Relation of age, exercise, anthropometric measurements, and diet with glucose and insulin levels in a population aged 70 years and over.

A community-based sample of people > or = 70 years from Mosgiel, New Zealand, was investigated to determine the relation of age, diet, exercise, drugs, and anthropometric measurements with glucose and insulin. From an initial sample of 856 subjects on August 1, 1988, 782 (91.4%) completed the questionnaires and physical examination. Glucose was estimated in 726 subjects and insulin in 607 subjects 2 hours after a standardized meal. In the multivariate analysis for women, glucose was related to age and exercise, and insulin was related to glucose levels, triceps skinfold thickness, and waist/hip ratios. In the multivariate analysis for men, none of the variables was related to glucose levels; insulin levels were related to glucose and waist/hip ratio. Impaired glucose tolerance in women was associated with high body mass index, waist/hip ratio, lower exercise levels, and the taking of thiazide drugs or oral steroids. In men, no significant model to identify those with impaired glucose tolerance could be developed. Glucose and insulin levels in women were related to age and external factors, particularly exercise and anthropometric measurements. In men, external factors were less clearly related to glucose and insulin levels, but this lack of association in men may be due to the smaller number of men in the sample, their younger age, and the narrower range of values found in the men.