We would like to commend Dinning and colleagues1 for their recent article regarding the prevalence of distal colonic neoplasia in patients with proximal lesions. This is an important issue meriting exploration, but we are concerned that some of their conclusions could be easily misapplied.
No one questions the need for colonoscopy in evaluating symptomatic patients; in this group, it is clear that the entire colon should be visualized. Patients presenting for screening, however, are, by definition, symptom free. If one assumes a prevalence of asymptomatic colon cancer in the population of 1% (which may be high) and a false-positive rate of 10% for 60-cm screening flexible sigmoidoscopy, even missing 25% of lesions still yields a negative-predictive value of 99.7%. Therefore, an asymptomatic patient who has a negative value of screening flexible sigmoidoscopy has only a 0.3% chance of harboring a significant lesion, not the 25% chance as could be interpreted