Abstract 18: Associations of Accelerometer and Questionnaire Measured Physical Activity With Incident Cardiovascular Events in Older Multiethnic Women

Purpose: Self-report misclassifies physical activity (PA) in older adults when compared to accelerometer measures of PA. Few studies have evaluated whether associations with CVD differ between accelerometer-measured and self-reported PA. Methods: We followed 5,719 women (mean age 79 years; 49.8% white, 33.3% black, 16.9% Hispanic) for incident adjudicated CVD (first myocardial infarction, stroke, or heart failure) in the Objective PA and Cardiovascular Health Study. Vector magnitude counts/15 sec epoch from a hip worn (2012-2014) ActiGraph GT3X+ accelerometer were used to define light PA (19-518 counts/15 sec) and moderate-to-vigorous (MVPA; ≥519). Self-reported light PA and MVPA were assessed with the CHAMPS questionnaire. Cox regression was used to estimate hazard ratios (HR) and 95% confidence intervals (CI) controlling for age, race-ethnicity, education, smoking, number of comorbidities, self-rated health, SF36 physical function score, and accelerometer wear time. Harrell’s c-statistic estimated discrimination of CVD event risk for models with accelerometer versus questionnaire PA exposure measures. Results: Mean min/d in PA determined with the accelerometer and questionnaire were 288.4 and 54.3 for light PA, and 51.3 and 32.1 for MVPA. Age-adjusted Spearman correlations were r = 0.13 for light PA and r = 0.29 for MVPA. There were 862 incident CVD events through February 2023 (follow-up mean 7.5 years). HRs (95% CIs) per 30 min/d increment for accelerometer vs questionnaire were 0.95 (0.92, 0.98) vs 0.98 (0.92, 1.02) for light PA, and were 0.82 (0.76, 0.89) vs 0.99 (0.98, 1.00) for MVPA. C-statistics showed similar discrimination of CVD event risk for accelerometer (c = 0.70) vs questionnaire (c = 0.69) models when either light PA or MVPA was the exposure. Associations across tertiles (T1 = ref) for accelerometer PA were 0.92 (0.79, 0.99), 0.78 (0.65, 0.94) for light and 0.76 (0.64, 0.89), 0.58 (0.48, 0.71) for MVPA; and for questionnaire PA were 0.89 (0.75, 1.05), 0.91 (0.77, 1.08) for light and 0.75 (0.64, 1.08), 0.77 (0.65, 0.92) for MVPA. Associations for accelerometer- and questionnaire-measured PA were consistent when stratified by race-ethnicity (white, black, Hispanic) and age (<80 vs ≥80 years). Conclusion: In this study, accelerometers captured greater amounts of daily light PA and MVPA than did questionnaire assessment in older ambulatory women. Inverse associations for light PA and MVPA with incident CVD were stronger and achieved statistical significance when measured by accelerometer as compared to questionnaire assessment where associations were generally weaker and nonsignificant. The modest correlations between accelerometer and questionnaire PA suggest that accelerometers are capturing different aspects of cardiovascular health-promoting movement in older women that are not captured in a widely used questionnaire designed to assess PA in older adults.