Low diagnostic accuracy of rigid sigmoidoscopy in the investigation of rectal cancer

We performed a retrospective observational study to investigate the diagnostic accuracy of rigid sigmoidoscopy (RS) in patients with rectal cancer (n=279). Fifty-six percent of the patients had performed an RS within three months before diagnosis and mostly by a primary care provider (93%). In 21% of the patients the physician determined that the examination was normal, in 50% a rectal tumor was suspected and in 29% of cases an unspecific pathology (e.g. luminal blood, mucosal abnormalities) was reported. A normal finding on RS was associated with a longer time between the first appointment and subsequent diagnosis (multivariate hazard ratio (HR) 0.50; 95th percentile CI 0.35-0.71) whereas a history of rectal bleeding (multivariate HR 1.49; 95th percentile CI 1.01-2.20) and adherence to new national guidelines (multivariate HR 1.46; 95th percentile CI 1.08-1.99) was associated with a shorter time to diagnosis. We conclude that RS only had modest diagnostic accuracy in the diagnosis of rectal cancer, at least in this mainly primary care-based setting.