Abstract
Background
Vitamin D deficiency is underdiagnosed and undertreated, especially among people living with HIV (PLWH). Recently, there has been an increased interest in the role of vitamin D in cardiovascular disease (CVD). While vitamin D deficiency has been associated with CVD in observational studies in the general population, there are limited data in PLWH. We therefore performed an analysis to assess the relationship of vitamin D and coronary atherosclerosis using coronary CT angiography (CCTA).
Methods
Women living with HIV (WLWH) without known CVD were included. Based on the median value of serum vitamin D levels, participants were dichotomized to either the <25 ng/ml (lower vitamin D group) or ≥25 ng/ml (higher vitamin D group). CCTA was used to assess plaque characteristics.
Results
Forty-three WLWH were included in the analyses (mean age 46 ±8 years, 56% African American, duration of HIV 15 ±6 years, 83% undetectable HIV viral load). WLWH in the lower vitamin D group (n=22) had significantly higher numbers of segments with any coronary plaque (2.27 ±3.01 versus 0.38 ±0.97; P=0.02) and segments with non-calcified coronary plaque (1.41 ±1.82 versus 0.29 ±0.64; P=0.03) compared with WLWH in the higher vitamin D group (n=21). After adjusting for Framingham CHD risk point score, body mass index, diabetes and race, the relationship remained significant.
Conclusions
Our study demonstrates a significant, independent relationship between lower vitamin D status and higher numbers of noncalcified coronary plaque segments in WLWH. Further studies are warranted to evaluate the effect of vitamin D on CVD in PLWH. Trial Registration Identifier: NCT00455793.
Introduction
Individuals living with HIV have an increased risk of coronary artery disease (CAD). In addition to traditional risk factors, such as hyperlipidaemia, diabetes mellitus, tobacco use and obesity, people living with HIV (PLWH) have non-traditional cardiovascular disease (CVD) risk factors associated with HIV [1–3]. These non-traditional risk factors include chronic inflammation, abnormal fat distribution and increased intestinal permeability [4–7]. Recent studies have implicated low vitamin D as a non-traditional risk factor for CVD in both the HIV-negative population and PLWH [8–11].
The high prevalence of vitamin D deficiency among PLWH and among women is well established [12,13]. Within these populations, African Americans have higher rates of vitamin D deficiency and are less likely to receive adequate replacement [14]. Multiple observational studies have linked vitamin D deficiency to cardiovascular events in the general population [8,9]. However, there are limited data on the relationship of vitamin D with CVD in PLWH. Lai et al. [11,15] have published on the association of low vitamin D with the severity of coronary stenosis and coronary calcium score in African American chronic cocaine users with HIV. To increase our understanding of the relationship between vitamin D and atherosclerosis, as well as vitamin D's relationship with traditional and non-traditional risk factors for CVD, we investigated the relationship between vitamin D and coronary plaque characteristics, and with inflammatory markers and traditional risk factors in women living with HIV (WLWH).
Methods
Study design
This study reports on new analyses from a previously published observational study [2]. The study was conducted from January 2011 to December 2012 and thus included seasonal variation. Participants between the ages of 18 and 65 were recruited from Boston community centres and infectious disease clinics. The HIV-negative controls were recruited from the same communities as PLWH. Some of the HIV-negative controls were family members or friends of PLWH in our cohort. This was done to ensure similarities in demographics and cardiovascular risk factors between PLWH and HIV-negative controls. Our analysis includes 43 women with chronic HIV infection and 24 HIV-negative women as controls. Participants with known cardiovascular disease (including cardiac disease or stroke) or symptoms consistent with angina were excluded. Participants with creatinine greater than 1.5 mg/dl were also excluded to reduce the risk of contrast nephropathy. Data on vitamin D and calcium supplementation was obtained by participant self-report. All participants provided informed consent prior to enrolment. The study was approved by the institutional review boards of the Partners Human Research Committee.
Coronary plaque measurement
Agatston calcium score was calculated using the non-contrast CT images by standardized techniques as previously described [2]. Coronary CT imaging was performed using a 64-slice CT scanner (Sensation 64; Siemens Medical Solutions, Forchheim, Germany). Assessment of coronary atherosclerotic plaque (including the number of total coronary segments with plaque, coronary segments with calcified plaque and coronary segments with non-calcified plaque) was determined by a consensus reading between two cardiac imaging specialists. Physicians analysing the scans were blinded to the participants’ clinical history, HIV or vitamin D status.
Metabolic, biochemical, immunological and inflammatory parameters
Total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL) and triglycerides were measured after an overnight fast using standard techniques. HbA1c and creatinine were determined using standard techniques. Serum 25-hydroxyvitamin D was measured using radioimmunoassay kit (DiaSorin Inc., Stillwater, MN, USA). Anti-Mullerian hormone (AMH) was measured by ELISA (Ansh Labs, Webster, TX, USA). High sensitivity IL-6 (hsIL-6) and monocyte chemoattractant protein (MCP)-1 were measured by ELISA (R and D). IL-10 was measured by ELISA (Invitrogen, Waltham, MA, USA). CD4+ and CD8+ T-cell counts were assessed by flow cytometry. HIV viral load was determined by ultrasensitive real-time PCR (lower limit of detection 48 copies/ml) at the Massachusetts General Hospital clinical laboratory. HIV testing was performed by ELISA and confirmed by western blot in HIV-negative participants. Total Framingham CHD risk point score was calculated using age, tobacco use, total cholesterol, HDL, systolic blood pressure and hypertension treatment [16]. Estimated glomerular filtration rate (eGFR) was calculated using Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.
Statistical analysis
The normality of distribution was determined by examining the distribution histogram. Continuous and normally distributed variables were analysed using Student's t-test for 2 group comparisons and were presented as mean ± standard deviation (
Results
25-Hydroxyvitamin D and traditional cardiovascular risk factors in women with HIV
In WLWH, the median 25-hydroxyvitamin D level was 24.6 ng/ml. Using this median as a dividing point, we dichotomized WLWH into less than 25 ng/ml or greater than or equal to 25 ng/ml group (Table 1). Twenty-two WLWH had vitamin D levels less than 25 ng/ml, and twenty-one WLWH had vitamin D levels greater than or equal to 25 ng/ml. There were no differences in race, HbA1c, blood pressure, total cholesterol, HDL, LDL, total Framingham point score, tobacco use or eGFR between the groups. There were no differences in menopausal status based on AMH and last menstrual period between the groups. Body mass index (BMI) tended to be higher among those in the lower vitamin D group. WLWH in the lower vitamin D group tended to have higher levels of triglycerides than the higher vitamin D group (97 [74–119] versus 76 [67–95] mg/dl; P=0.06). HIV parameters (such as duration of HIV diagnosis, CD4+ T-cell count, % with suppressed viraemia and ART regimen) were not significantly different between the groups. All participants in the lower vitamin D group were not taking vitamin D or calcium supplements. In the higher vitamin D group, two participants were taking vitamin D supplements and one participant was taking calcium supplements.
Comparison of clinical, metabolic, HIV and cardiac CT parameters in women living with HIV by vitamin D status
Data reported as mean (± standard deviation), median (IQR) or percentage. AMH, anti-Mullerian hormone; BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate; HbA1c, haemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; LMP, last menstrual period greater than 12 months; NNRTI, non-nucleoside reverse transcriptase inhibitors; NRTI, nucleoside/nucleotide reverse transcriptase inhibitors.
25-Hydroxyvitamin D and inflammatory markers in women with HIV
Among WLWH, hs-IL-6 was significantly higher in the lower vitamin D group relative to the higher vitamin D group (1.81 [1.07–2.85] versus 1.18 [0.69–1.70] pg/ml; P=0.02). MCP-1 trended to be higher compared with the higher vitamin D group (276 ±135 versus 209 ±91 pg/ml; P=0.06). In contrast, the anti-inflammatory cytokine IL-10 was not significantly different between the groups.
25-Hydroxyvitamin D and coronary plaque types in women with HIV
Among WLWH, the lower vitamin D group had significantly higher total coronary plaque segments compared with the higher vitamin D group (0.5 [0-4.25; 2.27 ±3.01] versus 0 [0-0; 0.38 ±0.97]; P=0.02; Table 1). Non-calcified coronary plaque segments were significantly higher in the lower vitamin D group compared with the higher vitamin D group (0 [0-3; 1.41 ±1.82] versus 0 [0-0; 0.29 ±0.64]; P=0.03). Calcified coronary plaque segments were similar between the groups (0 [0-0; 0.09 ±0.29] versus 0 [0-0; 0.05 ±0.22]; P=0.63).
To further assess the relationship between 25-OH vitamin D and atherosclerosis, we also analysed 25-OH vitamin D as a continuous variable and found consistent results using Spearman correlation coefficient to assess the relationship of 25-OH vitamin D level and total coronary plaque segments (ρ=-0.31; P=0.04) and non-calcified coronary plaque segments (ρ=-0.32; P=0.04) in WLWH.
Multivariable regression model assessing the adjusted relationship of 25-hydroxyvitamin D and coronary plaque types in women with HIV
To evaluate whether the observed relationship between vitamin D and coronary plaque types was independent of traditional CVD risk and HIV-specific risk factors, we performed a multivariable regression analysis to control for total Framingham CHD risk point score, BMI, diagnosis of diabetes and race. In this adjusted analysis, lower vitamin D remained significantly associated with total coronary plaque segments (β=0.98; P=0.009) and non-calcified coronary plaque segments (β=0.61; P=0.009; Table 2). In a separate analysis adjusting for CD4+ T-cell count and HIV viral load, the association between lower vitamin D and total coronary plaque segments (β=0.67; P=0.03) and non-calcified coronary plaque segments (β=0.46; P=0.04) also remained significant.
Multivariate regression analysis of the relationship of vitamin D status and coronary plaque parameters adjusting for Framingham risk point score (age, systolic blood pressure, smoking, total cholesterol and HDL), race, BMI and diabetes in women living with HIV
BMI, body mass index; HDL, high-density lipoprotein.
An additional multivariate model was performed using vitamin D as a continuous variable adjusting for the above-mentioned CVD risk factors. The relationship between vitamin D as a continuous variable and total coronary plaque segments (β=-0.08; P=0.03) and non-calcified coronary plaque segments (β=-0.06; P=0.02) remained significant in women with HIV.
Demographics and clinical characteristics in HIV-negative control women
Our study included 24 HIV-negative women. The HIV-negative controls were similar to the WLWH group with regards to age, race, ethnicity and smoking status (Table 3). The vitamin D levels were similar between WLWH and HIV-negative women (24.6 [17.7–31.4] versus 20.3 [14.9–38.7] ng/ml; P=0.31). BMI, HbA1c, blood pressure, eGFR and lipids were also similar between the groups. There was no significant difference in the total Framingham point score in WLWH and HIV-negative women. In contrast to women with HIV, among the HIV-negative women, 25-OH vitamin D as a continuous variable did not significantly correlate with total coronary plaque segments (ρ=0.21, P=0.34) or non-calcified coronary plaque segments (ρ=-0.17, P=0.42).
Baseline demographics, metabolic and clinical parameters in women living with HIV and HIV-negative women
Data reported as mean (± standard deviation), median (IQR) or percentage. BMI, body mass index; BP, blood pressure; eGFR, estimated glomerular filtration rate calculated using CKD-EPI formula; HbA1c, haemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein.
Discussion
In our study, lower 25-OH vitamin D level was related to coronary plaque in WLWH. WLWH in the lower vitamin D group had significantly higher total coronary plaque segments and non-calcified coronary plaque compared with those in the higher vitamin D group (Figure 1). This association remained significant after adjusting for total Framingham point score, race and BMI. Our data suggest that WLWH who have lower vitamin D may be more prone to having increased non-calcified plaque.
Vitamin D and log coronary segments with noncalcified plaque in women living with HIV
In the HIV population, there are limited studies evaluating the relationship between vitamin D status and CVD. Ross et al. [10] were one of the first to demonstrate an inverse relationship between vitamin D and CVD by assessing carotid intima media thickness in HIV. In another study, Lai et al. [15] showed that among African American men and women with HIV with high preponderance of cocaine use, calcium score and coronary stenosis were associated with vitamin D deficiency. Our findings further add to the field by assessing plaque composition including non-calcified plaque not previously assessed before and investigating the relationship between different plaque types and vitamin D in women. Our study demonstrated for the first time that WLWH who have lower vitamin D have increased non-calcified coronary plaque.
In the general population, vitamin D deficiency has been associated with CVD. One study found that carotid intima media thickness was significantly higher in vitamin D deficient individuals without known CVD [8]. In a large cohort study from NHANES III, low vitamin D was associated with EKG changes consistent with sub-clinical myocardial injury [9]. Verdoia et al. [17] found that women with vitamin D deficiency were more likely to have extensive CVD than those without vitamin D deficiency. In contrast, many randomized controlled trials with vitamin D supplementation have not shown significant reduction in CVD events, such as a recently published analysis from the VITAL group. In their cohort, however, baseline vitamin D was not low (30.9 ±10 ng/ml) [18]. Perhaps the disconnect between observational studies and placebo-controlled randomized clinical trials may be explained by the fact that for clinical equipoise, most individuals in the placebo-controlled randomized studies are not vitamin D deficient at baseline as it would not be ethical to randomize them to placebo, and thus further supplementation with vitamin D in those who are already vitamin D sufficient is less likely to show improvement in CVD outcomes. Furthermore, most randomized trials have a relatively short follow-up period, which may not allow enough time for the development of CVD events.
Perhaps one of the mechanisms by which low vitamin D contributes to the development of atherosclerosis may be through vitamin D's known role in lipid metabolism. Supplementation with vitamin D has been shown to improve triglyceride levels [19]. A meta-analysis of 81 randomized clinical trials evaluating vitamin D and CVD performed by Mirhosseini et al. [20] found that vitamin D supplementation was significantly associated with reduction in triglycerides and LDL and an increase in HDL [20]. Until recently, the mechanism by which vitamin D exerts its effects on cholesterol metabolism was unclear. Using a rat model of type 2 diabetes mellitus, Yang et al. [21] demonstrated that vitamin D increased expression of adiponectin receptor 2 (adi-poR2), p28 mitogen-activated protein kinase (MAPK) and lipoprotein lipase (LPL), which are involved in fatty acid oxidation and triglyceride catabolism in chylomicrons and very low density lipoprotein (VLDL).
Furthermore, vitamin D may play an intricate role in the development of non-traditional risk factors for CVD. Vitamin D is an important modulator of inflammation via the intranuclear vitamin D receptor (VDR) [22]. The VDR is found in many cells, including inflammatory cells such as macrophages [23]. Vitamin D has been shown to suppress the proinflammatory cytokines interleukin (IL)-1, IL-6 and tumour necrosis factor (TNF)-alpha by inhibiting the presentation of antigens on macrophages and dendritic cells via the anti-inflammatory cytokine IL-10 [24]. Vitamin D has also been shown to promote the activation of CD4+ T-regulatory cells via VDR [25]. In post-menopausal women, vitamin D was positively associated with IL-10, while an inverse correlation was observed with interferon (IFN)-gamma [26]. In the same study, women who received vitamin D for 9 months had significantly reduced levels of IFN-alpha, IFN-gamma, IL-17, IL-6 and IL-5, as well as higher levels of IL-10 [26]. In HIV, the proinflammatory markers IL-6 and TNF-alpha are associated with vitamin D deficiency [27]. In our study, participants in the lower vitamin D group had significantly higher levels of hs-IL-6 and a trend of higher MCP-1 relative to those in the higher vitamin D group. Many studies have shown that inflammatory markers (including IL-6 and MCP-1) are associated with non-calcified coronary plaque in HIV [28]. Our study suggests a potential physiologically important role of vitamin D in inflammation and the formation of non-calcified coronary plaque in WLWH.
Additional factors resulting in vitamin D deficiency in people with HIV have been linked to certain antiretroviral therapy regimens, such as those that include protease inhibitors and efavirenz [29]. Protease inhibitors have been associated with alterations in vitamin D metabolism. In an in vitro study, protease inhibitors led to a reduction in the conversion of 25-OH vitamin D to the bioactive form of 1, 25-OH2 vitamin D [30]. In PLWH, protease inhibitor use was related to higher 25-OH vitamin D [31,32]. In another study, the use of protease inhibitors was associated with lower 1, 25-OH2 vitamin D [33]. Efavirenz use has also been linked to changes in vitamin D metabolism and has been associated with decreased vitamin D [34]. Brown et al. [35] also assessed the effect of efavirenz on vitamin D status in a longitudinal study. They found a significant decrease in 25-OH vitamin D levels 6–12 months after the initiation of efavirenz-based ART. In a study in Sub-Saharan Africa, the use of an efavirenz-based ART regimen for 48 weeks was significantly associated with lower vitamin D levels relative to baseline [36]. A similar association was found in the ECHO trial, where participants randomized to an efavirenz-based ART regimen for 48 weeks had significantly lower 25-OH vitamin D levels compared with baseline [37]. In our study, there were no significant differences in either protease inhibitor use or efavirenz use between the low and high vitamin D groups.
Strengths of our study include the assessment of vitamin D and its relation to coronary plaque types in women, a population that is both generally understudied in HIV research and also at greater risk for vitamin D deficiency [12] and CVD [1]. The participants in our study had no prior history of known CVD, angina or chronic kidney disease. The lower and higher vitamin D groups were well-matched for demographic and clinical parameters. There are also limitations of our study, such as our small sample size. This study also included only women from the New England area and thus our findings may not be generalizable to men living with HIV or women living with HIV in other areas of the world. Furthermore, our cross-sectional design does not allow for determination of direction of causality. Vitamin D was measured at a single time point in our study, and variation in vitamin D levels have been associated with seasonal changes as well as differences in vitamin D receptor levels [38].
Our study highlights the importance of vitamin D and its relation to cardiovascular risk. We found that WLWH in the lower vitamin D group had significantly higher total coronary plaque and non-calcified plaque. These associations remained significant after adjusting for potential confounders of the relationship between vitamin D and CVD including Framingham risk score, BMI and race. Higher levels of triglycerides were observed among WLWH in the low vitamin D group. Additionally, women in the lower vitamin D group had increased levels of hs-IL-6 and a trend of higher MCP-1. Our findings suggest that vitamin D may contribute to traditional risk factors (by means of lipid metabolism) and non-traditional (via chronic inflammation) risk factors leading to CVD. Future prospective trials to assess vitamin D and its effect on lipids, inflammation and atherosclerosis in PLWH are warranted to further elucidate and extend our findings.
Footnotes
Acknowledgements
We wish to thank the participants and the Nursing and Bionutrition Staff of the MGH Clinical Research Center. Principal contributions of authors are: study conception (LTC, JL), study design (JL), participant recruitment, history-taking and physical examination (JL, KVF, SEL), data acquisition (JL, KVF, SEL, TLS, ML), statistical analysis and interpretation (LTC, JL), drafting of the manuscript (LTC, JL), critical revision of manuscript (LTC, CFS, TLS, KVF, SEL, ML, UH, JL) and supervision of study (JL).
This work was supported by NIH RO1HL123351 (JL), NIH K23HL092792 (JL), NIH P30 DK040561 (TLS), Bristol-Myers Squibb and NIH 5T32DK007028-44 (LTC). The project described was also supported by Grant Number 1 UL1 RR025758-04, the Harvard Clinical and Translational Science Center, and the National Center for Research Resources. Funding sources had no direct role in the design of the study, data analysis or the writing of the manuscript.
Trial Registration Identifier: NCT00455793.
TLS received unrelated grant support from Novo Nordisk, Inc. and Kowa Pharmaceuticals. UH received grant support from Siemens Healthcare, the American College of Radiology Imaging Network, and HeartFlow Inc. ML received consulting fees from PQBy-pass and research support from NVIDIA. JL participated in a Medical Affairs Advisory Board meeting for Gilead Sciences and served as a consultant for Viiv Healthcare. LTC, CFS, KVF and SEL have nothing to declare. All declaration of interests by the co-authors are unrelated to the design of this study and this manuscript.