Outcome of large-bowel perforation in patients with colorectal cancer

Abstract
PURPOSE: Perforation of the colon is seldom associated with malignant disease. Operative mortality varies widely in published studies and little is known about patterns of failure and long-term outcome. An observational study was undertaken to assess the outcome of colorectal cancer complicated by perforation. METHOD: we reviewed a series of 83 consecutive patients treated during a 14-year period at one institution. RESULTS: Fifty-four (65 percent) patients had perforation of the tumor itself, and 29 (35 percent) had diastatic perforation proximal to an obstructing tumor. Twenty-six (31.5 percent) patients had metastatic disease at laparotomy. Primary resection of the diseased segment was performed in 47 (87 percent) patients with perforation of the tumor itself and in 21 (72.4 percent) patients with diastatic perforation proximal to an obstructing tumor. However, only 57 patients (39 (72.2 percent) with perforation of the tumor itself; 18 (62 percent) with diastatic perforation proximal to an obstructing tumor; P=not significant) were potentially cured. Operative mortality was 16.7 and 48.3 percent, respectively (PPP=0.023). At a mean follow-up of 43 (median 31) months, 21 (46 percent) of the 46 potentially cured survivors were alive. The local recurrence rate was 22.9 percent in patients with perforation of the tumor itself and 18.2 percent in patients with diastatic perforation proximal to an obstructing tumor (P=not significant). Peritoneal seeding occurred in 17 and 0 percent (P=not significant); the mean disease-free interval was 33.9 and 49.9 months (P=not significant); and five-year cumulative disease-related survival probability was 0.51 and 0.90 (P=0.049), respectively. CONCLUSIONS: Diastatic perforation proximal to an obstructing tumor is associated with higher operative mortality and better cancer-related survival than a tumor perforating through the bowel wall. Early diagnosis in diastatic perforation and aggressive management of sepsis associated with radical surgical resection is recommended.

This publication has 15 references indexed in Scilit: