Objectives: To assess the association between Body Mass Index (BMI) and cause-specific mortality in older adults and to assess which BMI was associated with lowest mortality.
Design: Prospective study.
Setting: European towns.
Participants: 1,980 older adults, aged 70-75 years from the SENECA (Survey in Europe on Nutrition and the Elderly: a concerted action) study.
Measurements: BMI, examined in 1988/1989, and mortality rates and causes of death during 10 years of follow-up.
Results: Cox proportional hazards model including both BMI and BMI², accounting for sex, smoking status, educational level and age at baseline showed that BMI was associated with all-cause mortality (p<0.01), cardiovascular mortality (p<0.01) and mortality from other causes (p<0.01), but not with cancer or respiratory mortality (p>0.3). The lowest all-cause mortality risk was found at 27.1 (95%CI 24.1, 29.3) kg/m², and this risk was increased with statistical significance when higher than 31.4 kg/m² and lower than 21.1 kg/m². The lowest cardiovascular mortality risk was found at 25.6 (95%CI 17.1, 28.4) kg/m², and was increased with statistical significance when higher than 30.9 kg/m².
Conclusion: In this study, BMI was associated with all-cause mortality risk in older people. This risk was mostly driven by an increased cardiovascular mortality risk, as no association was found for mortality risk from cancer or respiratory disease. Our results indicate that the WHO cut-off point of 25 kg/m² for overweight might be too low in old age, but more studies are needed to define specific cut-off points.