Colovaginal and colovesical fistulae: the diagnostic paradigm

Abstract

Colovaginal and colovesical fistulae (CVF) are relatively uncommon conditions, most frequently resulting from diverticular disease or colorectal cancer. A high suspicion of a CVF can usually be obtained from an accurate clinical history. Demonstrating CVF radiologically is often challenging, and patients frequently undergo a multitude of investigations prior to definitive management. The aim of this study was to develop an algorithm for the investigation of suspected CVF in order to improve diagnosis and subsequent management. Thirty-seven patients from a single NHS Trust with a diagnosis of colovaginal or colovesical fistula were included in the study. Clinical records and imaging were reviewed retrospectively, and data on demographics, symptoms, investigations, management and outcome were collated. A total of 87.5% patients with a colovesical fistula presented with pathognomic symptoms of faecaluria or pneumaturia. The commonest aetiologies were diverticular disease (72.9%), colonic and gynaecological neoplasia (10.8% each). Computerised tomography (CT) was the most frequently performed investigation (91.9%) and was most sensitive in detecting the fistula (76.5%) and underlying aetiology (94.1%). Colonoscopy was most sensitive in detecting an underlying colonic malignancy (100%). Resectional surgery was performed in 62.1% of cases, although morbidity and 1-year mortality was significant, with rates of 21.7 and 17.4%, respectively. The diagnosis of CVF is predominately a clinical one, and patients with a suspected CVF are over-investigated. Investigations should be focused on determining aetiology rather than demonstrating the fistulous tract itself. We propose that, in the majority of cases, CT and lower gastrointestinal endoscopy should suffice.