Self-expandable metal stent (SEMS) placement or emergency surgery as palliative treatment for obstructive colorectal cancer: a systematic review and meta-analysis

aimed to exclusively analyze palliative treatment for primary obstructive colorectal cancer, with early complication rate as a primary outcome. A systematic literature search was performed on studies comparing palliative SEMS and emergency surgery. Corresponding authors were contacted for additional data. Eighteen studies were selected (1518 patients). Early complication rate was 13.6 % for SEMS and 25.5 % for emergency surgery (Odds Ratio (OR) 0.46, 95 % confidence interval (CI) 0.29 (cid:0) 0.74). Mortality was 3.9 % and 9.4 % (OR 0.44, 0.28 (cid:0) 0.69). Stomas were present in 14.3 % and 51.4 % of patients (OR 0.17, 0.09 (cid:0) 0.31). More late complications occurred after SEMS (23.2 % versus 9.8 %, OR 2.55, 1.70 – 3.83), mostly due to SEMS obstruction. In conclusion, SEMS placement seems the preferred treatment of obstructing colorectal cancer in the palliative setting.


Introduction
Colorectal cancer is a common malignancy worldwide with approximately 20 % of patients diagnosed with disseminated disease at presentation.The majority of these patients are treated with palliative intent (Suarez et al., 2010).Acute colonic obstruction might be the initial presentation of stage IV colorectal cancer, or can develop during the course of the disease (Jullumstro et al., 2011).
Patients are often in a poor clinical condition due to multiple days of reduced food intake and weight loss.Given the patients' limited expected life span and the desire to proceed to systemic chemotherapy as soon as possible, resection of the primary tumor may be of questionable benefit.Alternatively, a decompressing stoma can be constructed, but this may never be reversed with potential deteriorated quality of life as a result (Jansen et al., 2010).Furthermore, complications related to emergency surgery might delay start of systemic therapy.As an alternative to emergency surgery, self-expandable metal stent (SEMS) placement has been introduced as a minimally invasive decompressing intervention in patients with bowel obstruction.It has been suggested that SEMS placement results in lower mortality and morbidity rates and a lower chance of having a stoma compared to emergency surgery in the palliative setting (Faragher et al., 2008;Fernandes et al., 2016).In contrast to the curative setting, the oncological concerns about SEMS with a potentially increased risk of recurrent disease are not relevant if performed as a palliative procedure (Amelung et al., 2018).Several meta-analyses on palliative SEMS and emergency surgery have been published, with considerable heterogeneity of both inclusion criteria and results (Liang et al., 2014;Ribeiro et al., 2018;Takahashi et al., 2015;Z hao et al., 2013).Patients with extracolonic malignancies and colorectal cancer were often analyzed in one group, while several studies have shown lower technical and clinical success rates of SEMS for extracolonic malignancy (Kim et al., 2012(Kim et al., , 2011;;Kim et al., 2013;Moon et al., 2014;Shin et al., 2008;Trompetas et al., 2010).No distinction was made between acute and subacute obstructions in these reviews, despite the fact that the European guideline discourages prophylactic stenting.Furthermore, most meta-analyses are relatively outdated, and the most recent one exclusively included randomized controlled trials (RCT) with a total of only 125 patients (Ribeiro et al., 2018).
Therefore, the aim of the current systematic review and metaanalysis was to compare SEMS placement and emergency surgery as palliative treatment for bowel obstruction solely caused by colorectal cancer, with early complication rate as primary outcome measure.Sensitivity analyses were conducted for type of study, year of publication, type of surgery, and urgency of obstruction.

Materials and methods
This systematic review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (Moher et al., 2009).

Search strategy
With the aid of a clinical librarian, a systematic literature search was performed in MEDLINE (PubMed), EMBASE, the Cochrane Library, and Web of Science for studies comparing SEMS with emergency surgery as palliative treatment in patients with obstructive colorectal carcinoma (Supplementary text).The final search was performed on January 29th 2020.

Inclusion and exclusion criteria
Inclusion criteria were comparative studies in which patients were included with 1) acute or imminent large bowel obstruction, 2) caused by colorectal cancer, 3) located on either the left or right side, 4) treated with palliative intention, and 5) with data available on at least early complication rate.Exclusion criteria were studies in which patients were included with 1) an extracolonic malignancy without separate results on patients with a colonic obstruction caused by colorectal cancer, 2) a benign cause of large bowel obstruction without separate results on patients with colorectal cancer, 3) non-comparative studies, 4) age < 18 years, 5) animal studies, 6) studies not written in English, and 7) conference abstracts, reviews, letters, comments, and case reports.

Data extraction
Titles, abstracts, and subsequent full-text articles were independently scanned for eligibility by the first two reviewers (JV and DU).Discrepancies were resolved through discussion, and in case of any doubt resolved with the senior author (JvH).References of finally included articles were checked manually for additional studies.

Outcomes
The primary outcome was early complication rate.Secondary outcomes included technical success of SEMS, clinical success in decompressing the colon, major early complication rate, 30-day and/or inhospital mortality, stoma formation, hospital stay, interval to start or continuation of systemic therapy, late complication rate, major late complication rate, survival, quality of life, and treatment costs.

Definitions
Early complication rate included any complication occurring within 30 days after the first intervention, occurring either before or after discharge.Late complication rate included any complication occurring after 30 days.Major early and late complications were defined as complications requiring a surgical, endoscopic, or radiological reintervention.Technical success was defined as correct positioning of the stent, confirmed by either endoscopy or imaging.Clinical success was defined as clinical evidence of intestinal transit or passage of flatus or stools after the initial procedure.

Request letters
In case any of the outcomes were not reported in the included studies, the corresponding authors were contacted by e-mail and requested to deliver these data.In addition, for studies in which the emergency surgery group also consisted of decompressing stoma procedures, separate data were requested on decompressing stoma patients for subgroup analyses.Request letters were also sent to the authors of RCTs that were initially not eligible for the current study.All extracted data from the original publications supplemented with the requested data were used for statistical analysis.

Methodological quality assessment
Quality assessment was performed by two independent reviewers (JV and DU) according to The Oxford Centre for Evidence-Based medicine Levels of Evidence (Group OLoEW, 2011).Non-randomized articles were evaluated using The New-Castle-Ottawa Quality Assessment Scale for cohort studies (Wells et al., 2000).RCTs were screened according to the guidelines of the Cochrane Collaboration (Higgins et al., 2011) for potential bias by random sequence generation, allocation concealment, blinding of outcome assessment, blinding of participants, selective reporting, assessment of incomplete data outcome, and other potential sources of bias.In order to evaluate publication bias, a funnel plot of our primary outcome was created (Sterne and Egger, 2001;Sterne et al., 2000).

Statistical analysis
Odds Ratios (ORs) and weighted mean differences (MD) were calculated for dichotomous and continuous variables, respectively, both with 95 % confidence intervals (95 % CI), and outcomes below 1 favoring SEMS.Heterogeneity was assessed using the I 2 statistic (I 2 -value of ≥ 50 % represented significant heterogeneity).A random-effects model was used for analyses, considering the variability of surgical techniques and populations between the included studies.Sensitivity analyses for the primary outcome were performed based on study type (RCTs and prospective observational cohort studies versus retrospective studies), year of publication (< 2014 versus ≥ 2014), type of emergency surgery (only decompressing stoma, only resection, or all types of surgery combined), and urgency of the obstruction (subacute, also reported as 'imminent' in literature, acute or unspecified).A two-tailed P value < 0.05 was considered statistically significant.Data analysis was performed with Review Manager (RevMan version 5.3.Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) and MedCalc version 18.5 (MedCalc Software).

Quality of life and performance
Only one of the included studies reported on quality of life (Fiori et al., 2019) using the EQ-5D-FL questionnaire (© EuroQol Group, Rotterdam, the Netherlands) and Karnofsky performance scale for functional impairment.SEMS resulted in significantly better quality of life than resection at one month, while quality of life was similar at 3 months, and worse for SEMS than resection after 6 months.Karnofsky performance scale was better after SEMS at 1 month, but without significant differences at 3 and 6 months.

Treatment costs
Only one study assessed treatment costs (Abelson et al., 2017), but without clear definition.A median of 65.228 US dollars ) was calculated for SEMS versus 73.662 US dollars  for emergency surgery (p = 0.06).

Discussion
In contrast to previously published meta-analyses (Liang et al., 2014), the current meta-analysis specifically focused on patients with obstruction caused by primary colorectal cancer in the palliative setting.Based on 18 studies including three RCTs, early complication rate after SEMS was 50 % lower than after emergency surgery (OR 0.46), but with significant heterogeneity among the studies, and without significant difference in major early complications.Sensitivity analyses confirmed the favorable treatment effects following SEMS in the different subgroups.Thirty-day or in-hospital mortality and hospital stay were also in favor of SEMS.There was no difference in survival between SEMS and emergency surgery.Surgery was associated with fewer late complications.Fewer stomas were constructed in patients treated with SEMS.
Goals of palliative treatment of obstructive colorectal cancer essentially differ from the curative setting.Important outcomes include the prevention of complications, avoiding stoma formation, and limiting hospital stay, while survival is one of the main endpoints in the curative setting.The Dutch Stent-in I trial was designed to show superiority of SEMS for imminent obstruction in the palliative setting, but inclusion was discontinued in 2006, after 21 patients had been randomized.A high number of serious adverse events in the non-surgical arm occurred, consisting of SEMS perforation in three of nine patients (van Hooft et al., 2008).However, several studies afterwards, mostly dealing with acute obstruction, have shown more favorable results.The present meta-analysis confirms the initial hypothesis, showing that SEMS reduces the risk of short-term complications, avoids stomas, and shortens hospital stay.These results are in line with two prior meta-analyses (Takahashi et al., 2015;Zhao et al., 2013), but are in contrast to the meta-analyses by Ribeiro et al. and (Liang et al. (2014); Ribeiro et al., 2018).Both latter reviews included fewer patients and missed several eligible studies up to 2011 (Suarez et al., 2010;Faragher et al., 2008;Ptok et al., 2006;Karoui et al., 2007).This might explain contradictory findings, besides neglecting the degree of obstruction.Degree of obstruction was recently quantified by a Japanese group, who developed the ColoRectal Obstruction Scoring System (CROSS) (Group JCSSPR, 2012).Within this scoring system ranging from CROSS 0-2, a lower ability to eat soft solids results in a lower CROSS score.A recently pooled, post-hoc analysis of two prospective observational multicenter studies evaluated stricture degree in CROSS 0 (worse clinical state) versus CROSS 1 or 2 patients treated with SEMS as BTS (Ohki et al., 2020).Both clinical effectiveness and safety of SEMS were similar between CROSS 0 and CROSS 1 or 2 patients.Current sensitivity analysis was not able to show a significant impact of degree of obstruction on early complications.
It is important to notice that the emergency surgery groups in most of the included studies consist of both resection and decompressing stoma construction.However, a decompressing stoma can be constructed with a minimal surgical intervention by just making a small transverse incision in the upper abdomen, thereby avoiding a laparotomy.This translates into different clinical outcomes, as has been shown in the curative setting (Amelung et al., 2015).For this reason, SEMS should actually be compared with a similar surgical intervention that just aims to decompress the colon.In line with two previous meta-analyses, no significant difference in early complication rate was found between SEMS and decompressing stoma (Liang et al., 2014;Zhao et al., 2013), but the OR does suggest that there still might be an advantage of SEMS.A recent propensity score matched study on SEMS versus decompressing stoma in the curative setting did not reveal a clear preference and suggested an RCT comparing both techniques (Veld et al., 2020).
The role of resection of the primary tumor in stage IV colorectal cancer remains controversial.Recent comparative studies, including a meta-analysis, suggested improved survival after primary tumor resection (Venderbosch et al., 2011;Ha et al., 2018).However, selection bias may have influenced the results, and we have to await randomized studies (t Lam-Boer et al., 2014).Until proven otherwise, metastatic patients with an obstructing primary tumor should have the least invasive decompressing intervention to enable the earliest start of systemic therapy possible, and preferably not an emergency resection (Poultsides et al., 2009).
In contrast to fewer short-term complications, multiple metaanalyses reported more long-term complications for SEMS than emergency surgery (Liang et al., 2014;Takahashi et al., 2015;Zhao et al., 2013).This finding was confirmed in the current meta-analysis.One might question the relevance of this endpoint for the decision on the type of decompressing intervention in the emergency setting, especially considering the differences in clinical impact between certain complications.For example, stent migration can easily be managed with limited consequences for the patient, while late perforation requiring surgery is a severe complication.This requires a more balanced interpretation.
The substantial absolute difference in pooled short-term mortality between SEMS and emergency surgery confirms findings in some of the meta-analyses published earlier (Takahashi et al., 2015;Zhao et al., 2013).(Liang et al. (2014)) reported short-term mortality rates of 7.1 % and 11.6 %, respectively.In contrast, the meta-analysis of Ribeiro reported 30-day mortality rates of 6.3 % for SEMS and 6.4 % for emergency surgery, but only based on small RCTs with predominant subacute obstructions, and decompressing stoma as a surgical intervention (Ribeiro et al., 2018).
In line with earlier results (Ribeiro et al., 2018), a lower stoma rate was observed for SEMS than emergency surgery patients, also when solely analyzing studies on SEMS versus emergency resection.Stoma formation might negatively influence quality of life, which is especially important in the palliative setting (McCahill et al., 2002).In a small RCT on SEMS (n = 16) versus resection (n = 17) for subacute obstruction, quality of life was better after SEMS at 1 month, but better after resection at 6 months using the EQ-5D-5 L questionnaire (Fiori et al., 2019).Many re-obstructions observed after 30 days in the SEMS group may have contributed to this observation.However, the results of this small trial should be interpreted with caution.
Several limitations of the current meta-analysis must be taken into account.First, positive studies are more likely to be published, resulting in publication bias.However, visual inspection of the funnel plot suggested the absence of publication bias at least for early complication rate.Furthermore, results might have been influenced by strict inclusion criteria to increase homogeneity of the study populations.Only three RCTs with small numbers of patients fulfilled our inclusion criteria.Therefore, current level of evidence is almost exclusively based on cohort studies with all their inherent risks of bias.Finally, the emergency surgery groups still consisted of a wide variety of surgical procedures that may differ in invasiveness.Although an attempt was made to address this issue, separate analyses on SEMS versus decompressing stoma were hampered by relatively few studies and patients.
In conclusion, the current systematic review and meta-analysis on palliative treatment of colonic obstruction in patients with colorectal cancer suggests that SEMS results in better short-term outcomes than emergency surgery, with fewer stoma constructions and shorter hospital stay.

Declaration of Competing Interest
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Table 2
Morbidity, mortality, and survival for SEMS versus emergency surgery in the palliative setting.

Table 1
Study characteristics.In this study, sum of reported ASA scores results in 40 stent patients, although rest of the baseline characteristics are reported based on 38 stent patients.
DS = decompressing stoma, ER = emergency resection, ES = emergency surgery, SEMS = self-expandable metal stent, ES = emergency surgery, SD = standard deviation, NR = not reported.aRetrospective analysis of prospective database with ICD-codes.b Matched cohort.c Mean (range).d Type of surgery reported for only 49 of 70 patients, therefore shown percentages are based on a total of 49 e No ASA 4 patients.f ASA class unknown in 28 stent patients.g Median (standard deviation).h i Median of stent and emergency surgery patients combined.j ASA missing in 7 of 10 patients.k ASA missing in 4 of 24 patients.l Mean, no standard deviation provided.m Mean (standard deviation) of stent and emergency surgery patients combined.Pooled chemotherapy rate was 70.4 % (95 % CI 49.2-87.8)after SEMS and 69.5 % (95 % CI 46.4-88.

Table 3
Technical success, clinical success, hospital stay, stoma formation, and chemotherapy for SEMS versus emergency surgery in the palliative setting.
DS = decompressing stoma, ER = emergency resection, ES = emergency surgery, SEMS = self-expandable metal stent, ES = emergency surgery, SD = standard deviation, NR = not reported.a Mean (standard deviation or range not provided).b Mean (range).c Median, no range or interquartile range provided.d Patients with palliative chemotherapy were excluded from the study.