Determinants of recurrence after sigmoid resection for uncomplicated diverticulitis

This study aimed to evaluate the impact of surgery-associated variables on recurrence rates after sigmoid resection for diverticulitis.

Patients who underwent elective sigmoid resection for uncomplicated diverticulitis between 1992 and 2000 at two tertiary referral centers were followed up for recurrent disease as the primary end point. Recurrence after surgery was defined as left lower quadrant pain, fever, and leukocytosis, with consistent CT and/or contrast enema findings on admission and after six weeks. A logistic regression of the following variables was undertaken: patient demographics, duration of preoperative symptoms, previous admissions and abdominal surgery, surgical access (laparoscopic or open), postoperative complications, splenic flexure mobilization, anastomotic technique (handsewn or stapled), specimen length, inflammation at proximal resection margin, and anastomotic level (colosigmoid or colorectal). The last three variables were defined by the pathologist. Anastomosis level was based on muscle layer configuration (taeniae coli) at the distal resection margin.

Two hundred thirty-six patients (105 females) with a mean age of 60.4 (standard deviation, +/- 10) years were available for follow-up at 67 +/- 30 (range, 11-130) months. The median duration of preoperative symptoms was 18 (range, 12-120) months. All but one (99 percent) patient had at least one admission before surgery. One hundred forty (59 percent) and 96 (41 percent) patients underwent laparoscopic or open resection, respectively. The conversion rate was 13 percent (18 patients) in the former group and the 30-day complication rate was 23 percent, with 0.4 percent 30-day mortality and a 2.1 percent reoperation rate. The splenic flexure was mobilized in 109 patients (47 percent). Anastomoses were fashioned by stapler in 171 patients (73 percent) and were to the rectum in 143 patients (72 percent). Specimen length was 17.9 +/- 5.9 (range, 9-47) cm with inflammation at the proximal margin in 30 patients (14 percent). Twelve (5 percent) patients developed a recurrence at a mean of 78 +/- 25 (range, 34-109) months with reoperation in one (0.4 percent). The level of anastomosis was the only predictor of recurrence in regression analysis (P = 0.033). Patients with colosigmoid anastomosis had a four times higher risk of having a recurrence compared with patients with colorectal anastomosis (odds ratio, 95 percent confidence interval = 1.12, 14.96).

Colorectal (rather than colosigmoid) anastomosis was the single predictor of lower recurrence rates after elective sigmoid resection for uncomplicated diverticulitis.