Validation of DIABSCORE in screening for Type 2 Diabetes and prediabetes in Tunisian population

AIMS: To perform a validation of DIABSCORE in a sample of Tunisian adults and find out the optimal cut-off point for screening of Type 2 diabetes (T2D) and prediabetes. METHODS: 225 adults 18–75 years and a subgroup of 138 adults (18–54 years), with undiagnosed T2D from the region of Cap-Bon, Tunisia were included in the present study. The DIABSCORE was calculated based on: age, waist/height ratio, family history of T2D and gestational diabetes. Receiver operating characteristics (ROC) curves and areas under curve (AUC) were obtained. The T2D and prediabetes prevalences odds ratios (OR) between patients exposed and not exposed to DIABSCORE≥90 and DIABSCORE≥80, respectively were calculated in both age ranges. RESULTS: For screening of T2D the best value was DIABSCORE = 90 with a highest sensitivity (Se), negative predictive value (NPV) and lower negative likelihood ratio in participants aged 18–75 yr (Se = 97%; NPV = 97%) when compared to participants aged 18–54 yr (Se = 95%; NPV = 97%); for prediabetes, the best Se and NPV were for DIABSCORE = 80 in both age groups, but it showed a disbalanced sensitivity-specificity. The ROC curves for T2D showed a similar AUC in both age ranges (AUC = 0.62 and AUC = 0.61 respectively). The ROC curves for prediabetes showed a highest AUC in those aged 18–54 years than the older ones (AUC = 0.62 and AUC = 0.57, respectively). The prevalences OR of T2D for DIABSCORE≥90 was higher than for DIABSCORE≥80 in both age ranges. Nevertheless, the prevalences OR of prediabetes for DIABSCORE≥90 was half of the detected for DIABSCORE≥80 in both age ranges. CONCLUSION: The DIABSCORE is a simple clinical tool and accurate method in screening for T2D and prediabetes in the adult Tunisian population.