Biased Risk Factor Assessment in Prospective Studies of Peripheral Arterial Disease Due to Change in Exposure and Selective Mortality of High-Risk Individuals

Abstract

Background Our aim was to assess whether risk factor assessment in prospective studies of peripheral arterial disease (PAD) might be biased by change in exposure and selective mortality of individuals at high risk.

Methods The cohort ‘;Men born in 1914’ has been followed since the baseline examination 1969. PAD, (i.e. ankle-brachial pressure index < 0.9) at the age of 68 was related to risk factors measured at the age of 55 and 68.

Results Hypertension was present in 64% of the men. Of these, 62% were normotensive at 55. Men with hypertension at the age of 55 had lower odds for PAD (odds ratio 1.9; 95% Cl 0.8-4.5) than men who were normotensive at 55 (odds ratio 3.0; 95% Cl 1.5-6.1). One-third remained smokers at 68; 27% were ex-smokers. The odds for PAD was 3.0 times higher (95% Cl 1.4-6.3) in ex-smokers and 2.6 times higher (95% Cl 1.3-5.4) in current smokers than in non-smokers. Less than half remained in the same cholesterol quartile. The odds for PAD was 1.5 (95% Cl 0.8-2.8) in men with hypercholesterolaemia at 68 and 2.2 (95% Cl 1.1-4.7) in men with hypercholesterolaemia at both examinations. Smoking and hypertension at the age of 55 were both associated with an increased mortality rate: 26% of the smokers and 31% of the men with hypertension died before the age of 68.

Conclusion Risk factor assessment in prospective studies of PAD is affected by change in exposure and selective mortality of individuals at high risk.