Objective To report the causes of, and ages at, death of subjects in an English colorectal cancer screening trial.

Design and setting Analysis of 78 708 deaths occurring between 1981 and 2008, within the Nottingham randomised controlled trial of biennial faecal occult blood testing.

Main outcome measures Cause of death, age at death by sex and by cause.

Results Significantly more subjects died from verified colorectal cancer in the trial's control group than in the intervention group (3.2% vs 2.9%). For no other major cause of death was the difference in proportion across the two groups statistically significant. Age at death was lower for cancer than for other principal causes, except for ischaemic heart disease among women. However, mean age at death was higher for colorectal cancer than for other cancers, except for prostate cancer among men. Increasing levels of material deprivation significantly lowered the expected ages at death, independently of cause. For both men and women, the mean age at death from all causes for screening participants was higher than that of controls and non-participants. Mean deprivation was lowest among participants. Of those participating in screening, and dying from colorectal cancer, subjects receiving negative test results lived significantly longer than those who received positive test results. However, if dying from other causes, they died at an earlier age.

Conclusions The age at death from colorectal cancer is higher than that of most other cancers. Those accepting a screening invitation live longer than non-participants. In part, this difference is explained by relative deprivation. Among screening participants, the receipt of a positive, as opposed to a negative, test result is associated with a later age at death.