Identification of Genetic Variants Associated With Response to Statin Therapy
Objective— The purpose of this study was to test the association between polymorphisms in genes involved in either LDL cholesterol (LDL-C) metabolism or statin pharmacokinetics and LDL-C reduction with statins.
Methods and Results— 49 tagging and candidate polymorphisms in 9 genes were genotyped in 1507 post-ACS subjects randomized to atorvastatin or pravastatin. Two polymorphisms (rs7412, rs429358) that define the ε2, ε3, and ε4 isoforms of apolipoprotein E were significantly associated with percent reduction in LDL-C with atorvastatin (ε2 carriers 53.8%, ε3/ε3 48.1%, and ε4 carriers 46.4%, respectively, P=0.00039) and replicated in the pravastatin arm (ε2 carriers 22.1%, ε3/ε3 21.8%, and ε4 carriers 16.6%, respectively, P=0.00038). The proportion of subjects achieving an LDL-C ≤70 mg/dL at day 30 was higher for ε2 than ε4 carriers (P=1.3×10−5). In the pravastatin group, the triallelic rs2032582 variant (G2677T/A) in ABCB1 was associated with the percent reduction in LDL-C (GG 23.3%, non-G heterozygote 20.3%, and non-G homozygote 17.4%, P=0.042).
Conclusion— Carriers of APOE ε2 versus ε4 had significantly greater LDL-C reduction with atorvastatin and with pravastatin, and more frequently achieved a guideline-recommended LDL-C ≤70 mg/dL. Polymorphisms in triallelic G2677T/A variant in ABCB1 were associated with the degree of LDL-C lowering with pravastatin.
Lowering cholesterol with statins has been shown to reduce the risk of cardiovascular events. In multiple trials, intensive statin therapy has further decreased this risk and reversed the progression of coronary atherosclerosis. The reduction of low density lipoprotein cholesterol (LDL-C) in response to statin therapy, however, can vary by as much as 10% to 70% from person to person, with many individuals not reaching target goals.1
Whereas diet, concomitant medications, and comorbidities can account for some of this variation, genetic determinants may also influence interindividual lipid parameters and responses to lipid-lowering medications. Family studies have suggested that the genetic contribution to baseline cholesterol levels is as high as 50%,2 and genetic association analyses have demonstrated the impact of both rare and common genetic variants on baseline cholesterol levels.3 Several studies also point to genetic variation contributing to the response to statin therapy.4–8
We therefore conducted a focused pharmacogenetic study in more than 1500 subjects in a large clinical trial in which patients were randomized to treatment with pravastatin or atorvastatin. We genotyped tagging and candidate variants in 9 genes involved in LDL-C metabolism or statin pharmacokinetics, and assessed the association between these genetic variants and baseline and percent reduction in LDL-C in the setting of statin therapy, as well as achieved cholesterol goals, inflammatory biomarkers, and cardiovascular events.
PROVE IT-TIMI 22 enrolled patients hospitalized with an acute coronary syndrome (ACS) within the preceding 10 days and with a total cholesterol of ≤240 mg/dL.9 Study participants were randomly assigned to receive atorvastatin 80 mg/d or pravastatin 40 mg/d and followed for an average of 24 months. For the current pharmacogenetic analysis, we limited our evaluation to the 1507 self-classified white subjects who were not previously on chronic lipid-lowering therapy and provided a DNA sample. For analyses of change in LDL-C levels, we restricted our evaluation to the 1378 subjects who had >80% compliance with allocated study medication. The main trial and genetic substudy protocols were approved by relevant institutional review boards, and informed consent was obtained from all patients.
Lipid, Biomarker, and Genetic Samples
Baseline samples were drawn at enrollment. On-treatment cholesterol concentrations refer to values obtained at 30 days, and the percent changes refer to the changes in values from baseline to 30 days. C-reactive protein (CRP) was measured using a high-sensitivity assay at the TIMI Biomarker Laboratory and lipoprotein-associated phospholipase A2 (Lp-PLA2) activity was measured at GlaxoSmithKline. DNA was extracted and genotyping was performed at the Harvard Medical School-Partners Center for Personalized Genetic Medicine on the Sequenom massARRAY and Taqman ABI 7500 platforms.
Gene and Single Nucleotide Polymorphism Selection
We selected the following genes involved in the LDL-C metabolism pathway based on biological function: APOB, APOE, HMGCR, LDLR, and PCSK9. To generate a set of polymorphisms capturing the majority of common variation in each gene, haplotype tag single nucleotide polymorphisms (SNPs) were selected using Tagger (supplemental Appendix, available online at http://atvb.ahajournals.org). In addition, several candidate SNPs in these genes were included based on prior literature. We also evaluated genes involved in statin pharmacokinetics: ABCB1, CYP3A4, CYP3A5, and SLCO1B1, selecting SNPs based on prior comprehensive functional studies. Forty-nine SNPs (43 in LDL-C metabolism pathways, 6 in statin pharmacokinetics pathways) met quality control thresholds (supplemental Table I).
Associations between the 43 SNPs in the LDL-C pathway and baseline LDL-C levels were assessed using SAS Genetics v9.1 in a linear regression model adjusting for age, sex, body mass index (BMI), diabetes mellitus requiring pharmacotherapy, smoking status, and time from symptoms to randomization. Effect estimates were determined in an additive genetic model per copy of the variant allele. Adjustment for multiple hypothesis testing was conducted using false discovery rate Q values.
The assessment of change in LDL-C was restricted to subjects with >80% compliance with assigned statin therapy. Analyses were conducted separately in 2 subcohorts based on randomized statin (atorvastatin or pravastatin) treatment assignment because the percent LDL-C reduction differed between the two groups. In each group, associations between the 49 SNPs (LDL-C metabolism and statin metabolism pathways) and percent reduction in LDL-C were determined. To further isolate the genetic effect of a variant on percent reduction in LDL-C, beyond its influence on baseline LDL-C, we adjusted for baseline LDL-C in the linear regression models in addition to the aforementioned covariates.
To be considered significant, for the 43 SNPs in the LDL-C pathway, an association had to be directionally consistent in both statin cohorts and achieve a probability value <0.025 in one cohort and <0.05 in the other, which would yield a conjoint probability of 0.00125, a value that approximates a conservative Bonferroni correction. Based on our sample size and the mean amount and variability of LDL-C reduction, we had 80% power to detect a per allele effect of 5.2% in the atorvastatin arm and 6.2% in the pravastatin arm for a variant with a minor allele frequency of 10%. For the 6 SNPs in the pharmacokinetic pathway, the association between genotype and percent LDL-C reduction was conducted separately in the pravastatin and atorvastatin groups and not evaluated using conjoint probabilities.
Haplotypes were created using an expectation-maximization algorithm with SAS Genetics v9.1. For ABCB1, the triallelic rs2032582 and rs1045642 were used to generate haplotypes. For APOE, SNPs rs7412 and rs429358 were used to create haplotypes, and 3 haplotypes were found, which correspond to the 3 common isoforms of apolipoprotein E: ε2, ε3, and ε4. Crude lipid data are displayed with patients categorized, by convention, as ε2 haplotype carriers, ε3/ε3, and ε4 haplotype carriers. Multivariable adjusted associations between APOE haplotypes and baseline lipid parameters and change in LDL-C were assessed in regression models analogous to those described above. Heterogeneity of the effect of APOE genotype on lipid parameters between the 2 statin treatment arms was evaluated using a formal interaction term in a linear regression model. Event rates for cardiovascular death, myocardial infarction (MI), or stroke were determined using Kaplan–Meier failure rates, and hazard ratios (HR) and 95% confidence intervals (CIs) were estimated using a Cox proportional hazards regression model.
Baseline Characteristics and Lipid Parameters
At baseline, subjects were found to have a mean LDL-C of 114.7±30.3 mg/dL. Of the 43 SNPs tested in the LDL-C pathway, we found that 8 were associated with baseline LDL-C concentrations with a probability value <0.05 (Table 1; complete list in supplemental Table II). The 2 SNPs with the most significant values were in APOE and included rs7412 (P=7.9×10−9) and rs429358 (P=0.0032). The corresponding false discovery rate Q values were 1.8×10−7 and 0.037. Even after using a conservative Bonferroni correction, rs7412 was still significantly associated with baseline LDL-C (P=3.4×10−7). The adjusted effect estimates on baseline LDL-C concentration per copy of the minor allele were −13.24 mg/dL (95% CI −17.68 to −8.74 mg/dL) for rs7412 and +5.01 mg/dL (95% CI 1.72 to 8.39 mg/dL) for rs429358. Two of the other SNPs significantly associated with baseline LDL-C concentrations were also in APOE, 2 were in APOB, 1 was in PCSK9, and 1 was in HMGCR. The false discovery rate Q value for these top 8 SNPs was only 0.095, suggesting there would be <1 false-positive among these significant findings.
Change in LDL-C
Analyses of change in LDL-C were restricted to the 1378 subjects who had >80% compliance with statin therapy. Of the 686 subjects treated with atorvastatin 80 mg/d and 692 subjects treated with pravastatin 40 mg/d, the mean percent reduction in LDL-C from baseline to 30 days was 48.8±18.5% and 20.7±22.3%, respectively. The complete list of associations between SNPs and reduction in LDL-C are in supplemental Tables III and IV. After adjustment for baseline LDL-C and clinical features, 3 SNPs were significantly associated with the percent reduction in LDL-C in atorvastatin-treated patients (rs7412: P=8.3×10−6, rs429358: P=0.044, and rs769449: P=0.022) and were replicated in pravastatin-treated patients (rs7412: P=0.00065, rs429358: P=0.023, and rs769449: P=0.0047; Table 2).
Of these 3 SNPs, all were located in APOE. rs429358 and rs7412 are nonsynonymous variants, whereas the rs769449 polymorphism is intronic and in linkage disequilibrium with rs429358 (r2=0.75). Thus the 2 former SNPs were used to create APOE haplotypes. Three haplotypes were found and consisted of the T allele at rs429358 and the T allele at rs7412 (7% of study population), the T allele and the C allele (79%), and the C allele and the C allele (14%). These 3 haplotypes correspond to the 3 common isoforms of apolipoprotein E: ε2, ε3, and ε4.
There were no significant differences in baseline characteristics across APOE genotype (defined as carriers of the ε2 haplotype, n=155; ε3/ε3 wild-type, n=809; carriers of the ε4 haplotype, n=327) except for ε2 carriers being modestly older (supplemental Table V). Baseline LDL-C values were significantly associated with APOE genotype (P=1.5×10−6, Figure 1A). The effect estimate per copy of the ε2 haplotype was −12.77 mg/dL (95% CI −17.26 to −8.26 mg/dL, P=3.1×10−8) and of the ε4 haplotype was +4.00 mg/dL (95% CI 0.69 to 7.32 mg/dL, P=0.018). Moreover, e2/e2 (n=7) individuals had 27% lower baseline LDL-C values as compared with e4/e4 (n=25) individuals. APOE genotype also affected baseline concentrations of apolipoprotein B, HDL-C, and apolipoprotein E (supplemental Table VI).
The percent reduction in LDL-C levels from baseline to day 30 varied significantly across APOE genotype, being greatest in ε2 carriers, intermediate in ε3/ε3 individuals, and least in ε4 carriers, both in atorvastatin-treated patients (53.8±1.8%, 48.1±1.0%, and 46.4±1.5%, respectively, P=0.00039) and pravastatin-treated patients (22.1±2.8%, 21.8±1.1%, and 16.6±2.1%, respectively, P=0.00038; Figure 1B), after accounting for baseline LDL-C and clinical features. The relationship between APOE genotypes and percent reduction in LDL-C was similar with both atorvastatin and pravastatin. In all subjects, and adjusting for statin treatment, the effect estimate on percent reduction in LDL-C per ε2 haplotype was +7.84% (95% CI 4.94 to 10.73%, P=1.3×10−7) and per ε4 haplotype was −2.66% (95% CI −4.76 to −0.56%, P=0.013). After adjustment for baseline LDL-C and clinical factors, APOE genotype accounted for 3.8% of the remaining variance in LDL-C reduction, and no significant gene–environment interactions were observed with sex or smoking status (Pinteractions=NS).
In terms of on-treatment LDL-C at 30 days, levels were lowest for ε2 carriers, intermediate for ε3/ε3 individuals, and highest for ε4 carriers in both atorvastatin-treated patients (46.1±1.9, 56.9±1.0, and 61.5±1.6 mg/dL, respectively; P=1.2×10−7) and pravastatin-treated patients (72.1±2.4, 89.3±1.2, and 93.4±2.1 mg/dL, respectively; P=2.5×10−7). Analogously, the percent of subjects achieving an LDL-C goal of ≤70 mg/dL was higher for ε2 than ε4 carriers in both atorvastatin-treated patients (P=0.00037) and pravastatin-treated patients (P=0.0084; Figure 2). The relationship between APOE genotype and on-treatment LDL-C was similar with atorvastatin and pravastatin. In all subjects, and adjusting for treatment arm, APOE genotype was significantly associated with on-treatment 30-day LDL-C (P=2.6×10−13) and likelihood of achieving an LDL-C of ≤70 mg/dL (P=1.3×10−5).
With regard to inflammatory markers, similar to what was seen for on-treatment LDL-C (Figure 3A), on-treatment Lp-PLA2 activity at 30 days was lowest for ε2 carriers, intermediate for ε3/ε3 individuals, and highest for ε4 carriers in both atorvastatin-treated patients (P=0.0045) and pravastatin-treated patients (P=8.3×10−6; Figure 3B). In contrast, the opposite trend was seen for on-treatment hs-CRP levels at 30 days, with the highest levels seen in ε2 carriers, intermediate levels in ε3/ε3 individuals, and lowest levels in ε4 carriers, both in atorvastatin-treated (geometric means: 2.5±0.4, 1.8±0.2, and 1.2±0.2 mg/L, respectively; P=4.2×10−7) and pravastatin-treated patients (2.9±0.4, 2.6±0.2, and 2.1±0.3 mg/L, respectively; P=0.036; Figure 3C). The proportion achieving goal hs-CRP values varied significantly by genotype, with more ε4 carriers achieving a target of ≤2 mg/L (P=0.00026).
During an average of 2 years of follow-up, the rate of cardiovascular death, MI, or stroke for ε2 carriers, ε3/ε3, and ε4 carriers was 5.4%, 7.4%, and 7.0%, respectively. The HR for the risk of cardiovascular death, MI, or stroke for ε2 carriers versus ε3/ε3 individuals was 0.78 (95% CI 0.40 to 1.51) and for ε4 carriers versus ε3/ε3 individuals was 0.96 (95% CI 0.60 to 1.54). There was no interaction between genotype, statin treatment, and outcome.
SNPs in the Pharmacokinetic Pathway
Among subjects treated with pravastatin, the triallelic rs2032582 variant (G2677T/A) in ABCB1 was associated with the percent reduction in LDL-C when evaluated based on carriage of the non-G (T/A) allele (GG 23.3%, non-G heterozygote 20.3%, and non-G homozygote 17.4%, P=0.042). In terms of the individual effects of the T and A variant alleles, the T allele was associated with an effect estimate of −1.7% per allele (95% CI −3.86 to 0.46, P=0.12) and the A allele with an effect estimate of −9.4% per allele (95% CI −15.7 to −3.00, P=0.004). Adjustment for APOE genotype did not significantly affect the results. Another variant in ABCB1 was also tested (rs1045642, C3435T), and there was no significant association between this polymorphism and LDL-C reduction in subjects treated with pravastatin. Based on a prior report, haplotypes were then constructed using G2677T/A and C3435T.10 Using the G-C haplotype as a referent, the nonG-C haplotype was associated with an absolute 10.5% lesser reduction in LDL-C (P=0.0003) and the nonG-T haplotype a trend to an absolute −1.9% lesser reduction in LDL-C (P=0.10, supplemental Table VII). Further haplotype analyses evaluating the T and A at 2677 separately revealed that the A-C haplotype was associated with an absolute 12.4% lesser reduction in LDL-C (P=0.0007), the T-C haplotype with a trend to an absolute 7.8% lesser reduction in LDL-C (P=0.10), and T-T a trend to an absolute 2.0% lesser reduction in LDL-C (P=0.09).
Analogous analyses were conducted in subjects treated with atorvastatin, and no association was detected between the 2677 or 3435 variants and LDL-C reduction. Although the effect estimate for the T-C haplotype (−5.9%) was directionally consistent with what was observed with pravastatin, the result did not achieve statistical significance (P=0.17; supplemental Table VIII). None of the other pharmacokinetic SNPs were significantly related to the percent reduction in LDL-C in either treatment arm (supplemental Tables III and IV).
Statin therapy has become the mainstay for lowering LDL-C. However, there is a large degree of interindividual variability in the response to treatment.1 We therefore conducted a pharmacogenetic study in 1507 subjects to determine whether genetic polymorphisms influence LDL-C reduction in the setting of statin therapy. In this analysis, a set of tagging and candidate SNPs was used to assess 9 genes of interest in the LDL-C metabolism and statin pharmacokinetics pathways. We found that rs7412 and rs429358 in APOE were significantly associated both with baseline LDL-C levels and reduction in LDL-C with atorvastatin and with pravastatin. Additionally, there was an association between the G2677T/A variant in ABCB1 and the degree of LDL-C lowering with pravastatin. No associations between the other 46 SNPs and LDL-C reduction were detected.
The 2 genetic variants in APOE account for amino acid changes at positions 112 and 158 and haplotype analysis yielded the 3 common isoforms ε2, ε3, and ε4 of apolipoprotein E, an apolipoprotein that plays a critical role in receptor-mediated hepatic uptake of lipoproteins. The most common isoform, ε3, has a cysteine at position 112 and an arginine at position 158, whereas ε2 contains 2 cysteines and ε4 contains 2 arginines. In this study, carriers of ε2 had significantly reduced baseline LDL-C concentrations, with ε2 carriers having values that were 18.7 mg/dL lower than ε4 carriers. Moreover, e2/e2 individuals had 27% lower baseline LDL-C values as compared with e4/e4 individuals. These findings are consistent with prior studies, including a large meta-analysis of healthy subjects where ε2/ε2 individuals had 30% lower LDL-C values as compared to those with ε4/ε4.11
In terms of the association of APOE genotypes and LDL-C reduction with statin therapy, the majority of studies have reported greater LDL-C reduction in ε2 carriers and lesser reductions in ε4 carriers,6–8,12–14 although others have suggested either no significant relationship,15 gender-specific findings,16 or even the converse with lesser reductions in ε2 carriers and greater reductions in ε4 carriers.17 Our analysis benefited from a large sample size, consistent statin dosing (including use of atorvastatin, one of the most commonly prescribed statins in the United States currently) and timing of blood sampling, central laboratory measurements, monitored compliance with statin therapy, and standardized follow-up for clinical events.
In this setting, we found highly statistically significant and consistent associations between ε2 carrier status and greater percent LDL-C reduction, independent of baseline LDL-C, with ε2 carriers having greater LDL cholesterol percentage reduction than ε4 carriers. Our observations are biologically consistent with the known alterations in function with the different apolipoprotein E isoforms (supplemental Appendix). The APOE genotype accounted for 3.8% of the remaining variance in statin response, an effect on par with other genetic studies.3 Moreover, the APOE genotype classifications were associated with the proportion of subjects who achieved the ACC/AHA guideline recommended target LDL-C value. Although we would not advocate routine genotyping at this point because lipid-lowering therapy can be titrated to a target LDL-C goal, in the future there may be a panel of SNPs that could prove useful in tailoring statin therapy.18–20 The rates of cardiovascular death, MI, or stroke for ε2 carriers as compared to ε3/ε3 individuals were 5.0% versus 7.4% with an associated HR of 0.78 (95% CI 0.40 to 1.51). This difference was not statistically significant but is consistent with data from a meta-analysis of studies (not involving statin therapy) where the HR was 0.80 (95% CI 0.70 to 0.90).11
In terms of statin pharmacokinetic genes, ABCB1 (also known as the multi-drug resistence-1 gene) encodes P-glycoprotein. In several prior studies, atorvastatin has been shown to be a substrate of P-glycoprotein,21,22 and polymorphisms in this gene have been associated with the response to atorvastatin.23 In this study, there were no significant associations between variants in ABCB1 and percent LDL reduction among those treated with atorvastatin. Our observation may differ from prior ones because of the higher dose of atorvastatin given (80 mg/d).24 In a study by Chen and colleagues, pravastatin was a weak substrate for P-glycoprotein.25 In our analysis, ABCB1 G2677T/A was associated with less LDL-C reduction with pravastatin, with a more pronounced effect with the A allele. Haplotype analysis using G2677T/A and C3435T revealed the greatest effects with the A-C haplotype and trends for the T-C and T-T haplotypes (the A-T haplotype constituted <1% of the study population). Prior studies to date have yielded varied results for the functional impact of ABCB1 polymorphisms (supplemental Appendix). Additional studies will be helpful to continue to discern the functional effects of any of these variants or haplotypes.
Our study has several potential limitations that should be considered. First, our analyses were conducted exclusively in a white population. Second, we selected tagging SNPs with an MAF >10% in the LDL-C metabolism pathway; we cannot comment on the potential role of rarer variants. Third, in the case of variants involving statin pharmacokinetic genes, we did not use tagging SNPS and elected to focus on candidate variants because prior studies had sequenced these genes and functional variants had been established. Additional variants and the haplotypes they define may be important.26,27 Fourth, to be conservative and consistent, we a priori picked an additive model for all genetic comparisons to minimize multiple-hypothesis testing for different models. Fifth, our study population consisted of patients stabilized after ACS with upper limits of total cholesterol. Thus our baseline cholesterol levels are not generalizable to cholesterol levels in a healthy, ambulatory population cohort; however, the relative differences we observed in baseline cholesterol levels were in keeping with prior studies and our primary outcome of interest, the response to statin therapy, should be generalizable. Although ACS can affect cholesterol and inflammatory biomarker levels, the effect on the former is modest,28 and we only analyzed on-treatment inflammatory biomarker levels, not change. Finally, we used the random allocation to atorvastatin versus pravastatin to create 2 study populations to enable replication for SNPs in the LDL-C metabolism pathway. We acknowledge that it is possible that some genetic variants might alter response to statin therapy differently with different statins; thus, SNPs that achieved nominal statistical significance in one statin population but not in the other might still be worthy of further investigation.
In this pharmacogenetic study, that evaluated genetic variants in 9 genes, carriers of the ε2 haplotype were found to have lower baseline LDL-C values compared with ε4 carriers. ε2 carriers also experienced greater LDL-C reduction with both atorvastatin and pravastatin than did ε4 carriers. Moreover, the APOE genotype classifications influenced the proportion of subjects who achieved clinically meaningful LDL-C targets, with more than one-quarter of ε4 carriers failing to meet guideline recommended LDL-C goals even with intensive statin therapy. A polymorphism in ABCB1 was associated with the degree of LDL-C lowering with pravastatin. These findings add to the growing body of data evaluating the genetic mechanisms that influence statin efficacy.
Sources of Funding
The PROVE IT-TIMI 22 trial was funded by Bristol-Myers Squibb. Dr Mega was supported in part by a grant from the Leadership Council for Improving Cardiovascular Care (Schering-Plough). The TIMI Study Group reports receiving significant research grant support from AstraZeneca, Bayer Healthcare, Beckman Coulter, Biosite, Bristol-Myers Squibb, CV Therapeutics, Eli Lilly, Genentech, GlaxoSmithKline, Integrated Therapeutics Group, Johnson & Johnson, Merck, Nanosphere, Novartis, Pfizer, Roche Diagnostics, Sanofi-Aventis, Siemens Medical Solutions, Singulex, and Schering-Plough.
Dr Morrow reports receiving honoraria from Roche Diagnostics, Sanofi-Aventis, and Siemens Medical Solutions and is a consultant for Dade Behring, GlaxoSmithKline, and Sanofi-Aventis. Dr Cannon reports receiving research grant support from Accumetrics, AstraZeneca, Bristol-Myers Squibb, GlaxoSmithKline, Merck, Sanofi-Aventis, and Schering-Plough. Dr Sabatine reports receiving research grant support from Schering-Plough, Roche Diagnostics, and diaDexus and honoraria from Bristol-Myers Squibb.
Received December 8, 2008; revision accepted April 7, 2009.
Law MR, Wald NJ, Rudnicka AR. Quantifying effect of statins on low density lipoprotein cholesterol, ischaemic heart disease, and stroke: systematic review and meta-analysis. BMJ. 2003; 326: 1423.
Pilia G, Chen WM, Scuteri A, Orru M, Albai G, Dei M, Lai S, Usala G, Lai M, Loi P, Mameli C, Vacca L, Deiana M, Olla N, Masala M, Cao A, Najjar SS, Terracciano A, Nedorezov T, Sharov A, Zonderman AB, Abecasis GR, Costa P, Lakatta E, Schlessinger D. Heritability of cardiovascular and personality traits in 6,148 Sardinians. PLoS Genet. 2006; 2: e132.
Kathiresan S, Melander O, Guiducci C, Surti A, Burtt NP, Rieder MJ, Cooper GM, Roos C, Voight BF, Havulinna AS, Wahlstrand B, Hedner T, Corella D, Tai ES, Ordovas JM, Berglund G, Vartiainen E, Jousilahti P, Hedblad B, Taskinen MR, Newton-Cheh C, Salomaa V, Peltonen L, Groop L, Altshuler DM, Orho-Melander M. Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet. 2008; 40: 189–197.
Krauss RM, Mangravite LM, Smith JD, Medina MW, Wang D, Guo X, Rieder MJ, Simon JA, Hulley SB, Waters D, Saad M, Williams PT, Taylor KD, Yang H, Nickerson DA, Rotter JI. Variation in the 3-hydroxyl-3-methylglutaryl coenzyme a reductase gene is associated with racial differences in low-density lipoprotein cholesterol response to simvastatin treatment. Circulation. 2008; 117: 1537–1544.
Voora D, Shah SH, Reed CR, Zhai J, Crosslin DR, Messer C, Salisbury BA, Ginsberg GS. Pharmacogenetic predictors of statin-mediated low-density lipoprotein cholesterol reduction and dose response. Circ Cardiovasc Genet. 2008; 1: 100–106.
Thompson JF, Hyde CL, Wood LS, Paciga SA, Hinds DA, Cox DR, Hovingh GK, Kastelein JJ. Comprehensive whole genome and candidate gene analysis for response to statin therapy in TNT cohort. Circ Cardiovasc Genet. In press.
Nestel P, Simons L, Barter P, Clifton P, Colquhoun D, Hamilton-Craig I, Sikaris K, Sullivan D. A comparative study of the efficacy of simvastatin and gemfibrozil in combined hyperlipoproteinemia: prediction of response by baseline lipids, apo E genotype, lipoprotein(a) and insulin. Atherosclerosis. 1997; 129: 231–239.
Iakoubova OA, Sabatine MS, Rowland CM, Tong CH, Catanese JJ, Ranade K, Simonsen KL, Kirchgessner TG, Cannon CP, Devlin JJ, Braunwald E. Polymorphism in KIF6 gene and benefit from statins after acute coronary syndromes: results from the PROVE IT-TIMI 22 study. J Am Coll Cardiol. 2008; 51: 449–455.
Iakoubova OA, Tong CH, Rowland CM, Kirchgessner TG, Young BA, Arellano AR, Shiffman D, Sabatine MS, Campos H, Packard CJ, Pfeffer MA, White TJ, Braunwald E, Shepherd J, Devlin JJ, Sacks FM. Association of the Trp719Arg polymorphism in kinesin-like protein 6 with myocardial infarction and coronary heart disease in 2 prospective trials: the CARE and WOSCOPS trials. J Am Coll Cardiol. 2008; 51: 435–443.
Chen C, Mireles RJ, Campbell SD, Lin J, Mills JB, Xu JJ, Smolarek TA. Differential interaction of 3-hydroxy-3-methylglutaryl-coa reductase inhibitors with ABCB1, ABCC2, and OATP1B1. Drug Metab Dispos. 2005; 33: 537–546.
Kimchi-Sarfaty C, Oh JM, Kim IW, Sauna ZE, Calcagno AM, Ambudkar SV, Gottesman MM. A “silent” polymorphism in the MDR1 gene changes substrate specificity. Science. 2007; 315: 525–528.