I read with great interest the article by Brenner et al1 in the June 2009 issue of Archives. Recognizing the difficulties inherent to a retrospective analysis of a heterogeneous population of patients, I wished to comment about their intriguing findings.
First, I would highlight a correction to Table 2: the recurrence rate between 7 and 12 weeks is only 13%, not 27%. Also, the rate of fistula recurrence for operations earlier than 36 weeks is 9.6%, rather than 12%. Second, a clearer definition of an enterocutaneous fistula would be helpful. It is unusual that the authors report the need for emergency surgery in 10% of patients (7 with anastomotic leak and peritonitis and 7 with abscesses draining to the abdominal wall). Eleven of these 14 emergency procedures were performed within the early period before 36 weeks, representing 32% of that total. Perhaps their inclusion contributes to the better outcomes observed with earlier intervention. Some readers might suggest that an early anastomotic leak and peritonitis represent a technical failure from intestinal resection, rather than a true enterocutaneous fistula. Were radiology procedures to percutaneously control intra-abdominal sepsis less available in the earlier years of the study and a reason emergency surgery was required for patients with abscesses draining to the abdominal wall? Many would choose not to perform definitive surgery for an enterocutaneous fistula in the setting of uncontrolled infection. Finally, did the authors analyze their records to characterize whether a volume-outcome relationship of surgeon experience might explain the differences observed?