Background The effect of modifying lifestyle at middle age on mortality has been sparsely examined.

Methods Men and women aged 50–54 years randomised to the control group (no intervention) in the population-based Norwegian Colorectal Cancer Prevention trial were asked to fill in lifestyle questionnaires in 2001 and 2004. Lifestyle scores were estimated ranging from 0 (poorest) to 4 (best) based on health recommendations (non-smoking, daily physical activity, body mass index <25.0 kg/m² and healthy diet). Outcomes were all-cause, cancer and cardiovascular mortality before 31 December 2013.

Results Of the 6886 attainable individuals included in the study, 4211 (61%) responded to the baseline questionnaire in 2001. After a median follow-up of 12.3 years, 226 (5.4%) of the baseline questionnaire responders died; 110 (49%) from cancer and 32 (14%) from cardiovascular disease. For each increment in lifestyle score in 2001, a 21% lower all-cause mortality was observed (HR 0.79, 95% CI 0.67 to 0.94, adjusted for age, sex, occupational working hours and chronic disease or pain during 3 years before enrolment). A one-point increase in lifestyle score from 2001 to 2004 was associated with a 38% reduction in all-cause mortality (adjusted HR 0.62, CI 0.45 to 0.84). The group reporting lifestyle change from score 0–1 (unfavourable) in 2001 to score 2–4 (favourable) in 2004 had 4.8 fewer deaths per 1000 person years, compared with the group maintaining an ‘unfavourable’ lifestyle (adjusted HR 0.31, CI 0.13 to 0.70 for all-cause mortality).

Conclusions Favourable lifestyle changes at age 50–60 years may prevent early death.

Trial registration NCT00119912; pre-results.