External validation of the Finnish diabetes risk score in Venezuela using a national sample: The EVESCAM.

Abstract

To evaluate the performance of the Latin American Finnish Diabetes Risk Score (LA-FINDRISC) compared with the original O-FINDRISC in general population. To establish the best cut-off to detect unknown type 2 diabetes (uT2D) and prediabetes. The EVESCAM was a national population-based, cross-sectional, randomized cluster sampling study, which assessed 3454 adults from July 2014 to January 2017. Those with self-report of diabetes were excluded; a total of 3061 subjects were analyzed. Waist circumference adapted for Latin America was the difference between the LA-FINDRISC and the O-FINDRISC. The area under the curve (AUC), sensitivity, and specificity were calculated. The prevalence of uT2D and prediabetes were 3.3% and 38.5%. The AUC with the LA-FINDRISC vs. the O-FINDRISC were: for uT2D, 0.722 vs. 0.729 in men (p=0.854) and 0.724 vs. 0.732 in women (p=0.896); for prediabetes (impaired fasting glucose [IFG] + impaired glucose tolerance [IGT], 0.590 vs. 0.587 in men (p=0.887) and 0.621 vs. 0.627 in women (p=0.777); for IFG, 0.582 vs. 0.580 in men (p=0.924) and 0.607 vs. 0.617 in women (p=0.690); for IGT, 0.691 vs. 0.692 in men (p=0.971) and 0.672 vs. 0.671 in women (p=0.974). Using the LA-FINDRISC, the best cut-offs to detect uT2D were 9 in men and 10 in women and to detect IGT was 9 in both genders. LA-FINDRISC has similar performance than O-FINDRISC in Venezuelan adults and showed a good performance to detect uT2D and IGT, but not IFG. The best cut-offs to detect glucose alterations were established.