Plasma Selenium Concentrations Are Sufficient and Associated with Protease Inhibitor Use in Treated HIV-Infected Adults^,,

Abstract

Background: Selenium is an essential constituent of selenoproteins, which play a substantial role in antioxidant defense and inflammatory cascades. Selenium deficiency is associated with disease states characterized by inflammation, including cardiovascular disease (CVD). Although HIV infection has been associated with low selenium, the role of selenium status in HIV-related CVD is unclear.

Objectives: We sought to assess associations between plasma selenium and markers of inflammation, immune activation, and subclinical vascular disease in HIV-infected adults on contemporary antiretroviral therapy (ART) and to determine if statin therapy modifies selenium status.

Methods: In the Stopping Atherosclerosis and Treating Unhealthy bone with RosuvastatiN trial, HIV-infected adults on stable ART were randomly assigned 1:1 to rosuvastatin or placebo. Plasma selenium concentrations were determined at entry, week 24, and week 48. Spearman correlation and linear regression analyses were used to assess relations between baseline selenium, HIV-related factors and markers of inflammation, immune activation, and subclinical vascular disease. Changes in selenium over 24 and 48 wk were compared between groups.

Results: One hundred forty-seven HIV-infected adults were included. All participants were on ART. Median current CD4+ count was 613, and 76% had HIV-1 RNA ≤48 copies/mL (range: <20–600). Median plasma selenium concentration was 122 μg/L (range: 62–200). At baseline, higher selenium was associated with protease inhibitor (PI) use, lower body mass index, and a higher proportion of activated CD8+ T cells (CD8+CD38+human leukocyte antigen-DR+), but not markers of inflammation or subclinical vascular disease. Over 48 wk, selenium concentrations increased in the statin group (P < 0.01 within group), but the change did not differ between groups (+13.1 vs. +5.3 μg/L; P = 0.14 between groups).

Conclusions: Plasma selenium concentrations were within the normal range for the background population and were not associated with subclinical vascular disease in HIV-infected adults on contemporary ART. The association between current PI use and higher selenium may have implications for ART allocation, especially in resource-limited countries. Also, it appears that statin therapy may increase selenium concentrations; however, larger studies are necessary to confirm this finding. This trial was registered at clinicaltrials.gov as NCT01218802.

Introduction

Selenium is an essential micronutrient that incorporates into selenoproteins as selenocysteine. These selenoproteins are involved in multiple biological processes, including antioxidant defense and control of transcription factors involved in the inflammatory cascade (1). Selenium deficiency results in hierarchical loss of activity of selenoproteins and has been implicated in many disease states, including cardiovascular disease (CVD)⁸ (1, 2).

Some studies have found a link between HIV infection and selenium deficiency, possibly because of inadequate dietary intake, malabsorption, or overutilization in individuals who were not virologically controlled (3, 4). In this patient group, selenium deficiency has been associated with disease progression and mortality, even with adjustment for CD4+ T cell count (5–7). Presently, in the era of early aggressive antiretroviral therapy (ART) and virologic suppression, it is unclear whether HIV infection is associated with an alteration in selenium concentrations, because many patients are now well-nourished with a preserved immune system. There is also little known regarding the relation between selenium status and inflammation in this patient group, which is important, given that HIV infection is characterized by chronic immune activation even with virologic suppression (8, 9). Furthermore, despite mounting evidence that the heightened CVD risk seen with HIV infection (10, 11) is at least in part due to chronic inflammation and immune activation (12–14), there are few published reports on how selenium status relates to CVD risk in HIV (15). Last, it is not known whether statin therapy will affect selenium concentrations in HIV-infected individuals, as has been suggested in HIV-uninfected populations (16).

Therefore, the aim of this study was to evaluate selenium status in a cohort of HIV-infected adults on contemporary ART with good virologic control and to explore whether plasma selenium concentrations relate to markers of systemic inflammation, immune activation, or subclinical vascular disease measures in this patient group. A secondary aim was to evaluate for the first time in HIV the effect of rosuvastatin on plasma selenium concentrations over 24 and 48 wk. The hypotheses of this study were that higher plasma selenium would be associated with lower markers of systemic inflammation and immune activation and consequently lower CVD risk and that statin therapy would increase selenium concentrations.

Methods

The SATURN-HIV (Stopping Atherosclerosis and Treating Unhealthy bone with RosuvastatiN) study is a randomized, double-blind, placebo-controlled trial designed to determine the effect of rosuvastatin 10 mg daily on markers of CVD risk, skeletal health, and immune activation in HIV-infected adults on ART. As part of this study, plasma selenium concentrations were determined at baseline, week 24, and week 48. The eligibility criteria for this study have been previously described (17) and are listed at clinicaltrials.gov (NCT01218802). Briefly, all participants were ≥18 y of age, with chronic HIV-1 infection, on stable ART for ≥3 mo with a cumulative duration of ART of ≥6 mo and an HIV-1 RNA level <1000 copies/mL. Additional entry criteria included fasting LDL cholesterol ≤130 mg/dL and evidence of either heightened T cell activation and/or systemic inflammation [CD8+ T cells that express CD38 and human leukocyte antigen (HLA)-DR ≥19% and/or high-sensitivity C-reactive protein ≥2 mg/L]. Known coronary artery disease; diabetes mellitus; pregnancy; lactation; immunomodulating, hormonal, or anti-inflammatory medications; inflammatory condition besides HIV; creatinine clearance <50 mL/min by Cockcroft-Gault; or hemoglobin concentration <9 g/dL were exclusionary. Participants were randomly assigned 1:1 to receive rosuvastatin 10 mg by mouth daily or matching placebo. Random assignment was stratified by protease inhibitor (PI) use and presence or absence of coronary calcifications. The study was approved by the Institutional Review Board of the University Hospitals Case Medical Center, Cleveland, Ohio. All participants signed a written informed consent before enrollment.

Study evaluations

Demographics and medical history were obtained by self-report. A targeted physical exam and blood draw after a 12-h fast were performed at each visit (at screening and 0, 24, and 48 wk). Peripheral blood mononuclear cells were separated by centrifugation for 30 min, at 400 × g, at room temperature with Ficoll-Hypaque (GE Healthcare Life Sciences, Pittsburgh, Pennsylvania) and were cryopreserved until analyzed by flow cytometry in batch. Frozen plasma samples were stored at −80°C and analyzed in batch.

Plasma selenium concentrations.

Plasma selenium was measured as an assessment of selenium status in the study participants. The total selenium content of the plasma measures several components, including 2 selenoproteins (selenoprotein P and extracellular glutathione peroxidase 3) and protein- and nonprotein-bound selenium (18). The measurement of total plasma selenium content captures selenium status because in nondeficient individuals, selenoproteins are maximally expressed. Therefore, measurement of specific selenoproteins would be noninformative in these individuals (18). Plasma selenium was determined by automated electrothermal atomic absorption spectrophotometry with the use of a reduced palladium matrix modifier and an instrument equipped with L’Vov platforms and Zeeman effect background correction (19). Certified standards were used (Alfa Aesar, Perkin Elmer, and CPI). Calibration validation and calibration blanks were included at the beginning and end of the daily batch of samples, and at 10% intervals. Matrix effects were evaluated with the use of quantitative plasma standards (Seronorm and Utak) to assess percentage recovery of analyte from plasma.

Measures of subclinical vascular disease/CVD risk.

Common carotid artery intima media thickness and brachial artery endothelial function [flow-mediated dilation (FMD) and hyperemic velocity-time integral] were measured by ultrasound with the use of semiautomated edge-detection software (Medical Imaging Applications) as previously described (20). Carotid distensibility was measured from 10-beat ultrasound cine loops also as previously described (21). The coronary artery calcium (CAC) score was quantified by gated noncontrast computed tomography of the chest (20). Lipoprotein-associated phospholipase A₂, a measure of vascular inflammation, was measured with the use of ELISA (diaDexus).

Markers of inflammation and immune activation.

Soluble markers of inflammation (IL-6, soluble tumor necrosis factor receptors I and II, interferon γ–inducible protein 10, soluble intracellular adhesion molecule 1, and soluble vascular cell adhesion molecule 1) were measured by ELISA (R&D Systems), with the exception of high-sensitivity C-reactive protein, which was determined by particle-enhanced immunonephelometric assay on a BNII nephelometer (Siemens). Soluble markers of monocyte activation (soluble CD14 and soluble CD163) were measured by ELISA (R&D Systems) as well. All above biomarker assays were performed at the Laboratory for Clinical Biochemistry Research at the University of Vermont with the exception of interferon γ–inducible protein 10, which was performed at Case Western Reserve University.

Monocytes and T cells were phenotyped from peripheral blood mononuclear cells by flow cytometry as previously described (17). Three monocyte subsets, CD14+CD16+, CD14dimCD16+, and CD14+CD16−, were each quantified as a percentage of the overall monocyte population. T cell activation was quantified as the percentage of CD4+ or CD8+ cells that expressed both CD38 and HLA-DR. The flow cytometry was performed at Case Western Reserve University.

Other measures.

Insulin resistance was calculated from fasting glucose and insulin with the use of HOMA-IR (22). Fasting lipoprotein concentrations were determined in real time. CD4+ T cell counts and HIV-1 RNA levels were measured as part of routine clinical care.

Statistical analysis

Baseline demographics, HIV-related factors, CVD risk factors, and plasma selenium concentrations are described overall and by randomization group with the use of median and IQR for continuous variables and frequency and percentage for categorical variables. Baseline variables were compared between groups with the use of unpaired t tests or Wilcoxon Rank Sum tests as warranted by distribution for continuous variables and by chi-square tests, Fisher’s Exact tests, or Pearson Exact chi-square tests as appropriate for categorical variables.

Univariable followed by multivariable linear regression was used to explore associations with baseline plasma selenium concentrations, considering demographics (age, sex, and race), HIV-related factors [current CD4+ T cell count, nadir CD4+ T cell count, current HIV-1 RNA level (as a continuous variable and dichotomized by >48 or ≤48 copies/mL), known duration of HIV infection, cumulative ART duration, use of tenofovir, and use of PI], hepatitis B and C status, BMI, albumin concentration, estimated glomerular filtration rate, and all markers of inflammation and immune activation. Plasma selenium concentrations and all biomarkers were log-transformed before analysis. Variables with P < 0.25 in univariable analysis were considered for inclusion in the final multivariable model and backward elimination was used for selection. Albumin was kept in the final model, given that a significant proportion of plasma selenium is albumin-bound and concentrations of this protein affect total plasma selenium concentrations. The final model was checked to be sure the assumptions of linear regression were met. Next, Spearman correlation analysis was used to explore associations between plasma selenium concentrations and each measure of subclinical vascular disease (common carotid artery intima media thickness, FMD, hyperemic velocity-time integral, carotid distensibility, CAC score, and lipoprotein-associated phospholipase A₂ concentrations). Last, within- and between-group changes in plasma selenium were tested with the use of Wilcoxon Signed Rank and Rank Sum tests, respectively.

All statistical tests were 2-sided and considered significant with P < 0.05. Adjustments were not made in this significance level for multiple comparisons in this exploratory study. Analyses were performed with the use of SAS version 9.2.

Results

Enrollment in the SATURN-HIV study took place from March 2011 to August 2012. Of the 202 individuals screened, 147 were randomly assigned to receive rosuvastatin 10 mg daily (n = 72) or matching placebo (n = 75). Reasons for screening failure have been published previously (17). All 147 participants were included in this analysis. Participant flow diagram through 48 wk has been previously published (23). The baseline characteristics of the randomly assigned participants are displayed in Table 1 and were similar between treatment groups (all P > 0.05). Overall, median (IQR) age was 46 (40–53) y. Seventy-eight percent were men; 68% were African American; 8% had hepatitis C; and 63% were current smokers. Median (IQR) BMI was 27 (24–30) kg/m²; albumin concentration was 4.1 (3.8–4.3) g/dL; systolic and diastolic blood pressures were 121 (112–132) and 79 (72–84) mm Hg, respectively; HOMA-IR was 1.8 (1.1–3.3); and LDL and HDL concentrations were 97 (77–113) and 46 (37–57) mg/dL, respectively. Median (IQR) current and nadir (or lowest documented since HIV diagnosis) CD4+ T cell counts were 613 (425–853) and 178 (86–298) cells/mm³, respectively; known duration of HIV infection was 12 y. All participants were on ART by design, with 88% on tenofovir and 49% on a PI-containing regimen. Seventy-six percent had HIV-1 RNA levels ≤48 copies/mL (range: <20–600 copies/mL). Median (IQR and range) plasma selenium concentration was 122 (108–134 and 62–200) μg/L. Only 2 participants had plasma selenium concentrations <80 μg/L.

TABLE 1.

Baseline demographics; HIV-related, metabolic, and cardiovascular disease risk factors; and plasma selenium concentrations by random assignment group¹

Variable	Rosuvastatin	Placebo	P
Demographics
Participants, n	72	75
Age, y	45 (41, 51)	46 (39, 53)	0.98
Male	58 (81)	57 (76)	0.5
Race			0.93
Caucasian	20 (28)	23 (31)
African American	50 (69)	50 (67)
Other	2 (3)	2 (3)
HIV-related factors
Nadir CD4, cells/mm3	173 (84, 312)	189 (89, 281)	0.63
Current CD4, cells/mm3	608 (440, 848)	627 (398, 853)	0.76
HIV-1 RNA ≤48 copies/mL	55 (76)	57 (76)	0.96
HIV duration, y	11 (6, 17)	12 (6, 19)	0.45
ART duration, m	71 (39, 116)	63 (37, 119)	0.84
Metabolic and cardiovascular disease risk factors
Current smoker	43 (60)	50 (67)	0.38
BMI, kg/m2	27 (23, 30)	27 (24, 30)	0.88
Systolic BP, mm Hg	122 (112, 136)	120 (110, 132)	0.42
eGFRcr, mL/(min ⋅ 1.73 m3)	99 (85, 117)	101 (87, 118)	0.76
Glucose,2 mg/dL	79 (71, 87)	79 (73, 86)	0.77
HOMA-IR2	1.7 (1, 2.8)	2 (1.1, 4.4)	0.11
LDL cholesterol,2 mg/dL	96 (76, 107)	97 (77, 121)	0.3
HDL cholesterol,2 mg/dL	47 (38, 58)	46 (37, 57)	0.96
TGs,2 mg/dL	105 (77, 184)	121 (87, 184)	0.41
Total protein,2 g/L	236 (160, 346)	247 (147, 310)	0.86
Albumin,2 g/dL	4 (3.8, 4.2)	4.1 (3.9, 4.3)	0.1
Selenium concentrations
Plasma selenium, μg/L	122 (108, 139)	122 (110, 132)	0.93

¹Values are medians (IQRs) for continuous variables and frequency (%) for categorical variables, unless otherwise indicated. ART, antiretroviral therapy; BP, blood pressure; eGFRcr, creatinine-based estimated glomerular filtration rate.

²Serum.

Associations with plasma selenium at baseline.

In univariable analysis, higher selenium concentrations were associated with being Caucasian; being male; having a lower BMI, estimated glomerular filtration rate, and CD4+ T cell count; being on a PI; and having a higher percentage of CD8+CD38+HLA-DR+ T cells or activated CD8+ T cells. In the multivariable model, only being Caucasian, having a lower BMI, being on a PI, and having a higher percentage of activated CD8+ T cells (percentage CD8+CD38+HLA-DR+) remained independently associated with higher selenium concentrations. Other markers of inflammation and immune activation were not associated with selenium concentrations (Table 2).

TABLE 2.

Univariable and multivariable linear regression models with baseline plasma selenium as the dependent variable¹

	Univariable analysis		Multivariable model
Variable	Parameter estimate (SE)	P	Parameter estimate (SE)	P
Caucasian vs. other	0.111 (0.031)	<0.001	0.104 (0.032)	<0.01
Male vs. female	0.103 (0.035)	<0.01
BMI, kg/m2	−0.006 (0.002)	<0.01	−0.005 (0.002)	0.02
eGFRcr, mL/(min ⋅ 1.73 m3)	−0.001 (0.001)	0.04
Albumin, g/dL	0.097 (0.042)	0.02	0.021 (0.043)	0.62
Current CD4+ T cell count, cells/mm3	−9.571 × 10–5 (4.901 × 10–5)	0.05
On PI vs. not on PI	0.058 (0.029)	<0.05	0.055 (0.028)	<0.05
Log IP-10, pg/mL	0.028 (0.022)	0.22
Log CD8+CD38+HLA-DR+, %	0.062 (0.026)	0.02	0.063 (0.025)	0.01
Log sVCAM-1, μg/L	0.056 (0.045)	0.21
Log hsCRP, μg/mL	−0.015 (0.01)	0.16

¹All variables with P < 0.25 in univariable analysis are shown in this table. Selection with the use of backward elimination from these variables resulted in the final multivariable model shown (albumin included regardless of statistical significance). eGFRcr, creatinine-based estimated glomerular filtration rate; HLA, human leukocyte antigen; hsCRP, high-sensitivity C-reactive protein; IP-10, interferon γ–inducible protein 10; PI, protease inhibitor; sVCAM-1, soluble vascular cell adhesion molecule 1.

Relation between baseline plasma selenium and measures of subclinical vascular disease.

With the exception of brachial artery FMD, plasma selenium concentrations were not associated with any of the measures of subclinical vascular disease studied. In univariable analysis, higher plasma selenium concentrations were associated with lower brachial artery FMD in this cohort. Adjustment for usual CVD risk factors, including age, sex, race, BMI, smoking, and systolic blood pressure, attenuated this association and was no longer statistically significant (model not shown) (Table 3).

TABLE 3.

Correlation analysis of baseline plasma selenium concentration with measures of subclinical vascular disease and cardiovascular disease risk¹

Variable	Median value (IQR)	Spearman correlation coefficient	P
Brachial FMD, %	3.97 (2, 6.12)	−0.22	<0.01
Brachial hyperemic VTI, cm	77.39 (64.38, 94.68)	0.06	0.51
CCA IMT, mm	0.66 (0.62, 0.76)	0.15	0.06
Carotid distensibility, 10−6 ⋅ N−1 ⋅ m2	23.66 (18.84, 31.59)	0.03	0.74
Coronary calcium score	0 (0, 9)	0.06	0.48
Lp-PLA2, μg/L	167 (138, 201)	0.1	0.24

¹CCA IMT, common carotid artery intima media thickness; FMD, flow-mediated dilation; Lp-PLA₂, lipoprotein-associated phospholipase A₂; VTI, velocity-time integral.

Effect of rosuvastatin on plasma selenium concentrations.

At week 24, 136 participants had follow-up plasma selenium concentrations tested. Changes in selenium concentrations over 24 wk were not significant within or between groups at this time point [median (IQR) change in selenium concentration over 24 wk was +6 (−14, 15) μg/L with rosuvastatin (P = 0.51 within-group) vs. +7 (−15, 19) μg/L with placebo (P = 0.46 within group); P = 0.76 between groups]. At week 48, 128 participants had plasma selenium concentrations tested. Over 48 wk, selenium concentrations increased significantly in the rosuvastatin group [median (quartile 1, quartile 3) change in selenium concentration over 48 wk was +12 (−7, 34) μg/L; P < 0.01], but not the placebo group [+5.3 (−8, 19) μg/L; P = 0.12]. There was a trend toward a higher selenium concentration in the rosuvastatin group at week 48 [median plasma selenium concentration in the rosuvastatin group was 127 (116, 153) μg/L vs. 121 (110, 140) μg/L; P = 0.08], but changes in selenium concentrations over 48 wk were not different between groups (P = 0.14).

Discussion

To our knowledge, for the first time in a well-characterized, HIV-infected population on stable, contemporary ART, we investigated the relation between plasma selenium concentrations and markers of inflammation, immune activation, and subclinical vascular disease, and the effect of statin therapy on selenium concentrations in the setting of a randomized placebo-controlled trial. In this cohort, selenium deficiency was not apparent. Additionally, we found significant relations between plasma selenium and PI use, BMI, and T cell activation. Finally, over 48 wk, selenium concentrations increased significantly only in the rosuvatatin group, although between-group differences were not detected.

In contrast with our study, several cross-sectional studies have reported a significant association between low selenium concentrations and HIV infection (24–28). In these reports, various mechanisms have been postulated for the selenium deficiency observed, including inadequate intake, malabsorption, or overutilization and depletion of selenium that prevents adequate restoration (3, 4). A study conducted by Look et al. (29) compared HIV-infected adults classified into stages of HIV disease according to the CDC classification system with HIV-uninfected controls. Serum selenium concentrations were lower in patients with advanced clinical staging. Other studies have also reported a significant relation between CD4+ T cell count, opportunistic infections, HIV stage, and selenium concentrations. Jones et al. (30) showed that ART can result in an increase in selenium concentrations. Also, a recent study compared 3 groups of HIV-infected men; one group was antiretroviral treatment-naïve, one was on ART for <2 y, and the third was on ART for >2 y (31). The group with a longer time of exposure to ART and undetectable HIV-1 RNA levels did not show selenium deficiency, whereas the other 2 groups did, suggesting that selenium concentrations are related to better immune function and virologic control. Our findings support this hypothesis, because participants had a median CD4+ T cell count of 613 cell/mm³, cumulative ART duration of 5.3 y, and plasma selenium concentrations similar to what has been reported in the general population (18).

It appears that being on ART is associated with higher selenium concentrations (31). To add to this knowledge, specifically being on a PI-containing regimen was associated with higher plasma selenium concentrations in our study. The PI class of antiretrovirals has been associated with a lower risk of vitamin D deficiency (32), but, to our knowledge, this is the first report of the association between PIs and selenium. Plasma selenium concentrations capture selenium incorporated into the 2 selenoproteins (selenium-transporter selenoprotein P and glutathione peroxidase 3), as well as protein- and nonprotein-bound selenium (18). In this nonselenium-deficient cohort, it is to be expected that the 2 selenoproteins in the plasma are maximally expressed, leaving differences in total plasma selenium concentrations to changes in protein-bound selenomethionine (18). Interestingly, the association between PI use and plasma selenium was independent of the concentration of albumin, which makes up the majority of the protein in the plasma and would constitute most of the selenomethionine incorporated nonspecifically into protein in lieu of methionine. Therefore, the difference in plasma selenium concentration may reflect differences in the numbers of methionine positions actually filled by selenomethionine in albumin and, perhaps, other plasma proteins. One speculation is that PIs result in greater turnover of proteins in cells, which increases the availability of selenomethionine from tissue proteins for incorporation into plasma proteins during their synthesis. This being said, determining the mechanism underlying this association is outside the scope of this study.

In addition, having a high BMI was associated with lower plasma selenium in our study. Nutrient deficiencies historically have been associated with poor nutritional status, increased infectious disease burden, and overall poor health. However, recent data suggest that these deficiencies are also prevalent among the overweight and obese. Specifically, it has been shown that obese individuals have lower selenium concentrations than leaner individuals (33, 34). Further, Combs et al. (18) have suggested that there is a quadratic relation between selenium and BMI, with selenium concentrations being lower in individuals at the low and high ends of the BMI range.

In our study, a higher proportion of CD8+CD38+HLA-DR+ T cells, a marker of T cell activation, was associated with higher plasma selenium. Although this is counter to our hypothesis, selenium has been shown to affect the immune system in different manners, depending on the antigen and disease model. In the aging community, selenium may improve the immune response. In one study, in healthy aged adults, selenium increased the number of T cells and improved natural killer cell function (35). Another example is in the asthmatic population, in which there have been conflicting findings. Some studies suggest that selenium may play a role in reducing the severity of asthma (36, 37), whereas other studies have not shown this association (38, 39). This is likely secondary to the fact that although selenium is a potent antioxidant, it could also play a role in enhancing the immune response that drives the bronchoconstriction in the first place. One particular study found that selenium supplementation increased the number of CD3+HLA-DR+ T cells (40), which is consistent with what we have shown in this study.

Selenium deficiency has been associated with cardiomyopathy and cardiac dysfunction in several populations, including those with HIV infection. In a study in Rwanda, 18% of HIV-infected individuals had dilated cardiomyopathy, and, in multivariable analysis, low plasma selenium was associated with the development of cardiomyopathy (41). In a small study by Zazzo et al. (42), selenium supplementation in patients with AIDS and cardiomyopathy led to a return to a normal left ventricular shortening fraction within 21 d in 6 out of the 10 participants. However, the role of selenium in coronary heart disease is less clear. In the general population, a meta-analysis including several observational studies showed a statistically significant inverse association between selenium and coronary heart disease outcomes; specifically, a 50% increase in selenium concentrations was associated with a 24% decreased risk of coronary events (2). However, the same was not found in randomized controlled trials (43).There are limited data on the effect of selenium on CVD risk in HIV. As in our study, selenium was not associated with intima media thickness or CAC scores in HIV-infected adults in a study by Falcone et al. (15). Our study confirms and furthers this research, because we have evaluated several additional measures of subclinical vascular disease, as well as markers of inflammation and immune activation know to be elevated in HIV-related CVD (9, 12, 13).

To our knowledge, this is the first study to evaluate the effect of a statin on selenium in HIV infection. In the SATURN-HIV study, we have previously demonstrated that rosuvastatin reduces inflammation, including vascular inflammation, T cell and monocyte activation, and cystatin C (17, 21, 23, 44). Here, we have shown that whereas selenium concentrations did increase significantly over 48 wk only in the rosuvastatin group, the changes were not different between groups. Furthermore, an increase in plasma selenium of the magnitude seen in this study (+12 μg/L in the rosuvastatin group) is of unclear clinical significance, given that concentrations started in the sufficient range. Similar to our study, a short course of simvastatin, i.e., 4 wk, given to patients with dyslipidemia did not affect selenium concentrations (45). However, differences in selenium concentrations have been shown in hemodialysis patients on and not on a statin (16). It is possible that a longer duration of rosuvastatin would result in further increases in selenium concentrations, but additional studies are necessary to confirm this.

Our study has several strengths, including the double-blind, placebo-controlled, randomized trial design in addition to the comprehensive evaluation of inflammatory and subclinical vascular disease markers. We focused on a specific population of HIV-infected individuals with baseline heightened inflammation but normal LDL cholesterol concentrations; therefore, our findings may not be generalizable to all HIV-infected individuals.

In conclusion, in this cohort of HIV-infected adults on contemporary ART residing in the United States, plasma selenium concentrations were not deficient, which is encouraging, given previous associations between selenium deficiency and mortality in this patient group. Also, higher plasma selenium concentrations were associated with current PI use, a finding that may have implications for ART allocation, especially in resource-limited countries. Additional associations with higher plasma selenium concentrations included Caucasian race, lower BMI, and a higher proportion of CD8+CD38+HLA-DR+ T cells. Higher plasma selenium was not associated with improved measures of subclinical vascular disease or other inflammation markers, counter to our hypothesis. Finally, although plasma selenium concentrations did improve over 48 wk only in the rosuvastatin group, the changes were not significant between groups. Although these findings do not support the use of selenium supplementation for HIV-related CVD, follow-up studies are planned to evaluate associations between plasma selenium concentrations and changes in markers of inflammation and subclinical vascular disease with time.