methodological quality of the included studies. All four studies were found to have a high risk of bias. Blinding of outcome assessment was either not performed or not reported. Given the nature of the interventions being compared, blinding of the patient and surgeon is not feasible; however, outcome assessors could be blinded to reduce measurement bias. Since none of the studies reported or performed this aspect of methodological quality, studies were not excluded based on the lack of blinding. All four studies, however, did describe predefined discharge criteria used to determine patients’ readiness for discharge. This may somewhat reduce possible bias of surgeons deciding when to discharge patients.

Further affecting the quality of the evidence was the lack of well-defined outcomes15 and the lack of comparable groups at entry into the trial.12 All four trials reported length of stay as a primary outcome. Since there are often social barriers delaying discharge, length of stay (LOS) may not be the most objective outcome. For example, even if a patient has met the predefined discharge criteria, hospital discharge may be delayed for social reasons. In two of the studies included in this review, only patients who were living independently prior to surgery were included.13,14 This, therefore, limits the generalizability of the results and the ERAS programs. Furthermore, although LOS is an important outcome for cost-effectiveness and allocation of resources, it may not be the most important outcome for delivering quality care to patients. These studies were not powered to primarily assess more clinically relevant outcomes such as complications or even mortality. As well, patient satisfaction was not assessed. Further research in this field designed to evaluate these outcomes would be beneficial.

The review process was performed to minimize bias. The literature strategy used in this review was designed to be broad and capture evidence on any type of ERAS programs. However, there is a possibility that this review missed non-published studies. As well, meeting abstracts were not reviewed. Selection of articles for inclusion as well as data extraction and quality assessment were performed by two independent reviewers. Finally, the inclusion criteria for studies, quality criteria, outcomes, and analysis techniques were all designated a priori.

There is currently only one other review which compares ERAS to traditional perioperative care in patients undergoing elective colorectal surgery. Wind et al.4 performed a meta-analysis of six trials including 512 patients: three randomized trials which are included in our review and three non-randomized trials. One RCT15 included in this meta-analysis was not included in the meta-analysis by Wind et al.4 The meta-analysis revealed a shorter hospital stay when comparing patients enrolled in ERAS programs to traditional perioperative care. This review also reported no increase in mortality, morbidity, or readmission to hospital in patients enrolled in ERAS programs.4 In comparison, this review includes only randomized controlled trials and includes the most recently published trial by Khoo et al.15 The results of this meta-analysis are similar to those of the previously published meta-analysis. However, inclusion of only randomized controlled trials adds to the methodological rigor of this meta-analysis.

Larger randomized trials with improved methodological quality could add to the current literature in this field. In particular, trials must report on allocation concealment, blinding of outcome assessors, and include well-defined objective endpoints. Trials evaluating ERAS programs must set predefined discharge criteria for patients in both the ERAS arm and traditional perioperative care arm in order to further reduce bias. These predefined discharge criteria will ensure that only those patients who are clinically ready for discharge will be discharged from hospital regardless of the clinical pathway.

On the other hand, these trials are difficult to perform. Many of the individual components have already been adopted into standard care despite variable evidence of their effectiveness. So, performing a trial without contamination in the control group might be difficult. In all future studies, besides clinical outcomes, patient satisfaction must be considered. As well, feasibility and hospital resources and costs must be included. Finally, identifying which individual strategies are effective and should be included in the ERAS package should be an important consideration since many are resource intense.

Randomized controlled trials have shown that length of stay is reduced with laparoscopic colorectal surgery.1618 However, there has been no evaluation of ERAS versus traditional perioperative care in patients having laparoscopic surgery. Two randomized controlled trials have compared open to laparoscopic surgery in the setting of an ERAS program.19,20 In one trial which included 62 patients, Basse et al.19 reported rapid postoperative recovery and discharge with a median of length of stay of 2 days; however, there was no difference in this outcome when comparing the laparoscopic and open groups. In another similar trial also evaluating 62 patients, length of stay was 2.2 days shorter (p = 0.018) in patients undergoing laparoscopic surgery. Furthermore, patients undergoing laparoscopic surgery were also less likely to be readmitted with an odds ratio (OR) of 0.13 (95% CI = 0.02–0.79, p = 0.027).20 Currently, with evidence only from these two small trials, it is difficult to draw conclusions regarding the potential benefits of ERAS programs for patients undergoing laparoscopic colorectal surgery. A large randomized controlled trial on this topic would be beneficial.

Conclusions

Although the individual studies showed that ERAS programs were associated with shorter primary and total hospital length of stay, we chose not to pool these data for statistical reasons and thus cannot provide further evidence for these outcomes. Our review does indicate, however, that ERAS programs are associated with reduced total complications. Furthermore, it is also difficult to draw conclusions regarding reductions in mortality and readmission, as the pooling of these outcomes was still underpowered. There is some evidence to suggest that ERAS programs are better than traditional perioperative care, but a larger randomized controlled trial is necessary. This study, with more rigorous methodology, is required to more clearly define components for ERAS programs as well as assess more clinically relevant endpoints.

Acknowledgments

The authors would like to thank Dr. Prakesh Shah, Department of Pediatrics, University of Toronto, for his input during the conduct of the systematic review and preparation of the manuscript and Dr. Joseph Beyene, Department of Medicine, University of Toronto and Population Health Sciences Research Institute, Hospital for Sick Children for his input during the statistical analysis.

Copyright information

© The Society for Surgery of the Alimentary Tract 2009