Understanding Frailty among People Living with HIV in the United States

Abstract

In the recent era of HIV treatment, updated guidelines and antiretroviral therapy (ART) improvements have transformed HIV care into management of a chronic disease and people with HIV (PWH) are living longer and healthier lives as a result. Consequently, there has also been a rise in rates of aging-related conditions among this population, including frailty, a measure of physiologic reserve and vulnerability to health stressors. Frailty among PWH has been observed more often and at younger ages compared to non-HIV populations, leading to a need for accurate and efficient ascertainment integrated within HIV care as well as a better understanding of the risk factors and outcomes related to frailty. The field of research on frailty among PWH is rapidly growing, both expanding our knowledge and introducing new questions. At the forefront is understanding methods to measure frailty in a fast, low-burden, and accurate way to collect frailty information on many PWH. One of the most commonly used measures of frailty, Fried’s frailty phenotype, captures 5 functional status measures and classifies people as not frail, prefrail, or frail based on their results. Fried’s phenotype is often modified to conform to data availability, however, modifications warrant careful evaluation of their impact on measurement of frailty. Additionally, evaluating relationships between frailty and commonly used substances, such as tobacco and alcohol, could elucidate complex interplays between patterns in substance use behaviors and physical health. To answer these questions, we leveraged the Impact of Physical Activity Routines and Dietary Intake on the Longitudinal Symptom Experience of people living with HIV (PROSPER-HIV) study nested within the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS) cohort to evaluate the validity of a self-reported modified frailty phenotype used widely within CNICS. We then utilized the rich, comprehensive follow-up data within CNICS to assess relationships between smoking and alcohol with frailty. We found that the modified frailty phenotype developed and used within CNICS is a highly feasible and accurate measure of frailty, well-suited for ascertainment of frailty among PWH in clinical care and for application in epidemiologic studies. The modified phenotype classified PWH similarly to the gold standard comparator (i.e., Fried’s phenotype), with Cohen’s kappa values ranging from 0.64-0.75 depending on weighting scheme and receiver operator characteristics area under the curve values of 0.93 (95% CI: 0.91-0.96) for classifying PWH with frailty and 0.86 (0.83-0.89) for classifying prefrailty. Additionally, both definitions estimated comparable associations between frailty and having experienced a fall in the past year, highlighting the clinical relevance and utility of the CNICS modified frailty phenotype. Furthermore, among 8,608 PWH, 60% reported former (29%) or current (31%) smoking. We found strong evidence of a relationship between smoking tobacco and incident frailty, with a 79% greater risk of frailty associated with current smoking (95% CI: 1.54-2.08) and 12% greater risk associated with every 5 pack-years of smoking (95% CI: 1.09-1.16). We also found an association between current and cumulative (i.e., pack-years), but not former, smoking on the risk of deterioration of frailty, suggesting an important potential for smoking cessation to mitigate frailty risk in this population. Finally, in our investigation of the relationship between frailty status and alcohol consumption, we found that frailty is associated with a greater likelihood to quit and reduce drinking, including binge drinking. Among 8,174 PWH, most (69%) reported currently drinking alcohol, and 13% reported binge drinking at least monthly. We observed that frail PWH were 56% (95% CI: 1.14-2.14) and 33% (95% CI: 1.11-1.59) more likely to quit and reduce frequency of drinking, respectively, compared to not frail PWH. Additionally, frail PWH were 56% (95% CI: 1.19-2.04) more likely to reduce their frequency of binge drinking compared to not frail PWH. These findings are consistent with the sick quitting hypothesis, in which people may reduce alcohol consumption in response to declines in health status. In comparison to 20-35% lower risks of incident frailty associated with drinking (including former risky drinking), these findings highlight methodological concerns with estimating relationships between alcohol and frailty with observational data. Specifically, sick quitting may be confounding estimates for the risk of frailty associated with alcohol and future research should consider individual-level characteristics, such as health status, when evaluating this relationship. Overall, these findings expand the field of aging research among PWH by describing and evaluating a widespread frailty phenotype as well as applying it to answer important questions regarding substance use and frailty in this population. We describe issues of bias that are present in the existing literature and present areas for future work addressing remaining unanswered questions on this topic.