Geographic location was a strong predictor of falls among women with and without HIV in the Women's Interagency HIV Study. We examined regional variation in falls in a more geographically diverse cohort of older people with HIV (PWH) and explored whether physical activity, sex, or body–mass index modified these associations. PWH enrolled in the A5322 HAILO (HIV Infection, Aging, and Immune Function Long-Term Observational Study). Participants who reported falls in the 6 months before each semiannual visit and had ≥1 consecutive pair of fall assessments were included. We examined associations of geographic region [Northeast, Midwest, South, and West] with recurrent falls (≥2) over each 12-month period using repeated measures multinomial logistic regression models and assessed effect modification by adding an interaction term between geographic region and each potential effect modifier. A total of 788 men and 192 women with median age of 51 years at study entry contributed up to 240 weeks of data. U.S. regions included Northeast (22%), Midwest (29%), South (20%), and West (28%). In multivariable analyses, compared with the Western region, greater risk was seen among Midwestern (odds ratio [OR] = 2.35 [95% confidence interval (CI) = 1.29–4.28]) and Southern regions (OR = 2.09 [95% CI = 1.09–4.01]). Among those with higher physical activity, the Midwestern region had higher odds of recurrent falls than the Western region. Among obese individuals, the Southern region had higher odds of recurrent falls than the Western region. Sex did not modify the association between region and recurrent falls. Among older PWH, fall risk varied by geographic region. Associations between geographic region and recurrent falls appeared to be modified by physical activity and obesity. This may help identify subgroups of older PWH for targeted fall screening/interventions.

Introduction
Antiretroviral therapy (ART) has been effective in extending survival and narrowing the gap in life expectancy between people with HIV (PWH) and HIV-uninfected individuals.1–3 With increasing age, however, many PWH experience an increased burden of some comorbidities and impairments, including difficulties with balance, slow gait, weakness, and cognitive impairment.4–6 The increased burden of comorbidities combined with physical and cognitive impairments has contributed to a higher prevalence of falls in PWH compared with demographically similar HIV-uninfected adults.7–9

Falls are a major cause of mortality and morbidity globally, with fall-related injuries increasing with age. Every year in the United States, one in four older adults (65+ years) has a fall, resulting in over 800,000 fall-related hospitalizations and more than 27,000 fall-related deaths.10 Risk factors for falls among older adults include age, medications and comorbidities, gait and balance disorders, functional impairment, and cognitive impairment.11,12 Among PWH, HIV-related risk factors, including ART, toxicity of prior ART, and HIV-1 replication, may further contribute to falls.13,14

Two separate studies among PWH have also suggested that geographic location might be a risk factor for falls.13,15 In the Women's Interagency HIV Study (WIHS), a study of risk factors for falls in women with and without HIV, significant differences in fall occurrences were seen between study sites, with a higher prevalence of falls in the San Francisco site compared with the New York City site.15 Investigators hypothesized that the differences observed might have been due to reasons such as mode of transportation, weather conditions, seasonality, and physical activity.15 Similarly, in the Multicenter AIDS Cohort Study (MACS), the odds of a fall were twice as high in the Chicago and Pittsburgh sites, compared with the Baltimore site, and lowest in the Los Angeles site.13

These geographic differences may highlight important structural or environmental factors to consider in fall prevention strategies. However, the MACS and WIHS were limited to one city within each region and did not include participants from Southern states. To further investigate geographic differences in fall risk among both men and women with HIV from a variety of geographic locations, we investigated the geographic distribution of falls for participants across the 32 sites of the AIDS Clinical Trials Group (ACTG) study, A5322, also known as HAILO (HIV Infection, Aging, and Immune Function Long-term Observational Study).

We sought to identify factors that could influence the association between geographic region and falls with the goal of identifying subgroups of older PWH at highest risk for falls.

Methods
Participants
HAILO is a longitudinal study of long-term health outcomes among older (≥40 years at enrollment) men and women with HIV. All participants received their initial ART regimen through an ACTG clinical trial and were previously followed long term after their trial participation in another ACTG observational study, A5001.16 HAILO participants were enrolled at one of 32 clinical sites between 2013 and 2014.

Data collected at each semiannual visit include medical chart abstraction, targeted physical examination, fasting laboratory tests, and fall assessment, while frailty and neurocognitive function are assessed annually. All HAILO participants with at least one pair of consecutive fall assessments from study entry though study week 240 (year 5) were included in this analysis. The study was reviewed and approved by each site's local institutional review board an all participants provided written, informed consent.