Document Type



Rocco Micciolo


The associations between fat pattern and metabolic profile were evaluated , using cluster analysis, in 214 males and 244 fem ales (aged 21-60 years). The subjects were randomly selected from the patients of one general practice in Castel D’Azzano (Italy). The metabolic status was indexed by blood pressure, blood glucose, uric acid , serum triglycerides, total cholesterol, low-density lipo protein cholesterol, and high-density lipo protein cholesterol (both total and percentage). Fat pattern was evaluated using six skinfolds and three circumferences. The analyses were performed separately on anthropometric and metabolic variables for each sex. The analyses were repeated five times to evaluate the stability of the cluster solution . Anthropometric rather than metabolic variables provided a more stable solution for cluster analysis. Anthropometric clusters showed significantly different metabolic patterns. Most of the differences disappeared when a body mass index (BMI) effect was taken into account but not when an adjustment for waist to hip ratio (WHR) was made. On the other hand, neither BMI nor WHR was able to fully account for the differences in metabolic profiles of metabolic clusters. Cross-classification results of anthropometric and metabolic cluster mem bership showed a significant, although moderate, degree of association between the two classification sets. These results indicate that in both sexes an thropometric characteristics contribute to a limited degree to the metabolic profile and that the associations probably reflect a complex mechanism.