Acta Diabetologica

, Volume 50, Issue 2, pp 175–181

Transportability of the updated diabetes prediction model from Atherosclerosis Risk in Communities Study to a Middle Eastern adult population: community-based cohort study

Authors

  • Mohammadreza Bozorgmanesh
    • Prevention of Metabolic Disorders Research Center, Research Institute for Endocrine Sciences (RIES)Shahid Beheshti University of Medical Sciences
    • Prevention of Metabolic Disorders Research Center, Research Institute for Endocrine Sciences (RIES)Shahid Beheshti University of Medical Sciences
  • Fereidoun Azizi
    • Endocrine Research Center, Research Institute for Endocrine Sciences (RIES)Shahid Beheshti University of Medical Sciences
Original Article

DOI: 10.1007/s00592-010-0241-1

Cite this article as:
Bozorgmanesh, M., Hadaegh, F. & Azizi, F. Acta Diabetol (2013) 50: 175. doi:10.1007/s00592-010-0241-1

Abstract

We validated the transportability of the updated diabetes prediction model from Atherosclerosis Risk in Communities (ARIC) Study, to a Middle Eastern population. We investigated 3,721 participants of the Tehran Lipid and Glucose Study (TLGS) aged ≥20 years, free of diabetes at baseline. They underwent a standard 75gr 2-h post-challenge plasma glucose test that was repeated every 3 years using the same protocol. All the models were tested with respect to discrimination and calibration. We confirm the findings of Kahn et al. (Ann Intern Med 150(11):741–751, 2009) in a middle-aged, Middle Eastern population. We obtained the same predictive discrimination for the ARIC model (C statistic: men 0.790 and women 0.829) as for the TLGS’ own model (men 0.824 and women 0.847) and validated a good calibration for the updated ARIC diabetes prediction model in the TLGS sample. Among men, optimal cut-point was set to the score of 31 where the maximum value of sensitivity (71.6%) plus specificity (75.3%) was achieved. Among women, the optimal point was set to the score of 38 with sensitivity of 67.1% and specificity of 85.0%. The updated ARIC model predicted the individual diabetes risk with a high level of sensitivity and specificity in the TLGS population, which was comparable with that of original sample. More parsimonious model incorporating age, family history of diabetes, waist circumference, pulse rate, and fasting plasma glucose, which were significantly associated with the risk of incident diabetes in the TLGS population, could be equally effective in predicting diabetes.

Keywords

DiabetesPredictionValidationRisk factorScore system

Copyright information

© Springer-Verlag 2010