Cholesteryl Ester Transfer Protein and Mortality in Patients Undergoing Coronary Angiography
The Ludwigshafen Risk and Cardiovascular Health Study
Background— The role of cholesteryl ester transfer protein (CETP) in the development of atherosclerosis is still open to debate. In the Investigation of Lipid Level Management to Understand its Impact in Atherosclerotic Events (ILLUMINATE) trial, inhibition of CETP in patients with high cardiovascular risk was associated with increased high-density lipoprotein levels but increased risk of cardiovascular morbidity and mortality. In this report, we present a prospective observational study of patients referred to coronary angiography in which CETP was examined in relation to morbidity and mortality.
Methods and Results— CETP concentration was determined in 3256 participants of the Ludwigshafen Risk and Cardiovascular Health (LURIC) study who were referred to coronary angiography at baseline between 1997 and 2000. Median follow-up time was 7.75 years. Primary and secondary end points were cardiovascular and all-cause mortality, respectively. CETP levels were higher in women and lower in smokers, in diabetic patients, and in patients with unstable coronary artery disease, respectively. In addition, CETP levels were correlated negatively with high-sensitivity C-reactive protein and interleukin-6. After adjustment for age, sex, medication, coronary artery disease status, cardiovascular risk factors, and diabetes mellitus, the hazard ratio for death in the lowest CETP quartile was 1.33 (1.07 to 1.65; P=0.011) compared with patients in the highest CETP quartile. Corresponding hazard ratios for death in the second and third CETP quartiles were 1.17 (0.92 to 1.48; P=0.19) and 1.10 (0.86 to 1.39; P=0.46), respectively.
Conclusions— We interpret our data to suggest that low endogenous CETP plasma levels per se are associated with increased cardiovascular and all-cause mortality, challenging the rationale of pharmacological CETP inhibition.
Received April 24, 2009; accepted November 13, 2009.
Cholesteryl ester transfer protein (CETP) is a key player in the metabolic interaction between high-density lipoprotein (HDL) particles and triglyceride-rich lipoproteins, with its main function being to redistribute between lipoproteins the core lipids cholesteryl esters and triglycerides.1 CETP was considered a potential therapeutic target when rodents known to exhibit high levels of HDL and resistance to diet-induced atherosclerosis were found to lack plasma CETP activity. However, evidence supporting the potential benefit of CETP inhibition to prevent atherosclerosis is not straightforward.2 In rabbits, CETP inhibition reduced atherosclerosis, but studies in patients with CETP mutations as well as studies in transgenic mice were inconsistent. Pilot studies employing
Clinical Perspective on p 374
CETP inhibitors showed promising effects on lipids with highly elevated HDL cholesterol and reduced low-density lipoprotein (LDL) cholesterol levels without serious adverse effects3,4 However, a large, randomized, double-blind phase III clinical trial with the CETP inhibitor torcetrapib (Investigation of Lipid Level Management to Understand its Impact in Atherosclerotic Events [ILLUMINATE]) revealed that the increase in HDL cholesterol in the active treatment group was associated with an increased risk of cardiovascular events and death, causing premature termination of the trial.5 The authors speculated that the observed adverse clinical result may have been due to an off-target increase in blood pressure by torcetrapib, but CETP inhibition per se could not be excluded as cause. Thus, HDL-directed pharmacological intervention involving CETP has become the focus of debate.6–8 In the present study, we investigated the role of CETP in atherosclerosis further by relating endogenous CETP plasma levels to coronary artery disease (CAD) and mortality in the cohort of the Ludwigshafen Risk and Cardiovascular Health (LURIC) study, a prospective observational study of patients at intermediate to high cardiovascular risk.9
Study Design and Participants
We studied participants of the LURIC study.9 Inclusion criteria were as follows: German ancestry, clinical stability except for acute coronary syndromes, and the availability of a coronary angiogram. The indications for angiography in individuals in clinically stable condition were chest pain and/or noninvasive test results consistent with myocardial ischemia. Individuals suffering from acute illness other than acute coronary syndromes, chronic noncardiac diseases, or malignancy within the past 5 years and subjects unable to understand the purpose of the study were excluded. The study was approved by the Ethics Committee at the Aerztekammer Rheinland-Pfalz. Informed written consent was obtained from all participants.
CAD was assessed by angiography, with maximum luminal narrowing estimated by visual analysis. Clinically relevant CAD was defined as the occurrence of ≥1 stenosis of ≥20% in ≥1 of 15 coronary segments. Individuals with stenoses <20% were considered as not having CAD.
Diabetes mellitus was diagnosed when plasma glucose was >1.25 g/L in the fasting state or >2.00 g/L 2 hours after an oral glucose load10 or when antidiabetic medical treatment was prescribed. Hypertension was diagnosed when the systolic and/or diastolic blood pressure exceeded 140 and/or 90 mm Hg, respectively, or when a patient was on antihypertensive medication. Data of CETP plasma concentration, plasma lipids, and lipoprotein parameters, as well as coronary angiograms, were complete in all 3256 individuals included in this study.
Information on vital status was obtained from local registries. No patient was lost during follow-up. Of the 3256 persons studied, 754 deaths (23.2%) occurred during a median follow-up of 7.75 years. Cardiovascular death included sudden death, fatal myocardial infarction, death due to congestive heart failure, death immediately after intervention to treat CAD, fatal stroke, and other causes of death due to CAD. Cause of death of 24 individuals was unknown. These patients were included in calculations of all-cause mortality (n=754) but not in calculations considering different causes of death (n=730).
To perform all analyses, fasting blood samples were collected before angiography. The standard laboratory methods have been described.9 CETP was determined with the use of an enzyme-linked immunosorbent assay employing a CETP-specific recombinant single-chain antibody as coating antibody and an affinity-purified polyclonal anti-CETP antibody as detection antibody, respectively.11,12
Data normally distributed are presented as mean±SD. CETP, triglycerides, adiponectin, interleukin-6 (IL-6), and C-reactive protein (CRP) exhibited a skewed distribution and are presented as median and quartile 1 to quartile 3. Data not normally distributed were transformed logarithmically for statistical analyses. Age- and sex-adjusted differences between subjects with and without CAD were calculated by linear or logistic regression. The effects of cardiovascular risk factors, CAD status, intake of lipid-lowering drugs, and markers of inflammation on CETP levels were determined with the use of general linear models entering CETP as the dependent variable and sex, age, intake of lipid-lowering drugs, CAD status, body mass index (BMI), diabetes mellitus, metabolic syndrome, hypertension (blood pressure >140/90 mm Hg), smoking history (never, former, current), LDL cholesterol, HDL cholesterol, and triglycerides as covariates. LDL cholesterol, HDL cholesterol, LDL/HDL ratio, triglycerides, homocysteine, adiponectin, and IL-6 were categorized in quartiles.
Cox proportional hazard models were used to examine the effect of CETP on mortality. Multivariable adjustment was performed for age, sex, intake of lipid-lowering drugs, CAD status (none, stable CAD, unstable CAD, non–ST-segment elevation myocardial infarction, or ST-segment elevation myocardial infarction), BMI, hypertension, smoking status, LDL cholesterol, HDL cholesterol, triglycerides, and metabolic syndrome/type 2 diabetes mellitus.
We evaluated the combined role of CETP and inflammation markers as predictors of the risk of mortality using a likelihood ratio test to determine whether logistic regression models that included measurements of CETP and markers of inflammation provided a significantly better fit than did logistic regression models limited to markers of inflammation alone. Additionally, we computed the area under receiver operating characteristic (ROC) curves for prediction models based on different combinations of established risk factors, inflammatory markers, and CETP. All statistical tests were 2-sided; P<0.05 was considered significant. The SPSS 16.0 statistical package (SPSS Inc, Chicago, Ill) was used.
Clinical and biochemical characteristics of the study population are shown in Table 1. Besides BMI, all cardiovascular risk factors were more prevalent or severe in CAD patients. At baseline, CETP levels were lower in CAD patients compared with patients without CAD (P=0.002).
Association of CETP With Cardiovascular Risk Factors and Markers of Inflammation
CETP plasma concentration was significantly higher in women compared with men (Table 2). Lower CETP was found in diabetic but not in metabolic syndrome patients. CETP was positively related to LDL cholesterol and lower in patients using lipid-lowering drugs. An even stronger association was observed for the LDL/HDL ratio, with a 21.2% higher CETP in the fourth quartile (P<0.0001). No associations were found with age, BMI, hypertension, and triglycerides. Additionally, CETP was lower in smokers and patients with unstable CAD. CETP showed a negative correlation with high-sensitivity CRP (hsCRP) and IL-6 and a positive correlation with homocysteine and adiponectin (Table 3).
CETP and Mortality From All Causes
Among the 3256 persons studied, 754 deaths (23.2%) occurred during a median follow-up of 7.75 years. Compared with patients in the highest CETP quartile, the age- and sex-adjusted hazard ratio for death in the lowest quartile was 1.37 (95% CI, 1.10 to 1.70) (Table 4 and Figure, top left). CETP retained prognostic value after further adjustment for intake of antihypertensive, lipid-lowering, and antiplatelet therapy, CAD status, cardiovascular risk factors, and diabetes mellitus with a hazard ratio of 1.33 (95% CI, 1.07 to 1.65) in the lowest CETP quartile (Table 4, model 2). Subgroup analysis in 2560 subjects with angiographic CAD at baseline showed similar hazard ratios. No association with mortality was found for lipoprotein characteristics including LDL cholesterol, HDL cholesterol, and triglycerides (Table 5).
CETP and Mortality From Cardiovascular Causes
Among the 3256 subjects studied, 474 (15.5%) died of cardiovascular causes, 57 (1.8%) died of infection, 95 (2.9%) died of cancer, and 104 (3.2%) died of miscellaneous causes. Compared with patients in the highest CETP quartile, the age- and sex-adjusted hazard ratio for death from cardiovascular causes was 1.38 (1.05 to 1.82; P=0.021) in the lowest CETP quartile, 1.19 (0.88 to 1.61; P= 0.25) in the second quartile, and 1.20 (0.89 to 1.62; P=0.23) in the third CETP quartile and thus was similar to that obtained for mortality from all causes (Table 6 and Figure). Further adjustment for additional cardiovascular risk factors and inflammatory markers had only minor influence on the hazard ratios (Table 7). Again, subgroup analysis in subjects with angiographic CAD showed similar results (Table 7).
CETP and Inflammatory Markers
There were strong risk gradients for inflammatory markers, such as hsCRP and IL-6. The adjusted hazard ratios of death from all causes in the highest hsCRP and IL-6 quartiles were 2.05 (1.60 to 2.60) and 2.65 (1.95 to 3.56), respectively. To dissect the effect of CETP from that of inflammatory markers, we performed additional analysis of hazard ratios for death according to CETP. The association between CETP and mortality was retained even after adjustment for inflammatory markers including hsCRP, IL-6, homocysteine, and adiponectin, with a hazard ratio of 1.31 (1.05 to 1.64) in the lowest CETP quartile (Table 4, model 3).
Additionally, we computed the area under the ROC curves including CETP and inflammatory markers (Table 6). In these ROC analyses, the basic model, including established risk factors, yielded an area under the curve of 0.749. Addition of inflammatory markers or CETP led to a slight but not significant increase of the area under the curve. However, inclusion of both inflammatory markers and CETP to the basic model increased the area under the curve significantly (0.775 versus 0.749; P=0.045).
Early observations in Japanese subjects in whom CETP deficiency was associated with very high HDL cholesterol levels gave rise to the development of compounds inhibiting the function of CETP. However, evidence for CETP deficiency to confer protection against atherosclerosis in humans is conflicting.2 Results from different animal models are also inconsistent. Ultimately, CETP inhibition per se and HDL-directed pharmacological interventions in general have come under closer scrutiny, as it became evident that raising HDL cholesterol is not necessarily associated with a favorable cardiovascular outcome.5,7,8
A recent meta-analysis investigated the relationship of CETP polymorphisms with plasma lipid levels and coronary outcomes.13 Some common CETP genotypes were found to be associated with lower CETP mass and activity by 5% to 10% and with increased HDL cholesterol by 3% to 5%. However, no or only weak associations were found between CETP genotypes and coronary outcome.13 One explanation may be the fact that common genotypes are only modestly associated with CETP concentrations. In the study presented herein with the LURIC population, we also failed to find associations between CETP TaqI genotypes and cardiovascular or all-cause mortality (data not shown). However, measurement of CETP mass indeed allowed us to uncover this relationship. Direct measurement of CETP mass, known to be strongly correlated with CETP activity,11 appears to be more informative than the use of single CETP polymorphisms. Using measurements of CETP mass, we found low CETP plasma levels to be an independent risk factor for cardiovascular events and death. This finding supports and extends the results of the ILLUMINATE study5 and challenges the rationale of pharmacological CETP inhibition.
Much effort has been expended in raising HDL cholesterol for cardioprotection.2 However, only raising HDL cholesterol levels may not be sufficient for achieving this goal, when it comes at the cost of a decreased HDL function and thus reduced reverse cholesterol transport.8,14 The transport of peripheral cholesterol back to the liver for excretion relies on functional HDL and LDL particles and may be hampered by low CETP levels. Results from ILLUMINATE and our study may contribute to a paradigm change, shifting the focus from plasma concentrations of HDL and LDL particles to their function. In the study presented, high LDL cholesterol and low HDL cholesterol levels and in particular a high LDL/HDL ratio were associated with high CETP concentrations, suggesting an enhanced cholesterol transfer from HDL to LDL particles in the presence of high CETP. This lipoprotein pattern, considered widely undesirable, may, in the presence of high CETP levels, indicate an enhanced reverse cholesterol transport. Consequently, the therapeutic goal for future therapies may change from mere HDL elevation to the enhancement of reverse cholesterol transport.8
The relationship between plasma levels of CETP and HDL cholesterol in our study was rather modest. The triglyceride levels in patients and controls were within normal range for most patients, whereas the association of CETP and HDL cholesterol is found more clearly in hypertriglyceridemia and in the postprandial state.15
Participants of the LURIC study represent a population at intermediate to high cardiovascular risk because patients were recruited before coronary angiography. Thus, our data cannot necessarily be extrapolated to the general population. An increased rate of CAD and death observed in our study subjects may be due to lower CETP concentrations at baseline in men, current smokers, diabetics, patients with myocardial infarction, and patients on lipid-lowering drugs, all likely to have increased risk of death and CAD during a subsequent follow-up. Therefore, we adjusted for these potential confounders and continued to find an increased hazard ratio for death in the lowest CETP quartile. Furthermore, contribution of CETP to risk prediction persisted after adjustment for inflammatory markers. On closer examination, major effects were seen in quartiles 1 and 4, with only minor differences between quartiles 2 and 3 pointing to a nonlinear sigmoid shape of the relationship. Additionally, in subgroup analyses, we observed similar or even stronger associations in women, patients not taking lipid-lowering drugs, persons who had not experienced myocardial infarction, or nonsmokers. In patients with low CETP plasma levels, we observed an increase not only of cardiovascular but also of all-cause mortality. Interestingly, an increased rate of death from noncardiovascular causes was observed in the active treatment group of the ILLUMINATE study, suggesting perhaps additional functions of HDL particles by mechanisms extending beyond cholesterol transport. Among others, HDL particles have antioxidant and antithrombotic properties, and they are important for various functions of the vascular endothelium.16 In addition, HDL may turn out to be a component of the immune system because a large number of HDL-associated proteins have been identified to be involved in innate immunity, complement regulation, and inflammation.8 Accordingly, human CETP in vitro enhanced the lipopolysaccharide binding to plasma high-density lipoproteins. In addition, the expression of the HDL receptor SR-BI was demonstrated to protect mice against endotoxemia.17,18
To explore whether CETP added any predictive value to a fully adjusted model with markers of inflammation, we computed ROC curves including CETP and markers of inflammation. In these analyses, the use of CETP in addition to established risk factors did not significantly affect the outcome of ROC analysis. However, concomitant use of CETP and inflammatory markers significantly improved the basic model with established risk factors alone. In addition to C statistics, likelihood ratio tests were used to compare the fit of a predictive model based on known risk factors including markers of inflammation to a fit of the same model after addition of CETP. In these analyses, CETP increased the usefulness of the fully adjusted model including adiponectin, hsCRP, and IL-6 in prediction of risk (P<0.0001). Additionally, when we examined hazard ratios for other risk factors including lipoprotein parameters and markers of inflammation, no effect was found for LDL cholesterol, HDL cholesterol, and triglycerides, whereas markers of inflammation exhibited significant effects on relative risk for all-cause mortality in our population. However, the association between CETP and mortality persisted even after adjustment for inflammatory markers, including hsCRP, IL-6, homocysteine, and adiponectin. Taken together, our data support the view that CETP may be important for the anti-inflammatory function of HDL particles and/or may have additional lipoprotein-independent functions within the immune system.
Since the detection of CETP, its role in atherogenesis has been debated extensively but remains elusive.5,19–21 The increased mortality in the ILLUMINATE study may have been due, at least in part, to off-target effects of the CETP inhibitor torcetrapib because systolic blood pressure and aldosterone levels were increased in the treatment group. However, the relationship between changes in blood pressure and clinical outcome in the torcetrapib group was somehow inconsistent internally, insofar as the apparent increased risk of death was found in patients whose increase in systolic blood pressure was lower than the median. In our study, we found significantly increased hazard ratios for cardiovascular and overall death in patients of the lowest CETP quartile but no association between CETP plasma levels and blood pressure or aldosterone levels (data not shown). Therefore, our data strongly suggest that it was the low endogenous CETP plasma levels per se rather than an increase in blood pressure and aldosterone levels that caused the untoward results in the ILLUMINATE trial.
Our study clearly demonstrates for the first time that low endogenous CETP plasma levels constitute an independent risk factor for all-cause and cardiovascular mortality. In light of this result, it could appear unjustified to pharmacologically inhibit CETP activity in subjects already at risk because of low activity of CETP. Future directions for research in HDL-targeted interventions will have to focus on HDL function rather than on mere HDL plasma levels.
Sources of Funding
This work was supported by the Medizinische Forschungsfoerderung Innsbruck (No. 4316 to I.T.), by the Jubilaeumsfond der Oesterreichischen Nationalbank (No. 12156 to I.T. and A.R.), by the Fonds zur Foerderung der wissenschaftlichen Forschung (P19999-B05 to A.R.), and by the Center of Excellence Baden-Wuerttemberg — Metabolic Disorders (No. SFB518/A01 and GRK-1041 to B.O.B).
Barter PJ, Caulfield M, Eriksson M, Grundy SM, Kastelein JJ, Komajda M, Lopez-Sendon J, Mosca L, Tardif JC, Waters DD, Shear CL, Revkin JH, Buhr KA, Fisher MR, Tall AR, Brewer B. Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med. 2007; 357: 2109–2122.
Olsson AG. Is high HDL cholesterol always good? Ann Med. 2008; 41: 11–18.
American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care. 2006; 29 (suppl 1): S43–S48.
Ritsch A, Auer B, Foger B, Schwarz S, Patsch JR. Polyclonal antibody-based immunoradiometric assay for quantification of cholesteryl ester transfer protein. J Lipid Res. 1993; 34: 673–679.
Ritsch A, Ebenbichler C, Naschberger E, Schgoer W, Stanzl U, Dietrich H, Heinrich PC, Saito K, Patsch JR. Phage-displayed recombinant single-chain antibody fragments with high affinity for cholesteryl ester transfer protein (CETP): cDNA cloning, characterization and CETP quantification. Clin Chem Lab Med. 2004; 42: 247–255.
Föger B, Ritsch A, Doblinger A, Wessels H, Patsch JR. Relationship of plasma cholesteryl ester transfer protein to HDL cholesterol: studies in normotriglyceridemia and moderate hypertriglyceridemia. Arterioscler Thromb Vasc Biol. 1996; 16: 1430–1436.
Rye KA, Barter PJ. Antiinflammatory actions of HDL: a new insight. Arterioscler Thromb Vasc Biol. 2008; 28: 1890–1891.
Cholesteryl ester transfer protein (CETP) is the major player in reverse cholesterol transport, but its role in the development of atherosclerosis continues to be in question since its discovery nearly 20 years ago. After the ahead-of-schedule termination of the large phase III clinical trial Investigation of Lipid Level Management to Understand its Impact in Atherosclerotic Events (ILLUMINATE) of the CETP inhibitor torcetrapib, the dispute reached a fervent revival. The authors of the ILLUMINATE trial proposed 2 explanations for the higher mortality in the torcetrapib group: a side effect of the drug (eg, increased blood pressure) or suppressed CETP activity per se. In the work submitted herein, we present a large prospective observation on a very similar study population relating variation of CETP mass to mortality. Our data suggest that endogenous low CETP plasma levels constitute an independent risk factor for all-cause and cardiovascular mortality and thus strongly point to the latter explanation for the ILLUMINATE results (ie, CETP inhibition per se causing increased mortality). We believe that our study provides a serious caveat to CETP inhibition in general.