Long term risk factors for coronary heart disease and stroke: Influence of duration of follow-up over four decades of mortality surveillance

Aim While cohort studies have revealed a range of risk factors for coronary heart disease and stroke, the extent to which the strength of these associations varies according to duration of follow-up in studies with extended disease surveillance is unclear. This was the aim of the present study. Methods and results Initiated in 1967/70, the original Whitehall study is an on-going cohort study of 15,402 male UK government workers free of coronary heart disease when they took part in a baseline medical examination during which a range of standard risk factors was measured. In analyses in which we stratified by duration of follow-up, there was evidence of time-dependency for most risk factor-disease relationships. Thus, the associations of systolic and diastolic blood pressure, total cholesterol and cigarette smoking with coronary heart disease and stroke diminished in strength with increasing duration of follow-up, whereas the magnitude of the body mass index-coronary heart disease relation was unchanged. For example, the age-adjusted hazard ratios (95% confidence interval) for cigarette smoking (versus never smoked) in relation to coronary heart disease were: 2.49 (1.80, 3.44), 1.65 (1.34, 2.03), 1.36 (1.15, 1.61) and 1.32 (1.10, 1.58) for follow-up periods 0-10, 10-20, 20-30 and 30+ years, respectively. Conclusion Despite a general diminution in the strength of effect over time, even in the fourth decade of follow-up, classic risk factors retained some predictive capacity for coronary heart disease and, to a lesser degree, stroke.