Predicting diabetes risk among HIV-positive and HIV-negative women

AIDS. 2018 Nov 28;32(18):2767-2775. doi: 10.1097/QAD.0000000000002017.

Abstract

Objective: To assess the performance of an adapted American Diabetes Association (ADA) risk score and the concise Finnish Diabetes Risk Score (FINRISC) for predicting type 2 diabetes development in women with and at risk of HIV infection.

Design: Longitudinal analysis of the Women's Interagency HIV Study.

Methods: The women's Interagency HIV Study is an ongoing prospective cohort study of women with and at risk for HIV infection. Women without prevalent diabetes and 3-year data on fasting blood glucose, hemoglobin A1c, self-reported diabetes medication use, and self-reported diabetes were included. ADA and FINRISC scores were computed at baseline and their ability to predict diabetes development within 3 years was assessed [sensitivity, specificity and area under the receiver operating characteristics (AUROC) curve].

Results: A total of 1111 HIV-positive (median age 41, 60% African American) and 454 HIV-negative women (median age 38, 63% African-American) were included. ADA sensitivity did not differ between HIV-positive (77%) and HIV-negative women (81%), while specificity was better in HIV-negative women (42 vs. 49%, P = 0.006). Overall ADA discrimination was suboptimal in both HIV-positive [AUROC = 0.64 (95% CI: 0.58, 0.70)] and HIV-negative women [AUROC = 0.67 (95% CI: 0.57, 0.77)]. FINRISC sensitivity and specificity did not differ between HIV-positive (72 and 49%, respectively) and HIV-negative women (86 and 52%, respectively). Overall FINRISC discrimination was suboptimal in HIV-positive [AUROC = 0.68 (95% CI: 0.62, 0.75)] and HIV-negative women [AUROC = 0.78 (95% CI: 0.66, 0.90)].

Conclusion: Model performance was suboptimal in women with and at risk of HIV, while greater misclassification was generally observed among HIV-positive women. HIV-specific risk factors known to contribute to diabetes risk should be explored in these models.

Publication types

  • Multicenter Study
  • Research Support, N.I.H., Extramural

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Decision Support Techniques*
  • Diabetes Mellitus, Type 2 / epidemiology*
  • Female
  • HIV Infections / complications*
  • Humans
  • Longitudinal Studies
  • Middle Aged
  • Prospective Studies
  • ROC Curve
  • Risk Assessment
  • Risk Factors