This study provides an overview of several clinical outcomes following surgical management of deep infiltrating endometriosis in a tertiary hospital in France. We report a rate of severe post-operative complications of 2.42% which is consistent with previous studies7 9 8. Our findings are also in agreement with other studies, which evaluated rates of complications and their risk factors in endometriosis surgery. 18 19 Lermann et al.20 described a population of 134 patients who had surgical management for deep endometriosis and reported 3.7% and 12.7% rates of severe and minor complications respectively. These rates are comparable to our results, although, they excluded patients who required bowel resection. Of relevance, two of the patients in our study who presented with severe complications had extensive endometriosis involving their bowels requiring shaving or resection.
The main complication after surgery in our study was urinary tract infection, which occurred in 6.06% of our cohort. Urinary tract infections might be explained by the fact that patients had a urinary catheter during and following surgery to mitigate the risk of post-operative voiding dysfunction. It has been shown that bacteriuria occurred more often with increased length of catheterisation.21 However the risk of infection secondary to catheter insertion should be weighed against its benefits. Indeed, only 3 women included in our study had voiding dysfunction (1.82%) after surgery. In a comprehensive literature review, Campin et al. reported varying risk of post-operative voiding dysfunction ranging from 0.8 to 30%.22 Furthermore, this rate has been more recently reported to be as high as 50% by others.23
The rate of temporary protective ileostomy in our study was 3.64%, which is lower than previous reports. Indeed, in a national snapshot of the surgical management of deep infiltrating endometriosis of the rectum and colon in France in 2015, Roman et al reported a protective stoma rate of 19%.24 More recently, it was demonstrated that a protective stoma did not prevent the occurrence of a recto-vaginal fistula.25 This is consistent with our results where our restrictive protective ileostomy policy was still associated with a low rate of serious complications. To reduce the rate of complications after colorectal procedures, conservative surgery, as rectal shaving and discoid excision, have been proposed. Roman et al., on behalf of the FRIENDS group, conducted a large national retrospective study including 1135 patients undergoing surgery for deep endometriosis of the rectum and colon. They showed significantly higher rate of recto-vaginal fistula following rectal resection compared to conservative treatment.24 Findings from our centre also support this notion.
Among the severe complications in our study, 2 occurred at the beginning and 2 at the end of our studied period. However, with advances in surgical techniques, specialisation and centralisation of care the number of women who had surgical management of endometriosis have increased over the years. In a large multicenter retrospective study, Bendifallah et al.26 demonstrated that the rate of surgical complications when managing deep colorectal endometriosis was independently inversely linked to the annual number of procedures performed by centre and surgeon. In their study, the authors suggested that the optimal number of procedures required per surgeon per year to be between 7 and 13 operations. In our centre, the mean number of colorectal procedure performed per year was 7.
It has been reported that the main surgical risk factors during deep endometriosis surgery are the opening of the vagina, a low colorectal anastomosis and resection of endometriosis in relation to the urinary bladder and ureters.27 18 28 In our study, colpotomy was not identified as a risk factor for post-operative complications. Moreover, although rectal surgery was a significant factor on univariate analysis, it was not on multivariate analysis. However, it is possible that due to the small number of patients who had rectal surgery in our cohort, our study did not have enough power to demonstrate such association. Diabetes and obesity have been identified to be risk factors of wound and anastomosis healing29 30 31. The two patients with the most severe complications in our study did not have diabetes, but had body mass indices of 32 and 37. Nevertheless, body mass index (BMI) was not identified as a risk factor in our cohort, which could be because our study population were mostly not obese with a median BMI of 23 (IQR = 6).
Operative time was identified as a risk factor for post-operative complications in our study. Arguably, complex surgical procedures with extensive adhesiolysis and bowel resections require longer operative time; nonetheless, this was demonstrated as an independent factor on multivariate analysis. Poupon et al. reported that the ENZIAN score is a good predictor of complications.3 Other authors have studied the correlation between MRI and the ENZIAN score and showed good correlation between both.32 33 We were not able to assess the predictive effect of this score on surgical complication rates in our patients because it was not used in our unit during the study period.
This study has several strengths, which include the use of our comprehensive electronic patient record and clinical coding system to identify our study cohort minimising the risk of selection bias. Additionally, we mitigated the risk of contaminating our analysis by having clear and strict inclusion criteria. This has ensured that our findings are relevant to the group of women considered at higher risk for surgical management of their endometriosis. Therefore, we believe that our findings will help provide women with realistic information to enable them make an informed choice about their care. Finally, the use of an internationally recognised system to categorise post-operative complications adds to the external validity of our findings. However, we also recognise the limitations posed by the retrospective design of our study. Furthermore, the low rate of occurrence of severe complications could be perceived as a limitation. Nevertheless, our identified rates concur with other studies involving independent cohorts.
In conclusion, the rate of severe post-operative complications after deep endometriosis surgery in a specialised centre is low. Our study demonstrated that operative time and age were independent risk factors for postoperative complications. This study adds evidence to the importance of multidisciplanrity and centralisation of care in the surgical management of deep infiltrating endometriosis. This information should help clinicians when counselling women to enable them make an informed choice about their management.