What is the best way to predict who will develop T2D?

  • Lee CMY & al.
  • BMJ Open Diabetes Res Care
  • 01/01/2019

  • Miriam Tucker
  • Clinical Essentials
L'accesso ai contenuti di questo sito è riservato agli operatori del settore sanitario italiano L'accesso ai contenuti di questo sito è riservato agli operatori del settore sanitario italiano

Takeaway

  • All current pre-diabetes mellitus (pre-DM) definitions predict the 5-year risk of developing type 2 diabetes (T2D).
  • Choice may rest on available prevention resources.
  • The use of multiple risk factors may assist prediction.

Why this matters

  • Early identification of people at greatest risk for T2D facilitates prevention efforts.  

Study design

  • Comparison of predictive abilities of 5 different prediabetes definitions among 76,513 participants from the Obesity, Diabetes and Cardiovascular Disease Collaboration, encompassing populations from Asia, Australia, Europe, and North America.
  • Funding: National Health and Medical Research Council of Australia.

Key results

  • Across definitions, the progression from prediabetes to diabetes ranged from 23.7 to 79.4 per 1000 person-years.
  • The discriminatory ability for 5-year T2D risk was similar across all 5 definitions.
  • Based on adjusted spline models, optimal cut-points for incident diabetes were:
    • Fasting plasma glucose (FPG): 5.1 mmol/L; 
    • HbA1c 5.0% (31 mmol/mol).
    • With progression to T2D, they were:
      • 14.4/1000 person-years for FPG, and
      • 13.7/1000 person-years for HbA1c.
  • Via receiver operating characteristic curve analysis, optimal cut-points for 5-year T2D risk were:
    • FPG: 5.6 mmol/L;
      • Sensitivity, 64.1%; specificity, 77.0%. 
    • 2-hour postload plasma glucose: 7.0 mmol/L;
      • Sensitivity, 62.0%; specificity, 80.9%. 
    • HbA1c: 5.6% (38 mmol/mol);
      • Sensitivity, 72.8%; specificity, 79.6%.

Limitations

  • High between-study heterogeneity.
  • Only 1 study included oral glucose tolerance testing.
  • Lack of follow-up because of death unaccounted for.