Older Adults with Lowest Risk of Developing Type 2 Diabetes: Results from a Large and Diverse US Cohort (2003-2016) (OR22-06-19).

Academic Article


  • Objectives: High rates of early obesity substantially increase lifetime risk of developing diabetes. The objective of this research is to examine the impact of this shift on type-2 diabetes risk and aging in a large and diverse U.S. cohort. Methods: A nationally sampled, longitudinal survey, the REasons for Geographic and Racial Differences in Stroke (REGARDS) study (2003-2007), a collection of sociobiological data, were observed through 2013-2016. We included black and white men and women aged ≥45 without diabetes at baseline for whom follow-up data was available. To assess incident diabetes as a function of age, we used the R package Segmented and examined whether this risk was constant over the whole range of age or whether the risk was non-linear with a threshold inflection value. After inspection of the data, age as a continuous variable appeared to have a piecewise relationship with incidence diabetes. A logistic model to see the effect of age on incident diabetes adjusting for sex and race was run. We estimated changes in odds for developing diabetes in ten years by age using restricted cubic spline knots. We estimated the breakpoint statistically using the psi argument in 'segmented'. Odds ratios for developing diabetes on both sides of the breakpoint were calculated. Results: Baseline age appeared to show a non-linear association with future diabetes incidence with a breakpoint at age 55 (P = 0.005). After 55 years, incident diabetes declined (slope at age 45-54.9: 0.06; CI: 0.02, 0.10; slope age ≥55: -0.02; CI -0.04, -0.01). Odds ratio for developing diabetes in ten years for those 45-55 years (OR: 1.18, 95% CI 1.02-1.37, P = 0.024) confirm this relationship - those 45-55 years have 18% higher odds of developing diabetes than those over 55. Conclusions: This model estimated an age breakpoint of roughly 55 years of age when younger individuals were observed to have a progressive increase in incident diabetes while those 55 and older had a declining risk. Between 1992 and 2010, mean BMI in the US increased, resulting primarily from accelerated weight gain among persons initially normal and overweight - our findings argue for a careful data driven approach in considering the effects of age on incident diabetes to derive a more accurate prediction for diabetes. Funding Sources: Lua Wilkinson is supported in part by NIH grant T32 DK062710. Supporting Tables Images and/or Graphs:
  • Published In

    Digital Object Identifier (doi)

    Author List

  • Wilkinson L; Mehta T; Garvey WT
  • Volume

  • 3
  • Issue

  • Suppl 1