Objective
To test the hypothesis that the use of closed suction pelvic drains placed at time of large bowel resection (LBR) for ovarian cancer (OC) decrease morbidity following anastomotic leak (AL).
Methods
Consecutive cases of LBR for OC between 01/01/1994 and 06/20/2011 were retrospectively identified. Drains were routinely used until bowel movement. AL was defined as: 1) feculent fluid from drains/wound/vagina, 2) radiographic evidence of AL, or 3) AL found at reoperation. Descriptive statistics, Wilcoxon rank-sum, Pearson's chi-square and Fisher's exact test were used.
Results
43 cases met inclusion criteria. AL was characterized by method of diagnosis as follows: change in drain output only (DO, n=8); change in drain output associated with ambiguous clinical signs/symptoms (D-SSX, n=11); or clinical signs/symptoms only (SSX, n=24). The sensitivity of drains in diagnosing AL was 50%. Time to diagnosis was earlier in DO/D-SSX (median 7 vs. 11 days, P=0.003), however, no significant differences were observed in rates of reoperation, length of stay, time to chemotherapy (TTC), and 30- and 90-day mortality between DO/D-SSX and SSX. Comparing cases where no drains were placed (n=5) vs. those with drain (n=38), we observed no differences in outcomes. TTC though statistically significant (47 vs. 59 days, P=0.023) was not clinically significant.
Conclusions
Though a change in drain output correlated with earlier diagnosis, this did not appear to impact overall outcomes. We did not find strong evidence supporting routine prolonged drainage after LBR for OC. Additionally, absence of change in drain output does not rule out presence of AL.