Abstract
We aimed at comparing 2 a-priori -Mediterranean Adequacy Index (MAI), the Median Score (MED) - versus 2 a-posteriori, -Factor Analysis (FA2) and Principal Components analysis (PC2)- dietary scores in 1214 CHD-free men aged 45-64 belonging to the Italian Rural Areas of the Seven Countries Study examined in 1965 and followed-up for mortality during 40 years. CHD death was the end-point. Collection of dietary history allowed to define 17 main food groups expressed in gr/day and to compare dietary scores, each divided into 3 classes. Kaplan-Meier curves showed higher survival for classes 2 and 3 (healthy) versus class 1 (unhealthy), but the log-rank test was not significant for the 2 a-priori scores. Cox proportional hazards models showed similar significant findings comparing class 3 with class 1 in the a-posteriori scores FA2 and PC2, with hazards ratios of 0.48 and 0.43, that became 0.65 and 0.53 respectively after adjusting for six specific risk factors for CHD (age, cigarette smoking, systolic blood pressure, serum cholesterol, body mass index and physical activity). Food intake of class 3 in all 4 scores matched rather well the characteristics of the Mediterranean Diet. The poor performance of a-priori dietary scores was partly explained by the unexpected and direct relationship of fruit intake with CHD deaths. The superiority of a-posteriori dietary patterns could be in turn due to the specificity of the study population. External validations and comparisons of a-priori versus a-posteriori dietary patterns in larger cohorts, using the same 17 food groups explored here, are urgently needed.