Ileostomy versus fecal diversion device to protect anastomosis after rectal surgery: a randomized clinical trial

The inclusion criteria were as follows: 1) benign or malignant disease of the left colon and rectum, 2) extraperitoneal colorectal anastomosis, 3) age > 18 years, and 4) an American Society of Anesthesiologists (ASA) score of 1–3.

The exclusion criteria were as follows: 1) need for emergent surgery for obstruction or perforation, 2) abdominoperineal resection or total proctocolectomy, 3) no curative resection of the primary lesion, 4) anastomosis above the peritoneal reflex, and 5) refusal to participate.

The FDD, which consists of a head and tail portion, is a newly designed tube that is inserted into the colon to protect the anastomosis (Fig. 1). The head of the FDD consisted of an inner balloon, two external balloons, and a hole for colon cleansing (Fig. 2). The tail portion of the FDD consisted of a thin silicone tube, catheter complex for the balloons, and wash hole. The thin silicon tube passes through the inside of the anastomosis to prevent feces from coming into contact with the anastomosis.

We used a band and tension-measuring device to better affix the FDD to the colon. To prevent band erosion, an automatic tension-measuring instrument (ATMI; JSR Medical Inc., Daegu, Korea, Patent pending) was used (Fig. 3a). The ATMI measures the pressure acting on the colon wall and helps ensure that the absorbable poly lactic-co-glycolic acid mesh bands (NEOSORB MESH®^; Samyang Biopham. Co., Daejeon, Korea; Fig. 3b) are of adequate length. The band has a half-life of 6 weeks and helps affix the FDD.

All procedures were performed by experienced colorectal surgeons with an open or laparoscopic technique. Bowel preparation using a polyethylene glycol solution was performed the day before the surgery. After colon mobilization was confirmed, the diseased segment was resected. The colorectal anastomosis was performed using the double stapling technique or handsewn method, followed by a loop ileostomy or FDD procedure.

All stoma group patients underwent ileostomy using standard conventional methods [9]. All patients received enterostomal care education.

The FDD installation process was performed as follows. After the colorectal anastomosis, we marked an installation area in the colon 10–15 cm above the anastomosis site and used the ATMI to determine the appropriate band length. The band was placed through the mesocolon on the outer wall of the colon. The inner balloon of the FDD was inflated and inserted into the colon through the anus. The FDD was inserted where the band was located. The band was placed between the external balloons of the FDD and fixed. The external balloon of the FDD was then expanded and its tail portion was cut outside the body to an appropriate length (Fig. 4).

All patients were allowed to drink water immediately after the surgery. Once gas was released into the stoma or the anus, a soft diet was started. Patients were discharged if they were able to consume soft diets for 2 consecutive days without problems.

Colonic irrigation through the FDD began the day after the diet was started. Irrigation of the FDD was performed as follows. After the inner FDD balloon was inflated, tepid water (600–1000 mL) was placed through the irrigation catheter into the colon above the FDD. To evacuate the irrigated colonic contents, the patients deflated the inner balloon of the FDD and opened its inner lumen. The evacuation took 15–30 min. Twice-daily irrigation was recommended (Fig. 5).

All patients routinely underwent abdominopelvic computed tomography (CT) scans at 1 week and a contrast study at 3 weeks postoperative, while an additional contrast study was performed in the stoma group at 10–12 weeks to confirm anastomosis integrity prior to stoma reversal. AL was defined as a defect in the intestinal wall integrity at the anastomosis site regardless of clinical symptoms [10]. The defect was proven by a contrast study or CT scan. Clinical peritonitis without evidence of defects as well as perianastomotic fluid collection with or without gas were considered as AL. The FDD was removed when no evidence of AL was found in the contrast study at 3 weeks postoperative. If AL was found, the FDD was maintained for up to 3 additional weeks and the contrast study was repeated. If the anastomosis was normal on the last contrast study, the FDD was removed. The stoma takedown was performed when no evidence of AL was found in the contrast study at 10–12 weeks; otherwise, it was delayed by 1–2 months. If a persistent sinus was observed, stoma takedown was performed once it had stabilized. The clinical trial was ended after 6 months postoperative.

Patients were randomly assigned 1:1 to the two groups the day before surgery using a computer-generated simple randomization.

The primary endpoints of the study were FDD safety and effectiveness. FDD safety was defined as FDD-related morbidity and mortality. Complications were observed for 6 months. FDD effectiveness was defined as the capability of fecal diversion to prevent sepsis in cases of AL. The total hospital stay was defined as the postoperative hospital stay after the primary surgery. In the stoma group, the total hospital stay included hospitalization for ileostomy reversal. The total hospital costs included: 1) primary surgery and admission, 2) radiologic examination and sigmoidoscopy until FDD removal in the FDD group, 3) second admission including ileostomy takedown surgery in the stoma group, 4) stoma care products, and 5) treatment of complications. The total hospital costs did not include: 1) those not related to trials such as hepatectomy due to hepatic metastases or chemotherapy, and 2) examinations in regular follow-up appointments. The secondary endpoints included FDD and stoma use duration, operation time, and delay of starting the soft diet. The clinical AL results were reviewed. Data were collected during the preoperative study until FDD removal or stoma reversal, and at 6 months of postoperative follow-up.

A preliminary study was conducted to estimate the sample size for this study because the FDD had never been used before. The authors analyzed the clinical data of 10 patients who underwent the FDD or stoma procedure. In the preliminary study, no cases of FDD-related morbidity or mortality occurred. Several complications occurred, including wound site discharge and urinary tract infections. However, these complications were not considered directly related to the FDD or the stoma since they were common and might occur in other surgeries. The authors calculated the sample size as the total medical expense from the first operation to the FDD or stoma removal. For sample size estimation, G Power (version 3.1.9.2) was used; accordingly, 19 patients must be recruited in each arm to demonstrate a statistically significant difference between groups with 80% power at the 0.05 level of significance. Allowing for a dropout rate of 10% after the randomization, the sample size was estimated as 21 patients in each group.

All analyses were conducted using IBM SPSS version 23 (IBM, Armonk, NY, USA) and R version 3.5 with the rms and Hmisc package (web-r.​org). All analyses were performed using an intention-to-treat approach. Morbidity and mortality related to surgery, failure to protect the anastomosis, and other AL-related complications were compared using the Pearson χ² test or Fisher’s exact test. Total hospital days, total cost, duration of FDD or stoma, operation time, and delay in establishing soft diets were analyzed using Student’s t test. Values of P < 0.05 were considered statistically significant.

There were no cases of mortality in this study. Morbidity affected 20 (74.1%) patients and 18 (66.7%) patients in the FDD and stoma groups, respectively (P = 0.551) (Table 2). There were no statistically significant differences in Clavien-Dindo classification grade 2 or 3 complications between the groups (P = 0.584). In the stoma group, 10 patients (37.0%) underwent morbidity related to the stoma. One patient underwent partial necrosis of stoma, and two patients underwent prolapsed stoma. The other patients underwent skin erosion and peristomal pain. All patients did not need intervention, and their symptoms improved with conventional treatment. The overall AL rate was 27.8% (15 patients), but it did not differ significantly between groups (6 patients [22.2%] in the FDD group vs 9 patients [29.0%] in the stoma group, P = 0.555). The severity of AL is listed in Table 3 according to the grading proposed by International Study Group of Rectal Cancer [10]. The rates of postoperative obstruction, ileus, and wound infection were similar in the two groups. Two patients had FDD-related erosions, but neither had AL in the 3-week contrast test. Part of the band was noted at the mucosal defect on a sigmoidoscopy examination after FDD removal. Neither patient had generalized or localized peritonitis symptoms. Both were followed-up and did not require additional admission or further treatment.

The FDD was in place for the required duration, and the fecal diversion capacity was comparable to that of the stoma. There was no early drop out of FDD, and no conversion to stoma until the scheduled removal. The total hospital stays and medical costs are shown in Table 4. The mean postoperative hospital stay after the primary surgery was 13.9 ± 8.8 days in the FDD group versus 10.5 ± 4.2 days in the stoma group (P = 0.072). However, the stoma group needed further surgery and hospitalization for the ileostomy take-down procedure. The total hospital stays contained the postoperative hospital stay after the primary surgery and reoperation for ileostomy take-down. The mean total hospital stay of the FDD group was 16.4 ± 6.7 days, statistically shorter than that of the stoma group (23.4 ± 8.7 days) (P = 0.002). The mean total hospital cost was 12,726.8 ± 3422.8 USD in the FDD group and 17,954.9 ± 9040.3 USD in the stoma group (P = 0.008).

The mean operation time and intraoperative bleeding did not differ between groups (Table 5). The first bowel movement occurred at a mean of 3.1 ± 1.7 days in the FDD group versus 2.3 ± 1.1 days in the stoma group (P = 0.034). However, the delay in dietary progression was similar between groups (P = 0.204).

The clinical AL results are shown in Table 6. The mean fecal diversion procedure duration was 21.6 ± 6.1 days in the FDD group versus 114.9 ± 41.3 days in the stoma group (P < 0.0001). Of the 54 patients, 15 (27.8%) had AL; all cases were effectively managed by FDD and loop ileostomy. Three patients had FDD longer than 3 weeks. Of the 3 patients, 1 patient removed the FDD at 6 weeks, and the other two patients were converted to the stoma procedure due to contrast leakage at 6 weeks postoperative. AL was clinically detected in 14 of the 15 patients at the 1-week CT examination without a significant intergroup difference. Two patients in the stoma group had persistent AL on the 10-week contrast study. After 6 months postoperative, 3 patients of each group had a stoma. In the FDD group, 1 patient underwent transverse colostomy due to colonic necrosis. The other patients underwent stoma due to continuous anastomotic leakage. In the stoma group, two patients kept the stoma due to continuous anastomotic leakage. The other patient underwent delayed leakage after ileostomy take-down. Of the 6 patients, 5 patients still have a stoma after finishing the follow-up.