Despite a high prevalence of vitamin D deficiency (VDD) among a predominantly black human immunodeficiency virus (HIV)–infected cohort, deficiency did not explain an observed racial disparity in low bone mineral density. Prevalence of VDD among HIV-infected persons was not significantly different from non-HIV infected controls.
Abstract
Background. Low bone mineral density (BMD) is common among patients infected with human immunodeficiency virus (HIV) and present in higher rates in black subjects. This study assessed vitamin D levels in HIV cases versus noninfected matched controls to determine if deficiency was associated with BMD and HIV clinical outcomes.
Methods. In total, 271 military beneficiaries with HIV underwent dual energy x-ray absorptiometry (DEXA) screening in 2001–2. Serum 25OH-vitamin D levels were determined using stored serum from the time of DEXA and 6–18 months prior. Two non–HIV-infected controls for each active duty case (n = 205) were matched on age, sex, race, zip code, and season using the Department of Defense Serum Repository (DoDSR). Vitamin D levels <20 ng/mL were considered deficient. HIV-related factors and clinical outcomes were assessed using data collected in the DoD HIV Natural History study.
Results. In total, 165 of 205 (80.5%) active duty HIV cases had 2 matched controls available. HIV cases had greater odds of for vitamin D deficiency (VDD) compared with controls (demographics adjusted paired data odds ratio [OR], 1.46, 95% confidence interval [CI], .87–2.45), but this was not statistically significant. Blacks were disproportionately deficient (P <.001) but not relative to HIV status or BMD. Low BMD was associated with typical risk factors (low body mass index and exercise levels, alcohol use); given limited available data the relationship between tenofovir exposure and VDD or low BMD could not be determined. Analysis of HIV-specific factors and outcomes such as exposure to antiretrovirals, HIV progression, hospitalizations, and death revealed no significant associations with vitamin D levels.
Conclusions. VDD was highly prevalent in black HIV- infected persons but did not explain the observed racial disparity in BMD. Vitamin D deficiency was not more common among HIV- infected persons, nor did it seem associated with HIV- related factors/clinical outcomes.
Vitamin D deficiency (VDD) is common in both the general population and human immunodeficiency virus (HIV)-infected populations in the United States. The National Health and Nutrition Examination Survey (NHANES) noted lower serum 25-OH vitamin D levels in blacks compared to whites [1]. The Study to Understand the Natural History of HIV and AIDS in the Era of Effective Therapy (SUN) found deficiency to be highly prevalent (88.5%) in black HIV-infected persons. Risk factors included a lack of exercise, race, renal insufficiency, and efavirenz containing antiretroviral regimens [2]. Significantly lower vitamin D levels have been found in patients receiving protease inhibitors (PIs) or nonnucleoside reverse transcriptase inhibitors (NNRTIs) when compared to treatment-naive HIV- infected controls [3].
Osteoporosis is 3 times more prevalent in HIV-infected men compared to those without HIV [4]. Low body mass index (BMI), higher HIV viral load (VL), and increased duration of HIV infection have been associated with the presence of bone abnormalities [5]. Bone-related complications such as fractures have been reported at higher rates in HIV patient populations [6, 7].
During 2001–2, a cross-sectional study evaluated bone mineral density in 271 HIV-infected military beneficiaries enrolled in the HIV Natural History Study (NHS). Rates of osteopenia and osteoporosis were found to be 40% and 6.4%, respectively. Black ethnicity was significantly associated with low bone mineral density (BMD). The multivariable analysis of sociodemographic and clinical factors did not explain the observed ethnic disparity [8]. The purpose of the current study was to retrospectively analyze vitamin D levels and to use subsequent NHS follow-up data to determine if VDD was associated with low BMD or HIV- related outcomes.
MATERIALS AND METHODS
A hologic QDR 4500A bone mineral densitometer provided DEXA scans of the proximal femur and lumbar spine. Data were collected at the time of DEXA using coordinator administered questionnaires screening alcohol and smoking history, weight, exercise, lactose intolerance, concomitant medical disease, medication use, duration of HIV infection, and antiretroviral exposure. Inadvertent destruction of the original copies required questionnaire items to be reassessed among participants (n = 211) at follow-up clinical visits; a conventional statistical method for missing data was used to include the (presumed) random subsample not reassessed. NHS data repository review provided information regarding HIV- related factors [9, 10].
The active duty subset of the HIV-infected cohort (n = 205) was matched to an uninfected cohort from the DoDSR. Two predeployment HIV-uninfected military controls were matched to each case according to age ± 3 years, sex, race, season, date of serum collection ± 3 years, and first 3 digits of the zip code of the military unit. No clinical information was available regarding the matched controls.
Banked serum samples from the cohort were tested to determine 25OH vitamin D, parathyroid hormone, and total/free testosterone levels at the time of DEXA. Vitamin D levels were determined using the LIAISON 25-OH vitamin D immunoluminometric assay on serum samples from the cases at the time of DEXA and 6 or 18 months prior and on serum samples from each of the 2 matched controls. Averages of the 2 seasonally distinct case results and the 2 matched control results were used for the statistical analysis to reduce confounding from seasonal variation [11, 12]; this approach was the best possible given the limited capability of the DoDSR to provide multiple specimens from different seasons for a single control. Using averages resulted in a 1:1 exposure ratio, although controls were identified in a 2:1 ratio. Testosterone level was determined using an ELECSYS testosterone electrochemiluminescence immunoassay and free testosterone was calculated [13]. Parathyroid hormone was measured using an ELECSYS radioimmunoassay.
Review of the NHS data repository provided information regarding zip codes at time of DEXA, cumulative antiretroviral exposure, clinical outcomes, and laboratory data (comprehensive metabolic studies, lipids, CD4 count, VL, and hepatitis B/C serologies). Latitude was determined using online databases at www.zipinfo.com (primary) and http://federalgovernmentzipcodes.us. Clinical outcomes after DEXA were reviewed including death, hospitalizations, decline in CD4 count <200 cells/mm3, and AIDS-defining conditions [14].
Data Variable Definitions
VDD was defined as an average value of <20 ng/mL [15]. World Health Organization (WHO) criteria were used defining osteopenia as a BMD t score between −1 and −2.5 and osteoporosis as a t score of −2.5 or lower compared to a race/sex matched 30 year-old population [16]. CD4 counts and VL measurements at seroconversion and DEXA were the closest measurement within a ± 90-day window. Duration of antiretroviral therapy (ART) was defined as the sum of continuous time on regimens in months. Death and hospitalization were identified using first-time diagnostic codes for these events in the NHS data repository during 8 years of follow-up after the DEXA. Decline in CD4 was defined as having 2 occurrences of CD4 <200 cells/mm3 within 1 year. AIDS-defining illnesses were defined using the 1993 Centers for Disease Control and Prevention (CDC) classification [14]. Chronic hepatitis B was defined as having a repeatedly positive HBsAg over >6 months and chronic hepatitis C as having a single positive anti-HCV before the date of DEXA.
Statistical Methods
Baseline characteristics were tabulated using proportions for discrete-valued variables and medians for continuous variables (interquartile range [IQR]) calculated for the overall study population and individual vitamin D groups. Mean and standard deviation are reported for variables with a normal distribution. P values were computed using indicated tests. Fisher exact or χ2 tests were used for categorical variables. Wilcoxon tests were used for univariate analysis of continuous variables that did not meet normality criterion (Shapiro-Wilks test; P-value >.05 within each group); t tests were used for those that met criterion. Logistic regression was used to explore possible risk factors for low BMD including VDD. To include all available risk factor information for the subset of participants whose questionnaires were not readministered, a method for handling missing covariates was employed (EM algorithm by the method of weights, a one-time model-fitting alternative to multiple imputation) [17]. A stepwise approach was used to select the final model [18]. Analyses were conducted using R version 2, SAS version 9.2, and StatXact version 8; study design used PASS 2008.
RESULTS
Baseline characteristics of HIV-infected subjects and potential factors associated with VDD are summarized in Table 1. The cohort was predominantly male (85%) and black (61%) with low prevalence of underlying medical conditions such as renal disease, thyroid disease, and diabetes. The median age was 41 years (IQR, 35–47). The duration of HIV infection was a median 121 months (IQR, 71–163 months) and >90% were treated with antiretrovirals (most had <4 years exposure to highly active antiretroviral therapy [HAART]). The median CD4 count at the time of DEXA screening was 505 cells/mm3 (IQR, 336–704), whereas median VL was 2.6 log10 copies/mL (IQR, 2.6–3.8). The subjects lived primarily along the east coast although cases were distributed throughout the United States (Figure 1).
Table 1.
Demographic Characteristics and Risk Factors Relative to Vitamin D Deficiency in HIV Cases
| Characteristics | Sample Size | Overall | Vitamin D Deficient (<20 ng/mL) | Vitamin D Not Deficient (≥20 ng/mL) | P Value |
|---|---|---|---|---|---|
| Demographics | 270 | 270 | 187 | 83 | |
| Median age (IQR) at DEXA (years) | 270 | 41 (35,47) | 41 (35,47) | 41 (34,48) | .773a |
| Sex (male) | 270 | 229 (85%) | 157 (84%) | 72 (87%) | .685c |
| Ethnicity | 270 | <.001c | |||
| White | 79 (29%) | 34 (18%) | 44 (53%) | ||
| Black | 166 (61%) | 135 (72%) | 31 (37%) | ||
| Other | 26 (10%) | 18 (10%) | 8 (10%) | ||
| Body Mass Index (IQR) | 263 | 26 (23,28) | 26 (24,28) | 25 (23,27) | .039a |
| Normald | 100 (36%) | 61(34%) | 38 (46%) | ||
| Overweight | 130 (49%) | 94 (52%) | 36 (44%) | ||
| Obese | 34 (13%) | 26 (14%) | 8 (10%) | ||
| Active duty | 270 | 205 (76%) | 141 (75%) | 63 (76%) | .948c |
| Latitude group (degrees north) | 265 | .443c | |||
| 18–29 | 17 (6%) | 9 (5%) | 8 (10%) | ||
| 30–34 | 52 (20%) | 36 (20%) | 16 (20%) | ||
| 35–39 | 165 (62%) | 118 (64%) | 47 (57%) | ||
| 40–44 | 5 (9%) | 17 (9%) | 8 (10%) | ||
| >45 | 6 (2%) | 3 (2%) | 3 (4%) | ||
| Medical factors | |||||
| Glomerular filtration rate(mL/min) | 244 | 1.000b | |||
| ≤60 | 5 (2%) | 4 (2%) | 1 (1%) | ||
| >60 | 239 (98%) | 166 (98%) | 73 (99%) | ||
| Chronic hepatitis B | 270 | 24 (9%) | 17 (9%) | 7 (8%) | 1.000b |
| Chronic hepatitis C | 270 | 28 (10%) | 20 (11%) | 8 (10%) | 1.000b |
| Diabetes | 194 | 10 (5%) | 7 (5%) | 3 (5%) | 1.000b |
| Lactose intolerance | 194 | 31 (16%) | 23 (18%) | 8 (13%) | .531b |
| Triglycerides | 264 | 159 (103,258) | 159 (100,256) | 163 (109,273) | .49a |
| Cholesterol | 265 | 196 (166,229) | 199 (166,238) | 188 (166,209) | .005a |
| Diuretics | 195 | 31 (16%) | 21 (16%) | 10 (16%) | .892b |
| Calcium supplements | 195 | 15 (8%) | 7 (5%) | 8 (12%) | .091b |
| Exercise | 165 | .832b | |||
| <3 days/week | 29 (18%) | 20 (18%) | 9 (16%) | ||
| ≥3 days/week | 136 (83%) | 90 (82%) | 46 (84%) | ||
| Alcohol | 132 | .82b | |||
| ≤2 drinks/day | 119 (90%) | 82 (91%) | 37 (88%) | ||
| >2 drinks/day | 13 (10%) | 8 (9%) | 5 (12%) | ||
| Parathyroid hormone (ng/L) | 261 | 29 (23,39) | 33 (25,42) | 25 (18,30) | <.001a |
| HIV-related factors | |||||
| CD4 at DEXA,c/mm3 | 267 | 505 (336,704) | 510 (343,687) | 486 (330,726) | 0.369a |
| Viral load at DEXA, Log 10 copies/mL | 267 | 2.6 (2.6,3.8) | 2.6 (2.6,3.8) | 2.6 (2.6,3.8) | .614a |
| CD4 nadir, c/mm3 | 270 | 259 (125,369) | 261 (124,360) | 257 (132,371) | .940a |
| Year of HIV diagnosis | 264 | 1992 (1988,1996) | 1992 (1988,1996) | 1992 (1988,1996) | .933a |
| HIV diagnosis to DEXA, months | 264 | 307 (180,413) | 308 (186,404) | 304 (158,420) | .917a |
| On ART at DEXA | 270 | 254 (94%) | 174 (93%) | 79 (95%) | .597b |
| Time on ART at DEXA, months | 253 | 61 (44,85) | 62 (45,85) | 59 (44,86) | .947a |
| Time on HAART at DEXA, months | 245 | 46 (30,56) | 46 (32,56) | 46 (25,56) | .672a |
| Time on efavirenz, months | 140 | 20.6 ± 14.3 | 20.4 ± 13.8 | 21.4 ± 15.8 | .731a |
| Time on protease inhibitor, months | 206 | 40.9 ± 19.6 | 41 ± 8.7 | 40.8 ± 21.7 | .941a |
Data were tabulated with counts (proportions) for categorical variables and medians (IQR) for continuous variables. Of 271 original subjects, 244 had vitamin D levels obtained approximately six months apart, 26 subjects had only one vitamin D level (assumed to be equal to their season-averaged value), and only one subject did not have available serum, thus the overall sample size is reported as n = 270.
Abbreviations: ART, antiretroviral therapy; CD4, cluster of differentiation 4; c/mm3, cells per cubic millimeter; DEXA, dual energy x-ray absorptiometry; HAART, highly active antiretroviral therapy; HIV, human immunodeficiency virus; IQR, interquartile range; lbs, pounds; L, liter; mL, milliliter.
a Student t test.
b Fisher exact test.
c Pearson χ2 test.
d 3 underweight persons included.
Figure 1.
Geographic distribution of human immunodeficiency virus (HIV)–infected study participants in the 48 contiguous states* with summaries of season-averaged serum vitamin D levels by intervals of latitude; restricted to participants with 2 vitamin D measures with season-averaged serum levels summarized by – size No.: median [IQR], n = 266: 14.8 [9.1–22.4]. *Among the 270 HIV-infected subjects, 3 did not reside in the 48 contiguous states; 3 subjects were stationed in Germany, Panama, and Puerto Rico. Abbreviation: IQR, interquartile range.
A majority of HIV-infected cohort members (187 of 270) with available serum samples were VDD: 69% (95% confidence interval [CI], 63%–75%) versus 302 of 509 (59%) non- HIV infected controls. Race was significantly associated with VDD (P <.001; Table 1); blacks tended to be disproportionately deficient. Higher body mass index (BMI) and cholesterol were associated with VDD. Latitude was not significantly associated with deficiency (P = .244), although generally lower median levels were noted at higher latitudes (Figure 1). Several medications, including antiretroviral agents, were analyzed relative to deficiency. Subanalysis of cumulative exposure to specific classes of antiretrovirals (NRTI, NNRTI, and PI including ritonavir) revealed no significant associations to deficiency. Efavirenz exposure was not associated with deficiency (P = .161). Only 13 subjects had previous exposure to tenofovir (duration, 3.2 months; IQR, 1.8–5.7) at time of DEXA. To assess the association of VDD to HIV infection adjusting for demographics, samples of an active duty HIV-infected subcohort (n = 205) were matched using DoDSR controls. A total of 189 (92.2%) HIV-infected cases had matched control serum samples and 165 (80.5%) had 2 matched controls (n = 330 control serum samples tested). Matched case-control comparisons of vitamin D levels within each racial group are summarized in Figure 2; only those sets whose season-averaged levels were discrepant were informative to our chosen measure of association (OR for paired binary data). HIV cases appeared to be at greater risk of VDD compared to controls across racial groups but not to a statistically significant degree (likelihood ratio test, P = .15; adjusted OR, 1.46; 95% CI, .87–2.45).
Figure 2.
Active duty HIV-infected case versus matched non-HIV infected control season-averaged vitamin D levels within race categories. Race groups differ qualitatively in association measured by paired-data odds ratios (ORs); given this as well as only 59 case-control matched pairs with discordant vitamin D status contribute to an overall OR, there is inadequate precision to rule out “no association” (OR, 1.46; 95% confidence interval, .87–2.45). Abbreviations: CI, confidence interval; HIV, human immunodeficiency virus.
Risk factors for low BMD were also analyzed. A family history of osteoporosis was reported in 15 (12%) VDD HIV-infected subjects versus 14 (22%) nondeficient (P = .11). A self-reported history of fracture was noted in 38 (31%) deficient subjects versus 20 (38%) of the nondeficient (P = .43). There was no association between VDD and low BMD among HIV-infected cases for either race. Spearman correlation coefficients did not show a significant relationship between vitamin D levels and DEXA t-scores: the hip at 0.33 (95% CI, −.04, .26) or the spine at 0.077 (95% CI, −.002, .29). Elevated parathyroid hormone levels were associated with lower vitamin D levels; among n = 257 participants with both measurements Spearman correlation coefficient was estimated to be -0.35 (95% CI, −.45, −.24). Several clinical characteristics were associated with low BMD (see Table 2). Both ethnicity and lower BMI at time of DEXA were positively associated with low BMD (unadjusted P values of .017 and .037, respectively). Alcohol use (>2 drinks daily) was a risk factor (adjusted P value, .015), as was exercising <3 days per week (unadjusted P value, <.001) and family history of osteoporosis (unadjusted P = .001). Logistic regression adjusting for multiple predictors at once confirmed all of these as risk factors for low BMD. Table 2 lists all factors found significant in a stepwise multiple predictor model fitting that included the 4-category vitamin D levels using clinical cutpoints; a single term indicating VDD (<20 ng/mL) was substituted for ease of interpretation in the final step. Interestingly, lactose intolerance (unadjusted P = .045) was not significant in this final multiple predictor model.
Table 2.
Single and Multiple Predictor Variable Analysis of Selected Risk Factors Related to Low Bone Density
| Unadjusted Odds Ratio (95% CI) | Single-Predictor P Valueω | Adjusted* Odds Ratio (95% CI) | Multiple-Predictor P valueω | |
|---|---|---|---|---|
| Vitamin D level (clinical cutpoints) | (n = 270) | [.12]λ | (n = 269) | |
| >30 ng/mL | 1.000 (Ref) | |||
| ≤10 ng/mL | 3.01 (1.19, 8.4) | .026 | ||
| 10 to 20 ng/mL | 2.5 (1.02, 6.85) | .055 | ||
| 20 to 30 ng/mL | 2.86 (1.08, 8.27) | .041 | ||
| Vitamin D deficiency (<20 ng/mL) | (n = 270) | [.32]λ | ||
| >20 ng/mL | 1.000 (Ref) | 1 | (Ref) | |
| ≤20 ng/mL | 1.3 (.77, 2.19) | .32 | 1.48 (.72, 3.06) | .289 |
| Race/Ethnicity | (n = 271) | [.054]λ | ||
| White | 1.000 (Ref) | |||
| Black | 1.96 (1.13, 3.43) | .017 | 3.49 (1.58, 7.7)* | .002 |
| Other | 1.56 (.63, 3.85) | .331 | 1.71 (.55, 5.28) | .35 |
| Body Mass Index | (n = 263) | [.059]A | ||
| Normal | 1.000 (Ref) | .032 | .073 | |
| Overweight | 0.57 (.34, .95) | .089 | 0.56 (.3, 1.06) | .037 |
| Obese | 0.5 (.22, 1.1) | 0.36 (.24, .94) | ||
| Lactose intolerance | (n = 195) | [.04]λ | ||
| No | 1.000 (Ref) | |||
| Yes | 2.24 (1.03, 5.03) | .045 | ||
| Alcohol consumption | (n = 132) | [.11]λ | ||
| ≤2 drinks/d | 1 | (ref) | 1 | (ref) |
| >2 drinks/d | 2.5 (.78, 8.24) | .12 | 5.85 (1.40, 24.36) | .015 |
| Exercise | (n = 195) | [<.001]λ | ||
| <3 d/wk | 1.000 (Ref) | 1 (ref) | ||
| ≥3 d/wk | 0.226 (.088, .529) | <.001 | 0.17 (.05, .5)* | .001 |
| Family history of osteoporosis | (n = 192) | [<.001]λ | ||
| No | 1.000 (Ref) | |||
| Yes | 4.28 (1.85, 10.82) | .001 | 4.75 (1.64,13.78)* | .004 |
| Latitude (°north) | (n = 270) | [.07]λ | per 10° north | |
| 1.06 (1.1, 1.12) | .485 | 1.02 (.95, 1.11) | .485 | |
| Longitude (°west) | (n = 270) | [.018]λ | ||
| greater than 82°W (75%ile) | 0.27 (.1, .69) | 0.22 (.06, .74) | .014 | |
| >76°W, ≤82°W (w/in IQR) | 0.38 (.16, .88) | 0.323 (.116, .90) | .031 | |
| 76°W (25%ile) or less | 1 | (ref) | 1 | (ref) |
The hip or spine t score <−1 was compared to the race- and sex-matched 30-year-old population. All participants with available data were included (n ≤ 271), applying EM by methods of weights where noted (*). * is symbol used in table to show EM analysis used. Note that not all predictors are included in both columns; only those predictors whose significance warranted inclusion in the stepwise model-fitting process were used [15]. This is true for lactose intolerance, whereas the categorized serum vitamin D levels were maintained as a 4-level categorical variable until the final (refitted) model.
Abbreviations: CI, confidence interval; EM, expectation-maximization algorithm; IQR, interquartile range; %ile, percentile.
ω Wald test P value.
λ P values in [ ]s are from likelihood ratio tests.
Factors that were not significantly associated with low BMD included smoking; HIV- related factors including CD4 and VL at DEXA, CD4 nadir, duration of HIV infection, AIDS defining conditions; chronic hepatitis B or C; diabetes mellitus, hyperthyroidism; self-reported fracture; and medication use to include calcium, bisphosphonates, steroids, thyroid hormone, diuretics, and proton pump inhibitors and H2 blockers (some data not shown). An analysis of cumulative exposure to antiretrovirals as a whole and segregated by individual classes did not demonstrate a significant relationship to low BMD. Time on efavirenz was not related to low BMD. A total of 18 of 229 (7.9%) male subjects were found to have total testosterone levels <280 ng/mL. In sum, 10 of 18 (55.6%) were white and 8 of 18 (44.4%) were black. There was no significant association between free testosterone levels and hip and spine t-scores, with Spearman correlation coefficients of 0.735 (95% CI, −.09, .17) and 0.499 (95% CI, −.03, .22) respectively.
We analyzed VDD relative to subsequent all-cause mortality, hospitalizations, respiratory disease, non-skin malignancies, decline in CD4, and subsequent clinical AIDS events. The study did not have sufficient power to detect relationships at a 5% significance level between VDD and clinical outcomes other than bone loss while accounting for differential amounts of follow-up. Descriptive information is presented in Table 3.
Table 3.
Event Outcomes After Dual Energy X-Ray Absorptiometry Scan
| Total | Vitamin D Deficient | Not Vitamin D Deficient | |
|---|---|---|---|
| 270 (100) | 187 (69) | 83 (31) | |
| Decline in CD4 <200 c/mm3 (total at risk) | 61 (23) | 22 (12) | 39 (47) |
| Hospitalized | 94 (35) | 25 (13) | 69 (83) |
| AIDS defining condition | 25 (9) | 8 (4) | 17 (21) |
| All cause deaths | 26 (10) | 9 (5) | 17 (21) |
The exposure of interest (vitamin D deficiency) was not measured at a fixed time in each subject's postseroconversion follow-up but instead at time closest to dual energy x-ray absorptiometry scan; thus, we do not report rates or proportions as these are not interpretable without making unverifiable assumptions. Data are No. (%) of participants unless otherwise indicated.
DISCUSSION
VDD was highly prevalent in our generally healthy HIV-infected and antiretroviral exposed black population (74%) but equally prevalent in a healthy black active duty control group (78%). Single nucleotide polymorphism (SNP)-related genetic variation in vitamin D receptors have been proposed as an etiology for racial predisposition [19]. Two other studies have suggested less deficiency in HIV-infected subjects [20]. Our observation clarifies previously published variable results regarding vitamin D levels in male HIV patients and supports ethnicity as a risk factor for deficiency [1].
Vitamin D metabolism may be affected by exposure to specific classes of antiretrovirals, although no such associations were found in our cohort [21]. Exposure to efavirenz, black ethnicity, and lack of exercise were identified as risk factors in the aforementioned SUN cohort [2]. In a Swiss HIV cohort, vitamin D levels were higher in white patients and lower in those with active NNRTI use [22]. In HIV- infected and uninfected women in the United States, black patients had the highest rates of VDD with recent PI use as a predictor [23]. Low vitamin D levels were also seen in the HIV black volunteers screened in the MONET trial; greater increases in vitamin D levels were noted in those who discontinued efavirenz in favor of a darunavir [24]. Although longitudinal studies have demonstrated a 5 ng/mL reduction in 25OH vitamin D within 6–12 months of starting efavirenz, it is not known whether this effect is attenuated over time [25]. In our analysis 103 persons were on efavirenz at time of DEXA, evenly distributed by race, and we used a 20 ng/mL level for defining VDD (some studies finding an efavirenz association noted it with severe deficiency <10 ng/mL and in white predominant populations) [2, 22]. Our observations may differ because of insufficient sample size to determine a difference, overall high rates of VDD, and a population that was 71% persons of color.
The primary outcome in our study was skeletal: VDD invoked as an explanation for racial disparity in low BMD; however, our results did not support VDD as an etiology. VDD has been associated with bone loss in patient populations that have multiple predisposing factors for reduced BMD such as hemophiliacs [26]. HIV patients likely have significant abnormalities in the bone remodeling process. A meta-analysis found 884 (67%) HIV patients had reduced BMD with increased risk noted in those exposed to PIs. Few studies included in this analysis adjusted for variables such as disease severity or treatment duration [27]. Time on PIs, time on tenofovir, and current use of PIs have been suggested as risk factors for low BMD [28]. Other studies have not revealed an association between HAART and BMD; bone loss has been reported in therapy-naive HIV patients [5, 29]. HAART and PIs as a class were not associated with low BMD in our cohort, and tenofovir use was minimal. The role of VDD and bone loss in HIV-infected populations remains to be defined. In 74 HIV-infected cases no differences in vitamin D levels were noted between those with and without osteopenia; those without previous HAART exposure had lower vitamin D [29]. Although risk factors such as lower BMI, less exercise, increased alcohol use, black rather than white race, and a family history of osteoporosis were significantly associated with low BMD in our cohort; VDD was not. Lactose intolerance was identified as a risk factor for VDD in another HIV cohort [30]. Blacks generally report less dairy consumption and are more likely to experience physical symptoms of discomfort after ingesting dairy products [31]. Dairy foods are common sources of calcium/vitamin D, and reduced intake (regardless of self-reported lactose intolerance) could impact racial disparity in BMD.
There are few reports of extraskeletal clinical outcomes related to VDD in HIV-infected populations. Higher incidence rates of AIDS and death have been seen in patients with baseline 25-OH vitamin D levels <12 ng/mL [32]. Studies in non-HIV patients indicate that mortality from malignancy and cardiovascular disease varies according to increasing distance from the equator, suggestive of vitamin D effect [33]. Vitamin D has immune modulating properties (perhaps through autophagy), and deficiency has been linked to viral respiratory illnesses and tuberculosis [34, 35]. In HIV patients from Tanzania, low vitamin D status correlated with progression to WHO HIV disease stage III, although no significant relationship between vitamin D and CD4 count was noted. Higher vitamin D levels appeared to protect against all-cause mortality and anemia in infected women [36]. The modest number of events and lack of clearly delineated follow-up periods (eg, timing of DEXA/vitamin D) limited our estimation of interpretable clinical outcomes.
There are other limitations to our study. The cohort was geographically restricted to latitudes particularly at risk for VDD. Recall bias could have impacted data obtained regarding self-reported histories of osteoporosis and fractures in a population undergoing DEXA screening. The study design was subject to selection bias. This was an overall healthy HIV-infected population treated with antiretrovirals, with excellent access to medical care. The effect of lack of antiretroviral treatment could not be evaluated. Furthermore, tenofovir, with potentially deleterious effects on bone mineralization, was not generally available (approved by the Food and Drug Administration in 2001). Prospective data on the development of fractures and follow-up DEXA scan results were not collected. Data on vitamin D supplementation was not available nor was systematic assessment of subsequent supplementation in osteopenic patients, although it is notable that those identified at the time of DEXA as osteoporotic were referred for bisphosphonate therapy. Finally, the method of obtaining matched but de-identified control serum samples from the DoDSR did not allow for clinical information to be gathered on controls.
Strengths of our study are that the cohort was a relatively large population of color with a low rate of comorbidities that could potentially influence vitamin D and bone metabolism, generally stable socioeconomic status, and equal access to care. Use of a matched noninfected control group to assess the association of VDD with HIV infection strengthens the study. We were able to evaluate clinical outcomes nearly a decade after vitamin D levels were assessed. Our results support targeted behavioral interventions regarding alcohol reduction, exercise, and strategies to accommodate lactose intolerance and promote dietary dairy intake to maintain bone health in HIV populations.
CONCLUSIONS
In this predominantly black male cohort of HIV-infected patients, low BMD was highly prevalent. Risk factors related to low BMD were confirmed, although low BMD did not appear to be associated with VDD. Antiretroviral exposure did not impact bone density or the presence of deficiency. Given the limited data available we could not determine the relationship between tenofovir exposure and VDD or low BMD. With the data available we did not identify a relationship of VDD to HIV disease progression or all-cause mortality. Well-designed studies to evaluate vitamin D supplementation's effect on serum vitamin D levels and low BMD over time, as well as on HIV disease in the presence of antiretroviral therapy, should be conducted before empirically providing supplementation in an attempt to prevent adverse outcomes.
Notes
Acknowledgments. The authors would like to thank Dr Angie Eick of the Department of Defense Serum Repository, who was instrumental in performing matching and obtaining control serum samples, Greg Grandits, MS, and Patricia Grambsch, PhD, of the University of Minnesota who assisted in developing the statistical analysis, Rosemary McKaig, PhD, at the National Institute of Allergy and Infectious Diseases, the many individuals at the HIV Natural History Study Repository, Data Coordination, and Analysis Center, and the Walter Reed Army Medical Center and National Naval Medical Center clinical laboratories, who made this study possible (Gary Bosco, Chip Bradley, Bob Dholakia, Connor Eggleston, Conchita King, and Barbara Nagaraj), and finally the study coordinators and volunteers for their participation.
Disclaimers. The content and views expressed in this publication are the sole responsibility of the authors and do not necessarily reflect the views or policies of the National Institutes of Health, the Department of Health and Human Services, the Department of Defense or the Departments of the Army, Navy, Air Force, or the US government. Mention of trade names, commercial products, or organizations does not imply endorsement by the US government.
Financial support. Support for this work (IDCRP-000-03) was provided by the Infectious Disease Clinical Research Program (IDCRP), a Department of Defense program executed through the Uniformed Services University of the Health Sciences. This project has been funded in whole, or in part, with federal funds from the National Institute of Allergy and Infectious Diseases, National Institutes of Health, under Inter-Agency Agreement Y1-AI-5072.
Potential conflicts of interest. All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1.Ginde AA, Liu MC, Camargo CA. Demographic differences and trends of vitamin D insufficiency in the US population, 1988–2004. Arch Intern Med. 2009;169:626–32. doi: 10.1001/archinternmed.2008.604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Dao CN, Patel P, Overton ET, et al. Low vitamin D among HIV-infected adults: prevalence of and risk factors for low vitamin D Levels in a cohort of HIV-infected adults and comparison to prevalence among adults in the US general population. Clin Infect Dis. 2011;52:396–405. doi: 10.1093/cid/ciq158. [DOI] [PubMed] [Google Scholar]
- 3.Madeddu G, Spanu A, Solinas P, et al. Bone mass loss and vitamin D metabolism impairment in HIV patients receiving highly active antiretroviral therapy. Q J Nucl Med Mol Imaging. 2004;48:39–48. [PubMed] [Google Scholar]
- 4.Clay PG, Voss LE, Williams C, Daume EC. Valid treatment options for osteoporosis and osteopenia in HIV-infected persons. Ann Pharmacother. 2008;42:670–9. doi: 10.1345/aph.1K465. [DOI] [PubMed] [Google Scholar]
- 5.Annapoorna N, Rao GV, Reddy NS, Rambabu P, Rao KR. An increased risk of osteoporosis during acquired immunodeficiency syndrome. Int J Med Sci. 2004;1:152–64. doi: 10.7150/ijms.1.152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Mondy K, Tebas P. Emerging bone problems in patients infected with human immunodeficiency virus. Clin Infect Dis. 2003;36:S101–5. doi: 10.1086/367566. [DOI] [PubMed] [Google Scholar]
- 7.Triant VA, Brown TT, Lee H, Grinspoon SK. Fracture prevalence among human immunodeficiency virus (HIV)-infected versus non-HIV-infected patients in a large US healthcare system. J Clin Endocrinol Metab. 2008;93:3499–504. doi: 10.1210/jc.2008-0828. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Aronson N, McKaig R, Howard R, Hammock V, Hawkes C, Duncan W. Washington, DC: 2005. Frequent decreased bone mineral density is seen in a predominantly African-American male HIV-infected patient cohort. 45th Interscience Conference on Antimicrobial Agents and Chemotherapy. [Google Scholar]
- 9.Weintrob A, Fieberg A, Agan B, et al. Increasing age at HIV seroconversion from 18 to 40 years is associated with favorable virologic and immunologic responses to HAART. J Acquir Immune Defic Syndr. 2008;49:40–7. doi: 10.1097/QAI.0b013e31817bec05. [DOI] [PubMed] [Google Scholar]
- 10.Weintrob A, Grandits G, Agan B, et al. Virologic response differences between African Americans and European Americans initiating highly active antiretroviral therapy with equal access to care. J Acquir Immune Defic Syndr. 2009;52:574–80. doi: 10.1097/QAI.0b013e3181b98537. [DOI] [PubMed] [Google Scholar]
- 11.Tangpricha V, Pearce EN, Chen TC, Holick MF. Vitamin D insufficiency among free-living healthy young adults. Am J Med. 2002;112:659–62. doi: 10.1016/s0002-9343(02)01091-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Munger K, Levin L, Hollis B, Howard N, Ascherio A. Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis. JAMA. 2006;296:2832–8. doi: 10.1001/jama.296.23.2832. [DOI] [PubMed] [Google Scholar]
- 13.DeVan ML, Bankson DD, Abadie JM. To what extent are free testosterone (FT) values reproducible between the two Washingtons, and can calculated FT be used in lieu of expensive direct measurements? Am J Clin Pathol. 2008;129:459–63. doi: 10.1309/6PYTC60ALVQQ59RQ. [DOI] [PubMed] [Google Scholar]
- 14.1993 revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Recomm Rep. 1992;41:1–19. [PubMed] [Google Scholar]
- 15.Hanley DA, Davison KS. Vitamin D insufficiency in North America. J Nutr. 2005;135:332–7. doi: 10.1093/jn/135.2.332. [DOI] [PubMed] [Google Scholar]
- 16.Assessment of fracture risk and its application to screening for postmenopausal osteoporosis. Report of a WHO Study Group. World Health Organ Tech Rep Ser. 1994;843:1–129. [PubMed] [Google Scholar]
- 17.Ibrahim JG. Incomplete data in generalized linear models. J Am Statist Assoc. 1990;85:765–9. [Google Scholar]
- 18.Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New York: Wiley; 2000. [Google Scholar]
- 19.Engelman CD, Fingerlin TE, Langefeld CD, et al. Genetic and environmental determinants of 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D levels in Hispanic and African Americans. J Clin Endocrinol Metab. 2008;93:3381–8. doi: 10.1210/jc.2007-2702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ormesher B, Dhaliwal S, Nylen E, et al. Vitamin D deficiency is less common among HIV-infected African-American men than in a matched NHANES cohort. AIDS. 2011;25:1237–9. doi: 10.1097/QAD.0b013e3283474ef9. [DOI] [PubMed] [Google Scholar]
- 21.Childs K, Welz T, Samarawickrama A, Post FA. Effects of vitamin D deficiency and combination antiretroviral therapy on bone in HIV-positive patients. AIDS. 2012;26:253–62. doi: 10.1097/QAD.0b013e32834f324b. [DOI] [PubMed] [Google Scholar]
- 22.Mueller NJ, Fux CA, Ledergerber B, et al. High prevalence of severe vitamin D deficiency in combined antiretroviral therapy-naive and successfully treated Swiss HIV patients. AIDS. 2010;24:1127–34. doi: 10.1097/QAD.0b013e328337b161. [DOI] [PubMed] [Google Scholar]
- 23.Adeyemi OM, Agniel D, French AL, et al. Vitamin D deficiency in HIV-infected and un-infected women in the US. J Acquir Immune Defic Syndr. 2011;57:197–204. doi: 10.1097/QAI.0b013e31821ae418. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Fox J, Peters B, Prakash M, Arribas J, Hill A, Moecklinghoff C. Improvement in vitamin D deficiency following antiretroviral regime change: Results from the MONET trial. AIDS Res Hum Retroviruses. 2011;27:29–34. doi: 10.1089/aid.2010.0081. [DOI] [PubMed] [Google Scholar]
- 25.Yin M, Stein E. The effect of antiretrovirals on vitamin D. Clin Infect Dis. 2011;52:406–8. doi: 10.1093/cid/ciq169. [DOI] [PubMed] [Google Scholar]
- 26.Gerstner G, Damiano ML, Tom A, et al. Prevalence and risk factors associated with decreased bone mineral density in patients with haemophilia. Haemophilia. 2009;15:559–65. doi: 10.1111/j.1365-2516.2008.01963.x. [DOI] [PubMed] [Google Scholar]
- 27.Brown TT, Qaqish RB. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review. AIDS. 2006;20:2165–74. doi: 10.1097/QAD.0b013e32801022eb. [DOI] [PubMed] [Google Scholar]
- 28.Bonjoch A, Figueras M, Estany C, et al. High prevalence of and progression to low bone mineral density in HIV-infected patients: a longitudinal cohort study. AIDS. 2010;24:2827–33. doi: 10.1097/QAD.0b013e328340a28d. [DOI] [PubMed] [Google Scholar]
- 29.Ramayo E, González-Moreno MP, Macías J, et al. Relationship between osteopenia, free testosterone, and vitamin D metabolite levels in HIV-infected patients with and without highly active antiretroviral therapy. AIDS Res Hum Retroviruses. 2005;21:915–21. doi: 10.1089/aid.2005.21.915. [DOI] [PubMed] [Google Scholar]
- 30.Rodríguez M, Daniels B, Gunawardene S, Robbins GK. High frequency of vitamin D deficiency in ambulatory HIV-positive patients. AIDS Res Hum Retroviruses. 2009;25:9–14. doi: 10.1089/aid.2008.0183. [DOI] [PubMed] [Google Scholar]
- 31.Keith JN, Nicholls J, Reed A, Kafer K, Miller GD. The prevalence of self-reported lactose intolerance and the consumption of dairy foods among African American adults are less than expected. J Natl Med Assoc. 2011;103:36–45. doi: 10.1016/s0027-9684(15)30241-8. [DOI] [PubMed] [Google Scholar]
- 32.Viard JP, Souberbielle JC, Kirk O, et al. Vitamin D and clinical disease progression in HIV infection: results from the EuroSIDA study. AIDS. 2011;25:1305–15. doi: 10.1097/QAD.0b013e328347f6f7. [DOI] [PubMed] [Google Scholar]
- 33.Autier P, Gandini S. Vitamin D supplementation and total mortality: a meta-analysis of randomized controlled trials. Arch Intern Med. 2007;167:1730–7. doi: 10.1001/archinte.167.16.1730. [DOI] [PubMed] [Google Scholar]
- 34.Cannell JJ, Vieth R, Umhau JC, et al. Epidemic influenza and vitamin D. Epidemiol Infect. 2006;134:1129–40. doi: 10.1017/S0950268806007175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Yamshchikov AV, Desai NS, Blumberg HM, Ziegler TR, Tangpricha V. Vitamin D for treatment and prevention of infectious diseases: a systematic review of randomized controlled trials. Endocr Pract. 2009;15:438–49. doi: 10.4158/EP09101.ORR. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Mehta S, Giovannucci E, Mugusi FM, et al. Vitamin D status of HIV-infected women and its association with HIV disease progression, anemia, and mortality. PLoS One. 2010;5:e8770. doi: 10.1371/journal.pone.0008770. [DOI] [PMC free article] [PubMed] [Google Scholar]