European Journal of Epidemiology

, Volume 25, Issue 9, pp 619–625

Leisure-time physical activity, cardiovascular risk factors and mortality during a 34-year follow-up in men


  • Salla Savela
    • Oulu City Hospital
  • Pentti Koistinen
    • Oulu City Hospital
  • Reijo S. Tilvis
    • Department of Medicine, Geriatric ClinicUniversity of Helsinki
  • Arto Y. Strandberg
    • Department of Medicine, Geriatric ClinicUniversity of Helsinki
  • Kaisu H. Pitkälä
    • Unit of General PracticeHelsinki University Central Hospital
  • Veikko V. Salomaa
    • THL-National Public Health Institute
  • Tatu A. Miettinen
    • Department of Medicine, Geriatric ClinicUniversity of Helsinki
    • Department of Health Sciences/GeriatricsUniversity of Oulu (Oulun yliopisto) and University Hospital

DOI: 10.1007/s10654-010-9483-z

Cite this article as:
Savela, S., Koistinen, P., Tilvis, R.S. et al. Eur J Epidemiol (2010) 25: 619. doi:10.1007/s10654-010-9483-z


The inverse relationship between physical activity and mortality may be confounded by socioeconomic factors, cardiovascular risk factors and inverse causality. We investigated long-term association between self-reported regular physical activity and mortality in a socioeconomically homogeneous, initially healthy middle-aged (mean age 47) male cohort (the Helsinki Businessmen Study). In 1974, the men were assessed with questionnaires, clinical and laboratory examinations. Cardiovascular disease (CVD) risk factors (including body mass index [BMI], age, cholesterol, glucose, systolic blood pressure and smoking) and details of physical activity of 782 men were available. Leisure time physical activity was collapsed into 3 categories: low (n = 148), moderate (n = 398) and high activity (n = 236). Physical activity was also briefly assessed in questionnaire surveys in 1985–1986 and in 2000. Total mortality up to 2007 was retrieved from the Central Population Register. Altogether 295 men (37.7%) died during the 34-year follow-up, and leisure-time physical activity was significantly related to mortality in a step-wise manner: 45.9% (n = 68), 37.7% (n = 150), and 32.6% (n = 77) died in the low, moderate, and high activity groups, respectively (P < 0.001). With high activity group as referent and adjusted for midlife CVD risk, perceived health and fitness at baseline, hazard ratio for total mortality was 1.21 (95% confidence interval: 0.90, 1.62), and 1.61 (95% confidence interval: 1.13, 2.30) in the moderate and low activity groups, respectively. Conclusion: During the 34-year follow-up, leisure-time physical activity in initially healthy middle-aged men had a graded association with reduced mortality that was independent of CVD risk, glucose and BMI.


Body mass indexCardiovascular riskExerciseMortalityPhysical activitySocial class



Body mass index


Confidence interval


Cardiovascular disease


Hazard ratio


Standard error


Standard deviation

Copyright information

© Springer Science+Business Media B.V. 2010