Long term outcome of anastomotic leakage in patients undergoing low anterior resection for rectal cancer

The management of rectal cancer has substantially improved and standardized in the past twenty years. Preoperative chemoradiotherapy (CRT) followed by total mesorectal excision (TME) is now the gold standard for locally advanced rectal carcinoma [1‐4]. Indeed, TME has led to decreased local recurrence rates and increased cancer-specific survival rates [5]. More recently, technical advances include also an intersphincteric approach for resection [6, 7] which, combined with a neoadjuvant treatment [4], aims to reduce the permanent stoma rate in patients undergoing low rectal cancer surgery. Despite numerous improvements, anastomotic leakage (AL) remains the most dreaded complication following rectal resection [8‐10]. Its occurrence is associated with impaired postoperative functional results (anal incontinence, low anterior resection syndrome, [11] and ultimately an increased rate of permanent stomata (reaching nearly 30%)) [12‐14]. Furthermore, AL is also associated with impaired oncological results with increased local and/or distal recurrence rates as well as decreased overall survival rates [15‐17]. However, it is important to discriminate symptomatic from asymptomatic AL since these two conditions do not seem to have the same oncological and functional postoperative outcome [16]. Although several previous studies did not find a difference between without (AL) and asymptomatic anastomotic leakage (AAL) with regard to early postoperative functional and oncological outcome, these studies did not analyze and report long-term data from patients with AAL. Furthermore, these previous studies analyzed heterogeneous patient populations including also patients suffering from distant metastasis [11, 12, 18]. It is obvious that the inclusion of such patients has a negative impact on anastomotic healing rates, increases the risk for a permanent stoma and reduced recurrence-free and overall survival rates, respectively [14]. Therefore, the objective of our study is to report and analyse long-term oncological and functional outcome in a homogeneous population of patients with AAL and SAL who had undergone curative surgical resection for non-distant metastatic rectal cancer in a tertiary referral centre.

In this retrospective single center study we conducted a comparative analysis of asymptomatic and symptomatic AL in patients who underwent rectal resection with curative intent. Both short-term and long-term morbidity, functional and oncological outcomes in this homogeneous population were analyzed. Patients suffering from SAL were found not to be at risk for an increased number of local and/or distant tumor recurrence rates compared to the group of AAL and WAL patients. However, in the group of patients suffering from postoperative anastomotic leakage (AAL and SAL) we observed an impaired long-term outcome of postoperative functional results. Furthermore, there is an increased risk for a permanent stoma in AAL and SAL patients which is as high as 41% and compated to patients with no postoperative anastomotic leakage (9.3%) is significantly higher.

The effect of postoperative anastomotic leakage (AL) on the oncological outcome following rectal cancer surgery is still debated controversially and remains unclear. Some authors have reported about reduced local recurrence rates whereas distant recurrence rates and overall survival rates were found to be similar. In contrast, equivalent local, distant disease recurrence rates and overall survival rates were observed by others [15, 35, 36]. Furthermore, two meta-analysis reported that AL was associated with higher local disease recurrence rates and reduced long-term cancer specific survival. No impact was found on the incidence of distant disease recurrence rates [15, 35]. Moreover, five pooled randomized trials [36] published in 2009 concluded that the disease-free survival was not affected by the postoperative appearance of symptomatic AL. However, symptomatic AL was observed to be associated with an impaired overall survival in such patients. Such discrepancies between the reported results in the aforementioned studies might be be explained by considering two important facts about postoperative AL in more detail.

First, in most previously published studies, any clinical impact of postoperative AAL was not assessed since no discrimination between AAL and SAL was made. So far, only a few studies analyzed the subgroup of AAL and observed that the short- as well as the and long-term outcome differed from the SAL and WAL group, respectively [16].

Second, the diagnosic work-up, definition of AAL and also the discrimination from patients without AL are still not clearly defined and are controversially discussed. Therefore it is obvious that the rate of patients with diagnosed postoperative AL varies considerably between the different studies with a reported prevalence which is somewhere in the range between 5% [37‐39] to 28% [11‐16].

In this present study, we observed an incidence of AL of 18% (39/200) which is indeed somewhat higher compared to other studies [37‐39]. However, our findings could be explained by the fact that all patients in our study were subjected to AL screening which, in accordance with the recommendations of the working group of Panis et al., included routine blood testing and water soluble enema contrast CT-scan six weeks postoperatively and before stoma reversal [11‐16].

In contrast to the study published by Hain et al. [16] we observed no significant difference of the overall and disease-free survival between the AAL, SAL and WAL group, respectively. Indeed, we are aware that an important limitation of our study, besides that it is retrospective, is the limited number of included patients which hampers statistical data interpretation. Nevertheless, the homogeneity of our study population is an important quality criterion compared to other studies which also included patients suffering from distant metastatic disease [16, 38, 39]. Including such patients suffering from distant metastases carries a relevant risk to bias study data since such patients suffer from impaired postoperative anastomotic healing with all the corresponding known negative sequelae [14]. Finally, the management of rectal cancer has substantially improved in the last decade (i.e. rectal surgery, neoadjuvant treatment, diagnosis and treatment of AL). Therefore, data from previous studies (1971 to 2010) [37‐39] are not comparable to present data since they do not reflect current standards for diagnosis and treatment of rectal cancer. Hence, the interpretation of our current data in the context of such previous studies is probably of limited relevance.

Although we did not find a negative impact of SAL and AAL on the oncological outcome in our present study we observed an impaired short- and long-term functional outcome of AAL compared to patients without AL (WAL). This is again in contrast to a recently published study by Hain et al. which found no difference between the LARS score in the AAL, SAL, and WAL group, respectively. However, the authors of this study did not discriminate between WAL and AAL in their multivariate analysis of predictive factors of the postoperative functional outcome [11]. Although it is obvious that AL has a negative impact on the postoperative functional outcome [40‐45] it remains still unclear for the subgroup of patients with AAL. Similar findings as ours were found by Lim et al. who reported that the bowel function after ileostomy closure was equally impaired in the AAL and SAL group, respectively [40].

In our study, the postoperative time interval to stoma reversal was similar in SSL and AAL, but significantly longer compared to WAL patients. These results are consistent with findings in other previous studies [11, 40]. Furthermore, we observed that the permanent stoma rates were quite similar in the AAL (30%) and SAL (44.8%) group but significantly higher compared to the WAL group (9.3%). Some other authors have reported that all AAL patients can be managed conservatively with spontaneous healing whereas SAL patients showed only a 40% percent rate of spontaneous healing [40]. An important finding in our study is the fact that, even if AAL patients can be managed conservatively, the functional postoperative outcome can be impaired.

Our multivariate analysis revealed that SAL and R1 rectal tumor resection were the only independent risk factors for a permanent stoma. For the presence of AAL only a trend towards a higher stoma rate was found.

We are aware about some limitations of our study (small number, retrospective). Nonetheless, our study is the first which observed that AAL patients have a risk for worse postoperative functional long-term outcome and a high risk for a permanent stoma. Finally, previous studies are lacking of long-term outcome data for AAL and SAL and are heterogeneous since they also included patients suffering from distant metastatic disease [11‐18].

In conclusion, AL did not impair long-term oncological results (disease-free and overall survival) in patients with rectal adenocarcinoma. Despite the small numbers of patients, especially in the AAL group, long-term functional results were impaired by the occurrence of SAL and AAL with a similar permanent stoma rate in both groups.