Prediabetes tests in predicting risk of type 2 diabetes mellitus: a meta-analysis

There is a debate concerned with the best test for prediabetes screening. The aim of this meta-analysis is to provide an estimate of the relative risk (RR) of progression to diabetes in population with impaired fasting glucose (IFG, by fasting glucose), impaired glucose tolerance (IGT, by 2 hour glucose postload) or abnormal A1c on the basis of the current literature. A search of PubMed and Ovid databases for cohort studies published between November 12, 2003-December 31, 2012 yielded 23 independent studies. We estimated overall RR for each test reported in at least five studies. Study-specific RRs were extracted and summarized using a random-effects model. Compared to normoglycemic population the overall meta-analized RR and 95% confidence interval for diabetes was: 7.47 (5.11-10.91) with 100 mg/dl cut-off point for IFG, 7.14 (4.30-11.87) with 110 mg/dl cut-off point for IFG, 3.73 (1.92-7.25) with IGT and 5.27 (2.61-10.64) with abnormal A1c. For diabetes development, we found that IFG with both 100 mg/dl (IFG100) and 110 mg/dl (IFG110) cut-off points had similar RR; nevertheless, IFG100 can identify more people in risk. The A1c test had an intermediate RR. Although the IGT had the smallest RR, it can find additional people at risk. Among the most relevant diabetes risk factors, the largest RR is for body mass index. Future researchers can explore the combination for screening with overweight or obesity and the diagnostic tests for prediabetes and their cost-effectiveness.