| |
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 - pdf attached (this was presented in poster at CROI)
|
| |
| |
Download the PDF here
Clin Infect Dis. Feb 1 (2011) 52 (3): 396-405
"In summary, our study compares the prevalence of vitamin D insufficiency or deficiency among a large, diverse, contemporary cohort of HIV-infected adults with that in adults in the general US population.....Among 672 SUN participants with baseline serum 25(OH)D determinations who were not receiving vitamin D supplements, 70.3% (95% confidence interval [CI], 68.1%-74.9%) were vitamin D insufficient or deficient, compared with 79.1% (95% CI, 76.7-81.3) of US adults. Factors associated with vitamin D insufficiency or deficiency included black race (adjusted odds ratio [aOR], 4.51; 95% CI, 2.59-7.85), Hispanic ethnicity (aOR, 2.78; 95% CI, 1.31-5.90), higher body mass index (aOR, 1.04; 95% CI, 1.00-1.09), hypertension (aOR, 1.88; 95% CI, 1.10-3.22), lack of exercise (aOR, 3.14; 95% CI, 1.80-5.47), exposure to efavirenz (aOR, 1.98; 95% CI, 1.18-3.34), higher exposure to ultraviolet light (aOR, .78; 95% CI, .71-.86), renal insufficiency (aOR, .55; 95% CI, .36-.83), and exposure to ritonavir (aOR, .56; 95% CI, .35-0.89).....
.......Although the modestly lower prevalence of vitamin D insufficiency or deficiency among SUN participants was significant after adjustment for age, race, and sex, vitamin D insufficiency or deficiency remained highly prevalent, affecting >70% of HIV-infected persons, and was associated with exposures to certain antiretrovirals. The associations between low levels of 25(OH)D and many chronic conditions that are prevalent among HIV-infected persons highlight the need for routine screening for vitamin D insufficiency or deficiency. Future studies are needed to verify potential alterations in vitamin D metabolism caused by the exposure to antiretrovirals and their duration of use. Although one study has shown increases in 25(OH)D with vitamin D supplementation, larger studies are necessary to assess the impact of vitamin D supplementation on prevention of comorbidities among HIV-infected adults, such as osteoporosis or osteopenia and cardiovascular disease [49]......
.......Vitamin D deficiency and insufficiency or deficiency was highly prevalent among HIV-infected SUN participants but was less prevalent than among adults in the general US population. Similar to previous reports of vitamin D status in the general population, we observed higher odds of vitamin D insufficiency or deficiency among HIV-infected persons who were of non-Hispanic black or Hispanic race/ethnicity, had higher BMI, exercised less, or had hypertension [1, 6]. However, we did not observe associations between vitamin D insufficiency or deficiency and some other established predictors of 25(OH)D levels, including age and sex. .....In univariate analysis, factors significantly associated (P < .25) with higher odds of vitamin D insufficiency or deficiency included female sex, older age, both non-Hispanic black and Hispanic race/ethnicity, lower UV exposure, higher BMI, chronic HBV coinfection, HCV coinfection, GFR ≥90 mL/min/1.73 m2, hypertension, exercise <3 times per week, and exposure to efavirenz (Table 2).....In multiple logistic regression, non-Hispanic black race/ethnicity, Hispanic ethnicity, higher BMI, hypertension, lack of exercise, and efavirenz use were independently associated with significantly higher odds of vitamin D insufficiency or deficiency (Table 2). Higher UV exposure, GFR <90 mL/min/1.73 m2, and use of ritonavir were independently associated with significantly lower odds of vitamin D insufficiency or deficiency (Table 2). We did not observe significant associations between vitamin D insufficiency or deficiency and age or sex......Tenofovir and ritonavir (low- or standard-dose) exposure were associated with lower odds of vitamin D insufficiency or deficiency (Table 2).The duration of efavirenz and tenofovir exposure were associated with vitamin D insufficiency or deficiency (P = .003 and .08, respectively), but not duration of nevirapine or ritonavir exposure (P = .92 and .94, respectively).....Similar to some antiepileptic agents and alcohol, efavirenz probably reduces levels of circulating 25(OH)D by inducing 24-hydroxylase, a cytochrome P450 enzyme, which hydrolyzes 25(OH)D to its inactivate form, 24,25(OH)2D [27, 28, 42]......Our findings suggest that there is an association between ritonavir exposure and lower odds of vitamin D insufficiency or deficiency, which might reflect the underlying action of ritonavir on vitamin D metabolism. ......In our analysis, we did not observe any significant association between nevirapine use and vitamin D insufficiency or deficiency, a finding consistent with those of one small longitudinal study Immune parameters, such as nadir CD4 T cell count and AIDS diagnosis, were not significantly associated with vitamin D insufficiency or deficiency."
Christine N. Dao,1 Pragna Patel,1 E. Turner Overton,2 Frank Rhame,3 Sherri L. Pals,1 Christopher Johnson,1 Timothy Bush,1 John T. Brooks,1 and the Study to Understand the Natural History of HIV and AIDS in the Era of Effective Therapy (SUN) Investigators
1Division of HIV/AIDS Prevention, Centers for Disease Control and Prevention, Atlanta, GA; 2Washington University School of Medicine, Department of Infectious Diseases, St Louis, MO; 3Abbott-Northwestern Hospital, Department of Infectious Diseases, Minneapolis, Minnesota
Abstract
(See the editorial commentary by Yin and Stein, on pages 406-408.)
Background. We explored serum 25-hydroxyvitamin D (25[OH]D) levels and associated factors for insufficiency or deficiency in an adult human immunodeficiency virus (HIV) cohort and compared 25(OH)D levels with those in the general US population.
Methods. Using baseline data from the Study to Understand the Natural History of HIV and AIDS in the Era of Effective Therapy (SUN), a prospective, observational cohort study of HIV-infected adults enrolled at 7 HIV specialty clinics in 4 US cities from March 2004 to June 2006, we estimated the prevalence of vitamin D insufficiency or deficiency (defined as 25(OH)D levels <30 ng/mL), standardized by age, race, and sex. Using multiple logistic regression, we examined risk factors for vitamin D insufficiency or deficiency.
Results. Among 672 SUN participants with baseline serum 25(OH)D determinations who were not receiving vitamin D supplements, 70.3% (95% confidence interval [CI], 68.1%-74.9%) were vitamin D insufficient or deficient, compared with 79.1% (95% CI, 76.7-81.3) of US adults. Factors associated with vitamin D insufficiency or deficiency included black race (adjusted odds ratio [aOR], 4.51; 95% CI, 2.59-7.85), Hispanic ethnicity (aOR, 2.78; 95% CI, 1.31-5.90), higher body mass index (aOR, 1.04; 95% CI, 1.00-1.09), hypertension (aOR, 1.88; 95% CI, 1.10-3.22), lack of exercise (aOR, 3.14; 95% CI, 1.80-5.47), exposure to efavirenz (aOR, 1.98; 95% CI, 1.18-3.34), higher exposure to ultraviolet light (aOR, .78; 95% CI, .71-.86), renal insufficiency (aOR, .55; 95% CI, .36-.83), and exposure to ritonavir (aOR, .56; 95% CI, .35-0.89).
Conclusions. Similar to findings in US adults generally, vitamin D insufficiency or deficiency is highly prevalent among HIV-infected adults and is associated with known risk factors. Observed associations of vitamin D levels with renal insufficiency and with use of ritonavir- and efavirenz-containing regimens are consistent with both HIV-related and therapy-mediated alterations in vitamin D metabolism. Clinicians should consider screening all patients for vitamin D insufficiency or deficiency.
Vitamin D is essential for calcium homeostasis and bone metabolism [1]. Vitamin D deficiency is associated with a number of comorbidities, including hypertension, cardiovascular disease, insulin resistance, diabetes, dyslipidemia, impaired immune function, decreased neurocognitive function, and malignancies [2-6]. The primary determinant of vitamin D status is exposure to sunlight. With increasing urbanization and sunscreen use, vitamin D deficiency has become highly prevalent among the general population [7, 8].
Currently, serum concentration of 25-hydroxyvitamin D (25[OH]D) is considered the best indicator of vitamin D status, because it represents cumulative exposure to sunlight and dietary intake of vitamin D [8]. There is growing consensus that the necessary vitamin D level for calcium homeostasis and healthy bone metabolism is at least 30-32 ng/mL; below this level, secondary hyperparathyroidism develops to compensate for decreased absorption of calcium [9, 10]. A 25(OH)D level of >30 ng/mL indicates sufficient vitamin D, and levels of 21-29 ng/mL indicate an insufficiency [1, 8, 11, 12]. Using these definitions, it is estimated that 1 billion persons worldwide have vitamin D insufficiency or deficiency and that levels of vitamin D vary by sex, race, and age [1-3]. Low vitamin D levels can be caused by decreased cutaneous production of vitamin D3, decreased absorption of vitamin D in the intestine, increased hydroxylation of 25(OH)D to its active form, 1,25-dihydroxyvitamin D (1,25[OH]2D), or increased catabolism of both 25(OH)D and 1,25(OH)2D into biologically inactive calcitronic acid [8].
Middle-aged persons with human immunodeficiency virus (HIV) infection are at risk for numerous comorbidities typically seen in older aging populations, including frailty, metabolic syndrome, osteoporosis and fragility fractures, insulin resistance, diabetes, cardiovascular disease, and cognitive impairment [13-17], many of which have also been associated with vitamin D deficiency [5, 18-20]. Recently, low vitamin D levels have been associated with HIV disease progression and HIV-related complications [21]. Therefore, the role of vitamin D in preventing or mitigating these complications of HIV is of particular interest. Low vitamin D levels among HIV-infected persons have been described; however, these reports were either case series or studies from small cohorts of HIV-infected persons [22-28]. HIV infection and exposure to certain antiretrovirals might contribute to altered levels of 25(OH)D [25-30].
We assessed vitamin D status among HIV-positive adults in the Study to Understand the Natural History of HIV and AIDS in the Era of Effective Therapy (SUN). We compared the prevalence of vitamin D insufficiency or deficiency among HIV-infected adults with that in adults in the general US population. Finally, we examined potential risk factors for vitamin D insufficiency or deficiency among SUN participants.
RESULTS
The 672 SUN participants included in this analysis were predominantly male (76%), non-Hispanic white (58%), and men who have sex with men (55%) and had a median age of 41 years (interquartile range [IQR]: 35-47 years). They had a median CD4 T cell count of 471 cells/mm3 (IQR, 334-681 cells/mm3), 74% were virologically suppressed (HIV viral load, <400 copies/mL), and 79% were currently receiving cART (Table 1). Persons surveyed for the 2003-2006 NHANES represented nearly 170 million US adults, of whom 51% were male and 70% were non-Hispanic white, with a median age of 43 years (IQR, 31-56 years).
The age-, race-, and sex-adjusted prevalence of vitamin D deficiency (25[OH]D, <20 ng/mL) among SUN participants included in this analysis was 29.7% (95% CI, 26.3%-33.2%), compared with 38.8% (95% CI, 34.6%-43.1%) for US adults (Figure 1). The age-, race-, and sex-adjusted prevalence of vitamin D insufficiency or deficiency (25[OH]D, <30 ng/mL) among SUN participants was 70.3% (95% CI, 68.1%-74.9%), compared with 79.1% (95% CI, 76.7%-81.3%) for US adults (Figure 1). Among SUN participants and US adults, vitamin D insufficiency or deficiency was more prevalent among non-Hispanic blacks, Hispanics, and persons with lower UV exposure (Figure 2). Among US adults, the prevalence of vitamin D insufficiency or deficiency was equal among men and women (79% for both sexes); however, 80% of women in SUN were vitamin D insufficient or deficient, versus 69% of men (P = .01) (Figure 2). Vitamin D levels typically decrease with increasing age, as observed among US adults, but 82% of SUN participants 20-29 years of age were vitamin D insufficient or deficient, compared with 71% of those 30-39 years of age, 67% of those 40-49 years of age, and 76% of those ≥50 years of age (P = .04) (Figure 2) [2]. The difference in prevalence of vitamin D insufficiency or deficiency by age is explained by the overrepresentation of men and young, non-Hispanic black subjects among the SUN participants (45% participants aged 20-29 years were non-Hispanic black).
In univariate analysis, factors significantly associated (P < .25) with higher odds of vitamin D insufficiency or deficiency included female sex, older age, both non-Hispanic black and Hispanic race/ethnicity, lower UV exposure, higher BMI, chronic HBV coinfection, HCV coinfection, GFR ≥90 mL/min/1.73 m2, hypertension, exercise <3 times per week, and exposure to efavirenz (Table 2). The odds of vitamin D insufficiency or deficiency differed by the geographic location of the clinic and by season of specimen collection. Persons who had any cART exposure at baseline had a higher odds of vitamin D insufficiency or deficiency than those who were ART naive. Tenofovir and ritonavir (low- or standard-dose) exposure were associated with lower odds of vitamin D insufficiency or deficiency (Table 2). The duration of efavirenz and tenofovir exposure were associated with vitamin D insufficiency or deficiency (P = .003 and .08, respectively), but not duration of nevirapine or ritonavir exposure (P = .92 and .94, respectively). Immune parameters, such as nadir CD4 T cell count and AIDS diagnosis, were not significantly associated with vitamin D insufficiency or deficiency.
In multiple logistic regression, non-Hispanic black race/ethnicity, Hispanic ethnicity, higher BMI, hypertension, lack of exercise, and efavirenz use were independently associated with significantly higher odds of vitamin D insufficiency or deficiency (Table 2). Higher UV exposure, GFR <90 mL/min/1.73 m2, and use of ritonavir were independently associated with significantly lower odds of vitamin D insufficiency or deficiency (Table 2). We did not observe significant associations between vitamin D insufficiency or deficiency and age or sex. In the model that included duration of antiretroviral use, we observed the same associations as in our primary multiple logistic regression model (data not shown), with the exception that only duration of efavirenz exposure was associated with vitamin D insufficiency or deficiency (aOR, 1.02; 95% CI, 1.01-1.03), and duration of tenofovir exposure was not (aOR, .91; 95% CI, .73-1.12).
When we fit a multiple linear regression model with the same independent variables as our primary logistic regression model, we did not observe significant associations between serum 25(OH)D concentrations and hypertension (P = .17) or exposure to ritonavir (P = .07) (Table 3). However, exposure to tenofovir and higher BMI were independently associated with higher levels of 25(OH)D (Table 3). All other variables that were significant in our primary multiple logistic regression model remained independently associated with serum 25(OH)D concentrations (Table 3). When we included variables for duration of specific antiretroviral use in the multiple linear regression model, only the duration of efavirenz use was independently associated with lower levels of 25(OH)D (P < .0001).
DISCUSSION
Vitamin D deficiency and insufficiency or deficiency was highly prevalent among HIV-infected SUN participants but was less prevalent than among adults in the general US population. Similar to previous reports of vitamin D status in the general population, we observed higher odds of vitamin D insufficiency or deficiency among HIV-infected persons who were of non-Hispanic black or Hispanic race/ethnicity, had higher BMI, exercised less, or had hypertension [1, 6]. However, we did not observe associations between vitamin D insufficiency or deficiency and some other established predictors of 25(OH)D levels, including age and sex. Although we could not assess the association between levels of 25(OH)D and HIV infection, we did explore the associations between vitamin D insufficiency or deficiency and specific antiretroviral drugs [36].
Previous small cohort studies focused on vitamin D deficiency defined as 25(OH)D levels ranging between <10 ng/mL and <20 ng/mL [24, 26, 30]. We defined deficiency as 25(OH)D levels <20 ng/mL to identify both patients with severe and those with moderate deficiency. Our definition of vitamin D insufficiency or deficiency as levels <30 ng/mL was similar to that in previous NHANES assessments of vitamin D insufficiency or deficiency in the general US population and other reviews of the optimal 25(OH)D level for preventing several skeletal and nonskeletal health outcomes [2, 11, 37]. Based on these definitions, our findings corroborate other estimates of vitamin D levels among HIV-infected adults [24, 26, 30]. By focusing our analysis on vitamin D insufficiency or deficiency, we were able to explore risk factors among HIV-infected patients who were potentially at risk for bone disease, as well as other comorbidities associated with vitamin D status.
Previous small cohorts studies have reported associations between vitamin D deficiency and nonnucleoside reverse transcriptase inhibitors, and case studies have described low vitamin D levels among HIV-infected persons receiving efavirenz [24-30, 38-40]. In our analysis, we did not observe any significant association between nevirapine use and vitamin D insufficiency or deficiency, a finding consistent with those of one small longitudinal study [41]. The association we observed between efavirenz exposure and higher odds of vitamin D insufficiency or deficiency has been reported in 2 case studies that described the potential role of efavirenz in altering the metabolism of 25(OH)D and 1,25(OH)2D, leading to increased risk of vitamin D deficiency (25[OH]D, <10 ng/mL) [27, 38-40]. Similar to some antiepileptic agents and alcohol, efavirenz probably reduces levels of circulating 25(OH)D by inducing 24-hydroxylase, a cytochrome P450 enzyme, which hydrolyzes 25(OH)D to its inactivate form, 24,25(OH)2D [27, 28, 42].
Our findings suggest that there is an association between ritonavir exposure and lower odds of vitamin D insufficiency or deficiency, which might reflect the underlying action of ritonavir on vitamin D metabolism. Another study found higher levels of 25(OH)D in white HIV-infected persons exposed to protease inhibitors, compared with those exposed to nonnucleoside reverse transcriptase inhibitors or naive persons, but that study did not specifically explore associations with ritonavir exposure [30]. Ritonavir, a potent inhibitor of cytochrome P450 enzymes, has previously been shown in vitro to block the action of the 1α-hydroxylase that is responsible for the conversion of 25(OH)D to 1,25(OH)2D in the kidney [43]. Consequently, patients prescribed ritonavir might appear to have adequate levels of 25(OH)D but low levels of 1,25(OH)2D.
The association of tenofovir exposure with higher levels of 25(OH)D has biological plausibility. Tenofovir causes proximal tubule injury, inducing renal dysfunction, as measured by declines in GFR and increases in urine protein-creatinine ratio [44-46]. Given that the conversion of 25(OH)D to 1,25(OH)2D by 1α-hydroxylase occurs primarily in the proximal tubule, tenofovir-induced proximal tubule dysfunction might also reduce hydroxylation of 25(OH)D to 1,25(OH)2D, causing accumulation of 25(OH)D and leading to surreptitiously normal 25(OH)D levels despite low levels of 1,25(OH)2D [25, 47]. A recent study observed significantly higher levels of 1,25(OH)2D with tenofovir use but no association with 25(OH)D levels, highlighting the need for future studies to elucidate the effect of antiretrovirals on vitamin D metabolism [26].
Although HIV-infected persons are at higher risk for osteoporosis and fragility fractures [14, 16], we did not observe an association between BMD and vitamin D insufficiency or deficiency. However, in addition to 25(OH)D, BMD is determined by other key regulators of bone homeostasis, including parathyroid hormone, which regulates serum calcium levels through bone resorption and stimulation of 1α-hydroxylase activity to increase production of 1,25(OH)D and, therefore, increase absorption of calcium from the intestine. A study of 36 HIV-infected men showed elevated levels of parathyroid hormone among those who were prescribed tenofovir; among these persons, parathyroid hormone levels were much higher among men who were vitamin D insufficient than among vitamin D-sufficient men [48]. We did not measure serum concentrations of parathyroid hormone and therefore could not assess these levels among SUN participants receiving tenofovir. Future studies among HIV-infected persons are needed to assess the contributions of vitamin D status, parathyroid hormone, calcium, phosphorus, 1,25(OH)2D, and certain antiretroviral drugs, such as tenofovir, to low BMD.
Our study was subject to several limitations. SUN is only conducted in only 4 US cities; therefore, data presented here may not be generalizable to all HIV-infected persons in the United States. Moreover, this analysis was cross-sectional; thus, causal relationships cannot be determined. To assess potential alterations in vitamin D metabolism requires cohort and case-control studies that are sufficiently powered to evaluate serum concentrations of 25(OH)D and 1,25(OH)2D at the time HIV infection is diagnosed and throughout treatment with various ART regimens.
In summary, our study compares the prevalence of vitamin D insufficiency or deficiency among a large, diverse, contemporary cohort of HIV-infected adults with that in adults in the general US population. Although the modestly lower prevalence of vitamin D insufficiency or deficiency among SUN participants was significant after adjustment for age, race, and sex, vitamin D insufficiency or deficiency remained highly prevalent, affecting >70% of HIV-infected persons, and was associated with exposures to certain antiretrovirals. The associations between low levels of 25(OH)D and many chronic conditions that are prevalent among HIV-infected persons highlight the need for routine screening for vitamin D insufficiency or deficiency. Future studies are needed to verify potential alterations in vitamin D metabolism caused by the exposure to antiretrovirals and their duration of use. Although one study has shown increases in 25(OH)D with vitamin D supplementation, larger studies are necessary to assess the impact of vitamin D supplementation on prevention of comorbidities among HIV-infected adults, such as osteoporosis or osteopenia and cardiovascular disease [49].
METHODS
SUN Participants
The SUN is a prospective cohort study of 700 HIV-infected persons enrolled in 4 US cities (Denver, CO; Minneapolis, MN; Providence, RI; and St Louis, MO) from March 2004 to June 2006. The study design and methods have been described elsewhere [31]. Participants were generally healthy HIV-infected adults who were naive to antiretroviral therapy (ART) or had been exposed only to combination ART (cART) and were receiving routine HIV care at one of the study clinics. Demographic data, diagnoses, treatments (including dosage and duration), and laboratory data from all clinical visits were abstracted from medical records and entered into electronic databases. All enrolled participants provided written informed consent. The study protocol was approved and has been annually renewed by the institutional review boards of the Centers for Disease Control and Prevention (CDC) and each participating institution.
We reviewed behavioral, clinical, and laboratory data, including bone mineral density (BMD) and 25(OH)D levels, collected at study enrollment (baseline). SUN participants who had no documented vitamin D supplementation and baseline 25(OH)D assessments were included in the analysis. Serum vitamin D levels were measured using the 25(OH)D iodine 125 radioimmunoassay (Diasorin). Baseline comorbidities were defined as medical conditions clinically diagnosed and present at the baseline visit. Medications that participants were taking at enrollment and continued taking for at least 1 month thereafter were considered baseline concomitant medications. We estimated ultraviolet light (UV) exposure using data from National Weather Service for 2004-2006, calculating a 3-year average monthly UV exposure for the month at the location where the blood specimen was collected [32]. Participants were defined as chronically hepatitis B virus (HBV) infected if they had had a positive test for serum HBV surface antigen and as hepatitis C virus (HCV) infected if they had had a positive HCV serologic result or a detectable HCV plasma viral load, at or anytime before baseline in both cases.
National Health and Nutrition Examination Survey.
The National Health and Nutrition Examination Survey (NHANES) has been conducted annually by the National Center for Health Statistics at CDC since 1999. Each year, NHANES surveys a nationally representative sample of ∼5000 persons from 15 counties across the United States. Each person is interviewed to collect data on demographics, socioeconomic status, dietary habits including dietary supplement use, and health-related topics. Examinations are conducted on each participant to obtain medical and laboratory measures, including serum 25(OH)D level, which is measured using the 25(OH)D iodine 125 radioimmunoassay. NHANES data is released in 2-year cycles [33]. For this analysis, we used NHANES 2003-2004 and 2005-2006 data to describe levels of vitamin D in the general US population for comparison with SUN participants. The investigation followed the guidelines of the US Department of Health and Human Services regarding protection of human subjects. The study protocol was approved and renewed annually by each participating institution's ethical review board, and all study participants provided written informed consent.
Statistical Analysis.
We calculated the age, race/ethnicity, and sex-adjusted prevalence of vitamin D deficiency (defined as serum 25[OH]D levels <20 ng/mL), and vitamin D insufficiency or deficiency (defined as levels <30 ng/mL) among SUN participants, using estimates for the US adult population obtained from NHANES 2003-2006 data. We estimated the overall prevalence of vitamin D deficiency and insufficiency or deficiency in the US adult population using sample design variables and sample weights to account for the complex sampling scheme used for NHANES 2003-2006 [34]. We did not adjust the prevalence of vitamin D insufficiency or deficiency for UV index, because city- and state-level data on NHANES participants were unavailable.
We used univariate logistic regression to identify potential risk factors for vitamin D insufficiency or deficiency among SUN participants. Variables included (1) demographic and behavioral variables, including age, sex, race/ethnicity, exercise, HIV transmission category (men who have sex with men, injection drug users, heterosexual); (2) HIV-related variables including baseline CD4 T cell count, nadir CD4 T cell count, plasma HIV-1 viral load, ART exposure and duration of antiretroviral use; (3) environmental variables, including clinic site, UV exposure, and month of specimen collection; and (4) clinical variables including body mass index (BMI), BMD of the lumbar spine, neck, and hip measured using dual-energy x-ray absorptiometry, comorbidities, concomitant treatments, and select laboratory measures (eg, estimated glomerular filtration rates [GFRs]) (Table 2). Variables with a P value of <.25 in univariate analysis were included in the multiple logistic regression model [35]. In our primary logistic regression model, we only included antiretroviral exposure variables along with all other variables identified from univariate analysis. We constructed a second multiple logistic regression model with the same variables as the primary model but included variables for duration (in months) of antiretroviral use instead of antiretroviral exposure variables. Associations were assessed using adjusted odds ratios (aORs) and corresponding 95% confidence intervals (CIs). We fit a multiple linear regression model that included the same independent variables, with vitamin D as a continuous dependent variable.
| |
| |
| |
|
|
|
