Patients with malignant large bowel obstruction (MLBO) account for 8–34% of those with colorectal carcinoma, which is considered a life-threatening condition that requires immediate intervention [1,2,3]. Emergency one-stage resection for MLBO is associated with significantly higher mortality and morbidity rates than after elective surgery because patients with MLBO are in a poor clinical condition due to several days of reduced intake and impaired intestinal function [4]. Staged surgeries comprised bowel resection and stoma creation, and subsequent closure of the stoma have been performed in some cases [5,6,7]. Regarding left-sided MLBO, many studies have reported good short-term outcomes and safety of long-term prognosis of preoperative decompression with self-expanding metallic stent (SEMS) compared to emergency surgery, which is called a bridge to surgery (BTS) [8,9,10,11,12]. Based on these results, BTS for left-sided MLBO came to be recommended in the updated version of the guidelines from the European Society of Gastrointestinal Endoscopy (ESGE) in April 2020 [8].

In contrast, right-sided MLBO has been traditionally managed with resection and primary anastomosis, which is considered to be safe based on old prospective data showing equivalent morbidity and mortality between emergency and elective surgery in right-sided MLBO [13,14,15]. However, recent studies have demonstrated significantly higher operative risks for emergency resection compared to elective surgery [16,17,18]. Although approximately 32–54% of MLBO cases are located in the right-sided colon, it has been reported that only 5% of reported BTS cases involve the right-sided colon [19]. Studies evaluating the safety and efficacy of BTS for right-sided MLBO are sparse, which has hampered definitive conclusions [20,21,22,23,24,25,26]. Recently, some retrospective cohort studies have shown the promising results for BTS compared to emergency resection (ER). Thus, we conducted a meta-analysis to evaluate the safety and effectiveness of BTS in patients with right-sided MLBO on short-term outcomes with a large sample size (n = 5,136).

Material and methods

The current meta-analysis met the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement 2020 [27].

Literature retrieval and study selection

A literature systematic search was performed in MEDLINE (PubMed), Google Scholar, and the Cochrane databases to identify relevant studies published throughout December 2020 (Fig. 1). The search was limited to English language and human studies. The search terms were (‘‘colon cancer” OR “colonic obstruction”) AND (“proximal” OR “right-sided”) AND (“resection” OR “surgery” OR colectomy” OR “colostomy” OR “stoma” OR “stent”). The related article’s function was used to broaden the search. Reference lists of all relevant publications were manually searched for additional studies that were initially overlooked using this search strategy. Methodological index for non-randomized studies (MINORS) tool was used to assess the risk of bias for individual studies [28]. The grading of recommendations assessment, development and evaluation (GRADE) methodology was applied for assessing quality of evidence, and reported in the results with the help of the GRADE Pro Software (McMaster University and Evidence Prime Inc, Ontario, Canada) [29].

Fig. 1
figure 1

Flow chart of study selection according to PRISMA guidelines

Full size image

Inclusion and exclusion criteria

The inclusion and exclusion criteria were defined priori. The inclusion criteria were as follows: (1) patients with acute malignant proximal or right-sided colonic obstruction, and (2) a comparative study evaluating morbidity and mortality between patients who underwent BTS or ER. ER was defined as emergency one-stage tumor resection. Exclusion criteria were defined as (1) duplicated studies and (2) studies in which predefined outcomes were not reported or it was impossible to extract the number of outcome events.

Data extraction

Each retrieved publication was evaluated independently by two investigators (S.K. and A.M.) for inclusion or exclusion. The following data were extracted from the included studies: the primary author’s name, year of publication, country in which the study was performed, number of institutions, design and duration of the study, number of included participants, and their characteristics, including age, sex, tumor-related variables, performer of SEMS placement, interval from SEMS placement to surgery, and all available short-term outcomes.

Data synthesis and statistical analysis

Dichotomous variables were analyzed by assessing the odds ratio (OR) of short-term adverse outcomes in terms of primary anastomosis, stoma construction, mortality, overall morbidity, surgical site infection, anastomotic leakage, and postoperative ileus, occurring with BTS compared to ER along with 95% confidence intervals (95% CIs). An OR < 1 favored the BTS group, and the point estimate of the OR was considered statistically significant at P < 0.05, if the 95% CI did not include the value 1. The pooled OR was calculated using the Mantel–Haenszel fixed-effects model or DerSimonian–Laird random-effects model to combine ORs for outcomes of interest. Meta-analysis was performed using Review Manager (Version 5.1) for Windows’ (Nordic Cochrane Center, Cochrane Collaboration; Copenhagen, Denmark; http://www.cc-ims.net/RevMan). Cochran’s Chi square-based Q statistic test was used to assess between-study heterogeneity. I2 was used to test for heterogeneity among the included studies. Study heterogeneity was measured using the χ2 and I2 statistics, with χ2 P < 0.05, and I2 ≥ 50% indicating heterogeneity [30]. A fixed-effect model was used to estimate the overall effect if the OR was homogenous; if the OR was non-homogenous, a random-effect model was used [31]. Publication bias was assessed by statistical analysis using Egger’s linear regression test and the rank correlation test (Begg’s test) using ‘‘WINPEPI’’ software (available at http://www.brixtonhealth.com/pepi4windows.html)32,33.

Results

Literature review and included studies

In total, 255 citations were matched during the initial screening. After reviewing the titles and abstracts, 149 studies were excluded. Ninety-nine studies were excluded after full-article evaluation. Therefore, the seven remaining studies [20,21,22,23,24,25,26] (published from 2014 to 2020) were included in the meta-analysis (Fig. 1). The background characteristics of the included studies are listed in Table 1. Five studies [21, 23,24,25,26] originated from Asia and two [20, 22] from the Netherlands. Six [20, 21, 23,24,25,26] of these were retrospective, and one [22] is prospective observational studies. Only one study [25], which is the largest population-based data, applied propensity score matched analysis to adjust potential background confounders. The definitions of right-sided colon in each study were as follows, from the cecum (1) to the hepatic flexure [21], (2) to proximal transverse colon [23], (3) to the entire transverse colon [20, 22, 24,25,26]. Of the 5136 patients included in the meta-analysis, 1662 (32.4%) underwent BTS and 3474 (67.6%) underwent ER. The mean ± standard deviation of MINORS score of included studies was 14.57 ± 1.72, which indicates a fair quality of evidence for non-randomized studies [28]. A comprehensive list of MINORS for included studies is available in Table 2.

Table 1 Basic characteristics of the seven included studies
Full size table
Table 2 MINORS assessment of included studies
Full size table

According to the GRADE criteria, overall quality of evidence was very low for postoperative complications, primary anastomosis, anastomotic leakage, surgical site infection, and ileus. It was low for mortality, stoma construction and laparoscopic surgery. (Table 3).

Table 3 GRADE evidence profile
Full size table

Postoperative complications

Postoperative complications were reported in seven studies. The postoperative complication rates in the BTS and ER groups were 19.3% (320/1662) and 31.3% (1088/3474), respectively. The heterogeneity test indicated χ2 = 9.35, and I2 = 36%, demonstrating homogeneity. Therefore, the fixed-model was adopted, and the OR was 0.78 (95% CI: 0.66–0.92) (Fig. 2). This meta-analysis demonstrated that BTS contributed to a significant reduction in postoperative complications compared to ER. We found no significant publication bias by Egger’s test (P = 0.169), or Begg’s test (P = 0.179). Five [19, 20, 23,24,25] of seven studies classified the severity of postoperative complications. However, meta-analysis of severe postoperative complications, which is Clavien–Dindo grading ≥III or requiring reintervention, demonstrated no significant difference between the BTS and the ER groups (OR = 0.98, 95% CI 0.68–1.41).

Fig. 2
figure 2

Meta-analysis of postoperative complications between bridge to surgery and emergency resection. Odds ratio is shown with 95% confidence intervals. ER Emergency resection, BTS Bridge to surgery

Full size image

Mortality

Mortality was reported in seven studies. The postoperative mortality rates in the BTS and ER groups were 0.9% (15/1,662) and 5.2% (181/3474), respectively. The heterogeneity test indicated χ2 = 1.88 and I2 = 0%, demonstrating homogeneity. Therefore, the fixed-model was adopted, and the OR was 0.51 (95% CI: 0.28–0.92) (Fig. 3a). This meta-analysis demonstrated that BTS contributed to a significant reduction in mortality compared to ER. We found no significant publication bias by Egger’s test (P = 0.749), or Begg’s test (P = 0.497).

Fig. 3
figure 3

Meta-analysis of postoperative mortality between bridge to surgery and emergency resection. Odds ratio is shown with 95% confidence intervals. ER Emergency surgery, BTS Bridge to surgery

Full size image

Other outcomes

The other evaluated outcomes are listed in Table 4 and shown in Supplementary Figs. 1–6. The rates of primary anastomosis favored the BTS group over ER group (97.8% [134/137] and 85.9% [1653/1925], respectively), and stoma construction was preferable in the BTS group (2.0% [33/1637] and 11.0% [378/3425], respectively). Both meta-analyses demonstrated significant differences between the groups (OR = 0.31, 95% CI 0.10–0.96; primary anastomosis rate: OR = 0.45, 95% CI 0.25–0.83 for the stoma construction rate) without between-study heterogeneity (χ2 = 0.11, I2 = 0%, P = 0.95, and χ2 = 5.09, I2 = 21%, and P = 0.28, respectively).

Table 4 Meta-analysis of other short-term outcomes
Full size table

Although only 15.7% (547/3474) of ER patients underwent laparoscopic surgery, BTS enabled it in 48.5% (806/1662) of the patients. The OR was significantly higher in the BTS group than in the ER group (OR = 0.21, 95% CI 0.10–0.42). However, the analysis showed significant heterogeneity (χ2 = 18.36, I2 = 67%, P < 0.01).

Regarding detailed postoperative complications, anastomotic leakage, surgical site infection, and ileus were analyzed. Meta-analyses of anastomotic leakage and surgical site infection demonstrated significantly favorable results in the BTS group over ER group (OR = 0.66, 95% CI 0.45–0.96 for the anastomotic leakage: OR = 0.62, 95% CI 0.46–0.82 for the surgical site infection) without between-study heterogeneity. The meta-analysis showed a trend toward frequent ileus in the BTS group, although the difference was not significant (OR = 1.28, 95% CI 0.96–1.71).

Discussion

This meta-analysis was performed to verify an overview of the current literature on the outcomes of acute right-sided MLBO. This review article ultimately included seven studies consisting of 5,136 patients. The results suggested that BTS with subsequent elective resection of right-sided MLBO is associated with significantly lower mortality and postoperative complications when compared to ER.

Although patients with MLBO should be decompressed immediately, emergency resection of the carcinoma is associated with high morbidity and mortality. It may also deteriorate long-term outcomes in that there is no interval for staging workup or screening of synchronous proximal lesions [34]. Thus, staged surgery has been widely performed, which separates decompression of the proximal bowel from resection of the tumor by constructing a stoma [5,6,7].

SEMS was first used by Dohmoto et al. in 1990 for palliation and used for BTS since 1994 [35, 36]. Since then, various reports have indicated that preoperative decompression of bowel obstruction using SEMS reduces postoperative complications and mortality [11, 12, 37]. Recently, the reports of BTS for left-sided MLBO, demonstrating no significant differences in oncologic outcomes compared with ER, were published successively [4, 9, 10, 38]. Based on these reports showing at least equivalent short-term and long-term outcomes compared with ER, recently updated ESGE guidelines for BTS for left-sided MLBO have been recommended within a sufficient shared decision-making process [8].

On the other hand, patients with right-side MLBO are generally managed with emergency resection and primary anastomosis because it is considered relatively safe [39]. However, Kobayashi et al. reported that emergency right hemicolectomy has a significantly higher 30-day mortality rate than elective hemicolectomy (6.0% vs. 0.7%, P < 0.001) in the National Clinical Database of Japan [14]. In addition, previous studies reported high postoperative mortality (8.8–14.5%) and morbidity (32–54.3%) rates in patients who underwent ER for right-sided MLBO [40, 41]. Therefore, the management of patients with right-sided MLBO has become a critical issue. Campbell et al. first reported the use of SEMS in the proximal transverse colon in 1997 [42]. SEMS placement for the proximal colon might be technically challenging because of the long distance from the anus to the obstruction site, increased colon tortuosity, and inappropriate colonic preparation. Further, previous studies showed lower success rates of SEMS placement for the right-sided MLBO than for the left-sided [43]. A recent post hoc analysis of a multicenter trial in Japan demonstrated that the tumor site in the right colon was independently associated with the technical difficulty of SEMS placement (OR 2.5; 95% CI: 1.61–4.01, P < 0.0001) [44]. These factors could have considerable potential for selection bias, and preferable results of BTS compared to ER might be derived from the result of selection bias for “easy-to-stent cases” in all included non-randomized studies for this meta-analysis. However, it is noteworthy that the largest included study by Sakamoto et al. [25] applied propensity score matching and instrumental variable analysis to confirm the robustness of the results. Moreover, they finally demonstrated a lower postoperative complication rate and less stoma creation in BTS than in ER, which was consistent with the results of this meta-analysis. Furthermore, sensitivity analyses excluding the largest study by Sakamoto et al. [25] did not affect the significance in the postoperative complication and maintained the favorable mortality of BTS, although the statistical significance had disappeared (Supplementary Table 1). According to a systematic review by Amelung et al. [19], this is the first systematic review in which BTS for right-sided MLBO has a significantly lower postoperative complication rate (0.8% vs. 23.9%) and surgical mortality (0% vs. 10.8%) than ER. Although this report did not include a comparative study, many patients were included.

The influence to outcomes of the interval from SEMS placement to surgery is one of the concerns in the BTS strategy. Theoretically, longer interval could contribute to a reduction of postoperative complications due to the stabilization of patients’ condition. Our subgroup analyses, divided by median interval of 20 days, demonstrated significant reductions of postoperative complications (OR = 0.45; 95% CI: 0.22–0.92, P = 0.03 for less than 20 days, and OR = 0.81; 95% CI: 0.68–0.96, P = 0.02 for 20 days or more) regardless of length of the interval.

This study had several limitations. As with all systematic reviews, the strength of the conclusions depends on the quality of the primary studies, and there are no randomized controlled trials (RCTs) in our analysis. To draw a conclusion, a reliable RCT is preferred, but conducting RCTs with an adequate sample size on this topic is difficult due to lower incidence of right-sided MLBO compared to left-sided. Furthermore, there was heterogeneity between papers because the sample size, type of SEMS, technical quality of procedures, pathological staging, study design, and follow-up were different. There was also heterogeneity in the definitions of morbidity and mortality. These heterogeneities affect the results considerably.

In conclusion, it should be considered that the postoperative mortality and morbidity rates of surgery for right-sided colon cancer, especially emergency surgery, are higher than previously reported, and BTS may result in a reduction of postoperative complications and mortality for right-sided obstructive colon cancer.