Key Points
Question
Does the rs1967309 single-nucleotide polymorphism, reported to show a strong interaction for dalcetrapib, influence cardiovascular outcomes for patients treated with the cholesteryl ester transfer inhibitor evacetrapib?
Findings
In this nested case-control study, this single-nucleotide polymorphism was examined in 1427 cases and 1532 matched controls selected from the 12 092–patient evacetrapib cardiovascular outcome trial. The conditional logistic regression odds ratio for major adverse cardiovascular events for evacetrapib-treated patients with the AA genotype was not significant.
Meaning
Although directionally similar to the dalcetrapib analysis, there was no significant interaction between genotype and cardiovascular outcome with evacetrapib.
Abstract
Importance
A pharmacogenetic analysis of dalcetrapib, a cholesteryl ester transfer protein inhibitor, reported an association between a single-nucleotide polymorphism (SNP) in the ADCY9 gene (rs1967309) and reduction in major adverse cardiovascular events despite a neutral result for the overall trial.
Objective
To determine whether the association between the SNP in the ADCY9 gene and a reduction in major adverse cardiovascular events could be replicated for another cholesteryl ester transfer protein inhibitor, evacetrapib, in patients with high-risk vascular disease.
Design, Setting, and Participants
A nested case-control study examining the rs1967309 SNP in 1427 cases and 1532 matched controls selected from the 12 092–patient Assessment of Clinical Effects of Cholesteryl Ester Transfer Protein Inhibition with Evacetrapib in Patients at a High Risk for Vascular Outcomes (ACCELERATE) trial, a randomized, double-blind, placebo-controlled phase 3 trial conducted in patients with high-risk vascular disease randomized from October 2012 through December 2013. The genotyping was conducted from January 2017 to March 2017, and the data analyses were conducted from July 2017 to November 2017.
Exposures
Evacetrapib, 130 mg, or matching placebo.
Main Outcomes and Measures
The primary analyses used a conditional logistic regression model to assess the odds ratio (OR) for major adverse cardiovascular events for evacetrapib compared with placebo for each genotype. The basic model included adjustment for age, sex, and the top 5 principal components. An additional model included cardiovascular risk factors to adjust for potential bias in selecting control patients. The primary major adverse cardiovascular event end point was the composite of death from cardiovascular causes, myocardial infarction, stroke, coronary revascularization, or hospitalization for unstable angina.
Results
For patients with the AA genotype reported to demonstrate a beneficial effect from dalcetrapib, the OR for evacetrapib compared with placebo was 0.88 (95% CI, 0.69-1.12). For patients with the AG genotype, the OR was 1.04 (95% CI, 0.90-1.21). For patients with the GG genotype reported to show evidence for a harmful effect from dalcetrapib, the OR for evacetrapib was 1.18 (95% CI, 0.98-1.41). The interaction P value among the 3 genotypes was P = .17 and the trend P value was P = .06. When adjusted for cardiovascular risk factors, the OR for evacetrapib was 0.93 (95% CI, 0.73-1.19) for the AA genotype, 1.05 (95% CI, 0.91-1.22) for the AG genotype, and 1.02 (95% CI 0.85-1.24) for the GG genotype; interaction P = .71 and trend P = .59.
Conclusions and Relevance
Pharmacogenetic analysis did not show a significant association between the ADCY9 SNP (rs1967309) and cardiovascular benefit or harm for the cholesteryl ester transfer protein inhibitor evacetrapib.
This study explores whether the association between an SNP in the ADCY9 gene and a reduction in major adverse cardiovascular events could be replicated for another cholesteryl ester transfer protein inhibitor, evacetrapib, in patients with high-risk vascular disease.
Introduction
Large cardiovascular outcome trials have studied the effects of 4 drugs that inhibit cholesteryl ester transfer protein (CETP): torcetrapib, dalcetrapib, evacetrapib, and anacetrapib.1,2,3,4 These CETP inhibitors markedly raised circulating levels of high-density lipoprotein cholesterol (HDL-C), with increases ranging from approximately 30% to 130%. With the exception of dalcetrapib, these drugs also reduced low-density lipoprotein cholesterol (LDL-C), with decreases ranging from 21% to 37%. Dalcetrapib and evacetrapib had no significant effect on major adverse cardiovascular events (MACE), torcetrapib increased MACE, and anacetrapib produced a small reduction in MACE, which was consistent with the expected event reduction based on meta-analyses of trials of other agents that reduce levels of atherogenic lipoproteins.5
A post hoc pharmacogenetic analysis of the dalcetrapib trial suggested an association between a single-nucleotide polymorphism (SNP) in the ADCY9 gene on chromosome 16 (rs1967309) and cardiovascular outcome (P = 2.41 × 10−8).6 Patients with the AA genotype who received dalcetrapib showed a 39% reduction in the primary composite cardiovascular endpoint compared with placebo, whereas patients with the GG genotype showed a 27% increase in the primary end point. Results for the AG genotype were neutral. Supporting evidence for a pharmacogenetic relationship included an analysis of the relationship between genotype status and carotid intimal medial thickness in a separate clinical trial but for a different SNP (rs2238448) in the ADCY9 gene.6 A subsequent analysis suggested a beneficial relationship between the genotype and change in 2 biomarkers, high-sensitivity C-reactive protein (hsCRP), and cholesterol efflux capacity.7 On the basis of these findings, a new cardiovascular outcome trial (NCT02525939), currently underway, was designed to study the cardiovascular effects of dalcetrapib compared with placebo only in patients with the AA genotype.
In this analysis, we sought to replicate the pharmacogenetic findings from the dalcetrapib analysis in patients enrolled in the Assessment of Clinical Effects of Cholesteryl Ester Transfer Protein Inhibition with Evacetrapib in Patients at a High Risk for Vascular Outcomes (ACCELERATE) trial, which assessed the effect of evacetrapib on cardiovascular outcome in patients with high-risk vascular disease.3
Methods
The ACCELERATE trial
The ACCELERATE trial was a multicenter, randomized, double-blind, placebo-controlled, phase 3 trial conducted at 543 sites in 36 countries (NCT01687998). The trial design has been previously described.3,8 Approval was obtained from local institutional review boards and all patients enrolled in the ACCELERATE Trial provided written informed consent. Institutional review boards also approved the pharmacogenetics analyses. The genetic analyses were conducted in those patient samples for patients who provided written informed consent for participation in the genetics study. Enrolled patients had preexisting vascular disease with high-risk features, defined as the presence of at least 1 of the following conditions: an acute coronary syndrome within the previous 30 to 365 days, cerebrovascular atherosclerotic disease, peripheral arterial disease, or diabetes with coronary artery disease. Patients were randomly assigned in a 1:1 ratio to receive oral evacetrapib at a dose of 130 mg or matching placebo. The primary efficacy end point was MACE, defined as the first occurrence of any component of the composite of death from cardiovascular causes, myocardial infarction, stroke, coronary revascularization, or hospitalization for unstable angina (5-component MACE). A key secondary efficacy end point was the composite of death from cardiovascular causes, myocardial infarction, or stroke (3-component MACE). The trial was terminated prematurely for futility at the recommendation of the data monitoring committee after 82% of the planned 1670 primary events had occurred.
Pharmacogenetic Study Design
DNA samples were extracted from the frozen whole blood from the patients who consented for genetic analyses. All samples were genotyped using the Axiom Biobank genotyping array, version 2 (Affymetrix). Patient samples, along with duplicates and HapMap controls, were genotyped, and data on 717 345 SNPs were generated. Standard metrics for genome-wide association study (GWAS) quality control were used, and 634 382 SNPs passed quality-control screening. Genotype call rate for the key ADCY9 SNP of interest (rs1967309) was 99.2%, with a minor allele frequency of 41%. The rs1967309 SNP did not deviate from Hardy-Weinberg equilibrium. The pharmacogenetic sample information is illustrated in eFigure 1 in the Supplement.
The pharmacogenetic analysis used a nested case-control design that genotyped all patients with cardiovascular events and matched control patients. eFigure 2 in the Supplement shows the flow of patients in the analysis. From 779 patients in the evacetrapib treatment group with a positively adjudicated cardiovascular event for the primary end point, genotype data for genetic analysis were available in 719 patients. The frequency of the AA, AG, and GG genotypes were compared with 763 control patients without a primary cardiovascular end point matched for age, sex, and race/ethnicity. For the placebo treatment group, genotype data were available for 708 of 776 patients with a cardiovascular event and compared with 769 matched control patients. The primary end point for the analysis was the odds ratio (OR) for 5-component MACE comparing the evacetrapib treatment group with the placebo treatment group within each genotype. A sensitivity analysis assessed the OR for 3-component MACE, the key secondary end point in the ACCELERATE trial. Additional analyses assessed the OR for the effect per minor allele for evacetrapib and placebo for both 5-component and 3-component MACE (additive genetic model). The relationship between genotype and biomarkers associated with cardiovascular disease, including high-sensitivity C-reactive protein (hsCRP), was also determined.
Statistical Analyses
Conditional logistic regression methods were used to account for the matching of patients with events (cases) to patients without events (controls). To test the additive genetic effect for the minor allele, using 1° of freedom, the common alleles (GG) were coded as 0 (reference), heterozygotes (AG) were coded as 1, and minor allele homozygotes (AA) were coded as 2. To correct for any potential population substructure, principal component analysis was performed on linkage disequilibrium-pruned GWAS data, and top 5 principal components captured most of the genetic variability in our data set. Principal components plots showing the race/ethnicity distribution using principal component 1 and principal component 2 are shown in eFigure 3 in the Supplement, and a scree plot showing individual and cumulative percent genomic variance explained by top 10 principal components is shown in eFigure 4 in the Supplement. The primary analyses used a conditional logistic regression model that included treatment, genotype, and genotype-by-treatment terms and adjusted for the age, sex, and the top 5 principal components. An additional model included cardiovascular risk factors to adjust for potential bias in selecting control patients. The models were used to assess the effect of evacetrapib compared with placebo within each genotype category and to assess genetic effect within each treatment group. To understand the heterogeneity of treatment effects across the genotypes, a genotype-by-treatment interaction test was conducted using a likelihood ratio test with 2° of freedom. An alternative analysis examined the trend across genotypes after coding the genotypes as the number of minor alleles in the conditional logistic model, reporting the corresponding genotype-by-treatment interaction P values. A sensitivity analysis using only white patients was also conducted. For the primary outcome, the study was 87% powered to detect the treatment by SNP interaction hazard ratio (HR) reported for dalcetrapib. The P value level of significance was .05, and all P values were 2-sided.
ADCY9 Biomarker Analyses
Linear regression analyses were used to assess the genetic association of the ADCY9 SNP with the change from baseline of several biomarkers of interest (HDL-C, LDL-C, hsCRP, and systolic and diastolic blood pressure). The model was adjusted for baseline level, and evaluations were separately performed in both the evacetrapib and placebo treatment groups.
Results
Baseline Characteristics
Table 1 reports the baseline characteristics for both evacetrapib- and placebo-treated patients who experienced and who did not experience a primary end point within the ACCELERATE pharmacogenetic analyses. eTable 1 in the Supplement shows the baseline characteristics for patients who underwent genotyping and the full study population. These comparisons show similar characteristics with the exception of geographic region. For comparative purposes, baseline characteristics for the dalcetrapib outcomes (dal-OUTCOMES) genetic analysis patients are shown in eTable 2 in the Supplement. By trial design, all dalcetrapib-treated patients experienced a recent acute coronary syndrome, whereas slightly more than half of the ACCELERATE patients had experienced a prior acute coronary syndrome. In the evacetrapib study, a higher percentage of patients had diabetes compared with the dalcetrapib-treated patients. In the evacetrapib analysis nearly, all of the matched control patients resided in North America compared with approximately half of the dalcetrapib-treated patients.
Table 1. Baseline Characteristics of Patients Selected for Pharmacogenetic Analysis in ACCELERATE Trial.
| Characteristic | No. (%) | |||
|---|---|---|---|---|
| Patients With Events | Patients Without Events | |||
| Evacetrapib-Treated (n = 719) | Evacetrapib Placebo (n = 708) | Evacetrapib-Treated (n = 763) | Evacetrapib Placebo (n = 769) | |
| Age, mean (SD), y | 65.6 (9.6) | 65.7 (9.1) | 65.3 (8.5) | 66.0 (8.6) |
| Female sex | 158 (22.0) | 175 (24.7) | 166 (21.8) | 182 (23.7) |
| White | 612 (85.1) | 628 (88.7) | 650 (85.2) | 675 (87.8) |
| Geographic region | ||||
| North America | 396 (55.1) | 395 (55.8) | 689 (90.3) | 695 (90.4) |
| Europe | 199 (27.7) | 200 (28.2) | 34 (4.5) | 32 (4.2) |
| Asia | 36 (5.0) | 29 (4.1) | 15 (2.0) | 24 (3.1) |
| Other | 88 (12.2) | 84 (11.9) | 25 (3.3) | 18 (2.3) |
| BMI, mean (SD) | 30.9 (5.8) | 31.1 (5.9) | 32.0 (6.9) | 31.5 (5.8) |
| Hypertension | 673 (93.6) | 659 (93.1) | 697 (91.3) | 716 (93.1) |
| Diabetes | 523 (72.7) | 544 (76.8) | 583 (76.4) | 583 (75.8) |
| Current smoker | 126 (17.5) | 107 (15.1) | 104 (13.6) | 100 (13.0) |
| ACS | 448 (62.3) | 435 (61.4) | 414 (54.3) | 408 (53.1) |
| PAD | 184 (25.6) | 155 (21.9) | 121 (15.9) | 125 (16.3) |
| Previous MI | 470 (65.4) | 440 (62.1) | 420 (55.0) | 426 (55.4) |
| Previous PCI | 509 (70.8) | 517 (73.0) | 524 (68.7) | 487 (63.3) |
| Previous CABG | 227 (31.6) | 231 (32.6) | 281 (36.8) | 286 (37.2) |
| Any statin | 676 (94.0) | 670 (94.6) | 720 (94.4) | 740 (96.2) |
| HDL-C, mean (SD), mg/dL | 44.8 (11.4) | 44.8 (11.6) | 43.6 (11.3) | 43.5 (11.3) |
| LDL-C, mean (SD), mg/dL | 84.3 (30.0) | 83.7 (29.2) | 80.0 (26.5) | 80.8 (26.4) |
| ApoB/ApoA ratio | 1.82 (0.57) | 1.84 (0.58) | 1.87 (0.59) | 1.85 (0.56) |
| hsCRP, mean (SD), mg/L | 4.5 (8) | 4.1 (8) | 4.2 (13) | 3.2 (7) |
Abbreviations: ACCELERATE, Assessment of Clinical Effects of Cholesteryl Ester Transfer Protein Inhibition with Evacetrapib in Patients at a High Risk for Vascular Outcomes; ACS, acute coronary syndrome; ApoA, apolipoprotein A; ApoB, apolipoprotein B; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CABG, coronary artery bypass grafting; HDL-C, high density lipoprotein cholesterol; hsCRP, high-sensitivity C-reactive protein; LDL-C, low density lipoprotein cholesterol; MI, myocardial infarction; PAD, peripheral arterial disease; PCI, percutaneous coronary intervention.
SI conversion factor: To convert cholesterol levels to millimoles per liter, multiply by 0.0259; high-sensitivity C-reactive protein to nanomoles per liter, multiply by 9.524.
Genotype Frequencies
The genotype frequencies in the evacetrapib (ACCELERATE) and dalcetrapib (dal-OUTCOMES) trial for all patients who underwent genotyping are reported in eTable 3 in the Supplement. Frequencies are shown for placebo-treated and CETP inhibitor–treated patients who did and did not experience a cardiovascular event.
Logistic Regression Analysis
Table 2 shows the associations for the 3 genotypes based on conditional logistic regression analysis for evacetrapib (both 5-component MACE and the 3-component MACE sensitivity analysis). The ORs for 5-component MACE and 3-component MACE showed no significant difference between evacetrapib and placebo for any of the 3 genotypes. Patients with the AA genotype for the rs1967309 SNP that was reported to show evidence for a beneficial effect from dalcetrapib showed an OR of 0.88 (95% CI, 0.69-1.12) for 5-component MACE and an OR of 0.92 (95% CI, 0.67-1.27) for 3-component MACE. Patients with the AG genotype showed an OR of 1.04 (95% CI, 0.90-1.21) for evacetrapib treatment compared with placebo for 5-component MACE and 1.05 (95% CI, 0.86-1.28) for 3-component MACE. Patients with the GG genotype reported to show evidence for a harmful effect from dalcetrapib showed an OR of 1.18 (95% CI, 0.98-1.41) for evacetrapib treatment compared with placebo for 5-component MACE and 0.92 (95% CI, 0.72-1.16) for 3-component MACE. For 5-component MACE, the interaction P value for the primary analysis among the AA, AG, and GG genotypes was P = .17. An alternative analysis examining the trend across the 3 genotypes showed no significant interaction for 5-component MACE (P = .06). For 5-component MACE, analyses adjusted for cardiovascular risk factors also did not show significant differences for evacetrapib-treated patients compared with placebo for either the AA (OR, 0.93; 95% CI, 0.73-1.19) or GG genotype (OR, 1.02; 95% CI, 0.85-1.24); interaction P = .71; trend P = .12 (Table 2) For 3-component MACE, analyses adjusted for cardiovascular risk factors also did not show significant differences for the AA genotype (OR, 1.06; 95% CI, 0.77-1.46) or GG genotype (OR, 0.81; 95% CI, 0.63-1.03; interaction P = .17; trend P = .12).
Table 2. Treatment Effects of Evacetrapib Compared With Placebo Within Each Genotype of rs1967309.
| Genotype | Conditional Logistic Regressiona | Conditional Logistic Regression Adjusted for Cardiovascular Risk Factorsb | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Evacetrapib (5-Component MACE) vs Placebo | Evacetrapib (3-Component MACE) vs Placebo | Evacetrapib (5-Component MACE) vs Placebo | Evacetrapib (3-Component MACE) vs Placebo | |||||||||
| OR (95% CI) | P Value | OR (95% CI) | P Value | OR (95% CI) | P Value | OR (95% CI) | P Value | |||||
| Interactionc | Trendd | Interactionc | Trendd | Interactionc | Trendd | Interactionc | Trendd | |||||
| AA | 0.88 (0.69- 1.12) |
.17 | .06 | 0.92 (0.67- 1.27) |
.65 | .83 | 0.93 (0.73- 1.19) |
.71 | .59 | 1.06 (0.77- 1.46) |
.17 | .12 |
| AG | 1.04 (0.90- 1.21) |
1.05 (0.86- 1.28) |
1.05 (0.91- 1.22) |
1.08 (0.88- 1.32) |
||||||||
| GG | 1.18 (0.98- 1.41) |
0.92 (0.72- 1.16) |
1.02 (0.85- 1.24) |
0.81 (0.63- 1.03) |
||||||||
Abbreviations: MACE, major adverse cardiovascular events; OR, odd ratio; SNP, single-nucleotide polymorphism.
Adjusted for age, sex, and top 5 principal components in an interaction model with treatment and SNP main effects and treatment-by-SNP interaction effect.
Adjusted for age, sex, top 5 principal components, and risk factors, including baseline apolipoprotein B, history of cerebral vascular disease, history of peripheral arterial disease, history of prior percutaneous coronary intervention, history of prior myocardial infarction, region, race/ethnicity, and smoking status in an interaction model with treatment and SNP main effects and treatment-by-SNP interaction effect.
P value for heterogeneity of treatment effects across the genotypes, using the likelihood ratio test with 2° of freedom.
Trend test performed after coding the genotypes as the number of minor alleles in the conditional logistic model and the corresponding genotype-by-treatment interaction tested.
For comparative purposes, the HRs based on the Cox proportional hazards model reported for dalcetrapib (5-component MACE) are illustrated in eTable 4 in the Supplement. The Figure shows forest plots for the key analyses for the primary end point for the 3 genotypes for both dalcetrapib-treated and evacetrapib-treated patients, and eFigure 5 in the Supplement shows forest plots for evacetrapib-treated patients adjusted for cardiovascular risk factors. eTable 5 in the Supplement shows results for conditional logistic regression analysis in white patients, and eTable 6 in the Supplement shows genotype frequencies for various racial/ethnic subgroups. eTable 7 in the Supplement shows the summary of United Kingdom phenome-wide association study and data from public consortia for rs1967309, in which the P value was less than .05.
Figure. Treatment Effects of Cholesteryl Ester Transfer Protein (CETP) Inhibition Compared With Placebo for 5-Component Major Adverse Cardiac Events (rs1967309 Single-Nucleotide Polymorphism).
ACCELERATE indicates Assessment of Clinical Effects of Cholesteryl Ester Transfer Protein Inhibition with Evacetrapib in Patients at a High Risk for Vascular Outcomes; dal-OUTCOMES, dalcetrapib outcomes; HR, hazard ratio; OR, odds ratio.
aP value for heterogeneity of treatment effects across the genotypes, using the likelihood ratio test with 2° of freedom.
bTrend test performed after coding the genotypes as the number of minor alleles in the conditional logistic model and the corresponding genotype-by-treatment interaction tested.
Table 3 shows the results for conditional logistic regression analysis for the effect of each A allele in both evacetrapib and placebo treatment groups. No significant associations were demonstrated within the evacetrapib treatment group for either 5-component or 3-component MACE. However, the effect of each A allele in the placebo treatment group showed a nominally significant adverse effect on 5-component MACE (OR, 1.15; 95% CI 1.03-1.27), which was not significant when adjusted for cardiovascular risk factors (OR, 1.06; 95% CI, 0.95-1.19), and also not significant for 3-component MACE (unadjusted OR, 1.01; 95% CI, 0.88-1.17) (Table 3).
Table 3. Results From the Conditional Logistic Regression Analyses Within Treatment Group for the Additive Genetic Effect of the Minor Allele of rs1967309.
| Treatment | Genotype | Patients With Events | Patients Without Events | OR (95% CI) | |
|---|---|---|---|---|---|
| Unadjusted | Adjusteda | ||||
| 5-Component MACE | |||||
| Evacetrapib | AA | 125 | 130 | 0.99 (0.89-1.10) | 1.02 (0.91-1.14) |
| AG | 349 | 379 | |||
| GG | 245 | 254 | |||
| Placebo | AA | 143 | 129 | 1.15 (1.03-1.27) | 1.06 (0.95-1.19) |
| AG | 353 | 378 | |||
| GG | 212 | 262 | |||
| 3-Component MACE | |||||
| Evacetrapib | AA | 74 | 181 | 1.03 (0.90-1.20) | 1.02 (0.88-1.18) |
| AG | 191 | 537 | |||
| GG | 131 | 368 | |||
| Placebo | AA | 81 | 191 | 1.01 (0.88-1.17) | 0.87 (0.75-1.00) |
| AG | 191 | 540 | |||
| GG | 143 | 331 | |||
Abbreviations: MACE, major adverse cardiovascular events; OR, odd ratio.
Adjusted for sex, top 5 principal components, and risk factors, including baseline apolipoprotein B, history of cerebral vascular disease, history of peripheral arterial disease, history of prior percutaneous coronary intervention, history of prior myocardial infarction, region, race/ethnicity, and smoking status in an interaction model with treatment and SNP main effects and treatment-by-SNP interaction effect.
Biomarker Analyses
Table 4 shows the least square mean and 95% confidence intervals for changes for the 3 genotypes in biomarkers (LDL-C, HDL-C, and hsCRP) and blood pressure during the ACCELERATE trial in both the evacetrapib and placebo treatment groups. There were no significant differences in the changes in these variables for the 3 rs1967309 genotypes.
Table 4. Effect of rs1967309 Genotype on Biomarkers and Blood Pressure in the ACCELERATE Trial.
| Biomarker | Treatment Group | Total No. | GG/AG/AA, No. | LSM (95% CI) | P Value | ||
|---|---|---|---|---|---|---|---|
| AA Genotype | AG Genotype | GG Genotype | |||||
| Change, LDL-C, mg/dL | Evacetrapib | 1411 | 474/698/239 | −26.4 (−29.3 to −23.4) | −25.4 (−27.16 to −23.63) | −25.24 (−27.32 to −23.15) | .81 |
| Placebo | 1413 | 451/706/256 | 5.19 (2.37 to 8.01) | 3.05 (1.33 to 4.78) | 4.1 (1.95 to 6.24) | .40 | |
| Change, HDL-C, % | Evacetrapib | 1411 | 474/698/239 | 123 (117 to 130) | 126 (122 to 130) | 123 (118 to 127) | .50a |
| Placebo | 1413 | 451/706/256 | 0.4 (−2.3 to 3.2) | 0.2 (−1.5 to 1.9) | 0.01 (−2.1 to 2.1) | .97a | |
| Change, hsCRP, % | Evacetrapib | 1097 | 373/528/196 | 16 (3 to 32) | 14 (5 to 23) | 10 (1 to 20) | .74b |
| Placebo | 1078 | 341/537/200 | 6 (−6 to 19) | 8 (−0.2 to 16) | 6 (−3 to 16) | .95b | |
| Change, DBP, mm Hg | Evacetrapib | 1445 | 488/712/245 | 0.68 (−0.36 to 1.72) | 1.17 (0.55 to 1.8) | 0.55 (−0.18 to 1.29) | .38 |
| Placebo | 1443 | 464/714/265 | −1.15 (−2.14 to −0.16) | −0.12 (−0.73 to 0.5) | 0.34 (−0.41 to 1.1) | .06 | |
| Change, SBP, mm Hg | Evacetrapib | 1445 | 488/712/245 | 0.83 (−0.96 to 2.62) | 2.04 (0.96 to 3.11) | 1.49 (0.22 to 2.75) | .48 |
| Placebo | 1443 | 464/714/265 | −1.84 (−3.55 to −0.14) | −0.31 (−1.37 to 0.74) | −0.38 (−1.68 to 0.92) | .28 | |
Abbreviations: ACCELERATE, Assessment of Clinical Effects of Cholesteryl Ester Transfer Protein Inhibition with Evacetrapib in Patients at a High Risk for Vascular Outcomes; DBP, diastolic blood pressure; HDL, high-density lipoprotein cholesterol; hsCRP, high-sensitivity C-reactive protein; LDL-C, low-density lipoprotein cholesterol; LSM, least square mean; SBP, systolic blood pressure.
SI conversion factor: To convert cholesterol levels to millimoles per liter, multiply by 0.0259; high-sensitivity C-reactive protein to nanomoles per liter, multiply by 9.524.
A sensitivity analysis using Kruskal-Wallis rank sum test yielded a P value of .40 for testing the difference in change of HDL-C in the evacetrapib group and a P value of .68 in the placebo group.
A sensitivity analysis using Kruskal-Wallis rank sum test yielded a P value of .94 for testing the difference in change of hsCRP in the evacetrapib group and a P value of .49 in the placebo group.
Discussion
Genome-wide association study analyses are increasingly performed within large clinical outcome trials studying pharmacological therapies. A key rationale for conducting such studies is the possibility that a specific subgroup determined via GWAS might identify patients who show a more favorable response to the studied therapy or identify subgroups who experience adverse events. A GWAS analysis of the dal-OUTCOMES trial, a cardiovascular outcome trial studying the CETP inhibitor dalcetrapib, generated considerable scientific interest when it reported a 39% relative risk reduction for a variant within the ADCY9 gene (rs1967309) despite a neutral overall result for the trial. An unusual finding was the observation that the AA genotype was associated with benefit, whereas the GG genotype was associated with harm. On the basis of these findings, a new cardiovascular outcome trial (NCT02525939) was initiated to study only patients with the AA genotype. We conducted a cardiovascular outcome trial (ACCELERATE) with the CETP inhibitor evacetrapib, which also showed no cardiovascular benefits for the primary end point, 5-component composite MACE, compared with placebo (HR, 1.01; 95% CI, 0.91-1.11; P = .91).3 This trial provided the opportunity to determine whether the pharmacogenetic finding reported for dalcetrapib could be replicated for a different CETP inhibitor.
Our analysis showed no significant association between rs1967309 variant and cardiovascular outcomes using several analytical methods. For evacetrapib-treated patients compared with control patients, the primary analysis (5-component MACE) showed an OR of 0.88 (95% CI, 0.69-1.12) for patients with the AA genotype, 1.04 (95% CI, 0.90-1.21) for patients with the AG genotype, and 1.18 (95% CI, 0.98-1.41) for patients with the GG genotype (interaction P = .17). An alternative analysis examining the trend across genotypes for the relationship between rs1967309 variant and the association of evacetrapib with cardiovascular outcomes, showed a nonsignificant trend for 5-component MACE (P = .06) (Table 2 and Figure). Neither approach to interaction testing showed a significant relationship when the primary 5-component analysis was adjusted for cardiovascular risk factors, AA genotype (OR, 0.93; 95% CI, 0.73-1.19), and GG genotype (OR, 1.02; 95% CI, 0.85-1.24) (interaction P = .71; trend P = .59). A sensitivity analysis assessing 3-component MACE showed an OR of 0.92 (95% CI, 0.67-1.27) for the AA genotype, 1.05 (95% CI, 0.86-1.28) for the AG genotype, and 0.92 (95% CI, 0.72-1.16) for the GG genotype (interaction P = .65 and trend P = .83) (Table 2; eFigure 5 in the Supplement). The observed ORs in the ACCELERATE trial, although directionally similar, contrast with the significantly lower HR reported for dalcetrapib-treated patients with the AA genotype (0.61; 95% CI, 0.41-0.92) and significantly higher HR for patients with the GG genotype (1.27; 95% CI, 1.02-1.58) (Figure).
Additional evacetrapib analyses also did not show a significant association between the ADCY9 SNP and cardiovascular outcomes. A per allele analysis showed an OR of 0.99 (95% CI, 0.89-1.10) in the evacetrapib treatment group for each A allele for 5-component MACE and an OR of 1.03 (95% CI, 0.90-1.20) for the sensitivity analysis based on 3-component MACE. Analysis of biomarkers, including hsCRP and blood pressure, showed no association between evacetrapib rs1967309 genotype and change from baseline (Table 4).
There are several potential explanations for the differences between evacetrapib pharmacogenetic analyses and findings reported for dalcetrapib. Although both drugs are CETP inhibitors, there may be differences in mechanism of action with respect to either the on-target or off-target effects. Dalcetrapib is a weaker CETP inhibitor, showing approximately a 35% increase in HDL-C compared with 130% for evacetrapib, and dalcetrapib does not lower LDL-C. The patient populations studied in the 2 trials were somewhat different, including only recent acute coronary syndrome in dal-OUTCOMES and patients during a more chronic phase of the coronary disease in ACCELERATE. The 2 analyses used different statistical approaches. The dal-OUTCOMES analysis included 5749 patients selected from a 15 871–patient clinical trial and calculated an HR for each genotype based on a Cox proportional hazard model. The ACCELERATE analysis used a nested case-control design studying 1427 patients with events and 1532 matched control patients using conditional logistic regression analysis to calculate an OR for each genotype. The observed differences in the results from dal-OUTCOMES and ACCELERATE trials may be confounded by association of the SNP with some measured or unmeasured cardiovascular risk factors (mediation). To explore this possibility, we conducted a search of publicly available UK Biobank and consortia data sets. The association summary is shown in eTable 7 in the Supplement. There were no associations with genome-wide significance for any clinical phenotypes and rs1967309 in any of the data sets.
Statistical considerations represent an important issue in pharmacogenetic studies. The large number of analyzed SNPs in GWAS analyses (5.5 million in the dalcetrapib analysis) creates the potential for false discovery. Although the results from the dalcetrapib analyses within treatment reached genome-wide significance, the results comparing the effects within the genotypes and between treatments (genotype by treatment interaction) were only nominally significant. A best practice involves identification of potential genetic effects via a discovery cohort, with subsequent confirmation using a replication cohort. For dalcetrapib, the replication cohort was a small imaging study, and the end point was carotid intima medial thickness, not morbidity-mortality. The SNP identified in the replication cohort was different (rs2238448) from the discovery cohort, although it was in linkage disequilibrium with rs1967309 (r2 = 0.8). Because the number of patients with primary end point events in the dalcetrapib discovery cohort was low, the genetic substudy used a broader secondary end point that included unanticipated coronary revascularizations. Even using this broader end point, the significant genetic results from the dal-OUTCOMES study were driven by relatively sparse number of events in the dalcetrapib-treated AA genotype subgroup (38 events).
Strengths and Limitations
This analysis has strengths and limitations. Strengths include greater number of events in genotyped patients in the evacetrapib analysis (1427 events) compared with the dalcetrapib analysis (788 events), providing potentially more reliable estimates of the genotype frequencies. Limitations of our study include the retrospective nature of the analyses and the differences in the studied CETP inhibitors, the patient population, study design (nested case-control analysis), and relatively short follow-up for the evacetrapib study. Neither the dalcetrapib nor the evacetrapib analysis could be confirmed via a large replication study because only a single outcome trial exists for each drug.
Conclusions
In summary, our analysis does not demonstrate a significant association between the ADCY9 SNP (rs1967309) and clinical benefits from the CETP inhibitor evacetrapib. A nonsignificant trend was observed for the association between the ADCY9 SNP (rs1967309) and clinical benefits from the CETP inhibitor evacetrapib but was far less in magnitude than observed in the pharmacogenetic study with dalcetrapib. This trend did not persist when examining only harder cardiovascular outcomes or when adjusting for cardiovascular risk factors. The completion of the dalcetrapib pharmacogenetics outcome trial should clarify whether this is a false signal or a paradigm-shifting discovery.
eFigure 1. Flowchart Summarizing the Subject- and Variant-Levels Quality Control (QC) Information
eFigure 2. Flow of Patients Through the Pharmacogenetic Analysis
eFigure 3. Principal Components Analysis Results for Top 2 Components (PC1 vs PC2) Based on the Quality-Controlled Data
eFigure 4. Scree Plot Showing Individual and Cumulative Percent Genomic Variance Explained by Top 10 Principal Components
eFigure 5. Treatment Effects of CETP Inhibition Compared with Placebo for 5 Component MACE (rs1967309 SNP)
eTable 1. Baseline Characteristics for Patients Who Underwent Genotyping and All Patients Enrolled in the ACCLERATE Trial
eTable 2. Baseline Characteristics of Patients Selected for Pharmacogenetic Analysis in the dal-OUTCOMES Trial
eTable 3. Genotype Frequencies for the (rs1967309) SNP for the ACCELERATE and Dal-OUTCOMES Trials
eTable 4. Treatment Effects for Dalcetrapib Compared With Placebo Within Each Genotype of rs1967309
eTable 5. Results From the Conditional Logistic Regression Analyses Within Treatment Group for the Additive Genetic Effect of the Minor Allele of rs1967309 in White-only Patients
eTable 6. ADCY9 Variant (rs1967309) Allele Frequencies Across Different Ethnic Groups
eTable 7. Genetic Association Summary of rs1967309 from United Kingdom Biobank (UKB) and Other Consortia Datasets (European Ancestry Where P value was <.05)
References
- 1.Barter PJ, Caulfield M, Eriksson M, et al. ; ILLUMINATE Investigators . Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med. 2007;357(21):2109-2122. [DOI] [PubMed] [Google Scholar]
- 2.Schwartz GG, Olsson AG, Abt M, et al. ; dal-OUTCOMES Investigators . Effects of dalcetrapib in patients with a recent acute coronary syndrome. N Engl J Med. 2012;367(22):2089-2099. [DOI] [PubMed] [Google Scholar]
- 3.Lincoff AM, Nicholls SJ, Riesmeyer JS, et al. ; ACCELERATE Investigators . Evacetrapib and cardiovascular outcomes in high-risk vascular disease. N Engl J Med. 2017;376(20):1933-1942. [DOI] [PubMed] [Google Scholar]
- 4.HPS3/TIMI55–REVEAL Collaborative Group; Bowman L, Hopewell JC, Chen F, et al. Effects of anacetrapib in patients with atherosclerotic vascular disease. N Engl J Med. 2017;377(13):1217-1227. [DOI] [PubMed] [Google Scholar]
- 5.Business Wire . Merck provides update on anacetrapib development program. https://www.businesswire.com/news/home/20171011006286/en/Merck-Update-Anacetrapib-Development-Program. Published October 11, 2017. Accessed February 21, 2018.
- 6.Tardif JC, Rhéaume E, Lemieux Perreault LP, et al. Pharmacogenomic determinants of the cardiovascular effects of dalcetrapib. Circ Cardiovasc Genet. 2015;8(2):372-382. [DOI] [PubMed] [Google Scholar]
- 7.Tardif JC, Rhainds D, Brodeur M, et al. Genotype-dependent effects of dalcetrapib on cholesterol efflux and inflammation: concordance with clinical outcomes. Circ Cardiovasc Genet. 2016;9(4):340-348. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Nicholls SJ, Lincoff AM, Barter PJ, et al. Assessment of the clinical effects of cholesteryl ester transfer protein inhibition with evacetrapib in patients at high-risk for vascular outcomes: Rationale and design of the ACCELERATE trial. Am Heart J. 2015;170(6):1061-1069. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eFigure 1. Flowchart Summarizing the Subject- and Variant-Levels Quality Control (QC) Information
eFigure 2. Flow of Patients Through the Pharmacogenetic Analysis
eFigure 3. Principal Components Analysis Results for Top 2 Components (PC1 vs PC2) Based on the Quality-Controlled Data
eFigure 4. Scree Plot Showing Individual and Cumulative Percent Genomic Variance Explained by Top 10 Principal Components
eFigure 5. Treatment Effects of CETP Inhibition Compared with Placebo for 5 Component MACE (rs1967309 SNP)
eTable 1. Baseline Characteristics for Patients Who Underwent Genotyping and All Patients Enrolled in the ACCLERATE Trial
eTable 2. Baseline Characteristics of Patients Selected for Pharmacogenetic Analysis in the dal-OUTCOMES Trial
eTable 3. Genotype Frequencies for the (rs1967309) SNP for the ACCELERATE and Dal-OUTCOMES Trials
eTable 4. Treatment Effects for Dalcetrapib Compared With Placebo Within Each Genotype of rs1967309
eTable 5. Results From the Conditional Logistic Regression Analyses Within Treatment Group for the Additive Genetic Effect of the Minor Allele of rs1967309 in White-only Patients
eTable 6. ADCY9 Variant (rs1967309) Allele Frequencies Across Different Ethnic Groups
eTable 7. Genetic Association Summary of rs1967309 from United Kingdom Biobank (UKB) and Other Consortia Datasets (European Ancestry Where P value was <.05)