Propensity score‐matched comparison of stenting as a bridge to surgery and emergency surgery for acute malignant left‐sided colonic obstruction

Colorectal cancer (CRC) is the third most commonly diagnosed cancer and accounts for approximately 10 % of cancer-related deaths worldwide [1]. A total of 7–29 % of CRC patients present with acute large bowel obstruction [2]. These patients are mainly in advanced disease stages and exhibit poor clinical conditions, and most of them need urgent surgical interventions [3]. In addition to the high morbidity and mortality rates of acute malignant colonic obstruction (AMCO) itself in the emergency setting, conventional emergency surgery (ES) is frequently followed by severe complications, such as anastomotic leakage and surgical site infection [4, 5]. Moreover, ES is accompanied by a high risk of stoma creation, which often becomes permanent [6].

In the 1900 s, self-expanding metal stents (SEMSs) were first utilized in the treatment of palliative malignant rectal obstruction [7]. Since then, SEMSs have been increasingly used as an alternative option for AMCO treatment either as palliation or as a bridge to surgery (BTS). BTS provides a period of “optimization” of the patients’ clinical condition, allowing adequate oncological staging, accurate anesthetic assessment, optimal colonic preparation and early peri-operational chemotherapy [2, 5]. Therefore, decompression by SEMS insertion can transform the ES into an elective procedure, leading to subsequent medical stabilization and an increased rate of primary anastomosis. SEMS seems to be an effective and safe technique, and the benefits of short-term outcomes have been confirmed by recent meta-analyses compared to ES [8‐10]. However, SEMS as BTS is associated with annoying stent-related complications, including perforation, stent migration and reobstruction [6, 11]. In addition, BTS has the potential risk of tumor cell dissemination after stent insertion, leading to worse oncologic outcomes [12, 13]. Moreover, randomized controlled trials (RCTs) and related meta-analyses comparing BTS and ES are limited to specific populations and provide inconsistent results on surgical and oncological outcomes. Most studies evaluating the outcomes of BTS are conducted at single centers, and optimal matching is often lacking [2, 5, 14]. The objective of this study was to compare the surgical and oncological outcomes of BTS and ES in the management of AMLCO using propensity score-matched (PSM) analysis.

We performed a retrospective observational study of patients with AMLCO treated with curative intent, in three hospitals (Shanghai Tongji Hospital, Shanghai, China; Ningbo First Hospital, Ningbo, China; Ningbo Second Hospital, Ningbo, China) from April 2012 to December 2019. All patients who underwent resection for AMLCO were identified from in-hospital medical records. Data on baseline characteristics and short-term surgical outcomes were collected from medical records. Data on long-term oncological outcomes were collected upon follow-up treatment or survey. This study was approved by the institutional review board of Shanghai Tongji Hospital, Ningbo First Hospital and Ningbo Second Hospital.

This study is a multicenter trial that analyzed retrospective data from three hospitals with considerable follow-up periods where both BTS and ES were available. The results of this observational study suggest that SEMS placement as BTS in patients with AMLCO was associated with fewer primary stoma creations, higher total in-hospital costs and more perineural invasion. Patients in the BTS group had a higher recurrence rate and poorer 3-year and 5-year overall survival, which were closely correlated with surgical adverse effects and pTNM stage.

Successful placement of the stent relies on the severity of colonic obstruction and expertise of the endoscopist. The technical (89.8 %) and clinical success (85.7 %) rates in this study were similar to those in previous studies (84.2–100 % and 78.9–100 %, respectively) [18]. Established short-term advantages of bridged surgery include less temporary and permanent stoma creation as analyzed in many meta-analyses [18, 19]. Our study confirmed that BTS significantly increased the use of laparoscopy and decreased stoma creation.

However, recent studies have failed to show beneficial effects of stenting as BTS over emergency surgery, due to uncertainty of its impact on long-term oncological outcomes [20, 21]. This originates from concerns about tumor manipulation during stent insertion, guidewire perforations during stent placement [22], stent deployment force and eventual microperforations at the proximal and distal ends of the stent [23], which may induce tumor cell dissemination locally but also in the bloodstream [13]. In our study, we found that patients with BTS had a higher risk of perineural invasion, similar to Kim’s findings [24]. In a multivariate analysis by Leibig et al. [25], perineural invasion was thought to be an independent prognostic factor of oncological outcomes in colorectal cancer. In our analysis, perineural invasion was significantly associated with the overall survival of propensity score-matched patients by univariate survival analysis (data not shown). However, the correlation was not significant in multivariate analysis. Long-term large-scale studies are needed to better investigate the correlation of perineural invasion and oncological outcomes. Stent insertion was associated with more total recurrence in this study, although the difference was not significant in regional or distant recurrence alone. Similar to the findings in our study, a recent meta-analysis [26] of 7 randomized controlled trials demonstrated that BTS significantly increased the risk of recurrence, especially distant recurrence.

To date, very few studies have reported on long-term survival after SEMS placement as a BTS due to a scarcity of clinical data and the lack of comparable studies. Femke et al. [19] found that SEMS placement as BTS did not influence 3 and 5 year overall survival in a meta-analysis. This finding is similar to Sun’s findings [12], which suggest colonic stenting did not affect 5 and 10 year survival; however, the study population is relatively small and an accurate conclusion cannot be drawn. However, in Kim’s study [27], SEMS placement negatively affected 5 year overall survival and disease-free survival (DFS) in stage II and III CRCs (5-year OS: 44 % after SEMS versus 87 % after elective surgery for nonobstructing CRC). Sabbagh et al. [21] also found that 5 year overall survival was significantly lower in the BTS group, whereas 5 year cancer-specific mortality was significantly higher (48 vs. 21 %, p = 0.02). However, no significant differences in terms of 5 year DFS were noted. In our study, we found that 3 and 5 year overall survival were significantly lower in patients who underwent SEMS as BTS than in those who underwent ES.

Colonic stent insertion also affects patient survival in multiple aspects. Avlund et al. performed a 10-year follow-up study and concluded an association between SEMS-related perforations and decreased survival [28]. The interval from SEMS to resection surgery was thought to delay the surgery and increase the rate of recurrence and survival in the study by Broholm [20]; however, further larger studies are needed to confirm the results. Postoperative adverse effects, especially infectious complications, were associated with poorer survival in patients after colorectal cancer resection [29]. SEMS insertion was also reported to be associated with increased perineural invasion [30], which is a known prognostic factor in CRCs and correlates with the findings in our study. Many clinical factors could influence the prognosis of obstructing CRCs and overall survival. An analysis by Rodrigues et al. suggests that pTMN stage IV, number of lymph nodes harvested, adjuvant therapy and surgery-related complications could influence overall survival [2]. In our study, stenting along with surgical adverse effects and pTNM stage were associated with overall survival by multivariate analysis of propensity score-matched patients.

The present study had several limitations. First, its retrospective nature may introduce selection bias and affect the results. Second, because the population after propensity score matching was relatively small, the analysis of some variables showed a wide range. The effect of SEMSs as BTSs should be cautiously interpreted. Third, although the medical records were carefully reviewed and follow-up studies were thoroughly performed the causes of death were difficult to confirm in some cases, and disease-free survival was lacking. The strengths of our studies are the homogeneity between groups, due to the use of PSM analysis and the long follow-up period.

SEMS placement was associated with high technical and clinical success, similar to ES, as demonstrated by the higher primary anastomosis rate and lower stoma rates, with its possible positive effects on quality of life. However, SEMS placement as a BTS leads to more perineural invasion, higher recurrence rate and worse long-term overall survival compared to ES. These results suggest that SEMS placement should not be routinely performed in patients with potentially cured AMLCOs.