To determine the incidence and the rate of reversion of type 2 diabetes to a nondiabetic status in the 7- to 8-year follow-up of the San Antonio Heart Study, and to determine the influence of the recent 1997 American Diabetes Association (ADA) criteria for diabetes on these rates. Individuals who revert have been problematic for those developing criteria for the diagnosis of type 2 diabetes. Few studies have addressed this issue using 1979 National Diabetes Data Group/1980 World Health Organization (WHO) criteria.
We studied 3,682 Mexican-American and non-Hispanic white men and nonpregnant women who completed both the baseline and follow-up examination of the San Antonio Heart Study. Incidence and reversion rates were calculated using both the 1980 WHO and the 1997 ADA criteria. Risk factors for reversion were identified, and the best fitting model using multiple logistic regression was determined using both the 1980 WHO and the 1997 ADA criteria.
Using the 1997 ADA criteria, the age-adjusted incidences of type 2 diabetes for Mexican-American men and women were 10.8 and 12.2%, respectively. For non-Hispanic white men and women, the age-adjusted incidence rates were 5.5 and 5.1%, respectively. Similar age-adjusted incidences were recorded using the 1980 WHO criteria. The reversion rate for individuals with type 2 diabetes was 11.5% using the 1980 WHO criteria and 12.5% using the 1997 ADA criteria. These rates were not significantly different. Numerous risk factors for reversion were identified. The best fitting model, after controlling for age, sex, and ethnicity, included baseline 2-h glucose level, baseline HDL cholesterol, and previous diagnosis of diabetes. The models were the same for both the 1980 WHO and the 1997 ADA criteria.
There was no significant difference in the incidence or the reversion rates for diabetic subjects using either 1980 WHO or 1997 ADA criteria. In addition, the risk factors for reversion were very similar using either set of criteria. The revision of the ADA criteria did not have a significant influence on reversion in this study.