Abstract 19269: Relations and Predictive Value of Post-load and Fasting Glucose for Incident Cardiovascular Disease and Mortality in Older Adults: the Cardiovascular Health Study

Introduction: Elders have a high prevalence of post-load hyperglycemia, which may go undetected with standard screening. Post-load glucose has shown more robust associations with cardiovascular disease (CVD) and death than fasting glucose, but data in advanced old age are sparse. Whether post-load glucose improves risk prediction for CVD and death after accounting for fasting glucose has not been examined. Methods: Fasting and 2-hour post-load glucose were measured at baseline (1989) and follow-up (1996) visits in a prospective study of community-dwelling adults initially ≥65 years old (Cardiovascular Health Study). To evaluate if previously reported associations of fasting and post-load glucose with incident CVD from the baseline visit persist later in life, and apply to mortality, we focused on the 1996 visit (n=2394). To determine the incremental value of post-load glucose for risk prediction, we examined whether it could significantly reclassify baseline (1989) participants (≤75 years) into cholesterol treatment categories based on recent guidelines (n=2542). Results: Among participants in the 1996 visit (mean age 77), there were 543 incident CVD events and 1698 deaths during median follow-up of 11.2 years. In fully adjusted models, both fasting and 2-hour glucose were associated with CVD (HR per SD, 1.13 [1.03-1.25] and 1.17 [1.07-1.28], respectively) and mortality (HR per SD, 1.12 [1.07-1.18] and 1.14 [1.08-1.20]). After mutual adjustment, however, the associations for fasting glucose with either outcome were abolished, but those for post-load glucose remained unchanged. Among subjects ≤75 years old in 1989, there were 416 CVD events and 740 deaths at 10-year follow-up. Post-load glucose did not enhance reclassification using the 7.5% 10-year risk threshold, nor did it improve the C-statistic. Conclusion: In adults surviving to advanced old age, post-load glucose was associated with CVD and mortality independently of fasting glucose, but not vice versa, although there was no associated improvement in risk prediction. These findings affirm the robust association of post-load glucose with CVD and death late in life, but do not support the value of routine oral glucose tolerance testing for prediction of these outcomes in older adults.