Abstract
Little is known about longitudinal change in physical functioning of older African American/Black and White HIV-infected persons. We examined up to 10 years of data on African American (N = 1,157) and White (N = 400) men with HIV infection and comparable HIV-negative men (n = 1,137 and 530, respectively), age 50–91 years from the Veterans Aging Cohort Study Survey sample. Physical functioning was assessed using the SF-12 (12-Item Short Form Health Survey) physical component summary (PCS) score. Mixed-effects models examined association of demographics, health conditions, health behaviors, and selected interactions with PCS score; HIV biomarkers were evaluated for HIV-infected persons. PCS scores were approximately one standard deviation below that of the general U.S. population of similar age. Across the four HIV/race groups, over time and through ages 65–75 years, PCS scores were maintained; differences were not clinically significant. PCS score was not associated with race or with interactions among age, race, and HIV status. CD4 and viral load counts of African American and White HIV-infected men were similar. Older age, low socioeconomic status, chronic health conditions and depression, lower body mass index, and smoking were associated with poorer PCS score in both groups. Exercising and, counterintuitively, being HIV infected were associated with better PCS score. Among these older African American and White male veterans, neither race nor HIV status was associated with PCS score, which remained relatively stable over time. Chronic disease, depression, and lack of exercise were associated with lower PCS score. To maintain independence in this population, attention should be paid to controlling chronic conditions, and emphasizing good health behaviors.
Introduction
With the advent of highly active antiretroviral therapy (ART) and treatment initiation early in the course of the disease, HIV has become a chronic rather than a life-threatening condition. 1 Life expectancy is now closer to that of those without HIV although, as in the general U.S. population, less so for African Americans than for Whites. 2,3 Increased life expectancy, however, has come at the cost of higher rates of age-associated chronic conditions at younger ages among HIV-infected persons. 4 –8 This is a matter of concern, because chronic conditions lie along the pathway leading to impaired physical functioning and loss of independence. 4,9,10 The effect of HIV on physical functioning and the possible effect of an interaction between HIV and race on physical functioning are presently unclear.
Population-representative studies of older people typically indicate that certain persons—African American, lower socioeconomic status, older—have poorer physical functioning, and that over time physical functioning declines. 11 –18 Whether the same holds for older HIV-infected men remains to be determined. A review of studies comparing physical functioning in HIV-infected and HIV-negative persons is notable for the inconsistencies found. 4 In several cross-sectional studies, physical functioning was lower at higher ages in both HIV-infected and HIV-negative persons, but no statistically significant racial differences were found. 19,20 Examination of chronic conditions found that some conditions (e.g., cardiovascular and cerebrovascular disease and hepatitis C) were associated with poorer physical functioning, regardless of HIV status; others (diabetes and depression) had a greater negative impact on physical functioning in HIV-negative persons, or (pulmonary disorder) on HIV-infected persons; no statistically significant differences were found between African American and White persons for other conditions. 20 It remains to be established if such findings on the impact on physical functioning of age, chronic conditions, and particularly race, persist over time.
Our primary focus is on comparing change in physical functioning over time between African American and White men 50 years of age and older who differ in HIV status. We focus on physical functioning rather than physical performance, because physical functioning takes a broader view, indicating ability to take care of the self and to participate in society, whereas measures of physical performance target specific activities (e.g., grip strength and gait speed), which are important in their right, but have a less broad connotation. Among persons living with HIV, we explore the association of immune status and virological control with physical functioning. African American and White men were selected for two reasons: in the older population physical functioning has been shown to differ between these two groups, 13 –15,21 and in the United States HIV is most prevalent in these groups (seven times greater among African American men than among White men, and still increasing 22 ). We selected a minimum age of 50 years, because decline in physical functioning and the presence of chronic conditions are likely to be greater in this group. 23
We undertook this study with the expectation that level of physical functioning would be poorer in older sample members and in African Americans; would decline with increase in age and over time; and would be better in HIV-infected persons, than in HIV-negative persons.
Materials and Methods
Sample
Data were from men 50 years of age and older who participated in the Veterans Aging Cohort Study (VACS), 24 a prospectively enrolled, longitudinal, observational study of HIV-infected and HIV-negative veterans, matched by age, sex, race/ethnicity, and clinical site. Rolling enrollment at eight sites started in 2002, and for this study continued through 2012, with follow-up surveys every 12–18 months. VACS is approved by institutional review boards (IRBs) at the coordinating center (Veteran Affairs Connecticut Healthcare System), and by each participating Veterans Health Administration facility and corresponding university affiliates [in California, Georgia, Maryland, New York State (two sites), Pennsylvania, Texas, and Washington, DC]. 24,25 This study was approved by Duke University IRB. All participants provided informed consent.
Because of small sample sizes, women, Hispanics, and Asians were excluded. Of 3,233 African American/Black and White male enrollees 50–91 years of age in the database, 9 with unclear HIV status were excluded, leaving an analysis sample of 3,224 participants. They included 2,294 African American men (71% of the analysis sample; HIV infected: 1,157, 50%; HIV negative: 1,137, 50%), and 930 White men (29%; HIV infected: 400, 43%; HIV negative: 530, 57%). Most enrolled in the first 2 years (31% in 2002, 39% in 2003); enrollment tailed off to 0.5% of the sample in 2012.
Data gathered
Information relevant to this study was obtained using structured paper-and-pencil questionnaires completed by participants on each evaluation occasion, and clinical and administrative data abstracted from VA electronic medical records using methodology validated by chart review. Comorbid conditions were defined by ICD-9 (International Classification of Diseases, Ninth Revision) diagnostic codes and required at least one inpatient or two outpatient encounters to be considered present. 25 Also included was information on use of ART, measures of HIV status (CD4 cell counts and HIV-1 viral load), and selected laboratory tests.
Dependent measure
The dependent measure was the physical component summary (PCS) score of the 12-Item Short Form Health Survey (SF-12), 26 a validated, reliable, abbreviated version of the Medical Outcomes Study 36-Item Short Form Health Survey (SF-36), which is highly correlated (ρ = 0.74, p < .001), with, in the VACS study, the less consistently administered VACS function scale. 20
The PCS addresses mobility, intensity of activities performed, performance of physical role, bodily pain, and general health. Thus, it rapidly provides general information on functional capability. Scoring is on a 100-point scale, with a mean of 50 and standard deviation (SD) of 10 for the general U.S. population. Higher scores indicate a better level of performance. A difference of three points in score is considered clinically relevant. 27,28
Covariates
The covariates selected included demographic characteristics: age, education, and income.
Health conditions included eight physical health conditions and one mental health condition, selected based on the plausible relationship with physical function in HIV-infected and HIV-negative persons (Table 1).
8,20,29,30
The VACS Index, a weighted summary of age, CD4 count, HIV-1 RNA, hemoglobin, fibrosis-4 index (FIB-4), estimated glomerular filtration rate (eGFR), and hepatitis C (
Baseline Demographic and Health Characteristics of the Sample by Race and HIV Status
p-Value of simultaneous comparison of all four groups is based on chi-square and Wilcoxon tests. p-Value for pairwise comparisons is based on chi-square and t-tests. p < .05 is accepted as significant.
African American HIV-infected versus African American HIV-negative men.
White HIV-infected versus White HIV-negative men.
African American HIV-infected versus White HIV-negative men.
African American HIV-negative versus White HIV negative men.
SD, standard deviation; VACS, Veterans Aging Cohort Study.
Health behaviors included body mass index [BMI (kg/m2), trichotomized as <25 (including 1.8% with BMI <18.5), 25 to <30, and ≥30]; exercising (0 times/week, 1–2 times/week, and ≥3 times/week); alcohol use identification test (AUDIT-C), range 0–12, ≥4 indicating problematic use 36 ; and smoking status (never, current, or ever smoked).
The HIV-related measures included baseline CD4 cell counts, HIV-1 viral load, other components of the VACS Index (hemoglobin, FIB-4, serum creatinine with calculation of eGFR), and any use of ART.
Statistical analysis
With the exception of SF-12 PCS scores collected serially over 10 years, only baseline data were analyzed. Thus, this study indicates physical function over time, controlling only on baseline information. Descriptive statistics (percentages, Kruskal–Wallis, chi-square, and Wilcoxon rank sum tests, and t-tests) characterized the sample at baseline and compared HIV/race groups.
Mixed-effects models were used to examine the effect of baseline age (including age squared and age cubed to account for curvilinear association), HIV status, race, and their interactions on the repeated measures of SF-12 PCS, and including also income and education (model 1). The variance–covariance structure of the repeated measures was best fit by compound symmetry. Two additional models were examined by adding the individual health conditions and VACS Index to the basic model (model 2), and further adding health behaviors that differed between the groups (BMI, exercise, and smoking status) (model 3).
All analyses were conducted in SAS 9.4 (SAS Institute, Inc., SAS, Cary, NC).
Results
The median age of the total sample was 55 years. The mean number of visits per person was 3.89 (SD = 2.09, range = 0–8); HIV-infected had a higher return rate than HIV-negative persons [mean (SD), 4.0 (2.2) versus 3.8 (2.0), p < .023]; and Whites than African Americans [mean (SD), 4.3 (2.1) versus 3.7 (2.1), p < .0001]. African American men were, on average, 2 years younger than White men, and HIV-infected were a year younger than HIV-negative men (Table 1). Overall, African American men reported fewer years of education than White men, and a larger proportion of White HIV-infected, than White HIV-negative men, had post-high school education. Regardless of race, participants who were HIV infected had lower income than participants who were HIV negative.
Dependent variable
Baseline SF-12 PCS scores (Table 2) were comparable for all four groups, with the median score for each group nearly 1 SD below that for the general U.S. population at age 55–64 years. 26
Baseline Health Behaviors and SF-12 PCS Scores of the Sample by Race and HIV Status
p-Value of simultaneous comparison of all four groups is based on chi-square and Wilcoxon tests. p-Value for pairwise comparisons is based on chi-square and t-tests. p < .05 is accepted as significant.
African American HIV-infected versus African American HIV-negative men.
White HIV-infected versus White HIV-negative men.
African American HIV-infected versus White HIV-infected men.
African American HIV-negative versus White HIV-negative men.
BMI, body mass index; PCS, physical component summary; SF-12, 12-Item Short Form Health Survey.
Covariates
Baseline physical health conditions
Statistically significant race and HIV status differences with clinical import were found for hepatitis C, coronary artery disease, hypertension, and VACS Index score (Table 1, final column). With the exception of congestive heart failure and stroke, for which prevalence did not differ significantly across the four groups, the remaining health conditions considered were more prevalent in African Americans than in Whites, and among the HIV-negative men than among their race/ethnicity-corresponding counterparts.
Health behaviors
The median BMI levels of HIV-negative men were in the overweight range, and significantly higher than their HIV-infected counterparts (Table 2). The only statistically significant difference in frequency of exercising was among HIV-negative men, among whom a larger proportion of African Americans exercised. AUDIT-C scores were comparable across the four groups, and indicated that a quarter of each group reported alcohol misuse. A larger proportion of African Americans, particularly African American HIV-infected men, were current smokers (all differences between groups were statistically significant).
HIV-related laboratory findings
Baseline CD4 cell counts, HIV-1 viral loads, and eGFR levels were comparable for African American and White men (Supplementary Table S1). Regarding other biomarkers, African American men had poorer hemoglobin, FIB-4, and eGFR levels, and higher (poorer) VACS Index scores. Nearly all were taking ART (African American: 93%, White: 94%).
The unadjusted SF-12 PCS trajectory (Fig. 1), using age as the time scale, indicated that at the youngest age (50–55 years), the four HIV/race groups fell within ∼3 points of each other. They experienced comparable patterns of slightly improving performance with age until age 66–70 (improvement for African Americans: HIV infected ∼1.5 points, HIV negative ∼0.2 points; White HIV infected ∼1.5 points), and until age 71–75 for White HIV-negative men (∼2.9 points), after which there was a decline in score as age increased. Note, however, there were few observations over age 70, particularly among White HIV-infected men.
Number of unique subjects and number of observations, by age interval, race/ethnicity, and HIV status. A subject can provide as many observations in an age group interval as his age permits. AA, African American; PCS, physical component summary.
Initial multivariable regression analyses (Table 3, model 1) that included all demographic characteristics, as well as the two- and three-way interactions among age, race, and HIV status, and controlled for the curvilinear association of age, found no statistically significant difference by race. However, increased age, low income, less education, and HIV-negative status were associated with lower SF-12 PCS scores. Of the two- and three-way interactions among age, race, and HIV status, only the interaction of HIV and race was statistically significant (p = .04), with African American HIV-infected having a higher SF-12 PCS score than predicted simply by race and HIV status.
Association of Demographic Characteristics with SF-12 PCS Score (Model 1), After Addition of Health Conditions and VACS Index (Model 2), and After Further Addition of Health Behaviors (Model 3)
For a unit increase in VACS score, physical function decreases by .03, which translates to 0.15 decrease in function for 5-unit increase in VACS score.
CAD, coronary artery disease; CHF, congestive heart failure; PD, pulmonary disorder; PVD, peripheral vascular disorder; SE, standard error.
With the addition of the health conditions and VACS Index score (Table 3, model 2), the HIV × race interaction lost significance, but otherwise there was little change in the estimates of the demographic characteristics. Each physical health condition, with the exception of hypertension, reduced PCS score (by 0.8 to 3.3 points), as did higher VACS Index scores. Depression had the most notable negative association. Further addition of health behaviors (Table 3, model 3) yielded little additional change to the demographic estimates, or health condition estimates. Those with above “normal” BMI, who exercised more, and had never smoked had better PCS scores than their reference groups.
Discussion
This study compared physical functioning over time in African American and White men 50 years of age and older, as a function of HIV status. It was made feasible by the longitudinal VACS survey sample, 24 which unlike most major studies of HIV at the time, sought information on physical functioning, in contrast to physical performance. Contrary to general population-representative studies, and expectations regarding the impact of illness, we found neither racial nor HIV-associated differences in physical functioning. We also found stability, rather than the expected decline, with age.
Although national findings indicate that African Americans have poorer physical functioning than Whites, 13 –15 recently published findings indicate that physical functioning differences between older African Americans and Whites have narrowed, particularly among residents of low-income, integrated settings. 15 –17,37 –39 Similarly, and contrary to expectation, older persons who are HIV infected may not invariably have a poorer functional level than those who are HIV negative. 4
Our basic findings indicate that PCS score, as an indicator of physical functioning, was lower in this medically served sample than in age-comparable, nationally representative samples. In fully controlled analyses, higher PCS score was associated with better socioeconomic status, physical and mental health, and health behaviors, and counterintuitively, with being HIV infected, but not with race. Furthermore, trajectories of self-reported physical functioning of older African Americans were not consistently less favorable than among Whites, and performance of older HIV-infected men was not worse than that of HIV-negative men. Indeed, it was statistically better, although the difference was not clinically relevant (a 3-point difference is considered clinically important 27,28 ).
These results confirm findings based on earlier cross-sectional analyses of VACS data using a broader age sample, and the VACS function scale, a different (but highly correlated with PCS) functional measure. 8,20 They are also in line with the review by Erlandson et al. 4 who found no consistent difference in performance as a function of HIV status. Such stability of physical functioning, even at a low level of functioning, suggests a positive impact of health services, which arguably comes from use of the accessible and affordable VA medical care used by these VACS participants. It stands in contrast to objective findings on physical performance in HIV, which variously indicate that in major longitudinal studies (where care may come from a variety of medical sources), baseline level of performance is significantly lower in middle-aged and older nonwhite men, and in HIV-infected men, and that decline may be faster in these groups. The measures examined include the Short Physical Performance Battery, grip strength, and gait speed. 40 –42 Although unquestionably important as indicators of physical health status, with the exception of gait speed, which is a good predictor of health service use, change in health status, and mortality, 43,44 they do not indicate performance in everyday life. Although subjective, the SF-12 PCS has been shown to be valid and reliable 26 and indicates the respondents' everyday capabilities. This, and similar measures, are used in major longitudinal studies because the information they provide is easier, quicker, and cheaper to gather, while remaining valid and reliable.
We considered whether these findings might be attributable to the health conditions and health behaviors considered; the restricted sociodemographic range of participants; the health services available, provided, and used; and the VACS project itself. The PCS scores indicate that the general physical health status of this sample was poor, as might be expected of a clinically based group. Median scores for each of the four groups ranged from a half to ∼1 SD below that for reasonable comparison samples. The scores were similar to those for persons older than age 65 who averaged three chronic conditions, and to those of younger, low-income, HIV-infected persons seen at HIV service centers. 28,45 –48 Thus, although not doing well, over time they appeared to be performing better than those receiving treatment elsewhere.
Sociodemographic range
With fewer years of education and lower income, African American men could be expected to have poorer physical functioning than White men, which could be offset by their slightly younger age. We did not find this, and this finding is not unique. An earlier study of geographically representative samples reported that African Americans at age 65–74 had an increased risk of developing functional problems, partially explained by socioeconomic and health factors, and a crossover effect at age 85. 21 In a more recent nationally representative sample age 70–84, period-, cohort-, and sociodemographically adjusted analyses found that male African American/White differences were small, for instrumental activities of daily living (ADL) crossing over at age 74, and at age ∼80 years for basic ADL. 38 The restricted socioeconomic range in our sample may help to explain current findings.
Specific health conditions and multimorbidity
All health conditions, with the exception of hypertension, were significantly associated with PCS score. This is not unexpected. What is unexpected is the disease distribution. Although HIV is recognized as a risk factor for cardiovascular disease, 49,50 the prevalence of coronary artery disease, and such traditional risk factors for cardiovascular disease as hypertension and diabetes, and peripheral vascular disorder, were lower in the HIV-infected groups than in their HIV-negative counterparts. Given studies that show that even at the VA, discrimination in treatment of chronic health conditions may still be present 30,51,52 ; that HIV increases the risk of acute myocardial infarction by 50% 50 ; that ischemic heart disease, hypertension, and diabetes were nearly twice as prevalent in the Medicare HIV-infected population 49 ; and that the lower prevalence is unlikely to be attributable to ART, the most plausible explanations would seem to be poorer health status in the comparison sample, and increased preventive intervention by medical care providers. Possibly also what we are seeing is not a typical outcome in HIV, but an outcome that reflects access to, and provision and use of sound medical care.
The effect of depression should not be ignored. Its effect on PCS score was twice that of any other health condition. Psychotropic and psychotherapeutic interventions, alone or in combination, may be effective, and merit consideration. Similar findings have been reported in a small study of Latino men with HIV, where adjusted PCS score was on average 10 points lower in persons with depression. 53
Health behaviors
Higher BMI within the range we observed seems to indicate a better health status, whereas, as might be expected, smoking and lack of exercise are associated with poorer PCS score.
Race, changes in health, and the health care system
In the last decade, differences in longevity and in physical functioning between African Americans and Whites have declined. Longevity among African American men has increased, whereas it was found to be decreasing slightly among under-educated, middle-aged, White men. The crossover in mortality now comes at an earlier age. Analyses that take period, cohort, and sociodemographic characteristics and health insurance coverage into account find reduced or no differences in physical functioning between the two groups. 7,13,23,38 Our current sample may reflect these changes. In addition, consistent use of improved ART may avert the adverse functional impacts of HIV.
Limitations and strengths
Our findings are limited because they apply only to African American and White men who are veterans using VA medical services. Information at the uppermost ages should be accepted with caution given sparse numbers. Only baseline health conditions were considered, and their severity was not taken into account.
The strengths of this dataset include a sample of substantial size and duration; access to extensive demographic information and medical records that reduced problems with recall; and a system that provided sound, affordable care.
This study indicates that to date, in the VA setting, the effect of HIV on level of functioning is racially nondiscriminatory. The greater equality found may reflect similar access to medical care, which may be more uniform than elsewhere, and greater commonality of background. Level of functioning is affected by most chronic health conditions, including depression, and higher score on the VACS Index. In all treatment settings, intervention to improve health and physical functioning should be considered, so that independence is retained.
Footnotes
Acknowledgments
This work was supported by the National Institutes of Health, National Institute on Aging P30 AG08716 (Claude D. Pepper Older Americans Independence Center) for M.S.M., M.N.K., G.G.F.; Duke University Center for AIDS Research (CFAR) 5P30 AI064518 for M.S.M.; HIV and Aging Pilot Program RFA 1R24AG044325-01 (joint collaboration between Claude D. Pepper Older Americans Independence Centers and the Centers for AIDS Research) for M.S.M.; Veterans Aging Cohort Study funded by National Institute on Alcohol Abuse and Alcoholism [grants U10 AA013566 completed (J.P.T.), U24 AA020794 (J.P.T.), U01 AA020790 (J.P.T.), UO1-AA020795 (J.P.T.), UO1-AA020799 (J.P.T.)], VHA Public Health Strategic Health Core Group, Agency for Healthcare Research and Quality [grants U19-HS021112, R18HS023258, R18HS023464 (S.C.)], and National Institute of Mental Health 5F32MH105293 (K.N.-L.).
Disclaimer
The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs, the National Institutes of Health, the Department of Health and Human Services, or the United States government. Furthermore, this article was prepared while Karen Nieves-Lugo, PhD, MPH was employed at George Washington University.
Disclosure Statement
No competing financial interests exist.
Supplementary Material
Supplementary Table S1
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
