Background: To compare associations of conventional risk factors with cardiovascular death within couples and in the population as a whole.
Methods: We analysed baseline data (1995-97) from the HUNT2 Study in Norway linked to the national Causes of Death Registry. We compared risk within couples using stratified Cox regression.
Results: During 914776 person-years, 3964 cardiovascular deaths occurred, and 1658 of the deaths occurred among 1494 couples. There were consistently stronger associations of serum lipids and blood pressure with cardiovascular mortality within couples compared to the population as a whole. For instance, for systolic blood pressure (per 20mmHg), the hazard ratio (HR) within couples was 1.28 (95% confidence interval: 1.17, 1.40) compared to 1.16 (1.12, 1.20) in the total population, and for diastolic pressure (per 10mmHg), the corresponding HRs were 1.16 (1.07, 1.26) and 1.11 (1.08, 1.13). Anthropometric factors (BMI, waist circumference, waist-hip ratio) as well as diabetes, smoking, physical activity, and education, showed nearly identical positive associations within couples and in the total population.
Conclusions: Prospective population studies may tend to slightly underestimate associations of these factors with cardiovascular mortality.
Keywords: Cardiovascular mortality; Confounding; Couples; Population study; Risk factors.
Copyright © 2017 Elsevier B.V. All rights reserved.