Ecological Correlates of Differences in Mean Age at Death Across Nearly Extinct Cohorts: The Role of Dietary Habits

Objectives. Our objective was to study the ecological relationship of many risk factors and personal characteristics with mean age at death (AD) after a 50-year follow-up of nearly extinct cohorts. Material and Methods. There were 16 cohorts totaling 12,763 middle-aged men enrolled in the Seven Countries Study (SCS), and 58 variables were measured, including traditional risk factors, dietary nutrition and anthropometric variables. A follow-up of 50 years allowed the use of AD as the end-point. Analysis included simple linear regression correlation and multivariate modelling using Principal Component Analysis and regression and Ridge regression. Results. Out of 58 variables, only 11 (10 nutrition-dietary items plus age) showed a significant linear correlation coefficient (R) ≥ 0.50 and a p value ≤ 0.05. Linear regression was computed by using as a predictor the dietary factor score derived from a Principal Component Analysis of the 11 significant variables, which were used as independent variables, whose coefficients were significantly related with AD, and the final R2 was 0.52. The Principal Component regression and Ridge regression documented the direct relationship of food groups of vegetable origin (including olive oil) with the AD and the inverse relationship for food groups of animal origin. Conclusions. A few variables, all related to diet and nutrition, were able to statistically explain about 50% of the different AD in 16 cohorts of men followed up with nearly until death. Other variables, including traditional cardiovascular disease risk factors, did not contribute in a significant way for this purpose.