Document Type



Glucose tolerance tests on 157 adult Athapaskan-speaking Dogrib Indians residing in three villages in the Canadian Northwest Territories were carried out in April and October, 1979. The proportions of hyperglycemia in males and females, respectively, were: fasting (5= 140 mg/dl) 2% and 1%; hour-2 (> 160 mg/dl) 27% and 20%; hour-2 (2= 200 mg/dl) 13% and 7%. The regression of log (hour-2 glucose) on age was significant within each sex (males: y = 1.887 + .0044X; females: y = 1.993 + .0023X). Skinfold measurements were also obtained at seven body sites (4 on the trunk, 3 on the appendages) and compared between low glucose (hour-2: 160 mg/ dl) and high glucose (> 160 mg/dl) categories within each sex. Significant differences were observed at 2 sites in males only. Principal components analysis was also done on the skinfold measurements. Three components, which collectively account for 90 + % of the general variance, were identified. Hierarchical analysis of variance was used to test age-adjusted principal components for effects of sex, glucose level and localities sampled in different seasons. A significant effect was observed only on the second principal component, indicating that a trunkal deposition of body fat is significantly associated with elevated glucose level (F = 3.968, P < .05) and locality/season (F = 6.777, p .01). Examination of the same data in more rigorously defined glucose groups (low: < 200 mg/dl; high: 3= 200 mg/dl) increased the significance of these associations. Persons at risk for developing diabetes may be those who begin to deposit fat centripetally relative to age/sex appropriate members of their own population.


P. 499 Insert the following sentence at the beginning of the last paragraph: “To improve interpretability of the patterns, the components were subjected to one orthogonal (VARIMAX) rotation.”