Treatment of HIV infection may be associated with dyslipidemia [1,2]. Not only specific drug factors may account but also concomitant conditions may play an additional role. The extent of host genetics on the risk of developing antiretroviral therapy (ART)-associated dyslipidemia is poorly known.
In HIV-infected patients, abnormalities of lipid metabolism were described before the use of ART [3,4]. An early decrease in high-density lipoprotein cholesterol (HDL-C) and a later decrease in low-density lipoprotein cholesterol (LDL-C), followed by an increase in triglyceride levels were observed with the progression of HIV infection . Some studies reported the consistent finding that patients with advanced HIV infection or AIDS had higher levels of plasma triglyceride and lower levels of plasma HDL-C [4,5]. The decrease of HDL-C likely increases the risk of atherosclerosis more than the increase of LDL-C does. The effect is amplified because of the fact that HDL-C does not function optimally in the setting of infection or inflammation . Certainly, numerous studies [7–10] have reported dyslipidemia and an increased risk of cardiovascular disease (CVD) in HIV-infected persons receiving ART. The DAD study (Data collection on Adverse events of anti-HIV Drugs) reported an increase of 32% in the relative risk for CVD over the 5 following years since the initiation of ART, risk that appeared to be greater with the exposure of protease inhibitor relative to NNRTI [7–10].
The severity of dyslipidemia can vary according to the ART regimen used. However, not all patients exposed to the same ART regimen develop dyslipidemia, despite having had the same ART exposure, similar demographic characteristics and being immunologically and virologically comparable. This discrepancy may be due, at least in part, to host genetic factors. The risk of dyslipidemia has been described in patients carrying unfavourable genotypes for genes encoding apolipoproteins such as apolipoprotein E, apolipoprotein C3 and apolipoprotein A. In fact, this was more plausible in those patients exposed to a ritonavir (RTV) containing treatment [11,12]. Furthermore, a genome-wide association study (GWAS) in normal population identified other polymorphisms associated with dyslipidemia in HIV-infected individuals, suggesting that genetic information should be taken into account in addition to the dyslipidemic effects of ART . The main objective of this study was to assess the impact of 192 single nucleotide polymorphisms (SNPs), previously identified in six GWAS done in general population (Table_S1_supplementary, http://links.lww.com/QAD/A291), on plasma lipid levels in a cohort of 750 antiretroviral-HIV patients starting ART including a protease inhibitor or a nonnucleoside reverse transcriptase inhibitor (NNRTI). A second aim was to assess relevant nongenetic factors contributing to dyslipidemia when adjusting for genetic factors.
Materials and methods
Study design, patients and genetic variants
We studied an HIV-infected cohort prospectively followed-up at Hospital Clínic of Barcelona between 1 January 1997 and 31 December 2008.
The study population consisted of 750 naive HIV-infected patients (age between 18 and 75 years; 80% men) randomly selected who fulfill inclusion criteria of at least 1 year of follow-up without ART changes and no AIDS criteria. This project was approved by the ethics committees of our hospital. As part of the routine follow-up, these participants had measurements of serum levels of total cholesterol (T-Chol), LDL-C, HDL-C and triglyceride at baseline and after 1 month and each 3–6 months after ART initiation. All starting and stopping dates of each antiretroviral regimen and all cardiovascular medication were recorded in the database. No lipid-lowering agents or drugs that could have an influence on lipid levels were administered during the first year of treatment. Figure 1 provides the flow chart leading to the 750 DNA samples that were genotyped. In this study, 192 genetic variants (SNPs) in 87 genes from the lipid metabolism pathway were assessed. Participants gave written informed consent for genetic testing. All tested SNPs had significant genome-wide associations with serum lipid levels in previously published GWAS performed in the general population [13–18] with minor allele frequencies (MAFs) greater than 0.01 (Table_S1, http://links.lww.com/QAD/A291). Of all these variants, 26% were previously associated with T-Chol, 15% with LDL-C, 36% with HDL-C and 23% with triglyceride levels. DNA was isolated following the manufacturer's instructions with the QIAamp DNA Blood Mini Kit (QIAcube-QIAGEN; Qiagen Sciences Inc., Germantown, Maryland, USA). Genotyping was performed using the Illumina Golden Gate platform (Universitat Autònoma Barcelona, Spain). Duplicates from 10% of the samples were included either within or between plates as a genotyping quality control. The resulting genotypes were filtered prior to statistical analysis : individuals with genotyping call rate less than 80% (n = 23 samples) and SNPs with genotyping call rate less than 95% (n = 11 SNPs) were excluded. Ultimately, 727 patients and 181 SNPs in 83 genes were included in this study (Fig. 1). Because the clinical relevance of prolonged dyslipidemia is well documented by genetic causes of dyslipidemia [15,16], we included all lipid measurements available for each participant and calculated the observed proportions of participants with sustained dyslipidemia during the study period. The influence of ART and lipid-lowering agents, administered after the first year of ART initiation, on lipid levels was assumed to be rapid and reversible .
Dyslipidemia, as in previous reports [12,18], was defined according to the cut-off level defined by the National Cholesterol Education Program Third Adult Treatment Program (NCEP-ATPIII). Briefly, the cut-off lipid values considered were T-Chol greater than 6.2 mmol/l (240 mg/dl); LDL-C greater than 3.4 mmol/l (130 mg/dl); HDL-C less than 1.03 mmol/l (40 mg/dl); triglyceride greater than 2.26 mmol/l (200 mg/dl). Those individuals with an increase of T-Chol/LDL-C/triglyceride or a decrease of HDL-C levels with respect to the cut-off level during the first year of follow-up once ART started were considered cases.
Data were analysed longitudinally by modelling the individual effects of the different covariates on plasma lipid fraction (T-Chol, LDL-C, HDL-C and triglyceride) levels. Time-dependent covariates considered included age, CD4+ T-cells and HIV viral load. Fixed covariates included sex, ethnicity, hepatitis C virus (HCV) coinfection, first-line ART regimen either with a protease inhibitor or a nonnucleoside reverse transcriptase (NNRTI) and the genetic factors assessed. Medians were used to show central tendencies and interquartile ranges were calculated as measures of variability and statistical dispersion.
The study population was initially checked for Hardy–Weinberg equilibrium (P > 0.001) at all the tested SNPs and separately evaluated for different lipid fractions. PLINK1.07 was used for all association and statistical analyses . Study participants were treated as cases whenever an abnormal lipid value existed at 1 year of ART exposure. Stratified analysis controlling for first-line ART was performed using the Cochran–Mantel–Haenszel (CMH) test. Then, candidate SNPs were evaluated in a multivariate analysis on the basis of a logistic regression adjusted for all genetic and nongenetic variables for each lipid fraction. For the additive effects of SNPs, the direction of the regression represented the favourable or unfavourable effect of each extra minor allele. The power/effect of each variable was assessed using the odds ratio (OR) and the variance components calculated in the multivariate analysis. For multiple testing corrections, we used the Q-value R package  to apply a false discovery rate (FDR) of 10%, which corresponded to a significance threshold of a P value equal to 0.000318 for the LDL-C fraction and a P value equal to 0.002969 for HDL-C. For T-Chol, triglyceride and the sum of all lipid fractions, no significant thresholds could be calculated at 10% FDR.
Significant SNPs that overcame the 10% FDR correction were assessed in a Breslow–Day test of homogeneity of odd ratios, in order to check whether different antiretrovirals from the same family had different gene ORs. This was performed in both drug families: ritonavir-boosted protease inhibitor (PI/r) versus nonboosted protease inhibitor in the protease inhibitor group and efavirenz (EFV) versus nevirapine (NVP) in the NNRTI group. Statistically significant SNPs (P ≤ 0.05) for the Breslow–Day test were separately reassessed in a chi-square based basic association test in order to designate them as favourable or unfavourable with an OR less than 1 or an OR more than 1, respectively.
In order to check nonrandom association of alleles at two or more loci, linkage disequilibrium blocks were obtained from primary genotype data using Haploview [21,22]. The haplotype-based association study was restricted to genes including more than one associated genetic variant in the single-marker analysis. All the genotyped variants within these genes were considered. The best two-marker haplotype from all possible pairwise combinations was identified. Likewise, additional markers (up to three) were added in a stepwise manner to the initial two-SNP haplotype. Significance was estimated using 1000 permutations with the UNPHASED 3.2 software . As the expectation-maximization algorithm does not accurately estimate low haplotype frequencies , haplotypes with frequencies less than 0.05 were excluded.
Interaction analysis between SNPs from different genes was performed by comparing it to different regression models with a likelihood ratio test using the statistical package SPSS v15.0 (SPSS Inc., Chicago, Illinois, USA). In the first model, we used the abnormal serum lipid status as a dependent variable and the two risk alleles as predictive variables (affected status = a + bSNP1 + cSNP2). In the second model, we included the interaction between SNPs as an independent variable (Affected status = a + bSNP1 + cSNP2 + dSNP1*SNP2). For each gene, the inheritance model providing the best association results in the single-marker analysis was used. To test possible interactions between haplotypes from different genes, haplotype frequencies were estimated using PHASE software and then we followed the procedure described above.
Characteristics of the participants and lipid levels
Three hundred and eighty-one patients (52.4%) had a protease inhibitor and 346 (47.6%) an NNRTI prescribed in their first-line treatment.
Successful genotyping was accomplished for 727 out of 750 participants (96.9%). Their characteristics are summarized in Table 1. All of them were Spanish and whites. MAF values at all SNPs were similar to the ones previously described for ethnically equal populations (Table_S1, http://links.lww.com/QAD/A291) and all the tested SNPs were in Hardy–Weinberg equilibrium. During the study period, a median [interquartile range = upper quartile (Q3) – lower quartile (Q1)] of 8.5 (12), 5 (6), 7 (9) and 10 (13) measurements of T-Chol, LDL-C, HDL-C and triglyceride were determined in the protease inhibitor group and 9 (11), 6 (4.5), 8 (8) and 10 (13.75) in the NNRTI group.
Effect of antiretroviral regimen on plasma lipid levels
Before exposure to ART, all lipid fraction values were below the respective NCEP-ATPIII cut-off level. All lipid values except HDL-C increased once the ART treatment was initiated. At 1 year of starting treatment, 168 (44%), 193(50.6%), 283 (74.2%) and 258 (67.7%) patients from the protease inhibitor group showed abnormal values for T-Chol, LDL-C, HDL-C and triglyceride, respectively. In the NNRTI group, 151 (32.9%), 240 (52.2%), 333 (72.5%) and 223 (48.5%) patients showed abnormal values for T-Chol, LDL-C, HDL-C and triglyceride, respectively (Table 2). The increase of triglyceride levels after exposure to ART was higher in the protease inhibitor than in the NNRTI group. Nearly 67.7% of the patients presented an abnormal triglyceride level at 1 year of exposure to a protease inhibitor versus 48.5% of the patients exposed to an NNRTI. The percentage of individuals with abnormal lipid fraction levels was not significantly different between the protease inhibitor and the NNRTI groups (P = 0.26) (Table 2).
Contribution of single nucleotide polymorphisms to individual abnormal plasma lipid levels
A total of 51 SNPs were significantly associated (P < 0.05) with a lipid fraction in the univariate analysis stratified by ART (Table_S2, http://links.lww.com/QAD/A291). From all these variants, 40 SNPs remained significant (P < 0.05) in the logistic regression model-based multivariate analysis (Table_S3, http://links.lww.com/QAD/A291). In the final multi-SNP model adjusted for all genetic and nongenetic covariates, only 15 variants with a significance of P value less than 0.01 were taken into account (Table 3). Ultimately, one and six SNPs displayed a significant association with LDL-C and HDL-C respectively, after 10% FDR correction for multiple testing.
A total of nine SNPs (rs515135, rs541041, rs312985, rs506585, rs1260326, rs780094, rs904743, rs471364, rs1883025) in four genes (APOB, GCKR, TTC39B and ABCA1) maintained association (P < 0.05) with T-Chol levels in the logistic regression analysis (Table_S3, http://links.lww.com/QAD/A291).
In the final multi-SNP model, one SNP, rs471364 (TTC39B; P = 0.002927; OR, 0.515), contributed significantly to a decrease of plasma T-Chol levels (Fig. 2, Table 3, Table_S3, http://links.lww.com/QAD/A291).
Eleven SNPs (rs11206510, rs10198175, rs10495712, rs6754295, rs7557067, rs673548, rs754524, rs754523, rs2240466, rs714052, rs2967605) in four genes (PCSK9, APOB, BAZ1B and RAB11B) maintained association (P < 0.05) with LDL-C levels in the logistic regression analysis (Table_S3, http://links.lww.com/QAD/A291).
In the final multi-SNP model, three SNPs showed an unfavourable effect whereas one SNP showed a favourable effect. The SNPs that contributed significantly to an increase of plasma LDL-C levels were rs10495712 (APOB; P = 0.000318; OR, 2.013), rs754524 (APOB; P = 0.001263; OR, 1.864) and rs2240466 (BAZ1B; P = 0.009008; OR, 2.352). However, one SNP, rs6754295 (APOB; P = 0.009661; OR, 0.62) contributed favourably to plasma LDL-C levels (Fig. 2, Table 3; Table_S3, http://links.lww.com/QAD/A291). Despite the high degree of linkage disequilibrium (r2 > 0.9) between the studied SNPs in APOB, haplotype analysis did not reveal any significant association. Only SNP rs10495712 in APOB with an unfavourable effect (OR, 2.013) on LDL-C remained significant after 10% FDR correction.
A total of 13 SNPs (rs10487506, rs4149313, rs7903146, rs1800588, rs473224, rs261336, rs247615, rs12448528, rs173539, rs3764261, rs711752, rs708272 and rs7205804) in six genes (LEP, ABCA1, TCF7L2, LIPC, HERPUD1 and CETP) showed association (P < 0.05) with HDL-C in the logistic regression analysis (Table_S3, http://links.lww.com/QAD/A291).
In the multi-SNP model, six SNPs in three genes showed a favourable effect on HDL-C. SNPs rs4149313 (ABCA1; P = 0.000297; OR, 0.516), rs1800588 (LIPC; P = 0.002138; OR, 0.608), rs473224 (LIPC; P = 0.000307; OR, 0.483), rs261336 (LIPC; P = 0.002239; OR, 0.572), rs173539 (CETP; P = 0.002969; OR, 0.608), rs3764261 (CETP; P = 0.001521; OR, 0.588) contributed significantly to an increase in plasma HDL-C levels (Fig. 2, Table 3, Table_S3, http://links.lww.com/QAD/A291). The results remained significant after applying a 10% FDR correction. High degrees of linkage disequilibrium (r2 > 0.9) were found between the three SNPs in LIPC and also between the two SNPs in CETP. Haplotype analysis showed a protective haplotype (rs173539A/rs3764261A) in CETP (Table 4) associated with a favourable effect on HDL-C levels [P = 0.000167; OR, 0.638 (0.239–0.982)]. No associated haplotype was identified in LIPC. Interaction analysis was negative for all the possible gene pairs.
Eleven SNPs (rs1260326, rs780094, rs1919127, rs6544713, rs5908, rs264, rs4149313, rs1800682, rs10838852, rs174547 and rs174570) in nine genes (GCKR, C2orf16, ABCG8, HMGCR, LPL, ABCA1, FAS, OR4X1, FADS1 and FADS2) displayed association (P < 0,05) with triglyceride levels in the logistic regression analysis (Table _S3, http://links.lww.com/QAD/A291).
In the final multi-SNP model, three SNPs showed an unfavourable effect, whereas one SNP showed a favourable effect. The SNPs that contributed significantly to increase plasma triglyceride levels were rs1260326 (GCKR; P = 0.008063; OR, 1.361), rs780094 (GCKR; P = 0.003286; OR, 1.412) and rs174570 (FADS2; P = 0.005261; OR, 1.57). Conversely, rs4149313 (ABCA1; P = 0.00877; OR, 0.683) contributed significantly to decrease plasma triglyceride levels (Fig. 2, Table 3, Table_S3, http://links.lww.com/QAD/A291). The two SNPs in GCKR were having high linkage disequilibrium (r2 > 0.9). Haplotype analysis identified a risk haplotype in this gene (rs1260326A/rs780094A) that was associated with an unfavourable effect on triglyceride levels [P = 0.002276; OR, 1.391 (1.281–1.888)] (Table 4).
Contribution of nongenetic variables to plasma lipid levels (multivariate analysis)
Age was associated with plasma T-Chol [P = 1.05E−09; OR, 1.07 (1.05–1.09)], LDL-C levels [P = 0.00052; OR, 1.04 (1.02–1.07)] and HDL-C [P = 5E−06; OR, 0.95 (0.93–0.97)]. HCV coinfection showed a favourable association with T-Chol (P = 3.5E−06; OR, 0.31 (0.19–0.5)] and LDL-C [P = 0.00013; OR, 0.33 (0.2–0.58)]. However, a protease inhibitor containing ART regimen was unfavourably associated with plasma T-Chol [P = 0.009; OR, 1.6 (1.42–1.91)] and triglyceride [P = 0.00074; OR, 1.8 (1.39–1.99)] levels. First-line treatment with an NNRTI containing ART regimen [P = 8.9E–03; OR, 0.42 (0.19–0.5)] was favourably associated with plasma HDL-C levels. On the contrary, sex (male) showed an unfavourable association with both HDL-C [P = 2.3E–07; OR, 3.73 (1.16–1.44)] and triglyceride [P = 0.0034; OR, 1.81 (1.37–1.91)] (Fig. 2, Table_S3, http://links.lww.com/QAD/A291).
Subanalyses based on antiretroviral therapy drugs
Contribution of single nucleotide polymorphisms to plasma lipid levels depending on ritonavir-boosted protease inhibitor versus nonboosted protease inhibitor
From the global study, 72 individuals (18.9%) from the protease inhibitor group were treated with a PI/r and the other 309 (81.1%) were treated with a nonboosted protease inhibitor (Table 1). None of the SNPs showing a significant association (P < 0.01) in the global logistic regression model for T-Chol and triglycerides showed a significant association with an effect on these fractions depending on a nonboosted protease inhibitor versus PI/r in the Breslow–Daytest (data not shown).
Contribution of single nucleotide polymorphisms to plasma lipid levels depending on efavirenz versus nevirapine
From the global study, 227 individuals (65.8%) from the NNRTI group were treated with EFV and the other 118 (34%) were treated with NVP (Table 1). Three SNPs in LIPC [rs1800588; P = 0.0004198; OR, 0.39 (0.2337–0.6685)], rs473224 [P = 1.94E–05; OR, 0.27 (0.1485–0.5107)], rs261336 [P = 3.35E−05; OR, 0.31 (0.1789–0.5528)], which were associated with an increase of HDL-C in the global logistic regression, showed significant association with an increase of this fraction in those patients who started an EFV-containing NNRTI treatment. None of these SNPs were significant when a logistic regression analysis was performed for this subanalysis (data not shown).
In this longitudinal study of naive HIV individuals starting ART including one protease inhibitor or one NNRTI, SNPs that have consistently been associated with dyslipidemia in the general population GWAS [25–28] were validated. Moreover, previously published association studies in HIV population related to dyslipidemia have also been considered [11–13]. In this study, nongenetic factors have been taken into account in order to explain the interindividual variation of lipid levels once ART treatment with a protease inhibitor or an NNRTI has begun. We demonstrate that nongenetic covariates such as age, sex and HCV-coinfection are also playing an important role in these lipid variations. The additive logistic regression model showed the unfavourable effect of protease inhibitors on T-Chol and triglyceride levels. On the contrary, NNRTIs present a favourable effect on HDL-C levels. However, neither protease inhibitors nor NNRTIs apparently present a direct effect on LDL-C level variations. As in the general population, significant nongenetic traits such as sex (male) and age also influence lipid fraction levels . Surprisingly, a favourable association was found in our study between the presence of HCV coinfection and a decrease of T-Chol and LDL-C fractions. This could be explained by the ability of both HIV and HCV viruses, as many other infectious agents do, to impair cholesterol metabolism to satisfy their own cholesterol requirements at different stages of their life cycle . Furthermore, we assessed the influence of candidate genetic variables included in the logistic regression models performed for each lipid fraction. The only genetic variation found to be associated with T-Chol fraction is a polymorphism in the tetratricopeptide repeat domain 39B gene (TTC39B), which has previously been described in a general population GWAS  to be associated with HDL-C levels. This variation regulates the TTC39B expression. TTC39B seems, in general, to function in protein–protein interactions, but it actually has no known function in humans. This variant did not overcome10% FDR correction for multiple testing.
Regarding LDL-C levels, three polymorphisms in APOB (rs10495712, rs6754295, rs754524) and one in BAZ1B (rs2240466) reached significance in the final multi-SNP model. Only rs10495712 in APOB, with previously described association with cholesterol and triglyceride fractions in GWAS in the general population, presented a significant association after 10% FDR. Our results show the first evidence of association between the presence of this polymorphism and the increase of LDL-C levels in HIV patients with antiretroviral treatment.
The strongest association with triglyceride levels was found between a polymorphism in GCKR (rs780094), encoding the glucokinase regulatory protein, previously reported in a general population GWAS , and an increase of this fraction. This is the first evidence of association in treated HIV population between this SNP and an unfavourable effect on triglyceride levels in the presence of a protease inhibitor containing first-line antiretroviral regimen. This association should be validated in other HIV cohorts. There were SNPs in GCKR (rs1260326), in the fatty acid desaturase 2, FADS2 (rs174570), and in the ATP-binding cassette A, ABCA1,(rs4149313) that showed an association with triglyceride levels in the logistic regression model. An unfavourable effect was observed with an increase of triglyceride levels in the case of SNPs in GCKR and FADS2. In contrast, the effect was favourable in the case of rs4149313 in ABCA1 with a decrease of triglyceride levels. However, the association pattern of rs4149313 is contradictory with the association found in a previously reported multiple regression analysis , wherein the effect was described in the opposite direction as an unfavourable variant. However, we should consider that in our analysis, these SNPs did not reach significance after 10% FDR correction for multiple testing.
Finally, the logistic regression analysis for HDL-C fractions showed a significant association with six polymorphisms in the ABCA1, LIPC and CETP genes after 10% FDR correction. All these SNPs were associated with a favourable effect on the increase of HDL-C in the presence of an NNRTI-containing first-line treatment. Association between rs4149313 in ABCA1 and an increase of HDL-C levels confirmed previously reported results in HIV population [12,13]. Moreover, associations of rs173539 and rs3764261 in CETP showed the same effects as the ones previously published in general population GWAS [26,27] but are newly described in HIV population under ART. We describe a newly identified protective haplotype [A-A] between these two SNPs (P = 0.000167) in CETP. A previously reported association between rs1800775 in CETP and an increase of HDL-C levels in HIV population  was confirmed in our study only in the univariate analysis stratified by ART, and the significance was lost in the multivariate analysis. No gene interactions were found between any of the associated genes observed for each lipid fraction. Although SNPs in LIPC showed a favourable association with HDL-C when patients start an EFV-containing first-line ART regimen, these candidate polymorphisms lost association strength in the final logistic regression model performed for the NNRTI family subanalysis (EFV versus NVP). There was also a lack of association between candidate genes and protease inhibitor drug families (IP/r versus nonboosted protease inhibitor) regarding lipid variations.
In summary, our results show statistical evidence of specific polymorphisms, as well as nongenetic variables, being associated with an effect on lipid fraction levels depending on a protease inhibitor or an NNRTI containing first-line ART regimen. These variants should be validated in other cohorts and they could be used as biomarkers to identify patients at risk for dyslipidemia and to guide the rational selection of ART.
The strengths of this study in the field of HIV are its longitudinal design and the SNP selection based on previously reported GWA studies performed in the general population. Thus, our results allow validation of previously published results in genetic-dyslipidemia association studies in HIV population [12,13,28]. This is the first study of this kind in an HIV population with a first-line treatment, wherein complete lipid measurements of T-Chol, LDL-C, HDL-C and triglyceride have been considered. New models are required to confirm the variants implicated in ART-related dyslipidemia in HIV populations. In addition, functional studies are necessary to validate those already confirmed associations.
The authors thank the Retrovirology and Viral Immunopathology Laboratory of the IDIBAPS. Especially I Pérez for the statistical analysis advice and Dr E. Ros (Lipid Unit, Hospital Clínic of Barcelona) for his scientific support. A. Fernández improved the English of the text. The comments of the anonymous reviewers helped us to improve the manuscript and are greatly appreciated.
Conceived and designed the experiments: M.A., E.M., J.M.G. Performed the experiments: L.E., T.E. Analysed the data: M.A., L.E., B.C. Wrote the article: M.A., L.E., B.C., E.M.
This work was partially supported by grants from the Fondo de Investigación Sanitaria (FIS PI09/00396); Fondo Europeo para el Desarrollo Regional (FEDER); Fundación para la investigación y prevención del Sida en España (FIPSE 36715/08); Red de Investigación en Sida (RIS). Instituto de Salud Carlos III (ISCIII-RETIC RD06/0006/0000); Miguel Servet Program (M.A.06/0164) IDIBAPDS-ISCIII; (L.E.FI10/00174-ISCIII); AGAUR (SGR-0971).
These results were previously presented at 19th Retroconference (CROI, Seattle 2012) as an oral communication.
Conflicts of interest
J.M.G. and E.M. have been a consultant on advisory boards, have participated in speakers’ bureaus, have received research grants or have conducted clinical trials with Roche, Boehringer-Ingelheim, Abbott, BMS, GSK, Gilead, Janssen, Merck and Pfizer. M.A., L.E., B.C. and T.E. report no conflicts of interest relevant to this article. No other potential conflict of interest relevant to this article was reported.
1. Dube MP, Stein JH, Aberg JA, Fichtenbaum CJ, Gerber JG, Tashima KT, et al. Guidelines for the evaluation and management of dyslipidemia in human immunodeficiency virus (HIV)-infected adults receiving antiretroviral therapy: recommendations of the HIV Medical Association of the Infectious Disease Society of America and the Adult AIDS Clinical Trials Group. Clin Infect Dis 2003; 37:613–627.
2. Manuel O, Thiebaut R, Darioli R, Tarr PE. Treatment of dyslipidaemia in HIV-infected persons. Expert Opin Pharmacother 2005; 6:1619–1645.
3. Shor-Posner G, Basit A, Lu Y, Cabrejos C, Chang J, Fletcher M, et al. Hypocholesterolemia is associated with immune dysfunction in early human immunodeficiency virus-1 infection. Am J Med 1993; 94:515–519.
4. Grunfeld C, Kotler DP, Hamadeh R, Tierney A, Wang J, Pierson RN. Hypertriglyceridemia in the acquired immunodeficiency syndrome. Am J Med 1989; 86:27–31.
5. Grunfeld C, Pang M, Doerrler W, Shigenaga JK, Jensen P, Feingold KR. Lipids lipoproteins, triglyceride clearance, and cytokines in human immunodeficiency virus infection and the acquired immunodeficiency syndrome. J Clin Endocrinol Metab 1992; 74:1045–1052.
6. Grunfeld C. Dyslipidemia and its treatment in HIV infection. Top HIV Med 2010; 18:112–118.
7. Depairon M, Chessex S, Sudre P, Rodondi N, Doser N, Chave JP, et al. Premature atherosclerosis in HIV-infected individuals: focus on protease inhibitor therapy. AIDS 2001; 15:329–334.
8. Friis-Moller N, Sabin CA, Weber R, d’Arminio Monforte A, El-Sadr WM, Reiss P, et al. Combination antiretroviral therapy and the risk of myocardial infarction. N Engl J Med 2003; 349:1993–2003.
9. Friis-Moller N, Weber R, Reiss P, Thiebaut R, Kirk O, d’Arminio Monforte A, et al. Cardiovascular disease risk factors in HIV patients: association with antiretroviral therapy. Results from the DAD study. AIDS 2003; 17:1179–1193.
10. Group DADS, Friis-Moller N, Reiss P, Sabin CA, Weber R, Monforte A, et al. Class of antiretroviral drugs and the risk of myocardial infarction. N Engl J Med 2007; 356:1723–1735.
11. Tarr PE, Taffe P, Bleiber G, Furrer H, Rotger M, Martinez R, et al. Modeling the influence of APOC3, APOE, and TNF polymorphisms on the risk of antiretroviral therapy-associated lipid disorders. J Infect Dis 2005; 191:1419–1426.
12. Arnedo M, Taffe P, Sahli R, Furrer H, Hirschel B, Elzi L, et al. Contribution of 20 single nucleotide polymorphisms of 13 genes to dyslipidemia associated with antiretroviral therapy. Pharmacogenet Genomics 2007; 17:755–764.
13. Rotger M, Bayard C, Taffe P, Martinez R, Cavassini M, Bernasconi E, et al. Contribution of genome-wide significant single-nucleotide polymorphisms and antiretroviral therapy to dyslipidemia in HIV-infected individuals: a longitudinal study. Circ Cardiovasc Genet 2009; 2:621–628.
14. Tindall EA, Petersen DC, Nikolaysen S, Miller W, Schuster SC, Hayes VM. Interpretation of custom designed Illumina genotype cluster plots for targeted association studies and next-generation sequence validation. BMC Res Notes 2010; 3:39.
15. Cohen JC, Boerwinkle E, Mosley TH Jr, Hobbs HH. Sequence variations in PCSK9. low LDL. and protection against coronary heart disease. N Engl J Med 2006; 354:1264–1272.
16. Hoeg JM, Feuerstein IM, Tucker EE. Detection and quantitation of calcific atherosclerosis by ultrafast computed tomography in children and young adults with homozygous familial hypercholesterolemia. Arterioscler Thromb 1994; 14:1066–1074.
17. Martinez E, Leyes P, Ros E. Effectiveness of lipid-lowering therapy in HIV patients. Curr Opin HIV AIDS 2008; 3:240–246.
18. Periard D, Cavassini M, Taffe P, Chevalley M, Senn L, Chapuis-Taillard C, et al. High prevalence of peripheral arterial disease in HIV-infected persons. Clin Infect Dis 2008; 46:761–767.
19. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81:559–575.
20. Storey JD. A direct approach to false discovery rates. J Royal Stat Soc Series B 2002; 64 (Part 3):479–498.
21. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21:263–265.
22. Stephens M, Smith NJ, Donnelly P. A new statistical method for haplotype reconstruction from population data. Am J Hum Genet 2001; 68:978–989.
23. Dudbridge F. A survey of current software for linkage analysis. Hum Genomics 2003; 1:63–65.
24. Schork NJ, Nath SK, Fallin D, Chakravarti A. Linkage disequilibrium analysis of biallelic DNA markers, human quantitative trait loci, and threshold-defined case and control subjects. Am J Hum Genet 2000; 67:1208–1218.
25. Aulchenko YS, Ripatti S, Lindqvist I, Boomsma D, Heid IM, Pramstaller PP, et al. Loci influencing lipid levels and coronary heart disease risk in 16 European population cohorts. Nat Genet 2009; 41:47–55.
26. Sabatti C, Service SK, Hartikainen AL, Pouta A, Ripatti S, Brodsky J, et al. Genome-wide association analysis of metabolic traits in a birth cohort from a founder population. Nat Genet 2009; 41:35–46.
27. Kathiresan S, Willer CJ, Peloso GM, Demissie S, Musunuru K, Schadt EE, et al. Common variants at 30 loci contribute to polygenic dyslipidemia. Nat Genet 2009; 41:56–65.
28. Teslovich TM, Musunuru K, Smith AV, Edmondson AC, Stylianou IM, Koseki M, et al. Biological. clinical and population relevance of 95 loci for blood lipids. Nature 2010; 466:707–713.
29. O’Donnell CJ, Elosua R. Cardiovascular risk factors. Insights from Framingham Heart Study. Rev Esp Cardiol 2008; 61:299–310.
30. Diong C, Raboud J, Li M, Cooper C. Ontario HIVTNCSTHIV/hepatitis C virus and HIV/hepatitis B virus coinfections protect against antiretroviral-related hyperlipidaemia. HIV Med 2011; 12:403–411.
© 2013 Lippincott Williams & Wilkins, Inc.