Jump to
Abstract
Background—Despite favorable effects on high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol, the cholesteryl ester transfer protein inhibitor torcetrapib failed to slow atherosclerosis progression and increased mortality. We investigated the relationship between lipid changes and progression of coronary atherosclerosis.
Methods and Results—Intravascular ultrasound was performed at baseline and follow-up in 910 participants randomized to torcetrapib/atorvastatin or atorvastatin monotherapy. The relationship between changes in lipoprotein levels and the primary intravascular ultrasound end point, change in percent atheroma volume, was investigated. Compared with atorvastatin monotherapy, torcetrapib raised HDL-C by 61%, lowered low-density lipoprotein cholesterol by 20%, raised serum sodium (0.44±0.14 mmol/L, P=0.02), and lowered serum potassium (0.11±0.02 mmol/L, P<0.0001). Despite substantial increases in HDL-C, no effect was found of torcetrapib on percent atheroma volume. In torcetrapib-treated patients, an inverse relationship was observed between changes in HDL-C and percentage atheroma volume (r=−0.17, P<0.001). Participants with regression had greater increases in HDL-C (mean±SE, 62.9±37.4% versus 54.0±39.1%, P=0.002). Compared with the lowest quartile, torcetrapib-treated patients in the highest quartile of HDL-C change showed the least progression (−0.31±0.27 versus 0.88±0.27%, P=0.001). The highest on-treatment HDL-C quartile showed significant regression of percent atheroma volume (−0.69±0.27%, P=0.01). In multivariable analysis, changes in HDL-C levels independently predicted the effect on atherosclerosis progression (P=0.001).
Conclusions—The majority of torcetrapib-treated patients demonstrated no regression of coronary atherosclerosis. Regression was only observed at the highest HDL-C levels. Torcetrapib raised serum sodium and lowered potassium, consistent with an aldosterone-like effect, which may explain the lack of favorable effects in the full study cohort. Accordingly, other cholesteryl ester transfer protein inhibitors, if they lack this off-target toxicity, may successfully slow atherosclerosis progression.
Received May 5, 2008; accepted September 12, 2008.
Although inhibition of cholesteryl ester transfer protein (CETP) substantially increases levels of high-density lipoprotein cholesterol (HDL-C) and has favorable effects on levels of low-density lipoprotein cholesterol (LDL-C), recent imaging studies failed to demonstrate that CETP inhibition with the investigational agent torcetrapib slows atherosclerosis progression in the coronary or carotid arteries.1–3 A large morbidity–mortality trial studying torcetrapib was terminated prematurely after the data safety monitoring committee reported a significant increase in all-cause mortality.4 Subsequently, the company developing torcetrapib discontinued all further studies.5 Although the agent initially developed to inhibit CETP, torcetrapib, substantially increased blood pressure, the mechanism underlying the observed lack of benefit for torcetrapib has not been fully elucidated. Accordingly, uncertainty exists on whether the strategy of CETP inhibition is likely to yield a clinically useful pharmacological agent to raise HDL-C.
Editorial p 2491
Clinical Perspective p 2514
Two principal explanations for the failure of torcetrapib have been proposed. One hypothesis suggests that the HDL particles generated by CETP inhibition do not participate in reverse cholesterol transport or possess other protective properties.6 Alternatively, the specific agent tested in initial studies, torcetrapib, may have unique toxicity that counterbalances any antiatherosclerotic benefits arising from increasing HDL-C levels via CETP inhibition. The Investigation of Lipid Level management USing coronary ultrasound To assess Reduction of ATherosclerosis by CETP inhibition and HDL Elevation (ILLUSTRATE) trial assessed the effects of torcetrapib on the rate of progression of coronary atherosclerosis using intravascular ultrasonography (IVUS).1 In the present post hoc analysis, we sought to characterize the relationship between changes in lipid parameters and changes in atheroma burden in the ILLUSTRATE trial. In particular, we sought to determine whether increasing HDL-C with torcetrapib was associated with a beneficial effect, a detrimental effect, or no effect on disease progression. An increase in disease progression at higher levels of HDL-C would support the concept that CETP inhibition has an adverse impact on the protective properties of HDL. In contrast, a reduction in disease progression at higher levels of HDL-C would support the concept that HDL particles retain some degree of functionality in the setting of CETP inhibition. The objective of the analysis was to help answer the question, was the failure of torcetrapib related to the “molecule” or the “mechanism of action?”
Methods
ILLUSTRATE Study Protocol
The details of the ILLUSTRATE protocol have been described previously.1 After titration of atorvastatin to achieve a target LDL-C level <100±15 mg/dL (mean±SE; average dose 23 mg), patients were randomized to additional treatment with either torcetrapib 60 mg daily or placebo. IVUS images were acquired within a matched arterial segment at baseline and after 24 months of treatment. The leading edges of the external elastic membrane and lumen were defined by manual planimetry in images spaced precisely 1 mm apart. Atheroma area was calculated as the difference between areas occupied by the external elastic membrane and lumen. Percent atheroma volume (PAV) was determined as the sum of plaque areas for all cross sections divided by the sum of the external elastic membrane areas for these same cross sections.7 Biochemical and blood pressure measurements were performed 1 month after baseline and then at 3-month intervals.
Statistical Analysis
For the purpose of the present analysis, only subjects who underwent ultrasonic imaging at both baseline and follow-up were included. Consistent with the primary report,1 follow-up biochemical parameters were determined as the last observation carried forward, and blood pressure was reported as the average of all measures obtained during the treatment period. In the case of missing values (14% of samples), the nearest postbaseline result was used in the analysis. Student t tests (or Wilcoxon rank sum tests for nonnormally distributed data) and χ2 tests were performed to assess differences across study groups for continuous and categorical variables, respectively. Changes from baseline to follow-up within each treatment group were compared by 1-sample t tests (or Wilcoxon signed rank tests for nonnormally distributed data). Locally weighted scatterplot smoothing (LOWESS) plots were used to depict the relationship between changes in PAV, after controlling for baseline values, and both achieved levels of and percentage change in HDL-C in each treatment group. Spearman correlation coefficients were calculated to assess the relationship between achieved levels or changes in levels of biochemical parameters and the change in PAV. Relationships between changes in imaging end points and changes in HDL-C were also described by quartiles within each treatment group. Linear regression models were used to calculate least squares means of changes in PAV, controlling for the baseline value, in patients stratified according to the level of achieved HDL or its percentage change.
A multivariable linear regression model was performed to determine the independent predictors of changes in PAV in patients treated with torcetrapib. This model incorporated all univariate predictors of changes in PAV (geographic region, atorvastatin dose, baseline PAV, percent change in HDL cholesterol, and black race). Changes in systolic blood pressure and potassium were also included in the multivariable model. All analyses were performed with SAS version 8.2 (SAS Institute, Inc, Cary, NC). P<0.05 was considered significant.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Results
Changes in Biochemical Parameters and Atheroma Burden
Clinical characteristics, biochemical parameters, and measures of atheroma burden at baseline and follow-up are summarized in Tables 1 and 2⇓. In patients treated with torcetrapib plus atorvastatin, LDL-C decreased from 83.1 to 70.3 mg/dL (P<0.001), and HDL-C increased from 46.0 to 72.3 mg/dL (P<0.001). In patients treated with atorvastatin monotherapy, LDL-C increased from 84.3 to 88.1 mg/dL (P<0.001), and HDL-C decreased from 45.2 to 43.8 mg/dL (P<0.001). This resulted in a reduction in the LDL-C/HDL-C ratio from 1.9 to 0.9 in the torcetrapib-plus-atorvastatin treatment group (P<0.001) and an increase from 1.9 to 2.1 with atorvastatin monotherapy (P<0.001). Median levels of C-reactive protein decreased from 1.8 to 1.5 mg/L with atorvastatin monotherapy (P=0.001) and from 2.1 to 1.9 mg/L with torcetrapib plus atorvastatin (P=0.02). In patients treated with torcetrapib plus atorvastatin, sodium increased from 139.7 to 139.9 mmol/L (P=0.005), and potassium decreased from 4.22 to 4.14 mmol/L (P<0.001). In contrast, atorvastatin monotherapy was associated with a reduction in sodium from 139.8 to 139.6 mmol/L (P=0.03) and no change in potassium levels. Thus, compared with atorvastatin monotherapy, torcetrapib raised sodium (0.44±0.14 mmol/L, P=0.02) and lowered potassium (0.11±0.02 mmol/L, P<0.0001). Administration of torcetrapib plus atorvastatin did not have a beneficial impact on disease progression, with no significant changes in PAV from baseline to follow-up observed in either treatment group.
Table 1. Baseline Characteristics
Table 2. Biochemical Parameters and PAV at Baseline and 24-Month Follow-Up in Patients by Treatment Group
Relationship Between Changes in Biochemical Parameters and Atheroma Burden
The relationships between both baseline levels of and changes in biochemical parameters and changes in PAV are summarized in Table 3 and Figure 1. A significant correlation was observed between baseline levels of both triglycerides and apolipoprotein (apo) B and the change in PAV in both treatment groups. In patients treated with atorvastatin monotherapy, significant correlations were observed between the change in PAV and achieved levels of both triglycerides and the ratio of apoB to apoA-I. In contrast, patients treated with torcetrapib/atorvastatin showed significant correlations between the change in PAV and the achieved levels of HDL-C, change in HDL-C levels, apoA-I, and the LDL-C/HDL-C ratio (Table 3). Figure 1 shows the continuous relationship between absolute HDL-C levels or change in HDL-C levels and the rate of disease progression by IVUS. This relationship is apparent in the torcetrapib-treated group but not the atorvastatin monotherapy group.
Table 3. Correlation Coefficients Between Various Measures of Lipoprotein Levels and Change in PAV
Figure 1. Locally weighted scatterplot smoothing (LOWESS) scatterplots illustrating the relationship between change in PAV, after controlling for the baseline value, and either follow-up level of HDL-C (top, in mg/dL) or percentage change in HDL-C (bottom), stratified according to treatment group.
Biochemical Parameters and Regression Versus Progression
Biochemical parameters in patients experiencing regression or progression of atherosclerosis (decrease or increase in PAV, respectively) are summarized in Table 4. No significant differences were observed between regressors and progressors treated with atorvastatin monotherapy. In contrast, patients treated with torcetrapib plus atorvastatin who experienced regression had greater increases in HDL-C (P=0.0002) and apoA-I (P=0.01) and greater decreases in the LDL-C/HDL-C ratio (P=0.02) than patients who experienced progression. Lower baseline levels of triglyceride and greater increases in HDL-C were associated with an increase in the proportion of patients in the torcetrapib/atorvastatin group who underwent any degree of plaque regression (online-only Data Supplement Table). Changes in levels of apoB-containing lipoproteins did not predict the likelihood of regression or progression.
Table 4. Biochemical Parameters at Baseline and Follow-Up and Their Percentage Change in Patients Who Underwent Plaque Regression or Progression (Decrease or Increase in PAV, Respectively), Stratified According to Treatment Group
HDL-C and Changes in Atheroma Burden
The relationship between HDL-C and changes in atheroma burden is further illustrated in Figures 2 through 4⇓⇓. Despite considerable increases in HDL-C with torcetrapib, only patients who achieved the highest levels of HDL-C or the greatest percentage increase had regression of PAV (Figures 3 and 4⇓). Interestingly, no relationship was observed between either achieved levels or percentage change in LDL-C and disease progression (online-only Data Supplement Figures I and II). No relationship between achieved levels (P=0.57 for trend) or changes (P=0.26 for trend) in HDL-C and cardiovascular events was observed in torcetrapib-treated subjects, although the study was not powered to assess clinical outcomes. The lack of efficacy of torcetrapib in the entire study population is further supported by the observation of a greater degree of disease progression at all but the highest levels of HDL-C when stratified according to HDL-C quartiles in the entire cohort (Figure 4). With multivariable analysis, independent predictors of the change in PAV in patients taking torcetrapib included the percentage increase in HDL-C (P=0.001), atorvastatin dose (P=0.02), baseline PAV (P=0.0003), and black race (P=0.02). Of interest, the change in systolic blood pressure was not a significant independent predictor of the change in PAV (P=0.08), nor were change in triglycerides (P=0.51) or change in potassium (P=0.28; Table 5). Multivariable analysis of both treatment groups combined revealed an interaction between treatment, percent change in HDL-C, and change in PAV (P=0.08) that was not affected by inclusion of baseline HDL-C in the model (P=0.09).
Figure 2. Least squares mean change in PAV, after controlling for the baseline value, in atorvastatin monotherapy–treated subjects according to quartile of achieved level (top) and percentage change (bottom) of HDL-C. P values above and below bars for comparison between baseline and follow-up.
Figure 3. Least squares mean change in PAV, after controlling for the baseline value, in torcetrapib/atorvastatin-treated subjects according to quartile of achieved level (top) and percentage change (bottom) of HDL-C. P values above bars for comparison between baseline and follow-up.
Figure 4. Least squares mean±SEM change in PAV, after controlling for the baseline value, in patients treated with either atorvastatin monotherapy (solid bars), or torcetrapib/atorvastatin (open bars), stratified according to quartiles of percentage change in HDL-C in the combined treatment groups.
Table 5. Multivariable Analysis Indicating Independent Predictors of Changes in PAV in Patients Treated With the Combination of Torcetrapib and Atorvastatin
Discussion
Although torcetrapib treatment failed to slow the overall progression of coronary disease in the ILLUSTRATE trial, the current post hoc analysis demonstrated an inverse relationship between the extent of HDL-C increases with torcetrapib treatment and the rate of atherosclerosis progression as determined by IVUS. In particular, regression, not accelerated progression, of atherosclerosis was observed in those patients who achieved the highest levels of HDL-C with torcetrapib. Furthermore, the majority of patients in the lowest 3 quartiles of HDL-C did not undergo disease regression. These post hoc exploratory data support the concept that despite the failure of torcetrapib to slow disease progression, the achievement of very high levels of HDL-C via CETP inhibition has the potential to generate functional HDL particles that participate in reverse cholesterol transport. Accordingly, the concept of CETP inhibition as a means to limit coronary atherosclerosis may remain a potentially viable strategy that requires further study. These findings provide useful insights into the failure of torcetrapib and suggest that the absence of benefit may reflect molecule-specific toxicity rather than a mechanism-based failure of this pharmacological approach to elevation of HDL-C levels. However, it remains to be demonstrated conclusively that inhibition of CETP does not result in either cardiac or noncardiac toxicity. Nonetheless, other CETP inhibitors may succeed where torcetrapib failed.
Although torcetrapib substantially increases blood pressure, this was not an independent predictor of the change in atheroma burden. Neither ILLUSTRATE nor the prematurely terminated clinical events trial reported an increased rate of stroke,1,5 which is generally considered very sensitive to the effects of blood pressure changes. These observations suggest that the blood pressure increase is unlikely to be the principal explanation for the lack of an antiatherosclerotic benefit. It appears more likely that torcetrapib has some other toxicity that mitigated any beneficial effect of substantial increases in HDL-C levels. This is further supported by the finding that the numerical rate of progression was greater in patients treated with torcetrapib and atorvastatin, who achieved the least effective rise in HDL-C, and by the fact that the majority of patients at lower HDL-C levels did not regress (online-only Data Supplement Figure III). The observed increases in serum sodium and decreases in serum potassium levels, along with an increase in blood pressure, suggest an increase in aldosterone activity in response to torcetrapib treatment. Preclinical data have recently confirmed that torcetrapib functions as an aldosterone agonist, increasing systemic aldosterone levels.8
Increasing evidence implies a role of the renin-angiotensin-aldosterone system in atherosclerosis. Angiotensin II and aldosterone play a major role in blood pressure regulation and possess adverse oxidative and inflammatory effects.9–13 Furthermore, inhibition of renin-angiotensin-aldosterone system activity retards lesion initiation and progression in animal models of atherosclerosis.14 Given the observed changes in electrolytes and blood pressure, it now appears plausible that the aldosterone agonist effect of torcetrapib exerted a detrimental influence on atherosclerosis progression. Accordingly, this off-target toxicity may have counteracted any favorable effects derived from substantial increases in HDL-C produced by CETP inhibition.
After the initial description of genetic CETP deficiency,15,16 the potential for pharmacological inhibition of CETP to benefit patients with atherosclerosis remained controversial. Although it has been long recognized that CETP inhibitors can substantially raise HDL-C levels,17 considerable speculation has focused on the relative functionality of the HDL particles that are generated.18 Despite the recent report that the combination of torcetrapib and a statin did not reduce the burden of atherosclerosis in apoE(Leiden)CETP transgenic mice,19 beneficial effects of strategies that reduce CETP activity have been demonstrated in rabbits, a species that naturally expresses CETP.20–22 In the present study, we observed that patients with the largest increases in HDL-C and those who achieved the highest systemic levels demonstrated plaque regression. Furthermore, although the plots in Figures 2 and 3⇑ reveal a continuous relationship between changes in HDL-C and PAV in both treatment groups, they show a strong inverse slope only within the torcetrapib-treated subgroup. These data provide evidence that substantially raising HDL-C levels via CETP inhibition has the potential to slow or reverse coronary atheroma progression. Accordingly, the failure of torcetrapib to slow disease progression does not preclude the possibility that an alternative agent in this therapeutic class will result in benefit for patients with atherosclerosis.
The results of the present study have implications for understanding the potential impact of raising HDL-C via CETP inhibition on atherosclerosis in humans. Because the dosage of torcetrapib was limited by this agent’s dose-dependent adverse effect on blood pressure, maximal increases in HDL-C could not be explored.23 Greater CETP inhibition that results in higher HDL-C levels may produce more substantial antiatherosclerotic benefits. This is particularly important given that regression was not observed in the majority of patients in the lowest 3 quartiles of achieved HDL-C levels. Other CETP inhibitors that lack adverse effects on blood pressure or the vasculature can be used to explore greater levels of CETP inhibition. The finding of regression by IVUS for patients with the highest HDL-C elevations increases confidence that a potent CETP inhibitor may favorably affect clinical outcomes.
These observations provide useful insights into the impact of inhibiting CETP on potential HDL-mediated antiatherosclerotic effects in humans. Although some studies in patients with CETP deficiency have shown cardiovascular protection,24,25 others have reported the presence of atherosclerotic disease in patients with reduced CETP activity,26,27 although a low prevalence of CHD was observed when HDL-C levels were elevated.27 Several groups have demonstrated that lipid-depleted forms of HDL are the preferred acceptor of cholesterol effluxed from cells via the ABCA1 transporter.28,29 This observation raised the possibility that the generation of large, cholesterol-laden particles with CETP inhibition might result in impaired mobilization of cholesterol from macrophages within the vessel wall. However, the subsequent discovery that large HDL particles promote efflux via the ABCG1 transporter suggested that efflux would not be impaired after CETP inhibition.30 This hypothesis is supported by observations that the efflux activity of HDL31 and the excretion of fecal sterols32 are not impaired after administration of torcetrapib in humans and that CETP inhibition does not impair reverse cholesterol transport in a rabbit model.33 In addition, no evidence currently exists that CETP inhibition has an adverse impact on non–lipid-transporting functions of HDL. It is also of interest that greater disease progression was observed with the combination of torcetrapib and atorvastatin in patients with greater levels of apoB and triglyceride at baseline. This might suggest that CETP inhibition may be less effective or potentially detrimental in this clinical setting, although this requires further investigation with a CETP inhibitor that lacks any off-target toxicity.
Considerable evidence suggests that HDL can protect against the development of atherosclerosis. Observational studies consistently have demonstrated an inverse relationship between HDL-C levels and the subsequent risk of developing coronary heart disease.34 In animal models of atherosclerosis, an increase in HDL-C levels produces favorable effects on lesion size and composition.35–40 The greatest benefit of statin therapy typically is observed in patients with the lowest HDL-C levels.41–43 Modest elevation of HDL-C with statins,44 fibrates,45,46 and niacin47,48 is associated with reduced disease progression and morbidity. Two small studies in humans demonstrated that the infusion of reconstituted HDL promoted rapid regression of coronary atherosclerosis.49,50 Because relatively modest elevations in HDL-C levels may have contributed to these clinical benefits, considerable interest exists in developing therapies that raise levels more substantially.
A number of limitations of the present exploratory analysis should be noted. Administration of torcetrapib in combination with atorvastatin did not have a beneficial impact on disease progression in the entire cohort. The present analysis is exploratory and describes the relationship between changes in levels of HDL-C and PAV. Although the analysis does not evaluate the functionality of HDL particles directly, the finding of regression and not accelerated progression of disease in patients with the highest level of HDL-C suggests that the ability of particles to mobilize lipids from the artery wall would appear to be preserved. All patients presented for a clinically indicated angiogram and had documented coronary artery disease. It is unknown whether the present findings are applicable in primary prevention. It is uncertain whether substantial elevation of HDL-C with CETP inhibitors has a favorable impact on plaque composition, which may represent an important determinant of atheroma rupture. The translation of beneficial changes on IVUS to improved clinical outcome remains to be defined. As a result, it cannot be assumed that regression associated with substantial HDL-C elevations will result in improved clinical outcome. Nonetheless, the findings highlight the ability of IVUS to provide mechanistic insights into the impact of medical therapies on both atherosclerotic plaque and associated arterial wall remodeling.
The results of this post hoc exploratory analysis have implications for further attempts to develop therapeutic agents that raise HDL-C levels. Although torcetrapib did not slow disease progression in the entire cohort, the finding of disease regression at the highest levels of HDL-C raises the possibility that another CETP inhibitor, without compound-specific toxicity, may be clinically efficacious. This potential should be explored in further well-designed, randomized clinical trials. There remains hope that new agents promoting the protective properties of HDL will provide an adjunctive approach to achieve more effective prevention of cardiovascular disease.
Acknowledgments
We appreciate the technical expertise of the Intravascular Ultrasound Core Laboratory at the Cleveland Clinic.
Source of Funding
The ILLUSTRATE trial was sponsored by Pfizer Pharmaceuticals.
Disclosures
Dr Nicholls reports receiving honoraria from Pfizer, AstraZeneca, Takeda, and Merck Schering-Plough; consultancy fees from AstraZeneca, Roche, Novo-Nordisk, Pfizer, and Anthera Pharmaceuticals; and research support from LipidSciences. Dr Tuzcu reports receiving consultancy fees from Pfizer and honoraria from Pfizer and Merck. Dr Tardif holds the Pfizer and Canadian Institutes of Health Research chair in atherosclerosis. Dr Tardif has also received consultancy fees from Pfizer and AstraZeneca. Dr Nissen reports that the Cleveland Clinic Coordinating Center has received research support from Pfizer, AstraZeneca, Sankyo, Takeda, Novartis, Sanofi-Aventis, and Eli Lilly to perform clinical trials. Dr Nissen consults for many pharmaceutical companies but requires them to donate all honoraria or consulting fees directly to charity so that he receives neither income nor a tax deduction. D.M. Brennan reports no conflicts.
References
CLINICAL PERSPECTIVE
Considerable interest has focused on the development of therapies that raise systemic levels of high-density lipoprotein (HDL) cholesterol. Despite substantial elevation of HDL cholesterol, the cholesteryl ester transfer protein inhibitor torcetrapib failed to slow progression of atherosclerosis and was associated with excess mortality in clinical trials. This has fueled speculation that the functionality of HDL is impaired in the setting of cholesteryl ester transfer protein inhibition. Although torcetrapib substantially elevated HDL cholesterol but had no effect on percent atheroma volume, in this post hoc analysis an inverse relationship was observed between changes in levels of HDL cholesterol and percent atheroma volume. Disease regression at the highest levels of HDL cholesterol suggests that HDL particles can retain their functional activity, in terms of lipid mobilization, in patients treated with cholesteryl ester transfer protein inhibitors. The lack of efficacy is likely to reflect some form of off-target toxicity, such as activation of the renin-angiotensin-aldosterone axis. As a result, additional cholesteryl ester transfer protein inhibitors that lack such adverse effects may have a beneficial influence on cardiovascular disease risk. This possibility requires confirmation in prospective randomized clinical trials.
Footnotes
-
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
-
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.108.790733/DC1.
-
Clinical trial registration information—URL: http://www.clinicaltrials.gov. Unique identifier: NCT00134173.
This Issue
Jump to
Article Tools
Share this Article
- Cholesteryl Ester Transfer Protein Inhibition, High-Density Lipoprotein Raising, and Progression of Coronary AtherosclerosisCLINICAL PERSPECTIVE
