Improvement in prediction of incident type 2 diabetes using the 2003 criteria of impaired fasting glucose: results from a population-based 6.5-year follow-up study in Iran

ABSTRACT

To investigate whether using the 2003 criteria for impaired fasting glucose [IFG] improves the prediction of Type 2 diabetes mellitus [T2DM] in comparison to the 1997 criteria. A total of 5794 non-diabetic subjects, aged 42 +/- 13 years, were followed for 6.5 years. T2DM was defined based on the WHO 1999 criteria. Cox regression analysis was used to calculate the relative- risk [RR] of developing T2DM. Model discrimination was assessed by calculating the area under the receiver operating characteristic curve [AUC]. During the follow-up, there were 351 new cases of T2DM. In univariate analysis, the RR of the 1997 and 2003 criteria were 9.6 [7.5-12.3] and 8.4 [6.4-10.0] respectively. After adjustment for a full range of diabetes risk factors, the RR of the 2003 criteria was higher compared to that of the 1997 definition [RR [95%CI] 3.3 [2.6-4.2] vs. 2.4 [1.8-3.2] respectively]. Inclusion of the 2003 criteria in the multivariate model significantly improved discrimination in comparison to the 1997 definition [AUC [95%CI] 0.78 [0.74-0.83] vs. 0.74 [0.70-0.79], P<0.01]. When fasting glucose is used in community screening for pre-diabetes state, the 2003 IDF criteria yield significantly better results than the original criteria for prediction of future T2DM