Introduction

One of the most frequent complications after laparotomy is incisional hernia, which is reported in up to 20% of patients [1]. Temporary intestinal stomas are often created during emergency surgery or as a protective measure after colorectal resection. Patients with a temporary stoma after emergency surgery are at an increased risk of incisional hernia formation [2]. Repair of the incisional hernia in conjunction with stoma reversal as a one-stage procedure is tempting because of potential optimized time- and cost-effectiveness, though a recent study only found a marginally lower cost of a single-stage procedure [3]. In contrast, a two-stage procedure with postponement of the hernia repair may be a safer choice. Spillage of bowel contents during stoma takedown and construction of the anastomosis could potentially contaminate the surgical field and increase the rate of complications [4]. The most severe complication after stoma reversal is anastomotic leak that leads to increased morbidity, length of hospital stay, need for intensive care and thirty-day mortality [5,6,7]. The incidence of AL after stoma reversal depends on the type of stoma.

A previous study reported a 65% wound morbidity rate after a single-stage approach for enterocutaneous fistula takedown combined with large complex abdominal wall reconstruction [8]. Kugler et al. [9] found a 53% rate of surgical site occurrences (SSO) after a dual-stage approach to ventral hernia repair in potentially contaminated fields.

Until recently, we have routinely performed stoma reversal and incisional hernia repair as a single-stage procedure at our institution. The aim of the present study was to investigate whether a concomitant incisional hernia repair influenced the incidences of AL, 90-day overall complications and 90-day mortality in patients undergoing stoma reversal.

Materials and methods

This retrospective cohort study was conducted and reported in accordance with the STROBE guidelines [10]. We included all patients undergoing elective reversal of an intestinal end-stoma between October 1, 2010 and May 10, 2016 at the Digestive Disease Center, Bispebjerg University Hospital, Denmark. Patients were identified in the administrative surgical database by their relevant International Classification of Diseases 10th version procedure codes (Table 1). All patients undergoing reversal of loop ostomies were excluded, because these procedures do not require a laparotomy. Data on patient characteristics, surgical history, intraoperative data and 90-day postoperative complications were assessed by medical chart review. The following patient characteristics were retrieved: age, gender, body mass index (BMI), tobacco use, excessive alcohol intake (defined as > 168 g/week for women and > 252 g/week for men), American Society of Anesthesiologists’ (ASA) score, ischemic heart disease, chronic obstructive pulmonary disorder, diabetes, the use of immunosuppressive medication, previous intraabdominal surgery, duration of surgery, other procedures in addition to the stoma reversal, previous AL, type of anastomosis (colo-colonic, colorectal, ileo-colonic or ileo-ileal), length of stay and readmission. We also recorded the size of the hernia defect, number of defects, mesh type, mesh placement, mesh size and lateral release procedures, in patients undergoing concomitant incisional hernia repair. Exclusively patients with incisional hernias undergoing mesh repair were included in the hernia repair group, because non-mesh hernia repair was not considered to technically differ significantly from standard laparotomy closure.

Table 1 Surgical procedure codes included in the study
Full size table

The primary outcome was AL, defined as a grade C leakage requiring reoperation [11]. The secondary outcomes were any complication, complications with a Clavien Dindo score ≥ III, 90-day mortality and 3-year cumulative survival. The severity of postoperative complications was graded according to the Clavien Dindo classification [12]. The outcome “any complication” included the following four subgroups (1) AL, (2) SSO (superficial or deep surgical site infection, fascial dehiscence, seroma, enterocutaneous fistula or late bleeding/hematoma [4]), (3) pulmonary complications (pneumonia, atelectasis, hydrothorax, hemothorax or pneumothorax) and (4) other complications. Next, patients with Clavien Dindo scores below III were compared to patients with a Clavien Dindo score above or equal to III, to assess the severity of complications with and without concomitant hernia repair. Data on overall survival were extracted from the Danish Civil Person Register [13].

Statistics

Univariable analysis was performed on the entire cohort, comparing the recorded variables between (a) patients with and without a concomitant hernia repair; (b) patients with and without AL; and (c) patients with and without any postoperative complication. Continuous and categorical variables were compared using Student’s t test after a normality test and Chi-square test, respectively. Only a univariable analysis was performed on the primary outcome due to the small number of cases with AL. Multivariable analysis was performed on the composite outcome “any complication.” Potential confounding variables included in the multivariable analysis were selected according to a P value < 0.1 in the univariable analyses. The three-year cumulative survival was compared between the two groups using the log-rank test.

The statistical software used for all analyses was R 3.2 (Foundation for Statistical Computing, Vienna, Austria). P values < 0.05 were considered statistically significant. The study was approved by the Danish Data Protection Agency (j. no. 2012-58-0004).

Results

A total of 150 patients underwent stoma reversal during the study period. Eight patients were excluded from the cohort: Five patients underwent stoma reversal during an emergency procedure, two underwent loop colostomy reversal without laparotomy, and one patient was lost to follow-up, leaving 142 for analysis (Fig. 1).

Fig. 1
figure 1

Patient flowchart

Full size image

The indications for the initial stoma formation were benign condition (n = 88), cancer (n = 45), trauma (n = 5) and iatrogenic surgical injury (n = 4). In total, 18 (13%) patients underwent stoma reversal and concomitant incisional hernia repair. This patient group had higher BMI, ASA score, number of previous abdominal procedures, time from primary surgery to stoma closure and operative time (Table 2). The median size of the fascial defect of the hernias was 5.6 cm (range: 2–20 cm) horizontally and 7.5 cm (range: 3–18 cm) vertically. The mesh placement was retromuscular in 10 patients, intraperitoneal in seven patients and onlay in one patient. Fourteen of the implanted meshes were synthetic and four were biological. Colorectal consultant surgeons performed all colo-colic and colorectal anastomoses, whereas general surgeons also performed ileo-colic anastomoses.

Table 2 Characteristics of patients undergoing elective stoma reversal according to concomitant on hernia repair or not
Full size table

AL occurred in four of 18 (22.2%) patients in the hernia repair group and in three of 124 (2.4%) patients in the non-hernia repair group (P = 0.002) (Table 3). No patients in the cohort were diagnosed with a minor AL, which only required conservative treatment. The median time from stoma reversal to diagnosis of AL was 8 days (range 3–14). In the univariable analysis, increasing duration of surgery (mean 269 min vs. 174 min, P < 0.001), time from primary surgery to stoma closure (median 602 days [IQR 319–980] vs. 234 days [IQR 148–380], P = 0.018) and ischemic heart disease (2 out of 7 vs. 5 out of 135, P = 0.039) were significantly associated with AL. Detailed information about patients undergoing reoperation for AL is given in Table 4.

Table 3 Characteristics of patients undergoing elective stoma reversal according to development of anastomotic leakage
Full size table
Table 4 In-depth information regarding patient undergoing reoperation for anastomotic leakage
Full size table

Thirteen out of 18 (72.2%) patients in the hernia repair group experienced at least one complication compared with 28 of 124 (22.6%) patients in the non-hernia repair group, P < 0.001 (Table 5). In the hernia repair group, five out of 18 (27.8%) patients had a Clavien Dindo class III complication or higher, compared with 12 out of 124 (9.7%) in the non-hernia group, P < 0.001. Eight (44.4%) patients in the hernia repair group developed a SSO compared to 14 (11.3%) in the non-hernia repair group (P < 0.001) (Table 6). Factors associated with the development of at least one complication included increasing age (P = 0.002), higher BMI (P = 0.008), higher ASA score (P = 0.035), concomitant hernia repair (P < 0.001) and greater number of previous surgeries (P = 0.001) (Table 5). After multivariable adjustment for potential confounders, concomitant hernia repair was the only variable that significantly associated with the development of any postoperative complication after elective stoma reversal (OR 5.92, 95% CI 1.54–25.69, P = 0.012) (Table 7). Patients with any complications had a significantly longer postoperative length of hospital stay (median 12 days [IQR 7–19] vs. 5 days [IQR 3–5], P < 0.001) and higher rate of readmission (8 [19.5%] vs. 5 [5.0%], P = 0.016) (Table 6).

Table 5 Characteristics of patients undergoing elective stoma reversal according to development of complications within 90 days
Full size table
Table 6 Univariable analysis of complications after elective surgery for stoma reversal
Full size table
Table 7 Multivariable analysis of factors associated with any complication following elective surgery for stoma reversal after colonic resection
Full size table

No patients undergoing hernia repair concomitant to stoma reversal died during the 90-day follow-up, whereas three (2.5%) patients who only underwent stoma reversal died during this period. The fatalities occurred 13, 21 and 42 days postoperatively and were due to cerebral infarction, sepsis following AL and unknown reasons (no autopsy), respectively. The three-year cumulative survival for patients undergoing the combined procedure was 80.8% (56.8–100%) compared to 90.7% (84.7–96.7%) after stoma reversal only, P = 0.708.

Discussion

In the current study, we found concomitant hernia repair, longer duration of surgery, time to stoma closure and ischemic heart disease to be associated with AL. Further, concomitant hernia repair was associated with an increased risk of postoperative complications. To our knowledge, this is the first study to specifically evaluate the risks associated with stoma reversal and concomitant hernia repair.

Other studies have reported the risk factors for AL to include male gender, age beyond 60 years, obesity, duration of surgery > 180 min, ASA score ≥ 3, anastomosis close to the anal verge, excessive alcoholic intake, tobacco smoking, hypoalbuminemia, advanced tumor stage, emergency conditions, high intraoperative blood loss and transfusion therapy [14,15,16,17,18,19,20]. AL is relatively rare after stoma reversal and depends on the level of bowel segments joined in the anastomosis. The reported rates of AL vary from 0.4 to 2.9% after ileostomy reversal to 0.3–8.0% after colostomy reversal [5, 21,22,23]. Considering that the majority of anastomoses in our study were either colo-colonic or colorectal, the overall AL rate of 5% is low, whereas the 22% AL rate in patients undergoing concomitant hernia repair is unacceptably high.

The literature on the safety of combining stoma reversal with an additional procedure is limited. Lupinacci et al. [24] did not report an increased rate of AL after a combined surgical intervention with ostomy closure and liver resection in patients with hepatic metastatic disease from colorectal cancer. This finding is probably due to the exclusive inclusion of patients undergoing loop ileostomy closure as opposed to the present study, in which a majority of patients underwent colo-colonic or colorectal anastomoses and no patients underwent closure of a loop ostomy. It remains unknown whether incisional hernia repair in particular increases the incidence of AL. A combined approach induces a longer operative time and more pronounced perioperative stress, as hernia repair often entails wide parietal dissection and implantation of foreign body material. Interestingly, mesh repair increases the systemic inflammatory response compared to sutured repair of a hernia [25], and previous studies have demonstrated an association between the inflammatory activity and collagen turnover which might affect the strength of the anastomosis [26, 27]. Eight patients underwent repair of a fascial defect wider than 10 cm. An increase in intraabdominal pressure after large incisional hernia repair has previously been reported, potentially reducing anastomotic microcirculation [28].

The modified hernia grading scale predicts that the risk of a SSO is 46% in a grade 3 hernia (clean-contaminated, contaminated or dirty operative field) [4]. This is comparable to the 44% incidence of SSO in the present study, as all patients undergoing stoma reversal are categorized as grade 3. A previous study found that ventral hernia repair concomitant to another intraabdominal procedure more than double the risk of developing a SSO. This rate reached 86% in patients undergoing ventral hernia repair in combination with stoma reversal [29]. The incidence of SSO after hernia repair in the current study was rather low compared to studies reporting the results of ventral hernia repair in contaminated fields [8, 9, 29]. In contrast, a recent study examined the effects of prophylactic mesh on IH development at the former stoma site and the rate of SSOs after stoma takedown. Comparable rates of SSO and AL after ostomy takedown were found between patients with and without an additional retromuscular mesh [30]. This study differed from the present in that less than half of the patients had a midline IH present. Moreover, some of the patients only required a small peristomal incision rather than a full laparotomy.

The Ventral Hernia Working Group considers development of SSOs to be an important predictor of hernia recurrence [31], and we thus consider a 44% SSO rate in the hernia repair group to be unacceptably high. If patients require both stoma reversal and hernia repair undergoing a two-stage procedure, the hernia repair could be reduced from level 3 to a level 1 or 2, thus reducing the predicted SSO rate from 46 to 14–27% [4]. Furthermore, AL following stoma reversal and hernia repair including lateral release may require mesh removal and result in “burned bridges” in terms of future hernia repair. On the other hand, possible drawbacks of a dual-stage procedure are also potentially significant, including the risk of complications after exposure to two separate surgical and anesthesiologic procedures. Surgical site infection following mesh repair is a serious condition often requiring reoperation, mesh removal, prolonged hospitalization and higher healthcare expenses [32, 33]. Larger studies are required to evaluate whether a dual-stage approach is economically advantageous compared with a single combined approach.

There are limitations to this retrospective study. The distribution of patients into the two groups was prone to selection bias. Though this might be the case, the only significantly different demographic variables between the case and the control groups were BMI and number of previous surgeries, leaving the two groups relatively comparable. This particular study did not take into account the peri- and postoperative complications that could arise during a secondary hernia repair, nor did it address the issue of hernia-related complications in the waiting period of a dual-stage approach. Furthermore, the study lacked statistical power to allow for multivariable analysis on the risk of AL. Lastly, the low number of patients undergoing concomitant hernia repair in this study increases the risk of a type 1 error.

The findings of this study suggest that the risk of AL and overall complications after stoma reversal is increased by concomitant incisional hernia repair.