OBJECTIVES: to determine the association between the subjective rate of ageing and future health outcomes. DESIGN: prospective cohort study (Whitehall II study). At the third phase of the study (1991-1993), participants were asked at what age they think most people leave middle age. Participants were followed until the end of phase 5 (1997-2000), so that mean length of follow-up was 7 years. SETTING: London based office staff in 20 civil service departments. SUBJECTS: 5,262 male and 2,277 female civil servants aged 40-60. MEASURES: validated new cases of coronary heart disease and health function, measured by the SF-36 General Health Survey, at phase 5. RESULTS: perceived age of leaving middle age increased with age, self-rated health and grade of employment, and was higher in women. Adjusting for age and sex, people who believed middle age ends < or =60 years, compared to > or =70 years, were at higher risk for coronary heart disease (HR=1.43, 95% CI=1.05-1.94), fatal coronary heart disease and non-fatal myocardial infarction (HR=1.52, 0.95-2.42), and poor physical (OR=1.29, 1.10-1.50) and mental (OR=1.25, 1.07-1.45) functioning during follow-up. Adjustment for self-rated health, employment grade, health behaviours, social networks, control and baseline health status, respectively, did not eliminate these associations. CONCLUSIONS: the reported age at which middle age ends predicts future health outcomes. We hypothesise that perceived end of middle age acts as a general summary of the subjective rate of ageing.