| |
Vitamin D and clinical disease progression in HIV infection: results from the EuroSIDA study: Vit D levels associated with HIV progression & mortality
|
| |
| |
Download the PDF here
AIDS:
19 June 2011
"These results provide strong evidence that vitamin D deficiency is an important cofactor in HIV disease progression, even in the setting of widespread, efficient cART. Whether the relationship between vitamin D deficiency and events is causal must now be addressed, because of potential public health consequences."
"The present results confirm [15-19] that vitamin D insufficiency or deficiency is frequent in HIV-infected persons......In the US, 25(OH)D levels below 30 ng/ml and below 20 ng/ml were similarly prevalent in the SUN cohort of HIV-infected persons vs. in the general population (NHANES study), after adjustment for age, sex and race: 70.3 vs. 79.1% and 29.7 vs. 38.8%, respectively.....Conclusion: 25(OH)D deficiency was frequent in HIV-infected persons (83% on combined antiretroviral therapy), and was independently associated with a higher risk of mortality and AIDS events. Causality relationships should be examined, because of potential public health consequences.......an aggravating effect of vitamin D deficiency, possibly linked with immune dysfunction or increased inflammation, on different conditions.......Having a 25(OH)D level in the lowest tertile (<12 ng/ml) was strongly associated with the occurrence of AIDS events and all-cause mortality over a median follow-up of 5 years, after adjusting for a large number of variables, including season, ethnic and geographic origin, CD4 cell count and viral load (at baseline and time-adjusted), and there was no interaction between these variables and 25(OH)D levels for the association with events. Thus, a very low 25(OH)D level was associated with events, even in the case of virologically controlled HIV infection and immune restoration......There is increasing agreement that 25(OH)D levels below 30 ng/ml reflect insufficiency......We thus found a strong association between the lowest 25(OH)D tertile (<12 ng/ml) and the two main endpoints of all-cause death and AIDS events, but no differences between the medium and high 25(OH)D tertiles. Vitamin D deficiency therefore represents a new, independent, unfavorable prognostic marker in HIV infection, but without further research this cannot translate into clinical recommendations.......The cut-off of 30 ng/ml is not universally agreed upon as some consider that 20 ng/ml is more appropriate [8]. It is, however, consensual to consider that a 25(OH)D level below 10 ng/ml corresponds to vitamin D deficiency because such low levels are associated with rickets and osteomalacia, the recognized diseases caused by a lack of vitamin D.......Four hundred and seventy (23.7%) persons had 25(OH)D levels below 10, 1296 (65.3%) between 10 and 30, and only 219 (11%) above 30 ng/ml at baseline."
"The association of low 25(OH)D with clinical progression in HIV infection echoes the multiple functions now attributed to vitamin D, which, like other steroid hormones, can regulate gene activity in numerous cell types [2,5,6]. Vitamin D deficiency has been associated with cancer risk [1,2,5], with a large study suggesting that vitamin D supplementation could lower the risk of malignancies [33]. Studies have also associated vitamin D insufficiency with hypertension, insulin resistance and high BMI, thus establishing a link with cardiovascular risk [2], clinical cardiovascular disease [34,35], and cardiovascular and all-cause death [10,14]. Vitamin D regulates endothelial function [36], down-regulates the renin gene activity [37], and a placebo-controlled trial has shown that vitamin D administration lowers blood pressure in hypertensive persons [38]. Vitamin D has regulatory effects on immune functions [39]: the vitamin D receptor is expressed on adaptive and innate immune cells; vitamin D plays a role in innate immunity, particularly in the defence against mycobacteria, and in T-cell activation, with vitamin D insufficiency leading to altered T-cell proliferation [40]. Vitamin D deficiency has also been associated with inflammation [41]. In a study of patients referred for a coronary angiography, vitamin D deficiency was associated with all-cause and cardiovascular mortality, and correlated with high levels of markers of cell adhesion, oxidative stress and inflammation [C-reactive protein (CRP) and interleukin 6 (IL-6)] [10]. In a placebo-controlled study, vitamin D supplementation lowered the level of the inflammatory cytokine TNF-α and increased the level of the anti-inflammatory cytokine IL-10, in patients with congestive heart failure [42]. In a large cohort, low vitamin D levels have also been associated with increased levels of the coagulation activation markers tissue plasminogen activator and D-dimer [43].
Taken together, these results are of particular relevance to our findings, because of the well described increased vascular risk in HIV-infected persons, and because markers of inflammation (CRP and IL-6) and coagulation activation (D-dimers) have been associated with mortality and opportunistic diseases in HIV-infected persons [44,45]. In the present study, whereas non-AIDS events were not significantly associated with low 25(OH)D (possibly due to a lack of power), non-AIDS deaths were: this could reflect an aggravating effect of vitamin D deficiency, possibly linked with immune dysfunction or increased inflammation, on different conditions."
Viard, Jean-Paula; Souberbielle, Jean-Claudeb; Kirk, Olec; Reekie, Joanned; Knysz, Brygidae; Losso, Marcelof; Gatell, Joseg; Pedersen, Courth; Bogner, Johannes Ri; Lundgren, Jens Dc,j; Mocroft, Amandad; for the EuroSIDA Study Group
aCentre de Diagnostic et de Therapeutique, Hotel-Dieu, APHP, and EA 3620, Universite Paris Descartes, France
bService d'Explorations Fonctionelles, Hopital Necker, APHP, Paris, France
cCopenhagen HIV Programme, Panum Institute, Copenhagen, Denmark
dDepartment of Infection and Population Health, University College London Medical School, London, UK
eMedical University, Wroclaw, Poland
fHospital JM Ramos Mejia, Buenos Aires, Argentina
gHospital Clinic i Provincial, Barcelona, Spain
hOdense University Hospital, Odense, Denmark
iMedizinische Poliklinik, Munich, Germany
jCentre for Viral Diseases Diseases, Department of Infectious Diseases, Rigshospitalet, Copenhagen, Denmark.
Abstract
Background: We examined the association between vitamin D [25(OH)D] level and disease progression in HIV infection.
Methods: Within the EuroSIDA study, 2000 persons were randomly selected for 25(OH)D measurement in stored plasma samples closest to study entry. 25(OH)D results were stratified into tertiles. Factors associated with 25(OH)D levels and associations of 25(OH) levels with subsequent risk of all-cause mortality, AIDS and non-AIDS events were analyzed.
Results: Of 1985 persons with 25(OH)D levels available, 23.7% had 25(OH)D below 10, 65.3% between 10 and 30, and 11% above 30 ng/ml. At the time of 25(OH)D measurement, older persons, persons of black ethnic origin, living outside Southern Europe/Argentina, sampled during winter, and infected with HIV through nonhomosexual exposure were at higher odds of having low 25(OH)D levels, whereas persons receiving protease inhibitors were at lower odds. Compared to those in the lowest 25(OH)D tertile (<12 ng/ml), those in the middle (12-20) and higher (>20) tertiles had a significantly lower risk of clinical progression during subsequent follow-up. Adjusted incidence rate ratios for all-cause mortality were 0.68 (95% CI 0.47-0.99, P = 0.045) and 0.56 (95% CI 0.37-0.83, P = 0.0039), and for AIDS events were 0.58 (95% CI 0.39-0.87, P = 0.0086) and 0.61 (95% CI 0.40-0.93, P = 0.020), for the middle and higher tertiles, respectively. There was a similar, nonsignificant reduced incidence of non-AIDS events in the middle and higher tertiles.
Conclusion: 25(OH)D deficiency was frequent in HIV-infected persons (83% on combined antiretroviral therapy), and was independently associated with a higher risk of mortality and AIDS events. Causality relationships should be examined, because of potential public health consequences.
Introduction
In the general population, vitamin D insufficiency, assessed by the measurement of serum 25-hydroxyvitamin D [25(OH)D], is frequent. There is increasing agreement that 25(OH)D levels below 30 ng/ml reflect insufficiency [1-6]. This threshold was initially based on the relationship observed in apparently healthy populations between 25(OH)D serum levels and parathyroid hormone, whereas it now more consistently relies on the 25(OH)D concentrations obtained during randomized trials in which vitamin D has been shown to reduce the relative risk of peripheral fractures [7]. It is, however, probable that the extent to which low 25(OH)D levels are associated with various conditions differs. The cut-off of 30 ng/ml is not universally agreed upon as some consider that 20 ng/ml is more appropriate [8]. It is, however, consensual to consider that a 25(OH)D level below 10 ng/ml corresponds to vitamin D deficiency because such low levels are associated with rickets and osteomalacia, the recognized diseases caused by a lack of vitamin D [1-8]. In many observational studies, low 25(OH)D levels have been associated with osteopenia, cardiovascular disease, insulin resistance and diabetes, infections, cancer, renal abnormalities, inflammation and autoimmune diseases [1-7], and with mortality in elderly people [9], persons with high cardiovascular risk [10], renal failure [11], heart transplant [12], cancer [13] and diabetes [14]. Vitamin D deficiency could be viewed only as an end-marker of bad health, but the vitamin D receptor, a member of the nuclear receptor of steroid hormones family, is widely expressed, and active vitamin D, a steroid hormone, has been involved in the physiology of numerous cell types [2,5,6].
Vitamin D deficiency is frequent in HIV-infected persons [15-19], even on successful combined antiretroviral therapy (cART). Factors linked to HIV infection that could contribute to decreased 25(OH)D levels include low sun exposure and alimentary intake, poor absorption, altered storage in adipose tissue, abnormal vitamin D activation due to hepatic or renal impairment, and the interference of antiretroviral drugs with vitamin D metabolism [20]. In a group of untreated HIV-infected Tanzanian women, low 25(OH)D levels was associated with the risk of mother-to-child transmission, child death, and anemia, clinical disease progression and death in the mothers [21,22].
Results
Comparison of included and excluded persons
After adjustment for region, there were few differences between patients with available samples who were (n = 1985) and were not (n = 3450) selected for analysis. HCV antibody-positive persons were more likely to be included [adjusted odds ratio (aOR) 1.30; 95% confidence interval (CI) 1.11-1.52, P = 0.0010)], as were those with a higher viral load (aOR 1.07 per log10 higher; 95% CI 1.02-1.13, P = 0.0056). Older persons were less likely to be included (aOR 0.93 per 10 years; 95% CI 0.87-0.99, P = 0.016). Persons sampled during summer (aOR 0.78; 95% CI 0.66-0.92, P = 0.0043), autumn (aOR 0.81; 95% CI 0.67-0.98, P = 0.029) or winter (aOR 0.74; 95% CI 0.63-0.88, P = 0.0005) were less likely to be included, compared to those sampled during spring.
25(OH)D levels at date of sample
Four hundred and seventy (23.7%) persons had 25(OH)D levels below 10, 1296 (65.3%) between 10 and 30, and only 219 (11%) above 30 ng/ml at baseline. For statistical analyses, the population was stratified into tertiles, which defined the following groups: below 12, 12-20 and above 20 ng/ml (Table 1).
Stratification according to tertiles will be used throughout the result section.
There was a strong influence of season, with the highest proportions of persons with a high and a low 25(OH)D level having been tested in summer and in winter, respectively. Persons infected via nonhomosexual HIV exposure, those from regions other than Southern Europe, and older persons were more likely to have a low 25(OH)D level (Table 2). Dividing the sample date into quartiles, there was an increasing trend for those with later sample dates to be more likely to have a low 25(OH)D level. Compared to persons sampled in spring, those sampled in summer and autumn were less likely, whereas those sampled in winter were more likely to have low 25(OH)D levels. After adjustment for the other factors in Table 2, compared to those of white ethnic origin, persons of a black ethnic origin (n = 89) had almost four times the odds of having a 25(OH)D level in the lowest tertile (aOR 3.95; 95% CI 2.06-7.57, P < 0.0001). There was no increased odds for Asians (n = 55, aOR 1.22; 95% CI 0.59-2.50, P = 0.58), or for those of unknown ethnic origin (n = 122, aOR 1.24; 95% CI 0.75-2.09, P = 0.41). Compared to persons receiving no antiretrovirals at baseline, persons on protease inhibitor-based cART were at lower risk of having low 25(OH)D levels. No significant association was found with the use of efavirenz, nucleoside reverse transcriptase inhibitors (NRTIs) as a whole, or tenofovir.
Progression to clinical events
The event rates within the 25(OH)D strata are shown in Table 3. Kaplan-Meier curves of progression to these endpoints are shown in Fig. 1a-c. At 5 years after baseline, 10.3% of persons had developed AIDS in the lowest 25(OH)D group (95% CI 7.8-12.8), 6.0% in the middle 25(OH)D group (95% CI 3.9-8.1) and 5.1% in the upper 25(OH)D group (95% CI 3.3-6.9). The corresponding figures for death were 11.1% (95% CI 8.5-13.7), 6.6% (95% CI 4.7-8.9) and 5.6% (95% CI 3.7-7.5) and, for non-AIDS events were 8.8% (95% CI 6.3-11.3), 6.7% (95% CI 4.4-8.8) and 6.5% (95% CI 4.4-8.7), respectively.
Figure 2 shows the results of the univariate and multivariate analyses. The lowest 25(OH)D tertile was the reference group for all analyses. After adjustment, persons with 25(OH)D levels in the middle and upper tertiles had a statistically significant lower risk of AIDS, with an incidence rate ratio (IRR) of 0.58 (95% CI 0.39-0.87, P = 0.0086) and 0.61 (95% CI 0.40-0.93, P = 0.020), respectively. Similarly, persons with a 25(OH)D level in the middle tertile had a significantly reduced incidence of death (IRR 0.68, 95% CI 0.47-0.99, P = 0.045), as did those with a 25(OH)D level in the higher tertile (IRR 0.56; 95% CI 0.37-0.83, P = 0.0039). There were no differences between persons with 25(OH)D levels in the middle and higher tertiles. A similar trend was observed for the incidence of non-AIDS events: persons with 25(OH)D levels in the middle and higher tertiles had a nonsignificantly reduced incidence of non-AIDS events (IRR 0.73, 95% CI 0.48-1.12, P = 0.16, and IRR 0.79, 95% CI 0.52-1.20, P = 0.27, respectively). After adjustment, there were no significant differences in the incidence of the most frequent non-AIDS events, cardiovascular events and non-AIDS malignancies, according to 25(OH)D tertile (data not shown).
Deaths were stratified according to cause. Of 188 deaths, 160 were of known cause, of which 48 (30.0%) were judged to be AIDS-related and 112 (70.0%) non-AIDS related. Trends of IRR for both causes of deaths across 25(OH)D strata were consistent with those in Fig. 2, although the lower number of events reduced the power to detect differences. Persons in the middle and higher strata had a nonsignificantly reduced incidence of AIDS-related death (IRR 0.53, 95% CI 0.24-1.15, P = 0.11, and IRR 0.61, 95% CI 0.28-1.32, P = 0.21, respectively). The incidence of non-AIDS-related death was nonsignificantly reduced for persons with 25(OH)D values in the middle tertile (IRR 0.67, 95% CI 0.41-1.09, P = 0.10), but significantly reduced for those with values in the higher tertile (IRR 0.60, 95% CI 0.37-0.98, P = 0.043).
There was no evidence of an interaction between race and 25(OH)D tertile and any of the clinical events (P>0.1), although this analysis had limited power because of the small number of persons of non-white ethnic origin. Furthermore, there was no interaction between 25(OH)D level and HIV RNA viral load or use of cART for the prediction of AIDS, non-AIDS events or death. Further, adjusting for current RNA or CD4 value (i.e. taking account of measurements made during follow-up) did not significantly alter our findings. For example, for death, after additionally adjusting for CD4 and viral load as time-updated covariates, there was a significantly reduced incidence of death in both the middle (IRR 0.65, 95% CI 0.44-0.95, P = 0.025) and higher 25(OH)D tertiles (IRR 0.67, 95% CI 0.45-1.00, P = 0.048). Adjusting for cumulative time on treatment in the 1650 persons on cART at baseline did not significantly impact the analysis (data not shown).
Additional analyses
Hemoglobin was measured within 6 months of baseline in 1262 persons. There was very little correlation between 25(OH)D value and hemoglobin (correlation coefficient 0.087, P = 0.0020), across 25(OH)D tertiles. A multivariate model adjusting additionally for baseline or current anemia status showed consistent results to those in Fig. 2 (data not shown). In addition, serum creatinine values within 6 months of baseline were available for 355 persons. There was no correlation between estimated glomerular filtration rate (eGFR) and 25(OH)D (correlation coefficient -0.091, P = 0.086), and no patient had chronic kidney disease (two consecutive eGFR <60 ml/min/1.73 m2) at or before 25(OH)D evaluation.
Body mass index (BMI) was available for 1518 persons within 6 months of baseline, and there was no correlation with 25(OH)D levels (correlation coefficient 0.016, P = 0.53). After adjustment, compared to those with medium BMI, persons with a low BMI (≤18) had marginally increased odds of having a 25(OH)D level in the lowest tertile (aOR 1.71; 95% CI 0.92-3.17, P = 0.091), and persons with a high BMI (>25) had similar odds of having a 25(OH)D level in the lowest tertile (aOR 1.02; 95% CI 0.79-1.32, P = 0.88).
Discussion
The study examined 25(OH)D levels in a large population of well characterized HIV-infected persons. The objectives were to describe the prevalence and correlates of vitamin D deficiency, and to determine whether a 25(OH)D level was associated with subsequent all-cause death, AIDS-defining or non-AIDS events.
The present results confirm [15-19] that vitamin D insufficiency or deficiency is frequent in HIV-infected persons. Studies across Europe have found highly variable 25(OH)D levels in the general population, with levels below 10 ng/ml in 2-30% of adults [26]. In the US, 25(OH)D levels below 30 ng/ml and below 20 ng/ml were similarly prevalent in the SUN cohort of HIV-infected persons vs. in the general population (NHANES study), after adjustment for age, sex and race: 70.3 vs. 79.1% and 29.7 vs. 38.8%, respectively [27]. Therefore, vitamin D deficiency might not be more frequent in people living with HIV than in the general population. Unsurprisingly, the season of sampling was associated with 25(OH)D levels, and persons with less, or benefiting less from, sun exposure (living outside Southern Europe and Argentina, or of black ethnic origin) were more likely to have very low 25(OH)D levels. Nonhomosexual route of HIV-1 transmission was also associated with lower 25(OH)D levels. This is in agreement with an earlier study suggesting that intravenous drug use is a risk factor for low 25(OH)D [18]. Of note, the association with heterosexual transmission remained after adjustment for ethnicity. These differences across transmission groups may reflect other factors we were not able to adjust for, such as nutrition and sun exposure. As in the general population, increasing age was also associated with lower 25(OH)D levels. Exposure to cART [19] and the use of efavirenz [17,18,28,29] have been associated with lower 25(OH)D levels, but we were unable to reproduce this finding in the present study. Neither the use of NRTIs or protease inhibitors was associated with lower levels of 25(OH)D. The finding that protease inhibitor use was associated with higher 25(OH)D levels, as also shown in the SUN cohort [27], is of unclear biological relevance.
Having a 25(OH)D level in the lowest tertile (<12 ng/ml) was strongly associated with the occurrence of AIDS events and all-cause mortality over a median follow-up of 5 years, after adjusting for a large number of variables, including season, ethnic and geographic origin, CD4 cell count and viral load (at baseline and time-adjusted), and there was no interaction between these variables and 25(OH)D levels for the association with events. Thus, a very low 25(OH)D level was associated with events, even in the case of virologically controlled HIV infection and immune restoration. The prognostic value of vitamin D deficiency was also independent of anemia, another prognostic factor of HIV infection [30], that has been associated with low 25(OH)D levels in untreated HIV-infected African women [20]. In addition, although we only had information on a subgroup of persons, low 25(OH)D levels were not correlated with low eGFR, an indicator of cardiovascular risk in the general and HIV-infected populations [31], associated with all-cause and cardiovascular mortality [32].
We thus found a strong association between the lowest 25(OH)D tertile (<12 ng/ml) and the two main endpoints of all-cause death and AIDS events, but no differences between the medium and high 25(OH)D tertiles. Vitamin D deficiency therefore represents a new, independent, unfavorable prognostic marker in HIV infection, but without further research this cannot translate into clinical recommendations.
The association of low 25(OH)D with clinical progression in HIV infection echoes the multiple functions now attributed to vitamin D, which, like other steroid hormones, can regulate gene activity in numerous cell types [2,5,6]. Vitamin D deficiency has been associated with cancer risk [1,2,5], with a large study suggesting that vitamin D supplementation could lower the risk of malignancies [33]. Studies have also associated vitamin D insufficiency with hypertension, insulin resistance and high BMI, thus establishing a link with cardiovascular risk [2], clinical cardiovascular disease [34,35], and cardiovascular and all-cause death [10,14]. Vitamin D regulates endothelial function [36], down-regulates the renin gene activity [37], and a placebo-controlled trial has shown that vitamin D administration lowers blood pressure in hypertensive persons [38]. Vitamin D has regulatory effects on immune functions [39]: the vitamin D receptor is expressed on adaptive and innate immune cells; vitamin D plays a role in innate immunity, particularly in the defence against mycobacteria, and in T-cell activation, with vitamin D insufficiency leading to altered T-cell proliferation [40]. Vitamin D deficiency has also been associated with inflammation [41]. In a study of patients referred for a coronary angiography, vitamin D deficiency was associated with all-cause and cardiovascular mortality, and correlated with high levels of markers of cell adhesion, oxidative stress and inflammation [C-reactive protein (CRP) and interleukin 6 (IL-6)] [10]. In a placebo-controlled study, vitamin D supplementation lowered the level of the inflammatory cytokine TNF-α and increased the level of the anti-inflammatory cytokine IL-10, in patients with congestive heart failure [42]. In a large cohort, low vitamin D levels have also been associated with increased levels of the coagulation activation markers tissue plasminogen activator and D-dimer [43].
Taken together, these results are of particular relevance to our findings, because of the well described increased vascular risk in HIV-infected persons, and because markers of inflammation (CRP and IL-6) and coagulation activation (D-dimers) have been associated with mortality and opportunistic diseases in HIV-infected persons [44,45]. In the present study, whereas non-AIDS events were not significantly associated with low 25(OH)D (possibly due to a lack of power), non-AIDS deaths were: this could reflect an aggravating effect of vitamin D deficiency, possibly linked with immune dysfunction or increased inflammation, on different conditions.
The study has several limitations. The study population may not be representative of the whole cohort, with excluded persons being older and having been sampled during spring less often: however, this would only attenuate the effect of 25(OH)D level on events. There were also a relatively low number of non-AIDS events: the impact of vitamin D on these events should therefore be addressed in larger studies. Information on vitamin D supplementation is not collected in EuroSIDA, but it is unlikely that HIV-infected persons sampled around 2001 were widely receiving vitamin D supplementation, and persons who were included earlier in the cohort had higher 25(OH)D levels, as observed in the general population [46]. This study was based on a single 25(OH)D measurement, at entry into the cohort, which leaves open the questions whether the length of exposure to low 25(OH)D, or its level at a given time point, is the key factor, and whether 25(OD) level has a short or mid-term, rather than long-term, prognostic value. Lastly, this is an observational study, from which causal relations cannot be drawn. However, the association between vitamin D deficiency and immune dysfunction, inflammation and coagulation activation may suggest a coherent, testable, pathogenic link between low 25(OH)D levels and events in HIV-infected persons.
These results provide strong evidence that vitamin D deficiency is an important cofactor in HIV disease progression, even in the setting of widespread, efficient cART. Whether the relationship between vitamin D deficiency and events is causal must now be addressed, because of potential public health consequences.
| |
| |
| |
|
|
|
