Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials

The American Journal of Surgery xxx (xxxx) xxx

Contents lists available at ScienceDirect

The American Journal of Surgery journal homepage: www.americanjournalofsurgery.com

Review

Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials Mauro Podda a, *, Salomone Di Saverio b, R. Justin Davies b, Jenny Atzeni a, Francesco Balestra c, Francesco Virdis d, Isabella Reccia e, Kumar Jayant e, Ferdinando Agresta f, Adolfo Pisanu a a

Department of General, Emergency and Minimally Invasive Surgery, Policlinico Universitario di Monserrato, University of Cagliari, Italy Cambridge Colorectal Unit, Addenbrooke's Hospital, Cambridge University Hospitals NHS Foundation Trust, Cambridge, United Kingdom c Department of General, Emergency and Robotic Surgery, San Francesco Hospital, Nuoro, Italy d Department of General Surgery, Hillingdon Hospital NHS Foundation Trust, London, United Kingdom e Department of Surgery and Cancer, Hammersmith Hospital, Imperial College Healthcare NHS Trust, London, United Kingdom f Department of General Surgery, Adria Civil Hospital, Adria, Italy b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 February 2019 Received in revised form 21 March 2019 Accepted 16 May 2019

Background: Clinically evident Anastomotic Leakage (AL) remains one of the most feared complications after colorectal resections with primary anastomosis. The primary aim of this systematic review and meta-analysis was to determine whether Prophylactic Drainage (PD) after colorectal anastomoses confers any advantage in the prevention and management of AL. Methods: Systematic literature search was performed using MEDLINE, the Cochrane Central Register of Controlled Trials, and EMBASE databases for randomized studies comparing clinical outcomes of patients with Drained (D) or Undrained (UD) colorectal anastomoses performed for any cause. Results: Four randomized controlled trials comparing D and UD patients undergoing colorectal resections with primary anastomosis were included for quantitative synthesis. In total, 1120 patients were allocated to group D (n ¼ 566) or group UD (n ¼ 554). The clinical AL rate was 8.5% in the D group and 7.6% in the UD group, with no statistically significant difference (P ¼ 0.57). Rates of radiological AL (D: 4.2% versus UD: 5.6%; P ¼ 0.42), mortality (D: 3.6% versus UD: 4.4%; P ¼ 0.63), overall morbidity (D: 16.6% versus UD: 18.6%, P ¼ 0.38), wound infection (D: 5.4% versus UD: 5.3%, P ¼ 0.95), pelvic sepsis (D: 9.7% versus UD: 10.5%, P ¼ 0.75), postoperative bowel obstruction (D: 9.9% versus UD: 6.9%, P ¼ 0.07), and reintervention for abdominal complication (D: 9.1% versus UD: 7.9%, P ¼ 0.48) were equivalent between the two groups. Conclusions: The present meta-analysis of randomized controlled trials investigating the value of PD following colorectal anastomoses does not support the routine use of prophylactic drains. © 2019 Elsevier Inc. All rights reserved.

Keywords: Colorectal anastomosis Colorectal surgery Prophylactic drainage Anastomotic leakage Meta-analysis

Introduction Clinically evident Anastomotic Leakage (AL) remains one of the most feared complications after colon and rectal resections with

* Corresponding author. Department of General, Emergency, and Minimally Invasice Surgery, Policlinico Universitario “D. Casula”, University of Cagliari, SS 554, Km 4,500, 09042, Monserrato, Italy.
@aoucagliari.it E-mail addresses: [email protected], m.poddao (M. Podda).

primary anastomosis, as it is associated with higher loco-regional recurrence rates, cancer-related death, higher rates of stoma formation and morbidity, as well as with impaired long-term bowel function.1e3 Furthermore, patients developing AL are five times more likely to die postoperatively than those without (0.5%e3.5%), with anastomotic failure responsible for 25%e35% of all deaths after colorectal surgery.4e6 According to a recent meta-analysis on 110 272 patients, AL is significantly associated with male sex and preoperative

https://doi.org/10.1016/j.amjsurg.2019.05.006 0002-9610/© 2019 Elsevier Inc. All rights reserved.

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

2

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

irradiation.7 The introduction of Total Mesorectal Excision (TME) and neoadjuvant chemo-radiotherapy for rectal cancer treatment has dramatically enhanced oncological outcomes for patients undergoing low rectal resections, but on the other hand this needs to be balanced with significant rates of AL and pelvic sepsis, occurring with an incidence rate of 3%e28%.8e10 Historically, surgical dogma meant the correct positioning of a prophylactic drain, with the aim that it may both reduce the incidence of AL, and allow an early diagnosis of anastomotic failure when faecal contents or pus come through the drains.11 Furthermore, Prophylactic Drainage (PD) may facilitate the appropriate and prompt management of an AL, thus minimizing the need for reoperation in some cases or potentially facilitating a more rapid return to theatre in others, dependent on the clinical context.12 The level of evidence that supports this practice, however, is low. Some evidence suggests that PD does not decrease morbidity nor facilitate the conservative management of complications after colorectal AL, and may even be harmful.13e15 Surgeons who are against routine use of drains after colon and rectal anastomoses believe that PD of abdominal and pelvic cavities is associated with increased rates of wound infections, bowel perforation or obstruction, creation of a potential site of entry for intra-abdominal infections, and prolonged postoperative hospital stay.16,17 Experimental studies on animal models have even argued that PD may hamper anastomotic healing by causing infection in the anastomotic area.18 Several meta-analyses19e21, of randomized controlled trials14,22e24 and retrospective cohort studies25,26 have investigated the effectiveness of PD after colorectal anastomoses. However, the results of these studies are contradictory, mainly because of the lack of quality and statistical power. Furthermore, many of the randomized trials published to date have enrolled mainly patients with anastomoses above the peritoneal reflection.16,27,28 Only a few studies, similarly underpowered, focused on the safety and effectiveness of routine PD after extraperitoneal rectal resections, thus preventing any firm recommendation on PD for daily practice in operating theatres when low colorectal, coloanal, or ileoanal anastomoses are performed.22e24,29 As the dispute about whether to drain or not to drain the abdominal and pelvic spaces after colon and rectal anastomoses remains open, an evidence-based reassessment of the real value of PD in the light of more recently published high-quality randomized controlled trials is required. The primary aim of this systematic review and meta-analysis was to determine whether PD after colon and rectal anastomoses confers any advantage in the prevention and management of AL.

manually, and the “related articles” function in PubMed was used. The search strategy combined text words and MeSH terms related to drained versus undrained colon and rectal anastomoses: [drain, drainage colorectal, rectal, anastomosis, intra-abdominal, pelvic, intraperitoneal, extraperitoneal]. Corresponding search strategies were used for the Cochrane Central Register of Controlled Trials and EMBASE. No language restrictions were applied. The literature search was completed in December 2018. The detailed search strategy is freely accessible in the protocol (PROSPERO: CRD42019119138). Study selection For sensitivity reasons, only randomized controlled trials comparing clinical outcomes of patients with drained or undrained colorectal anastomoses performed for colorectal cancer were included in the systematic review and meta-analysis. All studies eligible for inclusion had to report a clear definition of the type of colorectal anastomosis performed. Only studies reporting at least the primary outcome or one of the secondary outcomes were included. The exclusion criteria were: observational studies; prospective non-randomized studies; cost-effectiveness studies; studies not reporting data on the selected outcomes of interest, or articles in which the outcomes of interest could not be calculated; studies not specifying the patients selection criteria; studies reporting outcomes of colonic anastomoses other than colorectal; studies not reporting the type of drain used for the drained group; non-human studies; studies that reported only percentages instead of absolute numbers or odds ratios; review articles; editorials; comments; letters and case reports. The two reviewers independently screened all studies retrieved from the search, and full text articles were obtained if inclusion criteria were fulfilled. Risk of bias assessment The risk of bias for the randomized controlled trials enrolled in the systematic review and meta-analysis was assessed using the Cochrane risk of bias tool, according to the Cochrane Handbook for Systematic Reviews of Interventions.31 Quality of evidence assessment The Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology was applied for assessing the quality of evidence, and reported in the results.32 Furthermore, the Oxford Quality Scoring System was used to independently assess the methodological quality of controlled trials.33

Methods Outcomes measures This systematic review and meta-analysis was conducted according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines30, and was specified in a registered protocol (PROSPERO: CRD42019119138). All stages of study identification, selection, quality assessment and data abstraction were carried out independently by 2 reviewers (M.P. And S.D.S). Any discrepancies were resolved by consulting a third reviewer (A.P.). Study identification MEDLINE (via PubMed), the Cochrane Central Register of Controlled Trials, and EMBASE were systematically searched for relevant studies. Reference lists of relevant studies were searched

The outcome measures were chosen to assess effectiveness and safety of routine PD after colorectal surgery with primary anastomosis. The primary outcome measure was: 1. Clinical anastomotic leakage, defined either as a communication between the intra- and extra-luminal compartments owing to a defect in the integrity of intestinal wall at the anastomotic site; or as abscess formation in the proximity of the anastomosis even if no communication with the bowel lumen could be demonstrated; or as fistula formation between the anastomosis and the abdominal wall; or as discharge of frank pus from the anus; or as discharge of pus, gas or faeces from a drain site or wound; or as formation of a localized abscess or generalized peritonitis with tenderness, fever and leucocytosis.

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

1.

2. 3. 4. 5. 6 7. 8 9

 Secondary outcome measures were evaluated to assess other potential benefits and drawbacks of PD, in terms of: Radiological anastomotic leakage, defined as a leak evident on the contrast enema, demonstrating less than a perfectly regular and uniform calibre at the level of the anastomosis with passage of contrast outside the bowel lumen. Mortality Overall morbidity Wound infection Pelvic sepsis, defined as the occurrence of generalized or localized peritonitis due to AL. Postoperative bowel obstruction Pulmonary complication, defined as the presence of clinical and radiographic signs of pneumonia. Reintervention for abdominal complication Postoperative hospital stay

Data extraction A predefined paper-based sheet was used for data extraction. Data collected for each article comprised the following predefined items: 1) Study identifier (first author, year of publication); 2) Essential study data (study period, study location, study population); 3) Treatment arms and number of enrolled patients; 4) Baseline characteristics of study patients (sex, age, rate of emergency operations, preoperative bowel preparation and antibiotic prophylaxis, operative technique, site of the anastomosis, type of anastomosis, presence of a diverting stoma, and duration of drainage); 5) General characteristics of eligible studies (exclusion criteria, sample size calculation, method for randomization, intention to treat analysis performance, type of drains used, duration of follow-up, number and percentage of patients submitted to surgery for cancer); 6) Outcomes of drained and non-drained patients, as described above. Missing data were requested from the authors of the included studies, when necessary. Statistical analysis Variables for pooled analysis were considered if they were previously evaluated by at least two studies. All statistical analyses were carried out using Reviewer Manager software (Review ManagerdRevMandversion 5.3.5, 2014, The Nordic Cochrane Centre, Cochrane Collaboration, www.cochrane-handbook.org). Data entries in the columns of forest plots were double-checked individually by the two reviewers to avoid errors. The odds ratio (OR) with 95% confidence interval (95% CI) was calculated for dichotomous variables, and the standardized mean difference (SMD) with 95% CI for continuous variables. When continuous data were presented as medians and range, the method of Hozo et al. to estimate respective means and standard deviations was applied.34 The point estimate of the OR value was considered statistically significant at P level of less than 0.05 if the 95% CI did not cross the value 1. The point estimate of the SMD value was considered statistically significant at P level of less than 0.05 if the 95% CI did not cross the value 0. Heterogeneity of the results across studies was assessed using the Higgins' I2 and Chi-Square tests. A P-value of Chi-square test less than 0.10 with an I2 value of greater than 50% were considered as indicative of substantial heterogeneity. Fixed-effects model (Mantel-Haenszel) was implemented if statistically significant heterogeneity was absent. Otherwise, a random-effects model was used for meta-analysis if statistically significant heterogeneity was found, according to the method of DerSimonian and Laird.35 Given that substantial differences in methodology and clinical settings were found among individual studies, subgroup analyses

3

were planned with the aim to explore inter-study heterogeneity according to the level of the anastomosis following colorectal resections (intra- or extraperitoneal). Results A total of 20956 references were identified through database searching. Five more references were identified by searching lists of retrieved studies [Fig. 1]. Thirty-four full text publications were finally assessed for eligibility, of which four randomized controlled trials comparing Drained (D) and Undrained (UD) patients undergoing colorectal resections with primary anastomosis were included for quantitative synthesis.22e24,29 Three studies enrolled both patients with intra- and extraperitoneal colorectal anastomoses22e24, and one study was conducted in patients with extraperitoneal anastomoses.29 The studies contained 59 to 469 patients. In total, 1120 patients were allocated to group D (n ¼ 566) or group UD (n ¼ 554). All the enrolled patients were diagnosed with colorectal cancer. No statistically significant differences were found between D and UD groups in terms of age (65.0 ± 1.0 versus 63.4 ± 3.1, P ¼ 0.74), operative technique (open approach: 58.7% versus 58.5%, P ¼ 0.97), site of anastomosis (intraperitoneal: 42.7% versus 42.2%, P ¼ 0.87), type of anastomosis (stapled: 42.8% versus 43.5%, P ¼ 0.72), and the presence of a diverting stoma (44.3% versus 42.3%, P ¼ 0.43). Drains remained in place for variables periods of time in different studies, ranging from three to seven days.22,29 Mean duration of drainage in days was 5.0 ± 1.6. General characteristics of patients as reported in the studies are shown in Table 1 [Table 1]. Study characteristics The articles included in the quantitative synthesis were published between 1995 and 2017. Of these studies, two had been conducted in France, one in the UK, and one in Singapore. The mean follow-up period was 2.3 ± 2.1 months (range 1e6 months), as reported in the included studies.22e24,29 Large heterogeneity was found among the included studies with regard to exclusion criteria. Sample size calculation was reported in three studies22e24, and an intention to treat analysis was described only in the study by Denost et al.24 [Supp. Digit. Content. Table 1]. Marked heterogeneity was also demonstrated with regard to diagnostic criteria for clinical AL, whilst the definitions of radiological AL, wound infection, pelvic sepsis and respiratory complication were homogeneous among the included studies [Supp. Digit. Content. Table 2]. Risk of bias and quality of evidence assessment Three of the four randomized controlled trials were judged at a low risk of bias22,23,29, and one at unclear risk.24 All but one trial24 generated random sequence adequately, and reported allocation concealment, resulting in a low risk of selection bias. None of the trials reported attempts at blinding patients, personnel, outcome assessors or data analysts. Three RCTs had an Oxford score of 322,23,29, and one RCT had an Oxford score of 224 [Table 2]. Overall quality of evidence, according to the GRADE criteria, was high for clinical AL, mortality, overall morbidity, wound infection, postoperative bowel obstruction, pulmonary complication, reintervention for abdominal complication, and postoperative hospital stay. Radiological AL and pelvic sepsis had a low quality of evidence [Table 3].

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

4

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

Fig. 1. The PRISMA flow diagram for search and selection of articles included in the systematic review and meta-analysis.

Clinical Anastomotic Leakage Overall, four studies reported the rate of clinical AL.22e24,29 The overall clinical AL rate was 8.5% (48/566) in the D group and 7.6% (42/554) in the UD group, with no statistically significant difference between the two groups at the pooled analysis (Sample size: 1120; OR 1.14, 95% CI 0.73e1.77, P ¼ 0.57; no heterogeneity was found for I2 ¼ 0%; P ¼ 0.62) [Table 4, Section A], [Fig. 2]. Radiological Anastomotic Leakage The radiological AL rate, as reported in three RCTs, was 4.2% (14/ 330) in the D group and 5.6% (18/321) in the UD group.22,23,29 The meta-analysis revealed no significant difference in the radiological AL rate between the two groups (Sample size: 651; OR 0.74, 95% CI 0.36e1.52, P ¼ 0.42; no heterogeneity was found for I2 ¼ 0%; P ¼ 0.62) [Table 4, Section A], [Fig. 2].

in the D group, and 18.6% (103/554) in the UD group experienced some postoperative complication, without any statistically significant difference (Sample size: 1120; OR 0.87, 95% CI 0.64e1.19, P ¼ 0.38; no heterogeneity was found for I2 ¼ 21%; P ¼ 0.29) [Table 4, Section A], [Fig. 2]. Wound infection The rate of wound infection was reported in three of the studies analysed.22,23,29 The overall wound infection rate was 5.4% (18/330) in the D group, and 5.3% (17/321) in the UD group. The metaanalysis showed that the incidence of wound infection was not associated with the insertion or not of a drain in the abdominal or pelvic cavities (Sample size: 651; OR 1.02, 95% CI 0.52e2.01, P ¼ 0.95; no heterogeneity was found for I2 ¼ 26%; P ¼ 0.26) [Table 4, Section A], [Fig. 2]. Pelvic sepsis

Mortality Three studies reported data on postoperative mortality. The overall mortality rate was 3.6% (12/330) in the D group and 4.4% (14/321) in the UD group. The meta-analysis showed that rates of death were equivalent in the two groups (Sample size: 651; OR 0.82, 95% CI 0.37e1.81, P ¼ 0.63; no heterogeneity was found for I2 ¼ 0%; P ¼ 0.99) [Table 4, Section A], [Fig. 2].

The rate of pelvic sepsis was reported in three studies23,24,29, but one of them did not observe any such event.29 Overall, the rate of pelvic sepsis after colorectal anastomoses was 9.7% (47/483) in the D group, and 10.5% (50/478) in the UD group, with no statistically significant difference (Sample size: 961; OR 1.21; 95% CI 0.37e3.94; P ¼ 0.75; heterogeneity was found for I2 ¼ 76%; P ¼ 0.04) [Table 4, Section A], [Fig. 2].

Overall morbidity

Postoperative bowel obstruction

Detailed data on the overall morbidity rate were provided by all of the included studies.22e24,29 In total, 16.6% (94/566) of patients

Three trials reported data on this outcome.23,24,29 The rate of postoperative bowel obstruction was 9.9% (51/514) in the D group

22,23,29

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

5

Pulmonary complication D ¼ drained; UD ¼ undrained; NR¼Not Reported; SD¼Standard Deviation; O¼Open Technique; L ¼ Laparoscopic Technique; IP¼Intra-peritoneal; EP ¼ Extra-peritoneal; S¼Stapled; HS¼Hand-sewn.

5 ± 1.6 6 8 566 554 314:221 308:218 242:324 234:320 220:294 220:286 228 214 445 443 (100) (11.5) (16.6) (100) (100) (58.7:41.3) (58.5:41.5) (42.7:57.3) (42.2:57.8) (42.8:57.2) (43.5:56.5) (44.3) (42.3) (100) 63.4 ± 3.1

124:109 128:108 37:196 33:203 15:218 15:221 NR NR e 64 ± 11.5 65.5 ± 11.4 e 158:75 236 233 158:78

e e 66 ± 14.1 64 ± 12.6 16:12 20:11 28 31

566 554 326:240 324:230 65 ± 1.0

NR 233

3 ± 1.4 28 28:0 0:31 NR

0:28

31:0

180 171 236 (76.3) (73.4)

5

27 147 (11.0) 16 31 (57.1) 68:177 184:63 e e 247 245 120:127 128:117 66 ± 17.8 66 ± 21.3

247 (100) 31 (100)

245 247:0 (100) 28 NR (100)

245:0

177:68

61:186

23 (9.3) 25 (80.6)

157

D UD D

31

UD

NR

D UD

NR

D

25:27

UD D

48:0

D UD D UD D

48 52

UD D

64 ± 20.1 58 ± 18.7

UD UD D D

28:24

22:26

8 52 48 52:0 6 (11.5) (16.6) (100) (100)

UD

20:28

NR

NR

25

7

and 6.9% (35/506) in the UD group, with no statistically significant difference (Sample size: 1020; OR 1.53, 95% CI 0.96e2.44, P ¼ 0.07; no heterogeneity was found for I2 ¼ 0%; P ¼ 0.56) [Table 4, Section A], [Fig. 2].

Sagar PM 1995 Merad F 1999 Brown SR 2001 Denost Q 2017 TOTAL

Site of Anastomosis IP:EP N (%) Operative Technique Emergency Bowel Operation N Preparation O:L N (%) and (%) antibiotic prophylaxis N (%) Age (Years): Mean ± SD Patients Sex (M:F) N. Author

Table 1 General characteristics of the patients as reported in the studies included for the Systematic Review and Meta-analysis.

Type of Anastomosis S:HS N (%)

Diverting Stoma: N Cancer (%) Indication N (%)

Duration of Drainage (Days) Mean ± SD

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

Three studies reported data on pulmonary complication.22,23,29 The rate of pulmonary complication was 4.8% (16/330) in the D group, and 5.9% (19/321) in the UD group. The meta-analysis showed no statistically significant difference between the two groups for this outcome of interest (Sample size: 651; OR 0.80, 95% CI 0.40e1.59, P ¼ 0.53; no heterogeneity was found for I2 ¼ 0%; P ¼ 0.97) [Table 4, Section A], [Fig. 2]. Reintervention for abdominal complication Three studies reported data on this outcome, with special attention to reoperation because of anastomotic complication.23,24,29 The rate of re-intervention for abdominal complications was equivalent between the D group (9.1%, 47/514) and the UD group (7.9%, 40/506), with no statistically significant difference (Sample size: 1020; OR 1.17; 95% CI 0.76e1.82; P ¼ 0.48; no heterogeneity was found for I2 ¼ 47%; P ¼ 0.15) [Table 4, Section A], [Fig. 2]. Postoperative hospital stay The pooled analysis of length of postoperative hospital stay included three studies.22,24,29 Overall, D and UD groups showed an equivalent length of postoperative hospital stay, without any statistically significant difference (Sample size: 628; 10.7 ± 2.6 Days in the D group versus 10.2 ± 1.9 Days in the UD group; SMD 0.92; 95% CI -0.81 e 2.66; P ¼ 0.30; heterogeneity was found for I2 ¼ 71%; P ¼ 0.03) [Table 4, Section A], [Fig. 2]. Subgroup analysis: extraperitoneal colorectal anastomoses The subgroup analysis of extraperitoneal colorectal anastomoses23,24,29 showed that Clinical AL was reported in 13.6% (45/330) of patients in both the D and UD groups (Sample size: 660; OR 1.00; 95% CI 0.64e1.56; P ¼ 0.99; no heterogeneity was found for I2 ¼ 0%; P ¼ 0.97). No significant difference was reported for overall morbidity (Sample size: 528; OR 0.92; 95% CI 0.39e2.21; P ¼ 0.86; heterogeneity was found for I2 ¼ 54%; P ¼ 0.14), or reintervention for abdominal complications (Sample size: 528; OR 0.94; 95% CI 0.54e1.64; P ¼ 0.83; no heterogeneity was found for I2 ¼ 47%; P ¼ 0.17). Conversely, the postoperative bowel obstruction rate was slightly higher in the D group (18.7%, 50/267) than in the UD group (12.6%, 33/261) although the P value of 0.05 did not reach statistical significance (Sample size: 528; OR 1.61; 95% CI 1.00e2.60; P ¼ 0.05; no heterogeneity was found for I2 ¼ 0%; P ¼ 0.61) [Table 4, Section B], [Fig. 3]. Discussion This comprehensive systematic review and meta-analysis of four randomized controlled trials, including a total of 1120 patients, currently supports with high-quality evidence that using a prophylactic drain following colorectal anastomosis does not prevent AL, has no role in the management of AL, and does not decrease the clinical sequelae of AL-related complications. Many series reported a variety of risk factors for AL after colorectal surgery, including age, preoperative nutritional status,

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

6

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

Table 2 Risk of Bias in the published Randomized Controlled Trials (By the Cochrane Risk of Bias Tool) and the Oxford Quality Scoring System. Study

Blinding of Patients, Personnel, Blinding of Study Random sequence Allocation Outcome and Outcome Assessors concealment type generation Assessment (Selection Bias) (Performance Bias) (Selection Bias) (Detection Bias)

Incomplete Outcome Data (Attrition Bias)

Selective Reporting (Reporting Bias)

Final Oxford Other Sources Judgement of Bias

Sagar PM 1995 Merad F 1999 Brown SR 2001 Denost Q 2017

RCT

Low Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

Low Risk Low Risk

3

RCT

Low Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

Low Risk Low Risk

3

RCT

Low Risk

Low Risk

Unclear Risk

Low Risk

Low Risk

Low Risk

Low Risk Low Risk

3

RCT

Unclear Risk

Low Risk

Unclear Risk

Unclear Risk

Low Risk

Low Risk

Low Risk Unclear Risk

2

RCT ¼ Randomized Controlled Trial. Oxford Quality Scoring System: 3 ¼ High range of quality score; 2 ¼ Low range of quality score.

Table 3 Quality of evidence according to the GRADE criteria (Clinical Anastomotic Dehiscence, Radiological Anastomotic Dehiscence, Mortality, Overall Morbidity, Wound Infection, Pelvic Sepsis, Postoperative Bowel Obstruction, Pulmonary Complication, Re-intervention for Abdominal Complication, Postoperative Hospital Stay). QUALITY ASSESSMENTa

Outcome N. of Studies

N. of Patients

Study Design

Risk of Bias

Inconsistency Indirectness Imprecision Publication Bias

Overall Quality of Evidence

Clinical Anastomotic Leakage

4

1120

4 RCT

Not Serious

Serious

Not Serious Not Serious

High Quality

Radiological Anastomotic Leakage

3

651

3 RCT

Serious

Serious

Serious

Not Serious

Low Quality

Mortality Overall Morbidity

3 4

651 1120

3 RCT 4 RCT

Not Serious Not Serious

Not Serious Not Serious Not Serious Not Serious Not Serious Not Serious

High Quality High Quality

Wound Infection

3

651

3 RCT

Not Serious

Not Serious Not Serious Not Serious

High Quality

Pelvic Sepsis Postoperative Bowel Obstruction

2 3

961 1020

2 RCT 3 RCT

Not Serious Not Serious

Serious Serious Not Serious Serious

Not Serious Not Serious

Low Quality High Quality

Pulmonary Complication

2

651

2 RCT

Not Serious

Not Serious Serious

Not Serious

High Quality

Re-intervention for Abdominal Complication Postoperative Hospital Stay

3

1020

3 RCT

Not Serious

Not Serious Serious

Not Serious

High Quality

3

628

3 RCT

Not Serious Not Serious Serious Not Serious Not Serious Serious Not Serious Not Serious Not Serious Not Serious

Not Serious

Not Serious Not Serious Not Serious

High Quality

a

Quality of Evidence according to the GRADE (Grading of Recommendations Assessment, Development and Evaluation) criteria.

cardiovascular and respiratory comorbidities, occurrence of adverse intraoperative events, and the presence of a defunctioning stoma. Tumour localization in the middle and lower third of the rectum, particularly with an anastomotic height of 5 cm or less from the anal verge, have also been considered clinical risk factors for AL.36,37 Reported clinical AL in our study was 8.5% and 7.6% in the drained and undrained groups, respectively, with no statistically significant difference at the pooled analysis. Our results were in keeping with those reported in the contemporary literature, in which clinical AL rates after colorectal resections have varied from 3% to 12%, depending on the timing of operations (emergency versus elective surgery), anastomotic level (intra- versus extraperitoneal), and surgeon's expertise in colorectal surgery.23,38e40 Although there is an evidence-based for avoidance of drains after intraperitoneal colorectal anastomoses unless excessive bleeding or other technical issues at the time of surgery occur19,41, on the other hand after low rectal resections pelvic drainage is still common place.42 Dissection around the rectum results in a large raw surface adjacent to an extraperitoneal anastomosis, and if the presacral space is entered fluid re-absorption and anastomotic healing may

be hampered as the surfaces involved are large, raw and nonperitonealized.22 These are all variables that, together with the reactive tissue hyperemia after preoperative radiotherapy, inhibit the physiological reabsorption of intra-abdominal fluids. Our systematic review and meta-analysis, stratified by site and level of anastomosis, showed no benefit of PD for extraperitoneal rectal anastomosis with regard to clinical AL rate (13.6% in both the drained and undrained groups), and the other secondary outcomes, such as overall morbidity (20.2% versus 24.9%), and the need for reinterventions for abdominal complications (10.1% versus 10.7%). With respect to the subgroup analysis of patients submitted to low anterior resections with extraperitoneal anastomosis, the postoperative bowel obstruction rate was slightly higher in the drained group than in the undrained group (18.7% versus 12.6%), although there was not strong evidence at the pooled analysis that routine PD has an effect on the incidence of this adverse outcome. The meta-analysis by Guerra et al. reported a higher risk of postoperative bowel obstruction after routine drainage positioning, but this finding must be interpreted with caution, as the small cohort of patients submitted to low anterior resection with extraperitoneal anastomosis enrolled in this systematic review and meta-analysis does not allow firm conclusions about the effect of PD on

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

NR 7.5 ± 7.2 12.2 ± 9 9.8 ± 2.3 NR 7 ± 7.8 12.2 ± 9 9.6 ± 2.6 NR e 28 (12.0) 28 (10.7) NR 3 (9.7) 24 (10.2) 27 (10.1) NR 2 (7.1) NR 2 (7.1) NR 2 (6.4) NR 2 (6.4) NR 1 (3.6) 32 (13.9) 33 (12.6) NR 3 (9.7) 47 (19.9) 50 (18.7) NR e 45 (19.3) 45 (19.3) NR e 35 (14.8) 35 (14.8) NR 3 (10.7) NR 3 (10.7) NR 5 (16.1) NR 5 (16.1) NR 6 (21.4) 59 (25.3) 65 (24.9) NR 10 (32.2) 44 (18.6) 54 (20.2) D ¼ drained; UD ¼ undrained; SD¼Standard Deviation; NR¼Not Reported.

NR 1 (3.6) NR 1 (3.6) NR 1 (3.2) NR 1 (3.2) NR 3 (10.7) NR 3 (10.7) NR 1 (3.2) NR 1 (3.2) 8 2 (7.1) 35 (15.0) 45 (13.6) 8 2 (6.4) 35 (14.8) 45 (13.6) 69 28 233 330 63 31 236 330 Merad F 1999 Brown SR 2001 Denost Q 2017 TOTAL

UD D UD D UD D UD D UD D UD D UD D UD D UD D

UD

D

UD D

11 ± 1.1 NR 7.5 ± 7.2 12.2 ± 9 12.2 ± 9.0 10.2 ± 1.9 13 ± 1.1 NR 7 ± 7.8 12.2 ± 9 12.2 ± 9.0 10.7 ± 2.6

Postoperative Hospital Stay Re-intervention for Abdominal Complication N (%) Pulmonary Postoperative Bowel Obstruction Complication N (%) N (%) Pelvic Sepsis N (%) Wound Infection N (%) Overall Morbidity: N (%) Mortality N(%) Patients Clinical Anastomotic Radiological Anastomotic N. Leakage Leakage N (%) N (%)

Section B ¼ EXTRA-PERITONEAL COLORECTAL ANASTOMOSES

Author

UD D UD

NR 12 (4.9) e 28 (12.0) 28 (12.0) 40 (7.9) NR 20 (8.1) 3 (9.7) 24 (10.2) 24 (10.2) 47 (9.1)

D UD

4 (8.3) 13 (5.3) 2 (7.1) NR NR 19 (5.9) 4 (7.7) 10 (4.0) 2 (6.4) NR NR 16 (4.8)

D UD

NR 2 (0.8) 1 (3.6) 32 (13.9) 32 (13.9) 35 (6.9) NR 1 (0.4) 3 (9.7) 47 (19.9) 47 (19.9) 51 (9.9)

D UD

NR 5 (2.0) e 45 (19.3) 45 (19.3) 50 (10.5) NR 12 (4.8) e 35 (14.8) 35 (14.8) 47 (9.7)

D UD

e 14 (5.7) 3 (10.7) NR NR 17 (5.3) 3 (5.8) 10 (4.0) 5 (16.1) NR NR 18 (5.4)

D UD

4 (8.3) 34 (13.9) 6 (21.4) 59 (25.3) 59 (25.3) 103 (18.6) 7 (13.5) 33 (13.4) 10 (32.2) 44 (18.6) 44 (18.6) 94 (16.6)

D UD

3 (6.2) 10 (4.0) 1 (3.6) NR NR 14 (4.4) 3 (5.8) 8 (3.2) 1 (3.2) NR NR 12 (3.6)

D UD

3 (6.2) 12 (4.9) 3 (10.7) NR NR 18 (5.6) 2 (3.8) 11 (4.4) 1 (3.2) NR NR 14 (4.2)

D UD

2 (4.2) 3 (1.2) 2 (7.1) 35 (15.0) 35 (15.0) 42 (7.6) 5 (9.6) 6 (2.4) 2 (6.4) 35 (14.8) 35 (14.8) 48 (8.5)

UD D

48 245 28 233 233 554

D

52 247 31 236 236 566 Sagar PM 1995 Merad F 1999 Brown SR 2001 Denost Q 2017 Denost Q 2017 TOTAL

Mortality: N (%) Patients Clinical Anastomotic Radiological Anastomotic N. Leakage Leakage N (%) N (%) Author

Section A ¼ INTRA & EXTRA-PERITONEAL COLORECTAL ANASTOMOSES

Table 4 Clinical outcomes of drained and undrained colorectal anastomoses.

Overall Morbidity: N (%)

Wound Infection N (%)

Pelvic Sepsis N (%)

Pulmonary Postoperative Bowel Obstruction Complication N (%) N (%)

Re-intervention for Abdominal Complication N (%)

Postoperative Hospital Stay N (%)

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

7

postoperative bowel obstruction.20 The observed variations of AL incidence worldwide may be partially explained by institutional and individual differences in patient selection, operative technique, and perioperative care. Furthermore, the lack of a generally accepted definition and severity of grading of AL has significant impact upon clinical research, and carries challenges when trying to compare different series of patients. According to the definition provided by the International Study Group of Rectal Cancer, AL following anterior resection of the rectum should be defined as a communication between the intraand extraluminal compartments owing to a defect of the integrity of the intestinal wall at the anastomosis between the colon and rectum or the colon and anus.43 Furthermore, a pelvic abscess adjacent to the anastomosis should be considered as AL, even if no communication can be demonstrated with the colonic lumen at the anastomosis.4,25,44 In contrast to previous evidence, the meta-analysis by Rondelli et al., enrolling both randomized and non-randomized studies, showed that the presence of a prophylactic drain reduced the incidence of extraperitoneal colorectal AL and the rate of reinterventions after anterior rectal resections. However, the protective value of PD was supported by the data from non-randomized studies only, as the subgroup analysis of randomized controlled trials did not show any benefit for the use of PD.45 Furthermore, the large Dutch trial published by Peeters et al. showed that, on multiple regression analysis, the absence of a pelvic drainage and a defunctioning stoma were the factors associated with anastomotic dehiscence. More specifically, the presence of one or more pelvic drains after surgery was strongly associated with a lower AL: 9.6% of patients with pelvic drainage had leakage, compared with 23.5% of patients without a drain, and 8.2% of patients with a defunctioning ileostomy or colostomy had a leak, compared with 16% without a stoma. Therefore, the authors concluded that, in an attempt to minimize the risk of clinical AL, the construction of a defunctioning stoma and the placement of one or more drains in the presacral space seems advisable for patients with both proximal and distal rectal tumours.25 However, data from this study were derived from a prospective trial that investigated the efficacy of short-term neoadjuvant radiotherapy in patients with rectal cancer, and it was not designed to investigate the role of PD on AL. Thus, the lack of randomization for presence of a drain, the presence of a performance bias, and the indirectness related to the use of substitute or surrogate endpoints in place of the planned outcome measure do not allow firm conclusions on the evidence for the value of PD in this study. In our study, mean duration of drainage in days was 4.6 ± 1.4. Maeda et al. recently assumed that an extended period of pelvic drainage for up to 6 days after operation might allow the detection of all cases of subtle AL prior to hospital discharge.46 However, a long drainage time is related with a higher risk of clinically significant infection in several reports, thus suggesting nosocomial contamination of surgical drains by coagulase-negative staphylococci, enterococci and anaerobes.47,48 Although it was expected that drains would represent a potential entry site for bacteria, our meta-analysis showed that the absence or presence of PD does not affect the rate of wound infections. Anastomotic leak may be diagnosed at different time points postoperatively, although it often becomes evident between the 5th and 9th postoperative day.49,50 Current screening methods for AL involve clinical observation, blood tests (such as CRP) and imaging studies such as CT scan and/or water-soluble contrast radiography.51 The reported sensitivity of CT scan in the early postoperative period is low, ranging from 15% to 52%, with a

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

8

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

Fig. 2. Intra & extra-peritoneal colorectal anastomoses. Meta-analyses of clinical anastomotic leakage [A], radiological anastomotic leakage [B], mortality [C], morbidity [D], wound infection [E], pelvic sepsis [F], postoperative bowel obstruction [G], pulmonary complication [H], Re-intervention for abdominal complication [I], postoperative hospital stay [L].

reported negative predictive value of 73% and false negative rates up to 53%.52,53 A recent study by Gessler et al. has demonstrated that CT scan was negative in almost 25% of patients who had been later diagnosed with AL. Moreover, when CT scan was positive for AL, it took a mean of 8.5 days before leakage was confirmed compared to 4.3 days in patients who were diagnosed during reoperations, thus advancing the hypothesis that a prompt second look in the operating theatre would be preferable when an AL is clinically suspected.50 Important limitations of the present study were the small number of randomized controlled trials that have reported on this subject to date, especially regarding extraperitoneal anastomoses, and the fairly extended period of time through which the enrolled studies were published. Therefore, the results coming from the subgroup analysis of extraperitoneal colorectal anastomosis must be interpreted with caution. Another potential limitation of our study derives from the low percentage of patients submitted to laparoscopic colorectal resections in both groups (22.8% and 23.5% in the drained and undrained groups, respectively). This issue may hamper any firm conclusion on the potential role of PD in the contemporary era, where, following the results of a number of major international randomized trials, minimally invasive approaches for colorectal resection are now incorporated into clinical practice in many settings.54,55 Indeed, feasibility of laparoscopic colorectal resections have been demonstrated, and its main benefits include lower intraoperative blood loss, faster return of bowel function, reduced length of stay with better cosmetic appearance compared to open surgery, without compromise to oncological outcomes.56e58 In a recent multicenter international study, two thirds of patients underwent a planned laparoscopic operation, showing the high

implementation of laparoscopic approach in contemporary practice.4,59 Furthermore, leak rates between open and laparoscopic approaches in these trials were not different, so it may be reasonable to extrapolate the findings of our study to the laparoscopic era. Conclusions The present systematic review and meta-analysis of randomized controlled trials investigating the value of PD following colorectal anastomoses shows that routine drain placement does not have beneficial or harmful effects on the incidence of anastomotic leakage and on the management of intra-abdominal complications. High-quality multicenter randomized controlled trials and subsequent clustered meta-analyses are needed to better assess the value of prophylactic drainage following laparoscopic low rectal resections with extraperitoneal anastomosis, and therefore further validate our findings. Our findings do not, however, support the routine use of PD after colorectal anastomoses. Contributions of authors MP: Study conception and design, literature search, acquisition, interpretation and analysis of data; drafting and critically revising the article for important intellectual content; and final approval of the version to be published. SDS: Study conception and design, literature search, acquisition, interpretation and analysis of data; drafting and critically revising the article for important intellectual content; editing and revising the English for the final version to be published; and final approval of the version to be published. JD: Interpretation and analysis of data; drafting and critically revising the article for important intellectual content; editing and revising the English for the final version to be published; and final

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

9

Fig. 3. Extra-peritoneal colorectal anastomoses. Meta-analyses of clinical anastomotic leakage [A], overall morbidity [B], postoperative bowel obstruction [C], Re-intervention for abdominal complication [D].

approval of the version to be published. JA: Study conception and design, literature search, acquisition, interpretation and analysis of data; drafting and critically revising the article for important intellectual content; and final approval of the version to be published. FB: Interpretation and analysis of data; drafting and critically revising the article for important intellectual content; and final approval of the version to be published. FV: Interpretation and analysis of data; critically revising the article for important intellectual content; and final approval of the version to be published. IR: Interpretation and analysis of data; critically revising the article for important intellectual content; and final approval of the version to be published. KJ: Interpretation and analysis of data; critically revising the article for important intellectual content; and final approval of the version to be published. FA: Interpretation and analysis of data; critically revising the article for important intellectual content; and final approval of the version to be published. AP: Study conception and design, literature search, acquisition, interpretation and analysis of data; drafting and critically revising the article for important intellectual content; and final approval of

the version to be published. Conflicts of interest and source of funding Mauro Podda, Salomone Di Saverio, Justin Davies, Jenny Atzeni, Francesco Balestra, Francesco Virdis, Isabella Reccia, Kumar Jayant, Ferdinando Agresta, and Adolfo Pisanu, have no conflict of interest to declare. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Ethical approval No ethical approval was required for this article. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.amjsurg.2019.05.006.

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

10

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx

References 1. Nesbakken A, Nygaard K, Lunde OC. Outcome and late functional results after anastomotic leakage following mesorectal excision for rectal cancer. Br J Surg. 2001;88:400e404. 2. Bell SW, Walker KG, Rickard MJ, Sinclair G, Dent OF, et al. Anastomotic leakage after curative anterior resection results in a higher prevalence of local recurrence. Br J Surg. 2003;90:1261e1266. €o €k O, Rutegård J, Simert G, Sjo €dahl R. Defunctioning 3. Matthiessen P, Hallbo stoma reduces symptomatic anastomotic leakage after low anterior resection of the rectum for cancer: a randomized multicenter trial. Ann Surg. 2007;246: 207e214. 4. 2017 European Society of Coloproctology (ESCP) collaborating group. Association of mechanical bowel preparation with oral antibiotics and anastomotic leak following left sided colorectal resection: an international, multi-centre, prospective audit. Colorectal Dis. 2018 Sep;20(Suppl 6):15e32. https:// doi.org/10.1111/codi.14362. 5. Boccola MA, Lin J, Rozen WM, Ho YH. Reducing anastomotic leakage in oncologic colorectal surgery: an evidence-based review. Anticancer Res. 2010;30: 601e607. 6. Alves A, Panis Y, Trancart D, Regimbeau JM, Pocard M, et al. Factors associated with clinically significant anastomotic leakage after large bowel resection: multivariate analysis of 707 patients. World J Surg. 2002;26:499e502. 7. Pommergaard HC, Gessler B, Burcharth J, Angenete E, Haglind E, et al. Preoperative risk factors for anastomotic leakage after resection for colorectal cancer: a systematic review and meta-analysis. Colorectal Dis. 2014;16:662e671. 8. Vallance A, Wexner S, Berho M, Cahill R, Coleman M, et al. A collaborative review of the current concepts and challenges of anastomotic leaks in colorectal surgery. Colorectal Dis. 2017 Jan;19(1):O1eO12. https://doi.org/10.1111/ codi.13534. 9. Eriksen MT, Wibe A, Norstein J, Haffner J, Wiig JN, Norwegian Rectal Cancer Group. Anastomotic leakage following routine mesorectal excision for rectal cancer in a national cohort of patients. Colorectal Dis. 2005;7:51e57. 10. Bertelsen CA, Andreasen AH, Jørgensen T, Harling H. Danish Colorectal Cancer Group. Anastomotic leakage after anterior resection for rectal cancer: risk factors. Colorectal Dis. 2010;12:37e43. 11. Tsujinaka S, Kawamura YJ, Konishi F, Maeda T, Mizokami K. Pelvic drainage for anterior resection revisited: use of drains in anastomotic leaks. ANZ J Surg. 2008;78:461e465. 12. Averbach AM, Chang D, Koslowe P, Sugarbaker PH. Anastomotic leak after double-stapled low colorectal resection. Dis Colon Rectum. 1996;39:780e787. 13. Sagar PM, Couse N, Kerin M, May J, MacFie J. Randomized trial of drainage of colorectal anastomosis. Br J Surg. 1993;80:769e771. 14. Johnson CD, Lamont PM, Orr N, Lennox M. Is a drain necessary after colonic anastomosis? J R Soc Med. 1989;82:661e664. 15. Galandiuk S, Fazio VW. Postoperative irrigation-suction drainage after pelvic colonic surgery. A prospective randomized trial. Dis Colon Rectum. 1991;34: 223e228. 16. Merad F, Yahchouchi E, Hay JM, Fingerhut A, Laborde Y, Langlois-Zantain O. Prophylactic abdominal drainage after elective colonic resection and suprapromontory anastomosis: a multicenter study controlled by randomization. French Associations for Surgical Research. Arch Surg. 1998;133:309e314. 17. Menahem B, Vallois A, Alves A, Lubrano J. Prophylactic pelvic drainage after rectal resection with extraperitoneal anastomosis: is it worthwhile? A metaanalysis of randomized controlled trials. Int J Colorectal Dis. 2017;32: 1531e1538. 18. Smith SR, Connolly JC, Crane PW, Gilmore OJ. The effect of surgical drainage materials on colonic healing. Br J Surg. 1982;69:153e155. 19. Zhang HY, Zhao CL, Xie J, Ye YW, Sun JF, et al. To drain or not to drain in colorectal anastomosis: a meta-analysis. Int J Colorectal Dis. 2016;31:951e960. 20. Guerra F, Giuliani G, Coletta D, Boni M, Rondelli F, et al. A meta-analysis of randomized controlled trials on the use of suction drains following rectal surgery. Dig Surg. 2018;35:482e490. 21. Karliczek A, Jesus EC, Matos D, Castro AA, Atallah AN, et al. Drainage or nondrainage in elective colorectal anastomosis: a systematic review and metaanalysis. Colorectal Dis. 2006;8:259e265. 22. Sagar PM, Hartley MN, Macfie J, Mancey-Jones B, Sedman P, May J. Randomized trial of pelvic drainage after rectal resection. Dis Colon Rectum. 1995;38: 254e258. 23. Merad F, Hay JM, Fingerhut A, Yahchouchi E, Laborde Y, et al. Is prophylactic pelvic drainage useful after elective rectal or anal anastomosis? A multicenter controlled randomized trial. French Association for Surgical Research. Surgery. 1999;125:529e535. 24. Denost Q, Rouanet P, Faucheron JL, Panis Y, Meunier B, et al. To drain or not to drain infraperitoneal anastomosis after rectal excision for cancer: the GRECCAR 5 randomized trial. Ann Surg. 2017;265:474e480. 25. Peeters KC, Tollenaar RA, Marijnen CA, Klein Kranenbarg E, Steup WH, et al. Risk factors for anastomotic failure after total mesorectal excision of rectal cancer. Br J Surg. 2005;92:211e216. 26. Yeh CY, Changchien CR, Wang JY, Chen JS, Chen HH, et al. Pelvic drainage and other risk factors for leakage after elective anterior resection in rectal cancer patients: a prospective study of 978 patients. Ann Surg. 2005;241:9e13. 27. Hoffmann J, Shokouh-Amiri MH, Damm P, Jensen R. A prospective, controlled study of prophylactic drainage after colonic anastomoses. Dis Colon Rectum.

1987;30:449e452. 28. Hagmuller E, Lorenz D, Werthmann K, Trede M. Effects and risks of drainage after elective colon resection. A prospective, randomized controlled clinical study. Chirurg. 1990;61:266e271. 29. Brown SR, Seow-Choen F, Eu KW, Heah SM, Tang CL. A prospective randomised study of drains in infra-peritoneal rectal anastomoses. Tech Coloproctol. 2001;5: 89e92. 30. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ. 2009 Jul 21;339:b2700. https://doi.org/10.1136/bmj.b2700. 31. Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.3.5. The Cochrane Collaboration; 2014. Available from: www.handbook.cochrane.org. 32. Balshem H, Helfand M, Schünemann HJ, Oxman AD, Kunz R, et al. GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol. 2011;64: 401e406. 33. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Contr Clin Trials. 1996;17:1e12. 34. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13. https://doi.org/10.1186/1471-2288-5-13. 35. DerSimonian R, Laird N. Meta-analysis in clinical trials. Contr Clin Trials. 1986;7: 177e188. €o €k O, Rutegård J, Sjo €dahl R. Intraoperative adverse events 36. Matthiessen P, Hallbo and outcome after anterior resection of the rectum. Br J Surg. 2004;91: 1608e1612. 37. Alberts JC, Parvaiz A, Moran BJ. Predicting risk and diminishing the consequences of anastomotic dehiscence following rectal resection. Colorectal Dis. 2003;5:478e482. 38. Eckmann C, Kujath P, Schiedeck TH, Shekarriz H, Bruch HP. Anastomotic leakage following low anterior resection: results of a standardized diagnostic and therapeutic approach. Int J Colorectal Dis. 2004;19:128e133. 39. Marusch F, Koch A, Schmidt U, Wenisch H, Ernst M, et al. Early postoperative results of surgery for rectal carcinoma as a function of the distance of the tumor from the anal verge: results of a multicenter prospective evaluation. Langenbecks Arch Surg. 2002;387:94e100. 40. Bokey EL, Chapuis PH, Fung C, Hughes WJ, Koorey SG, et al. Postoperative morbidity and mortality following resection of the colon and rectum for cancer. Dis Colon Rectum. 1995;38:480e486. 41. Urbach DR, Kennedy ED, Cohen MM. Colon and rectal anastomoses do not require routine drainage: a systematic review and meta-analysis. Ann Surg. 1999;229:174e180. 42. Kahokehr A, Robertson P, Sammour T, Soop M, Hill AG. Perioperative care: a survey of New Zealand and Australian colorectal surgeons. Colorectal Dis. 2011;13:1308e1313. 43. Rahbari NN, Weitz J, Hohenberger W, Heald RJ, Moran B, et al. Definition and grading of anastomotic leakage following anterior resection of the rectum: a proposal by the International Study Group of Rectal Cancer. Surgery. 2010;147: 339e351. 44. Buchs NC, Gervaz P, Secic M, Bucher P, Mugnier-Konrad B, et al. Incidence, consequences, and risk factors for anastomotic dehiscence after colorectal surgery: a prospective monocentric study. Int J Colorectal Dis. 2008;23: 265e270. 45. Rondelli F, Bugiantella W, Vedovati MC, Balzarotti R, Avenia N, et al. To drain or not to drain extraperitoneal colorectal anastomosis? A systematic review and meta-analysis. Colorectal Dis. 2014;16:O35eO42. https://doi.org/10.1111/ codi.12491. 46. Maeda H, Okamoto K, Namikawa T, Akimori T, Kamioka N, et al. Rarity of late anastomotic leakage after low anterior resection of the rectum. Int J Colorectal Dis. 2015;30:831e834. 47. Walker AP, Krepel CJ, Gohr CM, Edmiston CE. Microflora of abdominal sepsis by locus of infection. J Clin Microbiol. 1994;32:557e558. 48. Chylak J, Michalska W, Drews M, Marciniak R, Lange M. Comparison of bacterial flora found in the peritoneal cavity and drains after intraabdominal surgery. Med Sci Monit. 2000;6:285e290. 49. Hyman N, Manchester TL, Osler T, Burns B, Cataldo PA. Anastomotic leaks after intestinal anastomosis: it's later than you think. Ann Surg. 2007;245:254e258. 50. Gessler B, Eriksson O, Angenete E. Diagnosis, treatment, and consequences of anastomotic leakage in colorectal surgery. Int J Colorectal Dis. 2017;32: 549e556. 51. den Dulk M, Noter SL, Hendriks ER, Brouwers MA, van der Vlies CH, et al. Improved diagnosis and treatment of anastomotic leakage after colorectal surgery. Eur J Surg Oncol. 2009;35:420e426. 52. Nicksa GA, Dring RV, Johnson KH, Sardella WV, Vignati PV, Cohen JL. Anastomotic leaks: what is the best diagnostic imaging study? Dis Colon Rectum. 2007;50:197e203. 53. Khoury W, Ben-Yehuda A, Ben-Haim M, Klausner JM, Szold O. Abdominal computed tomography for diagnosing postoperative lower gastrointestinal tract leaks. J Gastrointest Surg. 2009;13:1454e1458. 54. Deijen CL, Vasmel JE, de Lange-de Klerk ESM, Cuesta MA, Coene PLO, et al. Tenyear outcomes of a randomised trial of laparoscopic versus open surgery for colon cancer. Surg Endosc. 2017;31:2607e2615. 55. Guillou PJ, Quirke P, Thorpe H, Walker J, Jayne DG, et al. Short-term endpoints

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006

M. Podda et al. / The American Journal of Surgery xxx (xxxx) xxx of conventional versus laparoscopic-assisted surgery in patients with colorectal cancer (MRC CLASICC trial): multicentre, randomised controlled trial. Lancet. 2005;365:1718e1726.  P, et al. Laparos56. Lacy AM, García-Valdecasas JC, Delgado S, Castells A, Taura copy-assisted colectomy versus open colectomy for treatment of nonmetastatic colon cancer: a randomised trial. Lancet. 2002;359:2224e2229. 57. Stevenson AR, Solomon MJ, Lumley JW, Hewett P, Clouston AD, et al. Effect of laparoscopic-assisted resection vs open resection on pathological outcomes in rectal cancer: the ALaCaRT randomized clinical trial. JAMA. 2015;314:

11

1356e1363. 58. Podda M, Saba A, Porru F, Pisanu A. Systematic review with meta-analysis of studies comparing single-incision laparoscopic colectomy and multiport laparoscopic colectomy. Surg Endosc. 2016;30:4697e4720. 59. 2017 and 2015 European Society of Coloproctology (ESCP) collaborating groups. The impact of conversion on the risk of major complication following laparoscopic colonic surgery: an international, multicentre prospective audit. Colorectal Dis. 2018 Sep;20(Suppl 6):69e89. https://doi.org/10.1111/ codi.14371.

Please cite this article as: Podda M et al., Prophylactic intra-abdominal drainage following colorectal anastomoses. A systematic review and meta-analysis of randomized controlled trials, The American Journal of Surgery, https://doi.org/10.1016/j.amjsurg.2019.05.006