STABILITY OVER TIME OF MODERN DIAGNOSTIC-CRITERIA FOR TYPE-II DIABETES

OBJECTIVE - To examine the long-term stability of the World Health Org anization criteria for diabetes in the 8-yr follow-up of the San Anton io Heart Study. A problem with older criteria for diagnosing diabetes was that many individuals classified as having borderline or chemical diabetes reverted to nondiabetic status on follow-up. Few studies have addressed this issue among diabetic patients who meet the more string ent, modem National Diabetes Data Group or World Health Organization c riteria. RESEARCH DESIGN AND METHODS - We studied 98 Mexican-American and 44 non-Hispanic white type II diabetic patients 8 yr after they we re diagnosed according to World Health Organization criteria in a popu lation-based epidemiological survey. Patients were classified as follo ws: whether they were on pharmacological treatment for diabetes at bas eline, or, if not, whether they had a prior diagnosis of diabetes at t he time of their baseline survey visit or were newly diagnosed. RESULT S - Of the 142 patients who met the study criteria for type II diabete s at baseline and whose status at follow-up was known, 20 (14.1%; 95% confidence interval 9.0-20.5%) no longer met criteria at follow-up 8 y r later. All but 2 of those who reverted to nondiabetic status were fr om the newly diagnosed group. Two of 20 had lost more than 5.0 kg over the 8 yr and an additional 5 had lost 2.5-5.0 kg. The 20 patients who reverted to nondiabetic status lacked many of the associated anthropo metric, physiological, and metabolic findings of diabetes such as obes ity, unfavorable body fat distribution, and dyslipidemia. Moreover, on ly one had proteinuria, and none had any grade of diabetic retinopathy at baseline. CONCLUSIONS - Epidemiological surveys that rely on a sin gle oral glucose tolerance test to determine the prevalence of type II diabetes may overestimate prevalence by as much as 16% (95% confidenc e interval 10.5-23.6%) (142/122 = 1.16). The actual overestimate could well be less because some patients who truly have diabetes at baselin e may be in remission at follow-up as a result of significant weight l oss. Despite these difficulties, the modem National Diabetes Data Grou p and World Health Organization criteria are more stable than earlier criteria.