HIV increases markers of cardiovascular risk: results from a randomized, treatment interruption trial
[from Jules: A key question is how preHAART CD4 and viral load affect inflammation markers AND the of HIV on inflammation markers effects after achieving high CD4s and undetectable viral load, and my understanding is successful HAART does not necessarily normalize inflammation markers. It appears and several recent published research studies show HIV increases inflammation markers, also that viral replication increases markers, and that despite undetectable HIV RNA inflammation markers may not be normalized. The key question is does this contribute to premature aging in HIV+ individuals. Although aging research is not well developed and characterized published literature suggests inflammation in HIV-negatives is considered a cause for accelerated aging.
"In the present study, we show that HIV-RNA replication is associated with increased levels of s-VCAM-1 and CCL2 and decreased levels of adiponectin and IL-10.....After 1 month of cART interruption in SMART, both IL-6 and D-dimer levels rose significantly but were stable in those who continued cART. Moreover, higher levels of IL-6 and D-dimer were associated with a large increase in the likelihood of death.....In the general population, the use of rosuvastatin in apparently healthy individuals with elevated high-sensitivity CRP levels reduced the incidence of major cardiovascular events. Intervention strategies should be studied for HIV-infected patients with high plasma levels of inflammatory markers and with ongoing HIV-RNA replication....mean values and their 95% CI by HIV-RNA strata for s-VCAM-1, CCL2, adiponectin and IL-10 across three strata of HIV-RNA (≤50, 50-10 000 and >10 000 copies/ml); s-VCAM-1 and CCL2 increased as HIV-RNA increased, and adiponectin and IL-10 plasma concentrations decreased as HIV-RNA increased. All these changes are highly statistically significant (P trend = 0.001, 0.003, 0.004 and 0.001, respectively)."
HIV increases markers of cardiovascular risk: results from a randomized, treatment interruption trial - (05/13/09)
Long-term cART increases the risk of acute cardiovascular events in HIV-infected adults . Continuous therapy (for years) with a protease inhibitor, but not with a nonnucleoside reverse transcriptase inhibitor, has been associated with an increased risk of MI. This is probably due in part to the dyslipidaemic effects of protease inhibitor therapy . On the contrary, a similar increased incidence of acute cardiovascular events occurs during interruption of antiretroviral treatment . This suggests that HIV itself increases the cardiovascular risk. The present study investigates the possible correlation between HIV replication and quantitative values of cardiovascular risk markers after interruption of cART.
In the present study, we show that HIV-RNA replication is associated with increased levels of s-VCAM-1 and CCL2 and decreased levels of adiponectin and IL-10. These strong associations persisted after adjustment for known cardiovascular risk factors. No significant correlation was observed between patients treated and those who stopped cART at week 12 of the trial for the other cardiovascular markers such as CRP, leptin, CCL3, IL-6, D-dimer and GM-CSF. All values reported in the present study were found within the normal ranges previously published among HIV-infected  and HIV-uninfected patients [31-37,23,38], except for s-VCAM-1, which was found to be as high among our HIV-infected study population as in patients with type II diabetes or end-stage renal failure [39,40]. Our work points to possible molecular mechanisms of the proatherosclerotic effect of HIV-RNA in the cascade leading to a cardiovascular event.
The biomarkers investigated in the present study are involved in endothelial activation (s-VCAM-1 and P-selectin), systemic inflammation (CCL2, CCL3, GM-CSF, CRP, IL-6 and IL-10), platelet (P-selectin), adipocyte (adiponectin and leptin) and the coagulation cascade (D-dimer).
Several studies have suggested a direct action of different HIV components in the modulation of inflammatory processes independent of CD4 cell counts. Virus proteins, such as Tat and Nef, influence monocyte functions (i.e. cytokine and chemokine production) and survival [41,42]. On the contrary, HIV induces endothelial activation [43,44]. Furthermore, HIV also interferes with adipose tissue homeostasis independent of antiretroviral drug-induced lipodystrophy. Hypertriglyceridaemia, associated with the wasting syndrome, occurs in advanced HIV disease .
Less is known about a possible direct effect of HIV in the coagulation cascade. Recent investigations in the Strategies for Management of AntiRetroviral Therapy (SMART) Trial  suggest that the coagulation product, D-dimer, is strongly related to all-cause mortality in HIV patients. HIV-induced dysregulation of the complex cross-talk between leukocytes, endothelial cells and adipose tissue promotes atherosclerosis up to and including acute ischaemic complications. The modifications of the soluble markers under investigation in our study (s-VCAM-1, CCL2, adiponectin and IL-10) may be key active players in the HIV-induced increase in cardiovascular risk.
The significant correlation between HIV-RNA and s-VCAM-1, CCL2, adiponectin or IL-10 persisted after adjustment for known confounders (lipids, glucose, age, etc.). We did not adjust our multiple regression models on treatment use because the correlation between HIV-RNA and treatment use was above 0.70, indicating multicollinearity. We also verified that treatment use was not an effect modifier in the association between the cardiovascular biomarker studied and HIV-RNA.
Inflammatory cells and soluble mediators, which play a central role in all phases of atherosclerotic inflammation, have been investigated to assess their role as independent cardiovascular risk factors [11,12,14,16,18-21,23,25,26]. Although their clinical use is still controversial and not recommended because of the low specificity and high cost , previous studies clearly show a strong correlation between these markers and the incidence of acute cardiovascular events [32,35,38]. On the basis of these premises, together with the absence of acute cardiovascular events reported at the end of STACCATO trial, the present study only focused on the association between cardiovascular biomarkers and HIV-RNA replication but not on the risk of cardiovascular events in HIV patients.
Among the biomarkers, CRP is also implicated in several conditions not related to atherosclerosis . We observed normal levels (about 1 mg/l) independent of cART use. The absence of CRP modification suggests that our STACCATO study population was not affected by clinical infections or other systemic inflammatory states, which might modify the other tested cardiovascular biomarkers .
CCL3 is a selective ligand for CCR5, the HIV coreceptor, and therefore, a potential HIV antagonist, but we did not observe any correlation between CCL3 and HIV-RNA . These data might be explained by a direct virus interference with CCL3 plasma levels, which could thus limit the use of CCL3 as a possible cardiovascular risk marker in HIV-positive patients.
D-dimers decreased after a median 8 months of cART and correlated with HIV-RNA, but we did not show any significant association with HIV-RNA replication in the adjusted analysis at week 12 for this marker. Our results can be compared to those of the SMART trial . In the SMART analysis , six biomarkers were assessed at study entry and 1 month following randomization in the interruption and continuous treatment arm, respectively. After 1 month of cART interruption in SMART, both IL-6 and D-dimer levels rose significantly with a median D-dimer increase of 0.05 µg/ml after 1 month off ART in the interruption group. In STACCATO, we also showed a significant increase in D-dimer off cART of a level similar to what was observed in SMART (+0.07 µg/ml, P = 0.01).
We did not report a similar finding for IL-6. However, we used a method with a detection limit of 0.0156 pg/ml, and we documented more than 75% of the patients with a value below our detection limit. With a more sensitive method, we might have found results similar to those reported in SMART.
After 1 month of cART interruption in SMART, both IL-6 and D-dimer levels rose significantly but were stable in those who continued cART. Moreover, higher levels of IL-6 and D-dimer were associated with a large increase in the likelihood of death. As noted, only two deaths and no MI occurred in STACCATO, precluding evaluation of cardiovascular events or deaths. We also had a majority of women, which may make comparison with the mostly male population in SMART difficult.
HIV could directly induce monocyte/macrophage, endothelial and adipocyte activation. These cells can increase atherosclerotic inflammation by upregulating the secretion of inflammatory molecules (CCL2) and downregulating anti-inflammatory factors (adiponectin and IL-10). The altered profile of these soluble mediators further contributes to endothelial activation and atherosclerosis development. Injured endothelial cells release the adhesion molecule, s-VCAM-1, in the vessel lumen, which should be considered as a marker of endothelial activation. Therefore, the present study shows a possible direct effect of HIV replication on inflammatory processes underlying atherosclerosis.
Our relatively small and young population limited the power of our study to correlate the biomarker changes with clinical events. Therefore, our data should be considered as an assessment of cardiovascular risk without clinical endpoints but validated by previously published studies [32,35,38]. These limitations, together with the selection of an exclusively Thai population, reduce the generalizability of our findings.
In conclusion, our study associates HIV-RNA replication and modifications of specific cardiovascular risk biomarkers (s-VCAM-1, CCL2, adiponectin and IL-10) independent of known confounders. In the general population, the use of rosuvastatin in apparently healthy individuals with elevated high-sensitivity CRP levels reduced the incidence of major cardiovascular events. Intervention strategies should be studied for HIV-infected patients with high plasma levels of inflammatory markers and with ongoing HIV-RNA replication .
15 May 2009 - Volume 23 - Issue 8 - p 929-939
Calmy, Alexandra; Gayet-Ageron, Angele; Montecucco, Fabrizio; Nguyen, Alain; Mach, Francois; Burger, Fabienne; Ubolyam, Sasiwimol; Carr, Andrew; Ruxungtham, Kiat; Hirschel, Bernard; Ananworanich, Jintanat; on behalf of the STACCATO Study Group
aHIV Unit, Switzerland
bFoundation for Medical Research, Cardiology Division, Geneva University Hospital, Geneva, Switzerland
cHIV Netherlands Australia Thailand Research Collaboration, Bangkok, Thailand
dSt. Vincent's Hospital, Sydney, New South Wales, Australia
eDepartment of Medicine, Faculty of Medicine, Chulalongkorn University, Thailand
fSouth East Asia Research Collaboration with Hawaii, Bangkok, Thailand.
Objectives: Plasma soluble inflammatory molecules are associated with the risk of ischaemic cardiovascular events. We investigated whether HIV replication modified the levels of these proteins in a combination antiretroviral therapy (cART) interruption trial.
Method and results: In 145 HIV-infected Thai patients (62% women, median CD4 cell count 271 cells/µl, median plasma HIV-RNA 4.66 log10 copies/ml) included in the Swiss-Thai-Australia Treatment Interruption Trial (STACCATO) trial, leptin, adiponectin, C-reactive protein, soluble vascular cell adhesion molecule-1 (s-VCAM-1), P-selectin, chemokine ligand 2, chemokine ligand 3, interleukin (IL)-6, IL-10, granulocyte macrophage colony-stimulating factor and D-dimer were measured before cART was initiated, after cART had suppressed HIV replication to less than 50 copies/ml plasma (median 8 months) and again 12 weeks after randomization to continued cART (n = 48) or interrupted cART (n = 97). Multiple linear regression and logistic regression were used to investigate the association between each cardiovascular marker and plasma HIV-RNA.
Initiation of cART resulted in significant declines in s-VCAM-1, P-selectin, leptin and D-dimer, whereas mediators with anti-inflammatory properties, such as adiponectin and IL-10, increased. At 12 weeks after randomization, we found positive associations between levels of s-VCAM-1 and chemokine ligand 2 with an increase in plasma HIV-RNA (r = 0.271, P = 0.001 and r = 0.24, P = 0.005, respectively), whereas levels of adiponectin decreased for each 1 log increase in plasma HIV-RNA (r = -0.24, P = 0.002). Detectable IL-10 was less likely (odds ratio = 0.64, 95% confidence interval = 0.43-0.96) for each 1 log increase in plasma HIV-RNA.
Conclusion: Plasma levels of several inflammatory, anti-inflammatory and endothelial activation markers of cardiovascular disease are associated with HIV-RNA replication.
Combination antiretroviral therapy (cART) has dramatically reduced both mortality and morbidity for infections associated with HIV infection . Non-AIDS events, including acute cardiovascular disease, account for a substantial proportion of all deaths in patients starting cART with CD4+ lymphocyte counts above 200 cells/µl [2,3]. The metabolic effects of cART can be expected to increase the risk of cardiovascular events. The risk of myocardial infarction (MI) increases with longer exposure to cART , although traditional risk factors still remain the main determinants of cardiovascular risk in HIV-infected adults .
Paradoxically, intermittent, CD4 lymphocyte count-guided cART was associated with a 50% greater risk of MI relative to continuous cART . One possible cause is HIV-induced endothelial activation .
Although established cardiovascular risk factors are widely used to assess cardiovascular risk in the general population, novel serum markers of endothelial activation and inflammation may also be involved in the pathogenesis of atherosclerosis . Subclinical arterial inflammation is a prognostic factor in atherosclerosis . Twenty-four weeks of cART improved endothelial dysfunction as measured by brachial artery flow-mediated dilation (FMD) . The only variable associated with an improvement in FMD was the reduction in plasma HIV viral load.
We hypothesized that HIV replication could modify the serum levels of cardiovascular endothelial and inflammatory markers. We selected 12 biomarkers that have been associated with proatherosclerotic processes, including endothelial activation [soluble vascular cell adhesion molecule-1 (s-VCAM-1) and P-selectin], systemic inflammation [chemokine ligand 3 (CCL3), CCL2, granulocyte macrophage colony-stimulating factor (GM-CSF), C-reactive protein (CRP), interleukin-6 (IL-6) and IL-10], platelet (P-selectin) and coagulation cascade (D-dimer) activation (Table 1) [10-26]. Adipose tissue-derived hormones (such as leptin or adiponectin), both of which are correlated with cardiovascular diseases in epidemiological studies, have also been analysed [24-26].
We designed a post-hoc analysis comprising a subgroup of participants in the Swiss-Thai-Australia Treatment Interruption Trial (STACCATO) trial who had no prior ART exposure at entry and measured plasma levels of soluble mediators associated with cardiovascular risk before cART initiation, on cART when plasma HIV-RNA was undetectable and then in the randomized phase of the trial when patients either continued cART or suspended cART in order to test whether and how the values of the different cardiovascular markers correlated with HIV replication .
Of the 216 Thai patients (67.1%) enrolled into STACCATO, we included all 145 patients with available samples. We did not find any difference regarding demographic and clinical variables between the 145 Thai patients with available samples and the 216 Thai naive patients (data not shown). Baseline (pre-cART) demographic and biological characteristics of the study population are shown in Table 2. Ninety patients were women (62%). Mean age was 33.9 years (±SD 8.1), and only four (3%) patients were classified as CDC stage C. Median baseline CD4+ lymphocyte count was 271 cells/µl (IQR = 231-328), and median log10 HIV-RNA was 4.66 (IQR = 4.17-5.12). At randomization to cART continuation or interruption, patients had been treated for a median of 8.0 months (IQR = 6.4-12.4), and the median CD4+ lymphocyte count was 442 cells/µl (IQR = 392-531).
Changes in cardiovascular markers after combination antiretroviral therapy and their correlation with HIV-RNA
The values of each cardiovascular marker before cART initiation and on cART (i.e. at randomization) are presented in Table 3. We showed a significant decrease in the values of endothelial activation markers (s-VCAM-1 and P-selectin), inflammatory hormone (leptin) and fibrin degradation product (D-dimer) after a median 8 months of cART (P ≤ 0.001, P < 0.001, P = 0.03 and P = 0.005, respectively). CCL2 decrease was of borderline statistical significance (P = 0.06).
Adiponectin, an adipose tissue-derived hormone with anti-inflammatory properties, showed a significant increase after cART initiation (P = 0.05). Additionally, the proportion of anti-inflammatory cytokine, IL-10, values above the limit of detection was significantly lower on cART than before treatment (P = 0.001).
Pearson's correlation coefficient between each cardiovascular marker before cART initiation and log10 HIV-RNA was not significant, except for D-dimer (P = 0.03) and CRP (P = 0.04) for which there was a slightly significant positive correlation with log10 HIV-RNA before treatment was initiated. Between randomization and week 12, we observed a significant increase in s-VCAM-1, CCL-2, P-selectin and D-Dimer (P < 0.001, P = 0.007, P = 0.01, P = 0.01, respectively) and a significant increase in adiponectin (P = 0.009) (data not shown).
Association of HIV-RNA with cardiovascular markers 12 weeks after randomization
Of the 145 patients included, 48 (33.3%) were randomized to the continued cART arm and 97 (66.7%) to the STI arm (cART interruption). Table 4 presents the median values and IQR for each cardiovascular marker (s-VCAM-1, adiponectin, CCL2, CCL3, P-selectin, leptin, CRP and D-dimer) 12 weeks after randomization and the proportion of patients with GM-CSF, IL-6 and IL-10 above the limit of detection 12 weeks after randomization for the STI and continued cART groups.
s-VCAM-1 and CCL2 were positively associated with each 1 log increase in HIV-RNA (r = 0.271, P = 0.001 and r = 0.238, P = 0.005, respectively); this association remains after adjustment for age, platelet, LDL cholesterol and white blood count for s-VCAM-1 and after adjustment for HIV route of infection and TC for CCL2. Adiponectin was negatively associated with each 1 log increase in HIV-RNA (r = -0.248, P = 0.002) after adjustment for sex and triglycerides. Detectable IL-10 was significantly less likely with each 1 log increase in HIV-RNA after adjustment for sex and route of HIV infection. We assessed the colinearity between cART and HIV-RNA and verified that cART was not an effect modifier in the relation between each cardiovascular marker and HIV-RNA.
Figure 1 shows mean values and their 95% CI by HIV-RNA strata for s-VCAM-1, CCL2, adiponectin and IL-10 across three strata of HIV-RNA (≤50, 50-10 000 and >10 000 copies/ml); s-VCAM-1 and CCL2 increased as HIV-RNA increased, and adiponectin and IL-10 plasma concentrations decreased as HIV-RNA increased. All these changes are highly statistically significant (P trend = 0.001, 0.003, 0.004 and 0.001, respectively).
Markers plasma concentration at study termination (interrupted combination antiretroviral therapy arm)
Patients randomized to the STI arm had a median of two cycles of treatment interruption during STACCATO trial. We compared the values of each cardiovascular biomarker between randomization (when patients had experienced a median 8 months of cART) and study termination, after a mean (SD) time of 20.4 (15.8) weeks of cART re-treatment. We found that s-VCAM-1, CCL2 and P-selectin were significantly higher at study termination (P < 0.001, P < 0.02, P < 0.04, respectively) than after randomization. Similarly, the protective factor, adiponectin, was significantly lower at study termination than after randomization (P < 0.001), and the proportion of IL-10 under the limit of detection was significantly higher at study termination than after randomization (P = 0.04).
Cardiovascular risk in the study population
Regarding paired values of lipids and glucose, TC (P < 0.001), triglycerides (P = 0.004), HDL cholesterol (P < 0.001) and glucose (P = 0.02) were significantly increased after cART initiation. TC, HDL cholesterol and LDL cholesterol had a significant negative correlation with HIV-RNA before treatment (r = -0.19, P = 0.02; r = -0.29, P < 0.001; r = -0.17, P = 0.04, respectively). None of these metabolic markers was significantly correlated with HIV-RNA at week 12 (data not shown).
The Framingham equation was calculated before cART was initiated. We showed that 49% of patients had a low risk, 6.2% had a moderate risk and no patient had a high predictive risk of coronary heart disease (Table 2).
Participants and methods
The STACCATO trial  included 490 patients from Thailand, Australia and Switzerland. Patients were treated with cART for at least 6 months until plasma HIV-RNA was below 50 copies/ml and CD4+ lymphocyte count was more than 350 cells/µl when they were randomized 1: 2 to continue cART (continued cART arm) or interrupted cART [structured treatment interruption (STI) arm]. Participant cART comprised ritonavir-boosted saquinavir and stavudine/didanosine until March 2003, when stavudine/didanosine was replaced by tenofovir/lamivudine or tenofovir/emtricitabine. To be eligible for the present analysis, participants had to be ART-naive and have frozen serum samples available from before initiation of cART, at randomization when all patients were on cART and 12 weeks after randomization to continued or interrupted cART. One hundred and forty-five patients, all from seven STACCATO sites in Thailand, were included in the present analysis (48 patients in the continued cART arm and 97 in the STI arm).
All participants were assessed on three occasions: before initiation of cART, when virologically suppressed at randomization and 12 weeks after randomization. Participants randomized in the STI arm also had samples analysed at the final visit in which they were re-treated with cART for 12 up to 24 weeks.
Real-time whole blood and ethylenediaminetetraacetic acid (EDTA) plasma were collected for measuring fasting total cholesterol (TC), triglycerides, glucose, CD4+ lymphocyte count and HIV-RNA. Plasma for storage was centrifuged at 800g for 10 min within 6 h of collection and then stored at -70°C for measurement of leptin, adiponectin, CRP, s-VCAM-1, P-selectin, CCL2, CCL3, IL-6, IL-10, GM-CSF, D-dimer and high-density lipoprotein (HDL) cholesterol.
Coronary risk profile was calculated using Framingham equation before the first antiretroviral treatment . Using the Framingham equation, patients were categorized at high (>20% risk or patients with a history of cardiovascular disease), moderate (10-20% risk), low (<10% risk) 10-year risk of coronary heart disease or an unknown 10-year risk when the predicted risk of coronary heart disease could not be calculated because of missing values.
Cardiovascular risk marker level detection
Plasma leptin, adiponectin, s-VCAM-1, P-selectin, CCL2, GM-CSF, IL-6, IL-10 and CCL3 levels were measured by colorimetric enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, Minnesota, USA). EDTA plasma D-dimer levels were measured by enzyme-linked fluorescent assay (Roche Diagnostics Systems, Basel, Switzerland). EDTA plasma D-dimer values were validated by statistical correlation with sodium citrate plasma D-dimer values in normal individuals (data not shown). CRP was measured by high-sensitivity nephelometric latex immunoassay (Roche Diagnostic systems, Basel, Switzerland). The limits of detection for leptin was 31.25 pg/ml, 62.50 pg/ml for adiponectin, 6.25 ng/ml for s-VCAM-1, 125 pg/ml for P-selectin, 15.6 pg/ml for CCL2, 31.25 pg/ml for GM-CSF, 0.156 pg/ml for IL-6, 0.78 pg/ml for IL-10, 7.8 pg/ml for CCL3, 45 ng/ml for D-dimer and 0.11 µg/ml for CRP. Glucose, triglycerides, TC, low-density lipoprotein (LDL) cholesterol and HDL cholesterol were measured at fasting state and expressed in milligram per decalitre.
Patient characteristics were described before treatment. Participants included in the present analysis were compared for demographical and clinical variables with the entire naive Thai population of STACCATO using χ2 test or Fisher's exact test for categorical variables and Student's t-test for means of continuous variables. When we compared cardiovascular markers between two time periods, we used paired Student's t-test or the Wilcoxon signed-rank test for continuous variables.
For most cardiovascular markers, results were expressed as medians [interquartile range (IQR)], except for GM-CSF, IL-10 and IL-6, each of which were presented as the proportion of patients with values above the limit of detection as normal distribution was not assessed for those three markers. We used a Wilcoxon signed-rank test or a McNemar's test to compare the values of each cardiovascular marker before and after cART and at randomization versus after final cART re-treatment.
Pearson's correlation coefficient was used to assess correlations between cardiovascular markers (s-VCAM-1, adiponectin, CCL2, CCL3, P-selectin, leptin, CRP and D-dimer) and log10 plasma HIV-RNA before cART and at week 12, as a linear relationship between those markers and log10 HIV-RNA before cART exists and normal distribution was also assessed.
The estimation of factors associated with changes in the levels of s-VCAM-1, adiponectin, CCL2, CCL3, P-selectin, leptin, CRP and D-dimer at week 12 was calculated using a multiple linear model adjusted on the main independent covariable, HIV-RNA, and on other well known demographical and clinical factors of HIV infection. We provide results of the unadjusted and adjusted ß slope for a 10-fold (1 log) difference in HIV-RNA and the associated P-value. Preliminary analyses were conducted to ensure no violation of the assumptions of normality, linearity, multicollinearity and homoscedasticity. We used a backward, stepwise procedure that examined all variables selected in univariable analysis with P < 0.25 and all clinically relevant interaction terms (between treatment use and log10 HIV-RNA and all interaction terms between log10 HIV-RNA and variables in the final model). First, all interaction terms with P > 0.05 were removed from the model. Then we eliminated the main covariates that were nonsignificant (P > 0.05) but kept in the model all potentially confounding variables. Residual analysis was used to assess the adequacy of the fitted, multiple, linear model.
We present the mean value of s-VCAM-1, adiponectin, CCL2 and IL-10 assorted with their 95% confidence interval (CI) by three HIV-RNA strata. We divided HIV-RNA into quartiles and defined three strata: undetectable HIV-RNA (<50 copies/ml or 25th percentile), detectable below the 75th percentile (50-10 000 copies/ml) or detectable above the 75th percentile (>10 000 copies/ml). A P-test for linearity of each mean marker across the three HIV-RNA strata was assessed using analysis of variance.
For GM-CSF, IL-10 and IL6, we used a logistic regression model to assess the risk of a value above the limit of detection for each 1 log difference of HIV-RNA at week 12. We controlled for demographical and clinical factors that are potentially associated with the assessed risk. We provided unadjusted and adjusted odds ratios with their 95% CIs. The Hosmer and Lemeshow goodness-of-fit test was interpreted as acceptable if the corresponding P-value of the Pearson's χ2 statistic with 8 DF was more than 0.05.
Regression models were adjusted on the following factors: as continuous variable, age, Centers for Disease Control and Prevention (CDC) stage at baseline (A stage as the reference), risk factors for HIV infection (heterosexual sex as the reference), sex (male as the reference), weight, white blood cell count, platelet count, CD4+ lymphocyte count, TC, glucose, LDL cholesterol, triglycerides and TC/HDL cholesterol ratio. All analyses were done with Statistical Package for the Social Sciences, version 11.0 (SPSS Inc., Chicago, Illinois, USA).