Although cholesteryl ester transfer protein (CETP) plays an important role in human lipoprotein metabolism, its relation to coronary artery disease remains controversial. The present study evaluated the relation between on-statin CETP mass and recurrent coronary events. The plasma CETP mass, measured after 4 months of statin therapy, was quantified in 3,218 patients enrolled in the Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction 22 (PROVE IT-TIMI 22) study. Of the 3,218 patients, 150 experienced the combined end point of recurrent myocardial infarction or death from coronary causes during a mean follow-up of 1.8 years. An increasing on-statin CETP mass was inversely related to the risk of coronary events in both unadjusted (hazard ratio [HR] per SD increase 0.77, 95% confidence interval 0.65 to 0.92, p = 0.005) and fully adjusted (HR per SD increase 0.81, 95% confidence interval 0.67 to 0.98, p = 0.027) analyses that included traditional cardiovascular risk factors. A similar trend was observed across increasing CETP mass quartiles (p trend = 0.07). A significant interaction between the CETP mass and on-treatment low-density lipoprotein (LDL) cholesterol was noted (p interaction = 0.007). A CETP mass greater than the median was associated with a decreased risk in patients with LDL cholesterol less than the median of 80 mg/dl (HR 0.52, 95% confidence interval 0.31 to 0.89, p = 0.02), but not in patients with LDL cholesterol greater than the median. In conclusion, an increasing on-statin CETP mass was inversely related to the coronary outcomes in this large clinical trial-based cohort, particularly among those with low LDL cholesterol levels. This finding is consistent with CETP facilitating reverse cholesterol transport in the setting of robust LDL clearance and might have important implications for efforts to optimally target patients with pharmacologic CETP inhibition.
Cholesteryl ester transfer protein (CETP) plays a major role in human lipid metabolism and has generated substantial interest as a therapeutic target in recent years. Plasma CETP promotes the transfer of cholesteryl esters from high-density lipoprotein (HDL) cholesterol to apolipoprotein B-containing particles, including low-density lipoproteins (LDLs) and very low-density lipoproteins. Pharmacologic inhibition of CETP with torcetrapib leads to markedly increased HDL cholesterol and decreased LDL cholesterol levels in humans. However, torcetrapib was associated with an increase in mortality and cardiovascular events in a large Phase III clinical trial. The results from imaging studies were similarly disappointing, leading to termination of the drug’s clinical development. Although these outcomes might have been in part due to off-target effects of torcetrapib, they raised concerns about whether the altered lipid profile seen with CETP inhibition is necessarily cardioprotective. Analyses in 2 large prospective cohorts have associated decreased levels of CETP mass or activity with increased cardiovascular risk. Pharmacologic CETP inhibition, if ever applied to clinical practice, would likely serve as an adjunct to current statin-based regimens. We, therefore, sought evidence for a relation between on-statin plasma CETP levels and the risk of recurrent coronary events. An additional hypothesis was that this relation would be conditioned by on-treatment LDL cholesterol levels, a surrogate for the efficiency of hepatic LDL clearance. These questions were addressed by measuring the on-statin plasma CETP mass in patients enrolled in the Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction 22 (PROVE IT-TIMI 22) trial, a study that randomized patients after an acute coronary syndrome to either intensive or moderate statin therapy.
Methods
The study population was derived from the PROVE IT-TIMI 22 trial (NCT00382460), the details of which have been previously reported. In brief, 4,162 patients hospitalized for an acute coronary syndrome were randomized to either intensive lipid-lowering therapy (atorvastatin 80 mg/day) or standard therapy (pravastatin 40 mg/day). As a part of the protocol, the plasma lipid and C-reactive protein levels were obtained 4 months after randomization. Measurement of the CETP mass was performed from the plasma from this 4-month blood sampling in 3,218 patients (77% of the original cohort). All laboratory measurements were done in a core facility. The CETP mass was measured using a commercially available enzyme-linked immunosorbent assay (Wako Chemicals, Richmond, Virginia). A composite clinical end point of recurrent myocardial infarction and death from coronary causes was used for the present analysis as described previously.
A 2-sample t test was used to compare whether the plasma CETP mass differed between the atorvastatin and pravastatin treatment groups. Spearman correlation coefficients were used to evaluate the relation between the on-statin CETP mass values and the achieved LDL, triglyceride (TG), and HDL levels obtained from the same blood sample.
A multistage process was used to test for a relation between the achieved CETP mass levels and recurrent coronary events. Cox regression analysis was initially performed with the CETP mass modeled as a continuous variable, with hazard ratios (HRs) per SD increase reported. Proportional hazard models that included (1) age, gender, and treatment group (atorvastatin 80 mg vs pravastatin 40 mg); (2) age, gender, treatment group, and LDL cholesterol level; and (3) age, gender, treatment group, body mass index, diabetic status, history of hypertension, smoking status, LDL cholesterol, and natural logarithm of C-reactive protein were used in subsequent multivariate analysis. Analysis of the Shoenfield residuals was performed to verify the proportional hazards assumptions for each covariate. The study population was then divided into quartiles of CETP mass levels. To assess whether the clinical characteristics varied across the quartiles, tests of trend were performed using a nonparametric test for trend across ordered groups and partitions of the Pearson chi-square test for ordered columns for continuous and discrete variables, respectively. Finally, the association between the CETP mass quartile and recurrent coronary events was investigated using the unadjusted and multivariate models specified.
Interaction terms within the Cox proportional hazard models were used to assess whether the relation between the CETP mass and recurrent events varied according to on-statin LDL cholesterol or TG levels. The interactions between the CETP mass (greater than vs less than the median) and LDL cholesterol and TG (modeled both as greater than vs less than the median and as continuous variables) were analyzed. Finally, Cox regression analysis was used to calculate the HR estimates for CETP mass (greater than vs less than median) in subgroups separated by LDL cholesterol (greater than vs less than the median).
All p values were 2-tailed, and all confidence intervals (CIs) were computed at the 95% level. Statistical analysis was performed using STATA/SE, version 9.2 (StataCorp, College Station, Texas).
Results
The mean age of the study participants was 58 years, and 79% were men; 17% had diabetes mellitus, 50% had a history of hypertension, and 37% were current smokers. The on-statin plasma CETP mass level was measured 4 months after an acute coronary syndrome in 3,218 patients ( Figure 1 ). The CETP mass levels were lower in the patients randomized to receive intensive statin therapy with atorvastatin 80 mg/day (n = 1,611) than in those who received pravastatin 40 mg/day (n = 1,607; mean ± SD 0.73 ± 0.26 vs 0.82 ± 0.32, respectively; p <0.0001). Scatterplot analyses revealed that an increasing CETP mass correlated with increasing on-statin LDL cholesterol (Spearman rho = 0.18, p <0.0001) and decreased on-statin TG (Spearman rho = −0.07, p = 0.0001). No relation was noted between the CETP mass and on-statin HDL cholesterol levels (Spearman rho = 0.0005, p = 0.98).
Additional clinical characteristics, stratified by CETP mass quartile, are listed in Table 1 . Patients with a CETP mass level in the higher quartiles were more likely to be women and less likely to have a history of diabetes or hypertension (p <0.05 for each). Furthermore, increased on-statin LDL cholesterol and decreased on-statin TG levels were noted across increasing CETP mass quartiles (p <0.0001).
| Variable | CETP Mass Quartile (μg/ml) | p Trend | |||
|---|---|---|---|---|---|
| Q1 (n = 806) | Q2 (n = 805) | Q3 (n = 805) | Q4 (n = 802) | ||
| Age (years) | 58.2 ± 10.9 | 58.2 ± 10.6 | 57.8 ± 11.2 | 58.0 ± 11.3 | 0.54 |
| Men | 654 (81.1) | 657 (81.6) | 638 (79.3) | 579 (72.2) | <0.0001 |
| Body mass index (kg/m 2 ) | 29.7 ± 5.7 | 29.8 ± 5.9 | 29.8 ± 5.7 | 29.4 ± 5.9 | 0.31 |
| Diabetes mellitus | 167 (20.7) | 157 (19.5) | 113 (14.0) | 113 (14.1) | 0.0001 |
| Hypertension | 411 (60.0) | 374 (46.5) | 431 (53.5) | 398 (49.6) | 0.04 |
| Current smoker | 320 (39.7) | 293 (36.4) | 287 (35.7) | 275 (34.29) | 0.14 |
| On-statin low-density lipoprotein cholesterol (mg/dl) | 77.5 ± 30.4 | 81.3 ± 31.1 | 84.8 ± 32.4 | 92.2 ± 32.2 | <0.0001 |
| On-statin high-density lipoprotein cholesterol (mg/dl) | 43.4 ± 12.6 | 42.3 ± 11.1 | 42.3 ± 11.1 | 42.9 ± 11.2 | 0.89 |
| On-statin triglycerides (mg/dl) | 167 (97–196) | 158 (92–191) | 156 (89–184) | 152 (89–176) | <0.0001 |
| On-statin C-reactive protein (mg/dl) | 1.7 (0.8–4.0) | 1.7 (0.8–3.7) | 1.6 (0.7–3.2) | 1.8 (0.8–4.1) | 0.59 |
A total of 150 patients experienced the combined end point of recurrent myocardial infarction or death from coronary causes during a mean follow-up of 1.8 years. Cox regression analyses that incorporated the CETP mass as a continuous variable demonstrated an inverse relation between CETP mass and recurrent coronary events ( Table 2 ). A univariate analysis of the CETP mass levels revealed an inverse relation between the CETP mass and the risk of recurrent events (HR/SD increase 0.77, 95% CI 0.65 to 0.92, p = 0.005; Table 2 ). This relation remained statistically significant in the proportional hazard models that incorporated age, gender, on-statin LDL cholesterol level, treatment group, and other traditional cardiovascular risk factors. Increased age, diabetes, and increased C-reactive protein were each associated with increased risk in the fully adjusted multivariate analysis (p <0.05 for each). Similarly, a tendency toward a decreased coronary risk was noted across increasing CETP mass quartiles ( Table 2 ). The HRs were minimally affected in a multivariable model that adjusted for age, gender, LDL cholesterol, and treatment group (p trend = 0.04) but were somewhat attenuated in a fully adjusted model that included additional cardiovascular risk factors (p trend = 0.23).
| Variable | CETP Mass (HR/SD Increase) | CETP Mass Quartile (μg/ml) | p Trend | |||
|---|---|---|---|---|---|---|
| Q1 (n = 806) | Q2 (n = 805) | Q3 (n = 805) | Q4 (n = 802) | |||
| Unadjusted | 0.77 | 1.0 | 0.89 | 0.85 | 0.64 | 0.07 |
| 95% CI | 0.65–0.92 | 0.59–1.38 | 0.55–1.32 | 0.40–1.02 | ||
| p Value | 0.005 | 0.62 | 0.48 | 0.06 | ||
| Model 1 ⁎ | 0.77 | 1.0 | 0.89 | 0.85 | 0.62 | 0.06 |
| 95% CI | 0.64–0.92 | 0.58–1.37 | 0.55–1.32 | 0.39–1.00 | ||
| p Value | 0.004 | 0.60 | 0.47 | 0.05 | ||
| Model 2 † | 0.75 | 1.0 | 0.89 | 0.83 | 0.60 | 0.04 |
| 95% CI | 0.63–0.90 | 0.58–1.36 | 0.54–1.28 | 0.37–0.97 | ||
| p Value | 0.002 | 0.58 | 0.40 | 0.04 | ||
| Model 3 ‡ | 0.81 | 1.0 | 0.98 | 1.05 | 0.68 | 0.23 |
| 95% CI | 0.67–0.98 | 0.62–1.53 | 0.66–1.66 | 0.41–1.15 | ||
| p Value | 0.027 | 0.92 | 0.84 | 0.15 | ||
⁎ Adjusted for age, gender, and treatment group.
† Adjusted for age, gender, treatment group, and LDL cholesterol.
‡ Adjusted for age, gender, treatment group, LDL cholesterol, body mass index, diabetes (yes vs no), history of hypertension (yes vs no), smoking status (current smoker vs nonsmoker), and logarithmically transformed C-reactive protein levels.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
