Age at death and the effect of lead-time bias in patients with colorectal cancer: a 10-year follow-up

Aim Studies addressing the benefit of early intervention are prone to lead-time bias, which results in an artificial improvement in cancer-specific mortality. We have previously compared the age at death for patients with colorectal cancer presenting on an emergency or elective basis. In this study, we aimed to repeat the analysis with a minimum follow-up of 10 years. Method A nonscreen-detected cohort of patients presenting with colorectal cancer to three Lanarkshire Hospitals between 2000 and 2006 were entered into a prospective database, with analysis performed on 28 November 2016. The following data were collected: age at death, presentation type (emergency/elective), operative intent (palliative/curative) and Dukes stage. Results are presented as [mean (95% confidence intervals)]. Statistical analysis was undertaken using Student's t-test and multivariate analysis performed using Cox proportional hazard models. Results One thousand six hundred and thirty-six patients were identified. Elective patients presented younger than emergency patients [67.9 (67.3-68.5) vs 70.9 (69.6-72.2) years; P < 0.0001]. Overall mortality was 71.1% at time of analysis; no difference was seen in the mean age at death between emergency and elective presentation [73.5 (72.4-74.8) vs 73.6 (72.3-74.9) years; P = 0.841]. Conclusion Current early detection strategies to diagnose colorectal cancer may improve cancer-specific survival by increasing lead-time bias. However, in our cohort of symptomatic patients, treatment on an elective or emergency basis does not influence overall survival. These data suggest that in selected patients, particularly where there is comorbidity, it may be reasonable to adopt a more expectant approach to investigate and treat colorectal symptoms.