Association of Cholesteryl Ester Transfer Protein Genotypes With CETP Mass and Activity, Lipid Levels, and Coronary Risk
- Alexander Thompson, MRes, MPhil;
- Emanuele Di Angelantonio, MD, MSc;
- Nadeem Sarwar, MRPharmS, MPhil;
- Sebhat Erqou, MD, MPhil;
- Danish Saleheen, MBBS, MPhil;
- Robin P. F. Dullaart, MD, PhD;
- Bernard Keavney, MD, FRCP;
- Zheng Ye, PhD;
- John Danesh, DPhil, FRCP
- Author Affiliations: Department of Public Health and Primary Care, University of Cambridge, Cambridge, England (Drs Di Angelantonio, Erqou, Saleheen, Ye, and Danesh, and Messrs Thompson and Sarwar); Department of Endocrinology, University Medical Center Groningen, University of Groningen, Groningen, the Netherlands (Dr Dullaart); and Institute of Human Genetics, Newcastle University, Newcastle, England (Dr Keavney).
-
Corresponding Author: John Danesh, DPhil, FRCP, Department of Public Health and Primary Care, University of Cambridge, Strangeways Research Laboratory, Worts Causeway, Cambridge, England CB1 8RN (john.danesh@phpc.cam.ac.uk).
Abstract
Context The importance of the cholesteryl ester transfer protein (CETP) pathway in coronary disease is uncertain. Study of CETP genotypes can help better understand the relevance of this pathway to lipid metabolism and disease risk.
Objective To assess associations of CETP genotypes with CETP phenotypes, lipid levels, and coronary risk.
Data Sources Studies published between January 1970 and January 2008 were identified through computer-based and manual searches using MEDLINE, EMBASE, BIOSIS, Science Citation Index, and the Chinese National Knowledge Infrastructure Database. Previously unreported studies were sought through correspondence with investigators.
Study Selection Relevant studies related principally to 3 common (TaqIB [rs708272], I405V [rs5882], and −629C>A [rs1800775]) and 3 uncommon (D442G [rs2303790], −631C>A [rs1800776], and R451Q [rs1800777]) CETP polymorphisms.
Data Extraction Information on CETP genotypes, CETP phenotypes, lipid levels, coronary disease, and study characteristics was abstracted from publications, supplied by investigators, or both.
Results Ninety-two studies had data on CETP phenotypes, lipid levels, or both in 113 833 healthy participants, and 46 studies had data on 27 196 coronary cases and 55 338 controls. For each A allele inherited, individuals with the TaqIB polymorphism had lower mean CETP mass (−9.7%; 95% confidence interval [CI], −11.7% to −7.8%), lower mean CETP activity (−8.6%; 95% CI, −13.0% to −4.1%), higher mean high-density lipoprotein cholesterol (HDL-C) concentrations (4.5%; 95% CI, 3.8%-5.2%), and higher mean apolipoprotein A-I concentrations (2.4%; 95% CI, 1.6%-3.2%). The pattern of findings was very similar with the I405V and −629C>A polymorphisms. The combined per-allele odds ratios (ORs) for coronary disease were 0.95 (95% CI, 0.92-0.99) for TaqIB, 0.94 (95% CI, 0.89-1.00) for I405V, and 0.95 (95% CI, 0.91-1.00) for −629C>A.
Conclusions Three CETP genotypes that are associated with moderate inhibition of CETP activity (and, therefore, modestly higher HDL-C levels) show weakly inverse associations with coronary risk. The ORs for coronary disease were compatible with the expected reductions in risk for equivalent increases in HDL-C concentration in available prospective studies.
- KEYWORDS:
- CHOLESTEROL, HDL
- CORONARY DISEASE
- GENETIC PREDISPOSITION TO DISEASE
- GENOTYPE
- LIPIDS
- RISK FACTORS
Observational and experimental studies in humans and animals have encouraged development of pharmacological agents that increase circulating levels of high-density lipoprotein cholesterol (HDL-C) in coronary disease prevention.1,2,3,4,5,6,7,8 Such agents include inhibitors of cholesteryl ester transfer protein (CETP), a protein that facilitates exchange of cholesteryl esters for triglycerides between HDL and triglyceride-rich lipoproteins.5 One CETP inhibitor, torcetrapib, increases HDL-C levels by at least 60%,9 but produced an excess of deaths and cardiovascular events in the ILLUMINATE trial.10 Although the exact reasons for the failure of torcetrapib remain uncertain,11,12,13,14,15,16 study of CETP genotypes may help to suggest whether further efforts to prevent coronary disease by CETP inhibition are warranted. Recent genome-wide association studies have reported that CETP genotypes are associated with HDL-C levels more strongly than any other loci across the genome.17,18 Furthermore, whereas the effect of a pharmacological agent on the CETP pathway may be difficult to disentangle from any “off-target” effects on other pathways, particular CETP genotypes should have more specific influences, notably on CETP mass and activity.
A previous meta-analysis19 of the association between CETP and coronary risk focused on only 1 CETP genotype in 10 studies involving a total of 13 677 participants, including 2857 coronary cases. The current reassessment of the associations of 6 CETP genotypes with CETP phenotypes, circulating lipid levels, and with coronary risk uses the following approaches to maximize power and minimize bias: (1) we report updated meta-analyses of CETP genotypes with CETP mass and activity, HDL-C, low-density lipoprotein cholesterol (LDL-C), triglycerides, and apolipoproteins A-I and B (involving data on up to 113 833 participants in 92 studies); (2) we report updated meta-analyses of CETP genotypes with coronary outcomes (involving data on up to 27 196 coronary cases and 55 338 controls in 46 studies), with tabular data sought from investigators to supplement and update published data; (3) we contacted principal investigators of larger genetic association studies of variants other than CETP to seek unreported data; and (4) we compared the association between genetically mediated increases in HDL-C concentrations and coronary risk with those expected for equivalent increases in HDL-C concentrations in available prospective studies.
METHODS
Study Selection
We sought studies published between January 1970 and January 2008 on CETP genotype associations (GenBank accession number NM_000078) with CETP mass, CETP activity, concentrations of HDL-C, LDL-C, triglycerides, or apolipoproteins A-I and B, or with risk of myocardial infarction (generally defined by World Health Organization Multinational Monitoring of Trends and Determinants in Cardiovascular Disease [MONICA] criteria)20 or angiographic coronary stenosis (generally defined as ≥50% of ≥ 1 major coronary arteries). For lipid markers, data were used from only apparently healthy individuals (ie, people without known coronary or other diseases or clinical lipid abnormalities) who had information on at least 1 relevant genotype. Electronic searches, not limited to the English language, were performed by using MEDLINE, EMBASE, BIOSIS, Science Citation Index, and the Chinese National Knowledge Infrastructure Database, and supplemented by scanning reference lists of articles identified for all relevant studies and review articles (including meta-analyses), by hand searching of relevant journals, and by correspondence with authors of included studies. The computer-based searches combined search terms related to CETP genotype (eg, cholesteryl ester transfer protein, cholesterol ester transfer protein, CETP, gene, genes, loci, polymorphi*, allel*, phenotyp*, SNP*, RFLP*, chromosom*, variant, mutat*), lipid phenotypes (eg, HDL, LDL, triglycerides, apolipoproteins), and coronary disease (eg, myocardial infarction, atherosclerosis, coronary heart disease, coronary stenosis) without language restriction (Figure 1).
CETP indicates cholesteryl ester transfer protein.
aBecause these studies tended to be smaller, they comprised a total of only approximately 3% of the overall number of coronary
cases included in this review, and a total of only approximately 8% of the overall number of participants included in the
analysis of lipid concentrations.
Data Extraction
Information was recorded on TaqIB (rs708272), I405V (rs5882), −629C>A (rs1800775), D442G (rs2303790), −631C>A (rs1800776), and R451Q (rs1800777). The following data were extracted independently by 3 investigators (S.E., D.S., and Z.Y.), using a protocol previously described21: genotype frequencies by categorical disease outcome; means and standard deviations of lipid markers by genotype; mean age of cases; proportions of men and ethnic subgroups (defined as people of white European continental ancestry, East Asian, or other [South Asian, Middle Eastern, South American, and North African]); and fasting status and assay methods. Discrepancies were resolved by discussion and by adjudication of a fourth reviewer (A.T.). We used the most up-to-date information in cases of multiple publications. We supplemented published data by a tabular data request to (1) authors of published reports, (2) investigators of 75 potentially relevant unreported studies involving at least 500 coronary cases or at least 1000 healthy participants listed in published meta-analyses21,22,23,24,25 of variants other than CETP, and (3) authors of published genome-wide association studies of relevant outcomes.26,27
Statistical Analysis
Analyses were performed by using only within-study comparisons to limit possible biases, involving studies that had used accepted genotyping and lipid assay methods and coronary outcomes (as described above). Principal analyses were prespecified to involve codominant genetic models. Summary odds ratios (ORs) for coronary disease and mean levels of lipid markers (and differences in mean levels in comparison with the common homozygotes) were calculated by using a random-effects model that included between-study heterogeneity (with sensitivity analyses involving fixed-effect models). To enable comparison of the magnitude of any associations with several different lipid markers and CETP phenotypes, associations were also presented as per-allele percentage differences (calculated in reference to the weighted mean level of each marker in common homozygotes). For studies that compared a single control group with both myocardial infarction and (nonoverlapping) coronary stenosis cases, we avoided any double counting by analyzing myocardial infarction and coronary stenosis cases separately before combining them into a single coronary disease group. Consistency of findings across studies was assessed by using the I2 statistic.28 Heterogeneity was assessed by using the Q statistic and by examining prespecified groupings of study characteristics. Evidence of publication bias was assessed by using funnel plots, the Egger test,29 and by comparing pooled results from studies involving at least 500 coronary cases (or ≥1000 healthy participants for gene-lipid investigations) with pooled results from smaller studies. Evidence of deviation from Hardy-Weinberg equilibrium was assessed by χ2 tests using the genotype frequencies in healthy individuals, with sensitivity analyses involving significance levels of both P < .05 and P < .01. Coronary heart disease risk estimates from the Prospective Studies Collaboration1 were used to calculate expected ORs for coronary disease corresponding to the per-allele increases in HDL-C levels observed with the CETP genotypes in this review, involving a χ2 test for consistency of genotype-coronary risk associations with HDL-C coronary risk associations. All analyses were performed by using Stata release 10 (StataCorp LP, College Station, Texas). All statistical tests were 2-sided and used a significance level of P < .05, except where indicated.
RESULTS
A total of 102 relevant studies reporting on 147 599 individuals were identified, including 38 studies that provided supplementary or previously unreported tabular data (Figure 1, eTable 1, and eTable 2).17,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161,162,163,164,165 Forty-five studies were based in Europe, 15 in North America, 29 in East Asia (predominantly China and Japan), 11 in other regions, and 2 studies were multinational. Twenty-one studies were prospective in design and 81 were either cross-sectional or case-control. The minor allele frequency in healthy white individuals was 0.42 for TaqIB (rs708272), 0.35 for I405V (rs5882), 0.48 for −629C>A (rs1800775), less than 0.01 for D442G (rs2303790), 0.08 for −631C>A (rs1800776), and 0.04 for R451Q (rs1800777) (Table 1). Previous studies have reported almost complete linkage of TaqIB with −629C>A, but the other CETP genotypes considered in this review appear to be only weakly correlated.166
Table 1. Description of CETP Genotypes Included in the Review
CETP Phenotypes and Lipid Levels
Ninety-two studies of CETP genotypes involving 113 833 individuals were identified that reported on CETP mass, CETP activity, circulating lipid levels, or all 3,17,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148 including 36 studies that provided supplementary or previously unreported data on 49 502 individuals (Table 2). Of 91 studies assessing associations with HDL-C, 7 used homogeneous assay methods to measure HDL-C levels, 49 used nonhomogeneous methods, and 35 did not report the assay method used. Overall, for each A allele inherited, carriers of the TaqIB variant had lower mean CETP mass (−9.7%; 95% confidence interval [CI], −11.7% to −7.8%), lower mean CETP activity (−8.6%; 95% CI, −13.0% to −4.1%), higher mean HDL-C levels (4.5%; 95% CI, 3.8%-5.2%), higher mean apolipoprotein A-I levels (2.4%; 95% CI, 1.6%-3.2%), lower mean LDL-C levels (−0.9%; 95% CI, −1.6% to −0.3%), lower mean apolipoprotein B levels (−0.5%; 95% CI, −1.1% to 0.1%), and lower mean triglycerides (−2.0%; 95% CI, −3.2% to −0.7%) than did common homozygotes (Figure 2). Overall, per G allele inherited, carriers of the I405V variant had lower mean CETP mass (−5.7%; 95% CI, −7.5% to −4.0%), lower mean CETP activity (−8.2%; 95% CI, −17.8% to 1.3%), higher mean HDL-C levels (2.5%; 95% CI, 1.8%-3.2%), higher mean apolipoprotein A-I levels (1.6%; 95% CI, 1.2%-2.0%), and lower mean triglycerides (−2.1%; 95% CI, −3.0% to −1.1%) than did common homozygotes but no discernible differences in LDL-C or apolipoprotein B levels. Overall, per A allele inherited, carriers of the −629C>A variant had lower mean CETP mass (−9.4%; 95% CI, −13.8% to −5.0%), lower mean CETP activity (−5.9%; 95% CI, −11.0% to −0.8%), higher mean HDL-C levels (4.9%; 95% CI, 4.3%-5.4%), higher mean apolipoprotein A-I levels (1.8%; 95% CI, 0.7%-2.8%), and lower mean triglycerides (−2.1%; 95% CI, −3.4% to −0.9%) than did common homozygotes but no discernible differences in LDL-C or apolipoprotein B levels. Data were insufficient for informative per-allele estimates in relation to D442G, −631C>A, and R451Q (Table 2); however, in dominant models, they were associated with mean differences in HDL-C of 13.4% (95% CI, 9.4%-17.4%), −0.7% (95% CI, −2.4% to 1.0%), and −8.8% (95% CI, −9.7% to −7.9%), respectively, compared with common homozygotes.
Table 2. Summary of Data Available on CETP Genotypes, CETP Phenotypes, Lipid Levels, and Coronary Outcomesa
CETP indicates cholesteryl ester transfer protein; CI, confidence interval; HDL-C, high-density lipoprotein cholesterol; LDL-C,
low-density lipoprotein cholesterol. To convert apolipoproteins A-I and B to mg/dL, divide by 0.01; to convert HDL-C and LDL-C
to mg/dL, divide by 0.0259; and to convert triglyercides to mg/dL, divide by 0.0113. Assessment of heterogeneity: I2 (95% CI) for CETP mass, CETP activity, HDL-C, apolipoprotein A-I, LDL-C, apolipoprotein B, and triglycerides, respectively,
were 66% (39%-81%), 71% (44%-86%), 75% (69%-80%), 66% (46%-78%), 51% (32%-65%), 14% (0%-51%), and 49% (30%-62%) for TaqIB; 0% (0%-71%), NA*, 56% (33%-71%), 0% (0%-68%), 24% (0%-58%), 16% (0%-60%), and 0% (0%-49%) for I405V; and 71% (17%-90%),
NA*, 37% (0%-61%), 36% (0%-78%), 29% (0%-63%), 0% (0%-90%), and 0% (0%-57%) for −629C>A. NA* indicates I2 statistics were not calculated when there were only 2 studies.
aPooled estimates calculated by random-effects models. Estimates calculated by fixed-effect models are shown in eTable 3.
bStandardized mean differences.
cCalculated with reference to the weighted mean level of each marker in common homozygotes.
There was evidence of heterogeneity in associations with HDL-C across studies (TaqIB: I2 = 75%; 95% CI, 69%-80%; I405V: I2 = 56%; 95% CI, 33%-71%; −629C>A: I2 = 37%; 95% CI, 0%-61%). This heterogeneity was partly explained by study level characteristics that had been recorded, including population source (TaqIB and −629C>A), ethnicity (TaqIB and −629C>A), study size (I405V), and whether data were obtained through correspondence with investigators or were extracted directly from published reports (TaqIB) (Figure 3). Significant deviation from Hardy-Weinberg equilibrium (P < .05) was detected in 10 studies for TaqIB (including 5 studies at P < .01), 2 studies for I405V (both studies at P < .01), and 4 studies for −629C>A (including 2 studies at P < .01), but their exclusion did not materially change the findings. Analyses by study size (Figure 3) and other standard tests (eFigure) did not reveal strong evidence of publication bias.
CETP indicates cholesteryl ester transfer protein; CI, confidence interval; HDL-C, high-density lipoprotein cholesterol. To convert HDL-C to mg/dL, divide by 0.0259. Sizes of data markers are proportional to the inverse of the variance of the weighted mean difference. For sex and ethnicity, studies may have contributed data to more than 1 category. Overall estimates were calculated using random-effects models (fixed-effect estimates are provided in eTable 3). Several recorded characteristics explained part of the heterogeneity observed, including ethnicity (P = .008), population source (P = .04), and data source (P < .001) for TaqIB; study size (P = .02) for I405V; and ethnicity (P < .001) and population source (P = .007) for −629C>A.
Coronary Outcomes
Forty-six studies reported data on 27 196 coronary cases and 55 338 controls,30,31,32,33,34,35,39,42,43,44,45,51,52,53,59,61,63,64,65,66,67,68,71,78,79,80,81,85,86,87,90,94,98,99,100,103,105,107,108,109,110,111,113,114,115,117,118,119,120,121,122,123,133,135,138,139,140,142,143,145,146,147,148,150,151,152,153,154,155,156,157,158,159,160,161,162,163,164,165 including 21 studies that provided supplementary or previously unreported data on 17 187 cases and 38 619 controls. Thirty-three studies recruited controls from general populations, 11 studies recruited controls from hospitals or were based in clinical trials, and 2 studies used both sources. The combined OR for coronary disease was 0.95 (95% CI, 0.92-0.99) per A allele of the TaqIB variant. There was possible modest heterogeneity among the 38 available studies (I2 = 18%; 95% CI, 0%-45%; P = .17), including somewhat more modest findings in the larger studies and those studies that had provided data through correspondence (P = .01 and P = .003, respectively, for heterogeneity) (Figure 4). The combined OR for coronary disease was 0.94 (95% CI, 0.89-1.00) per G allele of the I405V variant. There was possible modest heterogeneity among the 18 available studies (I2 = 39%; 95% CI, 0%-66%; P = .04), with most of it accounted for by differences in the selection of control groups (P < .001). The combined OR for coronary disease was 0.95 (95% CI, 0.91-1.00) per A allele of the −629C>A variant. There was possible moderate heterogeneity among the 17 available studies (I2 = 32%; 95% CI, 0%-62%; P = .10), but little of it was explained by the study characteristics recorded.
CETP indicates cholesteryl ester transfer protein; CI, confidence interval. Sizes of data markers are proportional to the inverse of the variance of the loge odds ratio. For ethnicity, source of controls, and outcome assessed, studies may have contributed data to more than 1 category. For ethnicity, results are not presented for 4 studies of TaqIB and 2 studies of I405V and −629C>A that were predominantly based in nonwhite, non−East Asian individuals. For outcome assessed in TaqIB, results are not presented for 1 study that did not provide genotype frequencies separately for cases of myocardial infarction and coronary stenosis. Assessment of heterogeneity: TaqIB (I2 = 18%; 95% CI, 0%-45%), I405V (I2 = 39%; 95% CI, 0%-66%), or −629C>A (I2 = 32%; 95% CI, 0%-62%). Observed heterogeneity could be partially explained by study size (P = .01) and data source (P = .003) for TaqIB and by source of controls (P < .001) for I405V (other comparisons P > .05 for each). Overall estimates were calculated using random-effects models; those calculated using fixed-effect models were 0.96 (95% CI, 0.93-0.99) for TaqIB, 0.95 (95% CI, 0.92-0.99) for I405V, and 0.95 (95% CI, 0.91-0.99) for −629C>A.
Significant deviation from Hardy-Weinberg equilibrium (P < .05) was detected in the controls of 6 studies for TaqIB (including 2 studies at P < .01), 1 study for I405V (P < .01), and 3 studies for −629C>A (including 1 study at P < .01). Again, exclusion of these studies did not materially change the findings. As was the case for studies of CETP phenotypes and lipid levels, there was not good evidence of publication bias from standard tests (eFigure), with the possible exception of TaqIB (Figure 4). Data were insufficient to provide informative risk estimates for the 3 rare CETP genotypes. Figure 5 displays estimates of associations between HDL-C and coronary risk derived from the Prospective Studies Collaboration,1 providing a comparison of ORs for coronary disease per-allele increases in HDL-C in genetic studies vs those ORs in prospective studies of HDL-C (for TaqIB, 0.95; 95% CI, 0.92-0.99 vs 0.92; 95% CI, 0.91-0.93; P for heterogeneity = .11; for I405V, 0.94; 95% CI, 0.89-1.00 vs 0.95; 95% CI, 0.94-0.97; P = .72; and for −629C>A, 0.95; 95% CI, 0.91-1.00 vs 0.92; 95% CI, 0.91-0.93; P = .19).
CETP indicates cholesteryl ester transfer protein; CI, confidence interval; HDL-C, high-density lipoprotein cholesterol. Sizes
of data markers are proportional to the inverse of the variance of the loge risk estimate. χ2 Test for difference: P = .11 for TaqIB, P = .72 for I405V, and P = .19 for −629C>A.
aPer-allele odds ratio for coronary disease as shown in Figure 4.
bHazard ratios for coronary heart disease calculated by using risk estimates from 153 798 participants in 61 studies1 for an increase in usual HDL-C levels equal to those observed per allele for TaqIB, I405V, and −629C>A (Figure 2).
COMMENT
Study of CETP genotypes that alter CETP mass and activity can help to clarify the relevance of the CETP pathway to lipid metabolism and coronary disease risk. However, because particular CETP genotypes are only modestly associated with CETP phenotypes and lipid levels, reliable genetic studies may require many thousands of participants, including large numbers of patients with coronary disease. In the absence of very large individual studies, we have conducted an updated meta-analysis that involves a total of more than 147 000 individuals (including more than 27 000 coronary cases). We demonstrated that common CETP genotypes decrease CETP mass and activity by approximately 5% to 10%, increase HDL-C and apolipoprotein A-I by approximately 3% to 5% (which is comparable with the observed differences in HDL-C between smokers and nonsmokers), and decrease triglycerides by approximately 2%. Associations with LDL-C and apolipoprotein B were even smaller or negligible. We showed that the same CETP genotypes that are modestly associated with increased HDL-C levels are weakly and inversely associated with coronary risk. The magnitude of per-allele risk reductions were compatible with those expected on the basis of associations observed between HDL-C levels and coronary risk in prospective studies (Figure 5).1 As discussed below, however, the quantity and quality of available genetic data require careful consideration.
Compared with a previous meta-analysis19 that focused solely on associations of the TaqIB variant with HDL-C levels and coronary disease risk, our review considers 6 CETP genotypes, assesses associations with several different lipid markers, and involves more than 10 times as much data. These data provide greater power than previously available to quantify the magnitude of any associations and, by including a considerable amount of previously unreported data, should reduce the scope for publication bias.21 However, although we did not detect strong evidence of selective publication, it is difficult to discount such bias entirely, particularly given the weak associations observed, the reliance of these associations on pooling results from both larger and smaller studies, and the general insensitivity of statistical tests for publication bias. (Ideally, with the availability of even larger numbers, pooled analyses would involve data from only the larger studies, which should be less liable to publication bias.) Because we did not have access to individual data, we could not control for population stratification,167 conduct mendelian randomization analyses,168,169 adjust for variables in possible intermediate pathways, explore heterogeneity by individual-level characteristics, or conduct haplotype analyses. These considerations highlight the need for very large individual studies of CETP genotypes with concomitant assessment of CETP phenotypes and lipid levels, perhaps involving more than 10 000 coronary cases and a similar number of controls. Furthermore, large-scale studies with lifestyle data will be needed to explore potential joint effects of CETP genotypes with environmental determinants of HDL-C levels (eg, exercise and alcohol) on risk of coronary disease. Even larger such studies will be needed for reliable assessment of the less common CETP genotypes.170
In contrast with available evidence from relatively short-term randomized trials of CETP inhibition,10,14 our analyses suggest that individuals with (presumably lifelong) increased HDL-C levels as a result of genetically mediated reductions in CETP may be at slightly reduced coronary risk. This apparent discrepancy may relate to “off-target” effects potentially specific to torcetrapib (notably interference with the renin-angiotensin system and blood pressure elevation).10 On the other hand, it has been suggested that HDL particles produced under conditions of CETP inhibition may be dysfunctional, with any apparent increase in HDL-C levels offset by compensatory HDL-C clearance through direct hepatic pathways11 and reduced apolipoprotein A-I–mediated removal of intracellular-free cholesterol from macrophages.11,12 Further trials of other CETP inhibitors171 and investigations of CETP genotypes in relation to blood pressure and other traits may help to address such mechanistic concerns.
CONCLUSION
In summary, 3 CETP genotypes are associated with moderate inhibition of CETP activity, modestly higher HDL-C levels, and weakly inverse associations with coronary risk. This study illustrates the need for larger studies to demonstrate the modest impact that single genetic variants have on complex outcomes such as coronary disease. Further studies are warranted to determine the value of CETP inhibition to coronary disease prevention.
Author Contributions: Dr Danesh and Mr Thompson had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Thompson, Di Angelantonio, Sarwar, Dullaart, Ye, Danesh.
Acquisition of data: Thompson, Di Angelantonio, Sarwar, Erqou, Saleheen, Dullaart, Ye, Danesh.
Analysis and interpretation of data: Thompson, Di Angelantonio, Sarwar, Erqou, Keavney, Ye, Danesh.
Drafting of the manuscript: Thompson, Di Angelantonio, Sarwar, Danesh.
Critical revision of the manuscript for important intellectual content: Thompson, Di Angelantonio, Sarwar, Erqou, Saleheen, Dullaart, Keavney, Ye, Danesh.
Statistical analysis: Thompson, Di Angelantonio, Sarwar, Erqou.
Obtained funding: Danesh.
Administrative, technical, or material support: Thompson, Di Angelantonio, Erqou, Saleheen, Dullaart, Ye, Danesh.
Study supervision: Danesh.
Drs Di Angelantonio and Erqou, and Messrs Thompson and Sarwar contributed equally to this article and are considered joint first authors.
Financial Disclosures: None reported.
Funding/Support: This work was supported by a British Heart Foundation program grant. Dr Danesh has been supported by the Raymond and Beverly Sackler Award in the Medical Sciences. Dr Di Angelantonio and Mr Thompson are supported by, and Mr Sarwar was supported by, a UK Medical Research Council PhD studentship. Dr Erqou is supported by a Gates Cambridge Scholarship and the UK Overseas Research Trust. Dr Saleheen is supported by the Yousef Jameel Foundation. Aspects of this work have been supported by an unrestricted educational grant from GlaxoSmithKline.
Role of the Sponsors: The sponsors had no role in the design and conduct of the study, in the collection, management, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript.
Additional Contributions: Rory Collins, FMedSci, FRCP (University of Oxford, Oxford, England), commented helpfully. Angela Harper (University of Cambridge, Cambridge, England) provided administrative support. The following investigators kindly provided additional information from their studies: Birgit Agerholm-Larsen, MSc, PhD, Herlev University Hospital, Herlev, Denmark; Nassr Al-Daghri, MD, Birmingham Heartlands Hospital, Birmingham, England; Rolf V. Andersen, MSc, PhD, Department of Clinical Biochemistry, Herlev University Hospital, Herlev, Denmark; Yasumichi Arai, PhD, Keio University School of Medicine, Tokyo, Japan; Gil Atzmon, PhD, Albert Einstein College of Medicine, New York, New York; Nir Barzilai, MD, Albert Einstein College of Medicine, New York, New York; Anja Bauerfeind, PhD, Humboldt University of Berlin, Berlin, Germany; Susanna E. Borggreve, MD, University Medical Center Groningen, Groningen, the Netherlands; Michiel L. Bots, MD, PhD, Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht, Utrecht, the Netherlands; Hannia Campos, PhD, Harvard School of Public Health, Boston, Massachusetts; Peter Clifton, MD, PhD, CSIRO Human Nutrition, Adelaide, Australia; Eliana C. de Faria, MD, MS, PhD, State University of Campinas, Sao Paulo, Brazil; Robin P. F. Dullaart (on behalf of the PREVEND Study Group), MD, PhD, University Medical Center Groningen, Groningen, the Netherlands; Moses Elisaf, MD, University of Ioannina, Ioannina, Greece; Jeanette Erdmann, PhD, Medizinische Klinik II, Lübeck University, Lübeck, Germany; Dilys Freeman, PhD, University of Glasgow, Glasgow, Scotland; Domenico Girelli, MD, PhD, University of Verona, Verona, Italy; Akitomo Goto, MD, Nagoya City University, Nagoya, Japan; John Griffin, PhD, The Scripps Research Institute, La Jolla, California; Wendy Hall, PhD, King's College, London, England; Mohamed Hammami, MD, Monastir University, Monastir, Tunisia; A. Geert Heidema, MSc, University of Maastricht, Maastricht, the Netherlands; Benjamin D. Horne (on behalf of the Intermountain Heart Collaborative Study Group), PhD, MPH, Intermountain Medical Center, Murray, Utah; Akihiro Inazu, MD, PhD, Kanazawa University, Kanazawa, Japan; Aaron Isaacs, DSc, Genetic Epidemiology Unit, Erasmus MC University Medical Center, Rotterdam, the Netherlands; Turgay Isbir, MD, University of Istanbul, Istanbul, Turkey; Yangsoo Jang, MD, PhD, Division of Cardiology, Cardiovascular Genome Center, Yonsei Medical Institute, Yonsei University, Seoul, South Korea; Majken K. Jensen, MSc, Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts; Bernard Keavney, MD, FRCP (on behalf of the ISIS Collaborative Group), Newcastle University, Newcastle, England; Kathy Klos, PhD, University of Texas–Houston Health Science Center, Houston, Texas; Jong Ho Lee, PhD, Yonsei University Research Institute of Science for Aging, Yonsei University, Seoul, South Korea; Robert W. Mahley, MD, PhD, University of California, San Francisco; Massimo Mangino, PhD, University of Leicester, Leicester, England; Nicola Martinelli, MD, University of Verona, Verona, Italy; Pamela McCaskie, BSc(Hons), University of Western Australia, Perth, Australia; Manjari Mukherjee, MD, India N.H. Hospital, Bangalore, India; Børge G. Nordestgaard, MD, DMSc, Herlev University Hospital, Herlev, Denmark; Kenji Okumura, MD, Nagoya University School of Medicine, Nagoya, Japan; Oliviero Olivieri, MD, University of Verona, Verona, Italy; Natalie Pecheniuk, PhD, The Scripps Research Institute, La Jolla, California; Jogchum Plat, PhD, Maastricht University, Maastricht, the Netherlands; Qin Qin, MD, Tianjin Chest Hospital, Tianjin, China; Eric B. Rimm, ScD, Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts; Nilesh Samani, MD, FRCP, University of Leicester, Leicester, England; José V. Sorli, MD, University of Valencia, Valencia, Spain; John F. Thompson, PhD, Pfizer Global Research and Development, Groton, Connecticut; Martin Tobin, PhD, University of Leicester, Leicester, England; Anne Tybjærg-Hansen, MD, DMSc, Copenhagen University Hospital, Copenhagen, Denmark; Yvonne van der Schouw, PhD, Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht, Utrecht, the Netherlands; Cornelia M. van Duijn, PhD, Genetic Epidemiology Unit, Erasmus MC University Medical Center, Rotterdam, the Netherlands; Mitsuhiro Yokota, MD, PhD, FACC, Aichi-Gakuin University, Nagoya, Japan; Shinji Yokoyama, MD, PhD, FRCPC, Nagoya City University, Nagoya, Japan; Mohammad Hadi Zafarmand, MD, Julius Centre for Health Sciences and Primary Care, University Medical Centre Utrecht, Utrecht, the Netherlands; Guo-Bing Zhang, MD, Shanghai First People’s Hospital, Shanghai, China. None of these individuals received any compensation.
REFERENCES
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.
- 9.
- 10.
- 11.
- 12.
- 13.
- 14.
- 15.
- 16.
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25.
- 26.
- 27.
- 28.
- 29.
- 30.
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40.
- 41.
- 42.
- 43.
- 44.
- 45.
- 46.
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53.
- 54.
- 55.
- 56.
- 57.
- 58.
- 59.
- 60.
- 61.
- 62.
- 63.
- 64.
- 65.
- 66.
- 67.
- 68.
- 69.
- 70.
- 71.
- 72.
- 73.
- 74.
- 75.
- 76.
- 77.
- 78.
- 79.
- 80.
- 81.
- 82.
- 83.
- 84.
- 85.
- 86.
- 87.
- 88.
- 89.
- 90.
- 91.
- 92.
- 93.
- 94.
- 95.
- 96.
- 97.
- 98.
- 99.
- 100.
- 101.
- 102.
- 103.
- 104.
- 105.
- 106.
- 107.
- 108.
- 109.
- 110.
- 111.
- 112.
- 113.
- 114.
- 115.
- 116.
- 117.
- 118.
- 119.
- 120.
- 121.
- 122.
- 123.
- 124.
- 125.
- 126.
- 127.
- 128.
- 129.
- 130.
- 131.
- 132.
- 133.
- 134.
- 135.
- 136.
- 137.
- 138.
- 139.
- 140.
- 141.
- 142.
- 143.
- 144.
- 145.
- 146.
- 147.
- 148.
- 149.
- 150.
- 151.
- 152.
- 153.
- 154.
- 155.
- 156.
- 157.
- 158.
- 159.
- 160.
- 161.
- 162.
- 163.
- 164.
- 165.
- 166.
- 167.
- 168.
- 169.
- 170.
- 171.
