BACKGROUND: The relationships of body mass index and skinfold thickness to all-causes mortality during a 10-year follow-up were assessed in 8,341 men and 1,100 women ages 30-69 years from different Italian population samples. RESULTS: Among men, both univariate and multivariate analyses showed a clear-cut parabolic (inverse J-shaped) relationship, whose left branch became less steep after the exclusion of smokers, people carrying severe diseases at entry, those who died during the first 5 years, or all of them. The minimum risk was almost always located around 28 units of body mass index, and it decreased to smaller levels of body mass index when the exclusions were adopted. The analysis of skinfold thickness showed similar but less clear-cut results. Among women, due to the limited number of fatal events, the analysis was unable to show any clear relationship of body mass index or skinfold thickness to all-causes mortality. The multivariate analysis showed a similar parabolic relationship of body mass index to all-causes mortality. The estimated multivariate risk of death broken down into five quintiles was unrelated to the mean level of body mass index for each quintile. From the multivariate model it was estimated that an excess of 10 kg in body weight (above the body mass index corresponding to the minimum risk level and everything else being equal) carries the same excess of risk produced by 5 mm Hg of systolic blood pressure or by less than 5 cigarettes smoked per day. CONCLUSION: These results suggest that high levels of obesity indicators are only slightly associated with an excess mortality and that overweight and obesity are health hazards only if they are accompanied by an elevation of other risk factors, mainly of blood pressure.
BACKGROUND: The relationships of body mass index and skinfold thickness to all-causes mortality during a 10-year follow-up were assessed in 8,341 men and 1,100 women ages 30-69 years from different Italian population samples. RESULTS: Among men, both univariate and multivariate analyses showed a clear-cut parabolic (inverse J-shaped) relationship, whose left branch became less steep after the exclusion of smokers, people carrying severe diseases at entry, those who died during the first 5 years, or all of them. The minimum risk was almost always located around 28 units of body mass index, and it decreased to smaller levels of body mass index when the exclusions were adopted. The analysis of skinfold thickness showed similar but less clear-cut results. Among women, due to the limited number of fatal events, the analysis was unable to show any clear relationship of body mass index or skinfold thickness to all-causes mortality. The multivariate analysis showed a similar parabolic relationship of body mass index to all-causes mortality. The estimated multivariate risk of death broken down into five quintiles was unrelated to the mean level of body mass index for each quintile. From the multivariate model it was estimated that an excess of 10 kg in body weight (above the body mass index corresponding to the minimum risk level and everything else being equal) carries the same excess of risk produced by 5 mm Hg of systolic blood pressure or by less than 5 cigarettes smoked per day. CONCLUSION: These results suggest that high levels of obesity indicators are only slightly associated with an excess mortality and that overweight and obesity are health hazards only if they are accompanied by an elevation of other risk factors, mainly of blood pressure.