Preventive drainage and anastomotic leakage after colorectal surgery

Abstract

BACKGROUND

Anastomotic leakage is a serious complication following colorectal surgery, associated with increased morbidity, reoperation, and poor long-term outcomes. Although preventive drainage is widely used, its effectiveness remains controversial. Previous meta-analyses have generally pooled data across different drainage modalities, potentially obscuring modality-specific effects. We hypothesized that the association between preventive drainage and anastomotic leakage varies by drainage modality.

AIM

To evaluate the association between preventive drainage modalities and postoperative anastomotic leakage after colorectal surgery.

METHODS

A systematic review and meta-analysis was conducted using PubMed and Web of Science databases up to April 25, 2026. Comparative studies evaluating preventive drainage vs no drainage following colorectal surgery were included. The primary outcome was postoperative anastomotic leakage. Pooled risk ratios (RRs) with 95% confidence intervals (CIs) were estimated using random-effects models. Subgroup analyses were performed according to drainage modality, and sensitivity and publication bias analyses were conducted.

RESULTS

Thirty-six studies (38 comparisons) involving 17690 patients were included. Preventive drainage was associated with a lower risk of anastomotic leakage than no drainage (RRs = 0.67, 95%CI: 0.54-0.83). Transanal drainage significantly reduced leakage risk (RRs = 0.58, 95%CI: 0.45-0.74), whereas pelvic drainage showed no significant benefit (RRs = 0.80, 95%CI: 0.53-1.19). Other drainage methods were also not associated with a significant reduction in leakage risk (RRs = 1.38, 95%CI: 0.89-2.13). Subgroup differences were significant (P = 0.0026). Sensitivity analyses showed stable results, with pooled RRs ranging from 0.65 to 0.69 after sequential study exclusion. Begg’s test showed no significant publication bias (P = 0.0576).

CONCLUSION

Preventive drainage is associated with reduced anastomotic leakage following colorectal surgery, primarily mediated by transanal drainage; the benefits of other drainage modalities remain unproven.

Key Words: Colorectal surgery; Anastomotic leakage; Prophylactic drainage; Transanal drainage tube; Pelvic drainage

Core Tip: This meta-analysis evaluated whether preventive drainage reduces anastomotic leakage after colorectal surgery and stratified outcomes by drainage modality. Thirty-six studies comprising 17690 patients were included. The overall reduction in leakage risk was primarily driven by transanal drainage. Pelvic drainage and other drainage methods showed no statistically significant preventive effect. These findings suggest that preventive drainage should not be considered a uniform strategy, and that the choice of modality should be based on drainage modality, anastomotic level, operative setting, and patient risk.

INTRODUCTION

Colorectal surgery is a cornerstone treatment for a range of intestinal disorders, including malignant tumors[1]. It is also indicated for a wide range of benign conditions, such as diverticular disease, inflammatory bowel disease, intestinal volvulus, ischemic bowel disease, traumatic injuries, and congenital anomalies[2-4]. When intestinal continuity must be restored, bowel anastomosis performed to restore both the structural integrity and functional capacity of the intestine[5,6]. Nevertheless, anastomotic leakage remains one of the most prevalent and severe complications of colorectal surgery, and its occurrence is intricately linked to multiple factors, including patient-related factors, anastomotic site, and intraoperative conditions[7,8]. If left untreated, it can lead to intra-abdominal infection, sepsis, multi-organ dysfunction, and even death[9]. Among the various strategies for preventing anastomotic leakage, the use of drainage remains a longstanding yet contentious practice[10,11]. Proponents argue that the potential advantages may facilitate evacuation of perianastomotic fluid, reduce bacterial load, enable early detection of leakage via drainage fluid characteristics or amylase levels, and decrease the risk of pelvic abscess formation[12,13]. Conversely, opponents contend that drainage tubes, act as foreign bodies that, may increase the risk of retrograde infection, cause pain, limit mobility, and prolong hospitalization, while numerous studies have failed to demonstrate any definitive benefit in reducing anastomotic leakage incidence[14-16]. The choice of drainage approach varies according to the surgical procedure undertaken, encompassing transanal drainage, pelvic drainage, and other novel or individualized approaches. This diversity makes the evaluation of prophylactic drainage particularly challenging[11,17,18].

Therefore, against the backdrop of considerable variation in current clinical practice, there is an urgent need for a systematic and comprehensive meta-analysis to determine the efficacy and safety of prophylactic drainage in preventing anastomotic leakage following colorectal surgery. This study synthesizes evidence from published observational research, systematically evaluating the impact of prophylactic drainage on the incidence of postoperative anastomotic leakage, while exploring potential differences across disease types, anastomotic locations, and drainage modalities, thereby providing robust, evidence-based support for clinical decision-making.

MATERIALS AND METHODS

Literature retrieval strategy

A comprehensive search was conducted across Web of Science and PubMed to identify records published up to April 25, 2026. The detailed search strategy is provided in Table 1. Additionally, we manually searched the reference lists of other meta-analysis or systematic reviews to identify additional eligible studies (Table 1).

Table 1 Literature search strategy.

Database	Search build	Results
PubMed	#1: “Colorectal” OR “rectal” OR “rectum” [MeSH] OR “Colon”[MeSH]; #2: “Drainage”[Mesh] OR “Drainage, Postural” OR “Manual Lymphatic Drainage” OR “Negative-Pressure Wound Therapy” OR “Paracentesis” OR “Negative Pressure Wound Therapy” OR “draining” OR “Suction” OR “drain”; #1 AND #2	3183
Web of Science	#1: ((((((((KP=(Drainage)) OR KP=(Drainage, Postural)) OR KP=(Manual Lymphatic Drainage)) OR KP=(Negative-Pressure Wound Therapy)) OR KP=(Negative Pressure Wound Therapy)) OR KP=(Paracentesis)) OR KP=(draining)) OR KP=(Suction)) ORKP=(drain) and Preprint Citation Index(Exclude -Database); #2: (((KP=(Colorectal)) OR KP=(rectal)) OR KP=(rectum)) OR KP=(Colon) and Preprint Citation Index(Exclude-Database); #1 AND #2	413

Inclusion and exclusion criteria

We applied predefined inclusion and exclusion criteria to screen relevant studies.

Inclusion criteria: (1) Participants: Patients undergoing colorectal-related surgical procedures, regardless of sex, age, or ethnicity; (2) Intervention: Postoperative drainage vs no drainage; (3) Primary outcome measure: Overall incidence of anastomotic leakage; and (4) Publication status: Original article.

Exclusion criteria: (1) Studies with insufficient or inaccessible data; (2) Review articles; (3) Animal studies; (4) Conference abstracts or proceedings; (5) Unavailable full texts; (6) Unextractable data; (7) Lack of detailed drainage method information; and (8) Duplicate publications or overlapping datasets.

Literature screening and data extraction

Two researchers independently screened the literature and extracted data, with cross-checking. Discrepancies were resolved through discussion with a third investigator. The extracted data included the first author’s name, year of publication, intervention details, and outcomes.

Statistical analyses

Meta-analyses were conducted using R software (version 4.3.1) and the meta package (version 8.0-2). The function metabin was used analysis with sm = “RR” (risk ratio)[19]. According to the results of heterogeneity quantifying and testing, the fixed-effect was used when the heterogeneity is low (P value from the χ² test > 0.05 and I² statistic value < 50%) and the random-effect was used when the heterogeneity is high (P value from the χ² test ≤ 0.05 or I² statistic value ≥ 50%). As no anastomotic leakage events occurred in the control groups of two studies[20,21], a continuity correction was applied, and sensitivity analysis was performed to examine the influence of these two studies on the results. To explore sources of heterogeneity, sensitivity, and subgroup analyses were conducted. Specifically, the metabin function was used to perform subgroup analysis with sm = “RR”, with subgroups defined by drainage modality, while the metainf function was used to conduct sensitivity analyses with pooled = “random”. The forest function was used to generate forest plots of meta-analysis, subgroup analysis, and sensitivity analyses. To assess publication bias, funnel plots were generated using the funnel function, and Begg’s test was performed using the metabias function. P < 0.05 on Begg’s test was considered indicative of statistically significant publication bias.

RESULTS

Search results

A total of 39 studies were initially identified investigating the relationship between drainage and anastomotic leakage following colorectal surgery. However, the single-center data reported by Luberto et al[22] and Crippa et al[23] from Italy may have been included in the multicenter research by Guadagni et al[24] on behalf of the Italian ColoRectal Anastomotic Leakage study group. Similarly, Kawada et al reported two studies in 2014[25] and 2018[26] with markedly similar datasets, suggesting possible overlap between them. Consequently, three articles, Luberto et al[22], Crippa et al[23], and Kawada et al[25], were excluded from further analysis.

Ultimately, 36 studies were retained for inclusion (Table 2)[27-54]. Among these, two studies, Challine et al[18] and Akiyoshi et al[55], reported outcomes for both transanal drainage and pelvic drainage regarding postoperative anastomotic leakage. Data from both drainage methods were incorporated into the meta-analysis. In total, data from 36 publications was analyzed, comprising transanal drainage (n = 28), pelvic drainage (n = 8), and other drainage modalities, including abdominal drainage (n = 1) and negative pressure drainage (n = 1; Figure 1).

Figure 1  Flow diagram of included studies.

Table 2 Distribution of drainage methods among included studies.

Ref.	Transanal drainage	Pelvic drain	Other drainage
Wang et al[27], 2025	Yes
Wang et al[28], 2024	Yes
Ho et al[29], 2024	Yes
Guadagni et al[24], 2024			Abdominal drain
Zhang et al[30], 2023	Yes
Sueda et al[31], 2023	Yes
Liang et al[17], 2022	Yes
Kuk et al[32], 2022	Yes
Zhao et al[11], 2021	Yes
Lee et al[33], 2021			A closed suction drain
Tamura et al[34], 2021	Yes
Challine et al[18], 2020	Yes	Yes
Carboni et al[35], 2020	Yes
Wang et al[36], 2020	Yes
Li et al[20], 2020	Yes
Kawada et al[26], 2018	Yes
Ito et al[37], 2017	Yes
Denost et al[38], 2017		Yes
Goto et al[39], 2017	Yes
Brandl et al[40], 2016	Yes
Yang et al[41], 2016	Yes
Matsuda et al[42], 2016	Yes
Hidaka et al[43], 2015	Yes
Lee et al[44], 2015	Yes
Kim et al[45], 2015	Yes
Adamova et al[21], 2014	Yes
Nishigori et al[46], 2014	Yes
Zhao et al[47], 2013	Yes
Xiao et al[48], 2011	Yes
Akiyoshi et al[55], 2011	Yes	Yes
Bülow et al[49], 2006	Yes
Peeters et al[50], 2005		Yes
Brown et al[51], 2001		Yes
Merad et al[52], 1999		Yes
Scott et al[53], 1996		Yes
Sagar et al[54], 1995		Yes

Drainage and postoperative anastomotic fistula risk

The heterogeneity assessment (Figure 2) indicated substantial variability among the included studies (I² = 60.21%, P < 0.0001), necessitating the use of a random-effects model. The meta-analysis (Figure 2) yielded a pooled RR of 0.67 68 [95% confidence interval (CI): 0.5455-0.83], demonstrating a statistically significant association (P = 0.0002).

Figure 2 Meta-analysis results from 38 studies in 36 papers. In the random effects model, risk ratio (RR) = 0.67, 95%CI: 0.54-0.83; P = 0.0002.

Subgroup analysis by drainage modality

Subgroup analyses were performed according to drainage modality. Heterogeneity was observed in the transanal drainage tube (TDT), pelvic drainage, and other drainage subgroups, with I² values of 47.1% (P = 0.0034), 62.5% (P = 0.0093), and 52.3% (P = 0.1478), respectively. Given the clinical and methodological differences among studies, including variations in patient characteristics, surgical procedures, and drainage strategies, random-effects models were used for all subgroup analyses.

TDT use was associated with a reduced risk of postoperative anastomotic leakage (RR = 0.58, 95%CI: 0.45-0.74). In the pelvic drainage subgroup, the pooled RR was 0.80 (95%CI: 0.53-1.19). Because the 95%CI crossed the null value, a statistically significant reduction in anastomotic leakage risk was not demonstrated. For other drainage methods, the pooled RR was 1.38 (95%CI: 0.89-2.13), which was also not statistically significant. However, the test for subgroup differences was statistically significant (P = 0.0026), indicating that the effect estimates may differ across drainage modalities (Figure 3).

Figure 3 Subgroup analyses of drainage models and anastomotic leakage after colorectal surgery risk. CI: Confidence interval; RR: Risk ratio.

Overall, the association between preventive drainage and a lower risk of anastomotic leakage was mainly observed with TDT. Although the point estimate for pelvic drainage was below 1, the evidence was insufficient to establish a definite protective effect because the CI included the null value. For other drainage methods, the pooled RR was greater than 1 and did not suggest a reduction in anastomotic leakage risk. Nevertheless, as the result was not statistically significant, an increased risk of anastomotic leakage could not be confirmed. Therefore, the true effects of pelvic drainage and other drainage methods remain uncertain.

It should also be noted that only a limited number of studies were included in the subgroup of other drainage methods. This subgroup included abdominal drainage, closed-suction drainage, and other drainage strategies, which differed in indication, technique, and perioperative management. Accordingly, the pooled estimate for this subgroup requires careful consideration.

Sensitivity analyses

We performed a sensitivity analysis to explore potential sources of heterogeneity. The results (Figure 4) showed that, after excluding any single study, the pooled estimates ranged from 0.65 (95%CI: 0.53-0.80) to 0.69 (95%CI: 0.56-0.86), and all analyses remained statistically significant. These findings indicate that the primary results are robust. Furthermore, exclusion of any individual study did not materially alter the pooled RR relative to the overall RR, supporting an association between drainage and reduced anastomotic leakage risk.

Figure 4 Sensitivity analyses were performed to investigate potential sources of heterogeneity and showed the main result was robust. CI: Confidence interval; RR: Risk ratio.

Publication bias

A funnel plot was constructed to assess potential publication bias, and Begg’s test was performed to examine bias among the included studies. The funnel plot, together with the Begg’s test result (P = 0.0576; Figure 5), indicated no evidence of publication bias, suggesting that the findings were not substantially influenced by such bias.

Figure 5 Funnel plot and Begg test plot. A: The funnel plot on the anastomotic leakage; B: Begg test and Begg test plot were performed to confirm that there was no significant publication bias between the drain and the occurrence of anastomotic leakage after colorectal surgery, P = 0.0576.

DISCUSSION

The value of prophylactic drainage in colorectal surgery has been extensively evaluated in various meta-analyses; however, these studies were limited by uncertainty regarding the drainage methods used, as well as by the relatively small number of eligible studies and patients included[56-58]. This likely contributed to heterogeneity arising from variations in drainage methods. Accordingly, this meta-analysis adopted a subgroup approach according to drainage modality, to separately evaluate transanal drainage, pelvic drainage, and other techniques, to elucidate potential differences in their efficacy in reducing the anastomotic leakage risk[59].

In the overall analysis, preventive drainage was associated with a lower risk of anastomotic leakage; however, this association was primarily driven by transanal drainage. Transanal drainage may reduce intraluminal pressure proximal to the anastomosis, facilitate evacuation of bowel contents, and decrease local mechanical tension, thereby providing more favorable conditions for anastomotic healing[43]. Nevertheless, differences in patient selection, anastomotic level, diverting stoma use, surgical approach, and perioperative management existed among the included studies. Accordingly, the findings support an association between transanal drainage and a lower risk of anastomotic leakage, but do not establish a uniform benefit for all patients undergoing colorectal anastomosis.

The pooled estimate for pelvic drainage was below unity, although the CI included the null value. Thus, a definite preventive effect of pelvic drainage on anastomotic leakage was not demonstrated, and its routine use in high-risk patients cannot be recommended based solely on the present findings[39,60]. In clinical practice, the potential role of pelvic drainage may be more closely related to the early detection and drainage of postoperative pelvic collections or infection than to the direct prevention of anastomotic leakage[61]. The decision to place a pelvic drain should therefore be based on anastomotic level, extent of pelvic dissection, intraoperative contamination, patient risk profile, and surgical judgment.

Other drainage methods likewise showed no reduction in anastomotic leakage risk. However, this subgroup included only a few studies, encompassing abdominal drainage, closed-suction drainage, and other strategies. Substantial heterogeneity existed in their indications, drain placement, duration of drainage, and perioperative management. These interventions should not be considered a homogeneous group, and no uniform clinical recommendation can be made. Furthermore, the observed effects of drainage may be influenced by tumor stage, anastomotic level, surgical approach, use of diverting stoma, definitions of anastomotic leakage, and follow-up protocols. Therefore, differences among drainage modalities should not be interpreted as evidence that one strategy is superior in all clinical settings[62,63].

Clinically, the present study focused on preventive drainage rather than management of established anastomotic leakage. Once anastomotic leakage occurs, treatment should be individualized based on the patient’s overall condition, the presence of peritonitis or sepsis, the extent of the leak, and the presence of intra-abdominal or pelvic collections. In hemodynamically stable patients with a contained leak and no generalized peritonitis, bowel rest, intravenous antibiotics, nutritional support, and image-guided percutaneous or transanal drainage may be considered. Patients with generalized peritonitis, persistent sepsis, hemodynamic instability, or failure of conservative or interventional treatment, timely surgical source control is required. Surgical options include abdominal lavage and drainage, proximal fecal diversion, anastomotic repair, or anastomotic takedown, depending on the degree of contamination, anastomotic condition, and the patient’s general status[64-66].

Anastomotic leakage is associated with postoperative infection, reintervention, prolonged hospital stay, a higher risk of permanent stoma, and reduced quality of life. Among patients with colorectal cancer, anastomotic leakage has also been associated with poorer long-term survival and a higher risk of local recurrence, particularly after rectal cancer surgery. However, the available evidence regarding long-term oncological outcomes is primarily from observational studies and may be influenced by tumor stage, surgical complexity, severity of postoperative infection, and delayed adjuvant treatment. Therefore, the association between anastomotic leakage and poor long-term outcomes should be interpreted as a clinical association rather than a causal relationship[67-69].

In summary, the association between preventive drainage and a lower risk of anastomotic leakage after colorectal surgery was primarily driven by transanal drainage, whereas sufficient evidence was not available to support a preventive effect of pelvic drainage or other drainage methods[70]. The observed differences in efficacy among drainage modalities underscore the importance of tailoring the choice to the specific surgical context, such as low rectal anastomosis, the precise location of the anastomosis, the surgeon’s expertise, and the patient’s individual risk profile[25,38]. While the results exhibit commendable robustness, the presence of moderate heterogeneity and the possibility of publication bias highlight certain limitations within the current body of evidence. Future research should therefore focus on large-scale, multicenter, rigorously designed randomized controlled trials directly compare the effectiveness of different drainage methods across varied patient populations and surgical settings. Such studies should also explore the long-term value of drainage in reducing AL, facilitating recovery, lowering reoperation rates, and improving overall prognosis. Moreover, attention should be paid to the potential drawbacks of drainage, such as infection, pain, and disturbances in bowel function, as well as its cost, benefit profile, thereby informing the development of more precise and individualized drainage strategies in clinical practice[9,62].

CONCLUSION

Preventive drainage was associated with a lower risk of postoperative anastomotic leakage after colorectal surgery; however, this association was mainly observed with transanal drainage. A statistically significant preventive effect was not demonstrated for pelvic drainage, and the clinical value of other drainage methods remains uncertain. Therefore, preventive drainage should not be applied routinely as a uniform strategy. Its use should be individualized according to the drainage modality, anastomotic level, operative setting, and patient-specific risk factors. Further well-designed multicenter randomized controlled trials are needed to clarify the effects of different drainage strategies on anastomotic leakage and other clinically relevant outcomes.

ACKNOWLEDGEMENTS

We are grateful to Zhang Yue and Zhang Su for their contributions in verifying the content of the article.