A high number (49%) of our patients suffered at least one complication after the anastomotic procedure and majority of these were managed conservatively. Complications after biliary-enteric anastomotic procedures have ranged from 3% to 43% in the previous reports [5, 10]. However, most of these studies are specific to BEA due to iatrogenic biliary tract injuries. The differing disease etiology, surgery performed and definition of 'complications' make it difficult to make logical comparisons. The largest and most recent of these studies by Sicklick et al consisted of 175 patients who underwent biliary reconstruction for the iatrogenic bile duct injury following laparoscopic cholecystectomy found a complication rate of 43% . In another study by Tocchi et al consisting of 84 patients undergoing hepaticojejunostomy for benign biliary stricture found a rate of complication of 21% . Even though, there were patients with biliary stricture due to a number of reasons including choledocholithiasis, trauma and choledochal cyst, a hepaticojejunostomy was performed in less than half (43%) of our patients. Also we have included a wider range of conditions in our definition of 'complications', which may have led to higher rates.
Our study included four patients who underwent Whipple pancreaticoduodenectomy due to a strong pre-operative suspicion for a malignant disease. All four of these patients had complications during hospital stay. Our patient population consisted of all patients who underwent BEA with a final diagnosis of a benign disease condition and therefore those four that underwent the Whipple's procedure were included in the analysis. The authors believe that this is a more 'real life' representation as pancreaticoduodenectomies for benign conditions is not a rarity in the literature, and a recent study reviewing 459 pancreaticoduodenectomies for a suspected malignant disease found 11% of them to have a final benign histology . Type of surgery which included a category for 'Whipple's procedure' was not found to be significantly associated with incidence of early complications, however this may be due to lack of power. Furthermore we did perform sensitivity analysis after excluding patients that underwent Whipple's procedure and the results were similar with the incidence of early and local complications being 48% and 39% respectively.
Our study demonstrated a higher incidence of wound infection (23%) as compared to other studies . In both studies by Sicklick et al and Tochi et al, wound infection was the most frequent complication, 8% and 12% respectively [3, 5]. Biliary leak occurred in 10% of our patients while it has been reported to occur in 3-8% of patients in previous studies [3, 8]. Patients in Pakistan generally present late, are malnourished and thus sicker when compared to patients in the Western world. This is a possible reason for the higher rates of surgical complications in our study. In fact, 87% of our patients had low pre-operative albumin levels (< 3.5 g/dl) and multivariate analysis demonstrated low albumin levels to be an independent predictor of post-operative complications. Previous studies have also shown hypoalbuminemia to influence on in-hospital complications, mortality and long term survival [6, 7].
The association between higher ASA classes and the occurrence of complications to the best of our knowledge has not been studied for BEA. ASA class is a measure of functional status of the patient and higher ASA class has been known to lead to a higher rate of complication in other gastrointestinal surgeries [12, 13]. ASA class, but not the Charlson score, was shown to be significantly associated to the outcome in our study; we believe this may be reflective of inadequate optimization of patients prior to surgery. To note, none of the other demographic, clinical or operative variables were significantly associated with the outcome. This is similar to other reports .
The mortality in our study was 5%. Mortality after BEA for benign biliary stricture has fallen over the years from 10% to 1% in the recent years [3, 5, 6, 8, 10]. Half of the deaths in our study occurred in patients who underwent Whipple's pancreaticoduodenectomy; this has not been included in the previous studies. The remaining deaths occurred in patients with multiple complications including renal failure and chest infection. Both of these patients had low pre-operative albumin levels. Due to the small number of deaths (n = 4) we were unable to determine independent factors associated with mortality.
There is no dedicated hepatobiliary service at present at AKUH and all surgeries were performed by the general surgery service. The surgeries were performed by 16 different surgeons over the course of 22 years. Limitations in both financial and human resources make it difficult for institutions in underdeveloped countries to maintain dedicated sub-specialty surgical services. However this may be an important factor contributing to the higher rate of complications as most other reports are from dedicated hepatobiliary centers. Higher volumes of surgery has been associated with better outcomes . A dedicated service offers higher volume to the surgeon and the team. We are currently in the process of developing a number of subspecialty services including pancreatobiliary surgery. It will be interesting to see how outcomes change in the following decades.
The inherent limitation of the retrospective study design made it difficult to account for all potential confounders or study every possible variable that may be related to rate of complications. For example we were unable to retrieve data on time since biliary injury, local ischemia, evolution of surgical technique or other similar factors. However, as we have discussed above, previous studies have also not found an association between other operative, clinical and demographic factors. It would have been interesting to look at factors responsible for specific complications for example wound infection and biliary leak. However in our analysis we did not have enough power to test for these associations. A related limitation is that of the small sample size, we were only able to report on 79 cases; however this is still the largest report of this kind from an underdeveloped country. A larger study looking for associations with specific complications should be conducted.