THE BIDIRECTIONAL ASSOCIATION BETWEEN FRAILTY AND CARDIAC DYSFUNCTION

Abstract

Abstract 1

Background: Frailty is associated with greater risk of developing heart failure (HF) in late-life, and prevalent HF is associated with a high prevalence of frailty. However, little is known regarding the temporal and bidirectional relationship between subclinical alterations in cardiac structure and function and the development of frailty.

Methods: This analysis included 2,574 participants in the community-based Atherosclerosis Risk in Communities (ARIC) study who attended study Visit 5 (2011-2013), 6 (2016-2017), and/or 7 (2018-2019), and were free of HF at visit 5. Frailty was assessed using the Fried frailty phenotype at each study visit, and protocol echocardiography was performed at Visits 5 and 7. The associations of frailty at visit 5, and longitudinal changes in frailty status from visits 5 to 7, with longitudinal changes in cardiac structure and function from visits 5 to 7 were assessed using multivariable linear regression. Among 1,648 HF-free robust participants at Visit 5, we further assessed the association of measures of cardiac structure and function at visit 5 with progression in frailty status using multivariable ordinal logistic regression. Models were adjusted for demographics and comorbidities.

Results: Among 2,574 HF-free participants with frailty and echocardiographic assessments at Visits 5 and 7 (mean age 74 ± 4 years at Visit 5 and 81 ± 4 years at Visit 7; 57% women; 22% self-reported Black race), 3% (n=83) were frail. Frailty at Visit 5 was significantly associated with greater left atrial volume index (LAVi) and E/e’ ratio at both Visits 5 and 7, but not with greater changes in cardiac structure and function between visits. Participants who transitioned from robust at visit 5 to frail at visit 7 demonstrated greater increases in left ventricular (LV) mass index, LAVi, and E/e’ over the same period. Among 1,648 robust participants at Visit 5, 49 developed frailty at Visit 6. Greater LVMi and MWT, lower TDI e’, and higher E/e’ ratio were associated with a higher odds of progression in frailty status.

Conclusion: Among robust, HF-free, older adults in the community, a bidirectional relationship exists between the development of frailty and subclinical LV remodeling and diastolic dysfunction. As both frailty status and cardiac structure and function are modifiable, future studies should evaluate whether interventions to modify one may exert beneficial effects on the other.

Abstract 2 Background The incidence and prevalence of heart failure (HF) and frailty is high among older individuals, and frequently co-exist. While HF and frailty have several common pathobiologic features, limited date exist regarding potential shared mechanisms between the two conditions.

Methods Among participants in the Atherosclerosis Risk in Communities (ARIC) study, an ongoing community-based cohort study, 4,877 plasma proteins were measured using an aptamer-affinity assay (SomaScan v4) at study Visit 3 (V3; 1993-1994; n=10,368, age 60 ±6 years; 822 incident HF events) and at study Visit 5 (V5; 2011-2013; n= 3,908, age 75 ±5 years; 336 incident HF events), and frailty was assessed using Fried criteria at V5 and Visit 6 (V6; 2016-2018; n= 2,358, age 79 ±5 years, 152 incident frailty events). Multivariable Cox proportional hazard regression models were used to identify proteins consistently associated with incident HF post-V3 and post-V5. Among these HF-associated proteins, multivariable logistic regression was used to identify proteins associated with both prevalent frailty at V5 (n=223 cases) and incident frailty between V5 and V6 (n=152 incident cases). All models adjusted for age, sex, race, hypertension, diabetes, cardiovascular disease, BMI, and atrial fibrillation. HF- and frailty-associated proteins were evaluated for associations with cardiac function at V5, incident HF with preserved (HFpEF) or reduced ejection fraction (HFrEF) post-V5, individual frailty components at V6, enriched biologic pathways using the g:profiler toolkit, and for potential causal associations using two-sample Mendelian randomization.

Results: A total of 289 proteins were associated with incident HF post-V3 at p 1.0 x 10-5 (0.05/4,877) and 84 were significantly associated with incident HF post-V5 at p 1.7x10-4 (0.05/289). Among 4,131 HF-free participants at V5, 48 of these 84 HF-associated proteins associated with prevalent frailty at p 5.9 x 10-4 (0.05/84). Among 3,908 frailty and HF-free participants,18 of these proteins were significantly associated with incident frailty at p 1.0 x 10-3 (0.05/48). At V5, these candidate proteins associated most consistently with larger left ventricular and atrial size and higher E wave velocity and tended to be more strongly associated with incident HFpEF than HFrEF post-Visit 5frailty than HF. Mendelian randomization identified potential causal associations between WFDC1, and frailty and HF. Pathway enrichment analysis identified pathways linked to collagen metabolism and inflammation as enriched.

Conclusions We identified 18 proteins that independently associate with incident HF and incident frailty in late life, supporting the presence of shared biologic pathways underlying frailty and HF in late life. These proteins highlight collagen metabolism and immune pathways as potential shared biologic pathways for frailty and HF and deepen our understanding of the interplay between frailty and HF in late-life.