At-home physical activity and mortality risk in patients with heart failure and obesity: insights from wrist-worn accelerometry in NHANES
A Sharma, R Ceskoutse, G Perlman, O KoseAbstract
Background
Obesity is highly prevalent among persons with heart failure (HF) and may influence habitual physical activity behavior as well as its prognostic significance. Wrist-worn accelerometry enables objective assessment of free-living physical activity. We have previously demonstrated that near-maximal volitional movement (or 'activity intensity') as measured by a wrist-worn accelerometry device was associated with mortality and CV mortality in patients with HF [1]. However, whether activity levels differ by obesity status and whether obesity modifies the association between physical activity and mortality in HF remain incompletely understood.
Objectives
To compare at-home physical activity levels between persons with and without obesity among individuals with HF, and to evaluate whether obesity modifies the prognostic association between physical activity and mortality.
Methods
We analyzed data from 300 adults with HF participating in NHANES with valid wrist accelerometry recordings. Physical activity was quantified using total Monitor Independent Movement Summary (MIMS) (reflecting total movement) and Signal Vector Magnitude (SVM) percentiles (reflecting near-maximal volitional movement or 'activity intensity'). SVM90 reflects the 90th percentile of distribution of SVM. Unadjusted group differences were assessed using non-parametric tests. Multivariable linear regression models adjusted for age, sex, and accelerometer wear time examined associations between obesity status and activity metrics. Cox proportional hazards models were used to evaluate all-cause mortality, including interaction terms between obesity status and physical activity.
Results
In unadjusted analyses, median physical activity levels were similar between persons with and without obesity across all SVM percentiles and MIMS metrics (all p > 0.24; Figure 1), and accelerometer wear time did not differ between groups. After multivariable adjustment, obesity status was associated with a reduction in SVM90 (β = −2.4 × 10−4, p = 0.022; Figure 2). In Cox models, no significant interaction between physical activity and obesity status was observed (pinteraction = 0.91) for the outcome of mortality and CV mortality.
Conclusions
Among patients with HF, overall free-living activity volume remains largely preserved among those with and without obesity; however, increasing BMI is associated with reductions in activity intensity after accounting for key demographic and measurement-related confounders. Obesity did not modify the association between activity intensity and mortality. While future studies are needed, activity intensity as measured by SVM90 may represent a novel marker to track patients with HF and obesity using wrist-worn accelerometry.For image description, please refer to the figure legend and surrounding text.For image description, please refer to the figure legend and surrounding text.