Analysis of survival of subjects with colorectal cancer diagnosed by different modalities can provide insight into the mechanism by which screening has an effect. It can also give an indication of the feasibility of using prognostic indicators as surrogate outcome measures to predict mortality in future studies.


This paper examines the survival of individuals with colorectal cancer diagnosed in the Nottingham trial and explores the role of selected prognostic factors as possible surrogate outcome measures.


Survival was significantly better in subjects with screen-detected cancers than in controls, even after adjusting for tumour stage and accounting for lead-time bias. Survival was inversely related to stage of tumour, with patients with stage A tumours having the best survival. Subjects with well or moderately differentiated tumours had a significantly better survival than those with poorly differentiated tumours.


Screening for colorectal cancer by means of faecal occult blood testing improved survival among subjects with screen-detected cancers. Differences in prognostic factors largely explain the differences in survival between both non-responders and subjects with interval cancers and those in the control group, but not the improved prognosis for patients with screen-detected cancers. The use of such factors as surrogate outcome measures may therefore be inappropriate. © 1999 British Journal of Surgery Society Ltd