A total of 5082 men and women in the Framingham Heart Study population who were free of any glucose abnormality and aged 33 to 67 years were followed prospectively over 14 years for the occurrence of glucose intolerance. The diagnosis of glucose intolerance was defined as developing documented hyperglycemia or being placed on justified treatment by a physician. The 14-year incidence was 6.7% in men and 5.5% in women. Multivariate analysis was used and future glucose intolerance in men and women was highly associated with casual blood glucose, Metropolitan Relative Weight and very low density lipoproteins at the baseline exam. Other factors showed only sex-specific or univariate associations. Obesity and lipoprotein abnormalities were shown to be independent markers in the prediction of future glucose intolerance.