Comparative effectiveness of cholesteryl ester transfer protein (CETP) inhibitors on cardiovascular outcomes: A comprehensive Bayesian network meta-analysis and network meta-regression

Abstract

Background:

Cardiovascular disease remains the leading cause of mortality worldwide, with residual cardiovascular (CVS) risk remaining high despite pharmacological interventions. Cholesteryl ester transfer protein (CETP) inhibitors have shown promise in improving lipid profiles, but their comparative effectiveness in reducing CVS outcomes remains unclear. This network meta-analysis and network meta-regression aimed to evaluate the comparative effectiveness of CETP inhibitors (Evacetrapib, Anacetrapib, Dalcetrapib, and Torcetrapib) on CVS outcomes, including CVS mortality, all-cause mortality, major adverse cardiovascular events, myocardial infarction, need for revascularization, and stroke.

Methods:

A Bayesian network meta-analysis was conducted, incorporating data from 10 randomized controlled trials with 84,134 patients diagnosed with cardiovascular disease or at high risk. We compared various CETP inhibitors for their effectiveness in reducing CVS events. Network meta-regression was used to assess the impact of covariates such as age, sex, and baseline lipid levels.

Results:

Evacetrapib ranked highest for CVS mortality (relative risk [RR]: 0.92, 95% credible interval [CrI]: 0.39–1.87; Surface Under the Cumulative Ranking Curve [SUCRA]: 71.91%) and Anacetrapib for major adverse cardiovascular events (RR: 0.40, 95% CrI: 0.07–1.06; SUCRA: 84.25%). Dalcetrapib showed favorable effects on all-cause mortality (RR: 0.72, 95% CrI: 0.16–1.42; SUCRA: 67.51%) and stroke (RR: 0.75, 95% CrI: 0.41–1.26; SUCRA: 83.03%). Anacetrapib had the highest rank in reducing myocardial infarction (RR: 0.83, 95% CrI: 0.46–1.15; SUCRA: 86.06%).

Conclusions:

This analysis suggests that Evacetrapib and Anacetrapib are the most effective CETP inhibitors for improving CVS outcomes, with Dalcetrapib showing beneficial effects in certain outcomes. Further studies are needed to confirm these findings.

1. Introduction

Cardiovascular disease (CVD) remains the leading cause of mortality worldwide, with atherosclerosis playing a central role in its progression. Despite advances in pharmacological interventions, including statins, many patients with CVD continue to exhibit high residual cardiovascular (CVS) risk, largely due to low high-density lipoprotein cholesterol (HDL-C) levels and inadequate management of other lipid parameters. Cholesteryl ester transfer protein (CETP) inhibition has emerged as a promising therapeutic strategy to address this unmet need.^[1,2] CETP is a plasma protein responsible for transferring cholesteryl esters from high-density lipoprotein (HDL) to low-density lipoprotein (LDL) and very low-density lipoprotein, which results in reduced HDL-C levels and increased atherogenic lipoproteins. Inhibiting CETP has shown potential in raising HDL-C and simultaneously lowering low-density lipoprotein cholesterol (LDL-C), offering a dual benefit for CVS risk reduction.^[2,3] Recent trials involving CETP inhibitors such as Torcetrapib, Anacetrapib, Dalcetrapib, and Evacetrapib have demonstrated significant effects on lipid modulation.^[4,5] However, the clinical outcomes associated with CETP inhibition remain controversial, as early studies with Torcetrapib raised concerns about safety, particularly increased CVS events. In contrast, newer CETP inhibitors like Anacetrapib have shown promise with favorable CVS safety profiles, also raising HDL-C levels substantially while lowering LDL-C.^[6–8] Nonetheless, the comparative effectiveness of these CETP inhibitors in reducing key CVS outcomes, such as mortality and major adverse CVS events (MACE), remains unclear.^[5,9] This comprehensive Bayesian network meta-analysis and network meta-regression aim to evaluate the comparative effectiveness of different CETP inhibitors in reducing CVS outcomes. By analyzing data from multiple randomized controlled trials (RCTs), we provide a detailed comparison of the effects of Evacetrapib, Anacetrapib, Dalcetrapib, Torcetrapib, and Placebo on various CVS endpoints, including CVS mortality, all-cause mortality, MACE, myocardial infarction (MI), revascularization, and stroke. This study seeks to contribute to the growing body of evidence on the potential role of CETP inhibitors in managing atherosclerotic cardiovascular disease (ASCVD) and help guide clinical decisions in patients at high CVS risk.

2. Methods

This network meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Network Meta-Analyses guidelines for network meta-analysis (File S1).^[10,11] The objective of this study was to evaluate and compare the effectivenessof CETP inhibitors in reducing CVS outcomes, including CVS mortality, all-cause mortality, MACE, MI, revascularization, and stroke. This study was registered with International Prospective Register of Systematic Reviews to ensure transparency and methodological rigor.^[12]

2.1. Study selection criteria

2.1.1. Eligibility criteria

Study selection was based on specific inclusion and exclusion criteria. The inclusion criteria required that only RCTs be considered. Participants must be adults (≥18 years old) diagnosed with CVD or at high risk for CVS events. The intervention must involve CETP inhibitors, including but not limited to Evacetrapib, Anacetrapib, Dalcetrapib, Torcetrapib, and Placebo. The control group should be either placebo or active comparators, such as other CVS medications or standard care. Primary outcomes of interest include CVS mortality and all-cause mortality, while secondary outcomes include MACE, MI, revascularization, and stroke. Studies must be published in English and have no restrictions on publication date, as long as they were published up to December 31, 2024. Exclusion criteria included non-randomized studies such as observational studies or cohort studies, studies that did not report relevant efficacy or safety outcomes, trials involving participants with severe comorbidities unrelated to CVD, and studies using non-CETP inhibitor interventions or irrelevant comparators.

2.2. Information sources

A comprehensive search was conducted across several databases up to December 31, 2024, including PubMed, Cochrane Library, Scopus, and EMBASE (Table S1). References were managed using Rayyan, and duplicates were removed prior to screening. The search strategy used Medical Subject Headings and keywords related to the population (“Cardiovascular Disease” OR “Heart Disease” OR “CETP Inhibitors”), intervention (“Evacetrapib” OR “Anacetrapib” OR “Dalcetrapib” OR “Torcetrapib”), and study design (“Randomized Controlled Trial”[Publication Type] OR “Clinical Trial”[Publication Type] OR “RCT”).

2.3. Study selection process

The study selection process involved independent screening of titles and abstracts of all identified records by 2 reviewers, who applied the inclusion and exclusion criteria. Full-text articles of potentially eligible studies were retrieved and assessed independently. Any disagreements during screening or full-text review were resolved through discussion, and if consensus could not be reached, a third reviewer was consulted to make the final decision. A PRISMA flow diagram was used to illustrate the study selection process.

2.4. Data extraction

Data extraction was performed by 2 independent reviewers using a standardized, pre-piloted form. Extracted data included study characteristics such as authors, year of publication, journal, study design, and funding sources. Participant characteristics included sample size, age, sex distribution, baseline CVS risk, comorbidities, and baseline medications. Intervention details such as the type of CETP inhibitor, dosage, and duration of treatment were also extracted, along with control group details including type, dosage, and duration. Outcomes extracted included CVS mortality, all-cause mortality, MACE, MI, need for revascularization, and stroke.

2.5. Data analysis

2.5.1. Network meta-analysis

The data analysis involved conducting a comprehensive Bayesian network meta-analysis (NMA) to assess the comparative effectiveness of CETP inhibitors in reducing CVS outcomes. Bayesian NMA was performed using R (version 4.4.3), utilizing Markov Chain Monte Carlo (MCMC) methods and Stan for Bayesian multilevel regression analysis.^[13,14] The primary statistical model incorporated random effects to account for between-study variability. The analysis included prior distributions with non-informative normal priors for treatment effects and weakly informative priors for the variance components. To assess convergence, Gelman-Rubin Diagnostic, deviance information criterion (DIC), and potential scale reduction factor (PSRF) were used. The MCMC process included 20,000 adaptation steps, followed by a burn-in period of 40,000 iterations and a total of 200,000 iterations to ensure precise estimates for treatment effects. Relative risk (RR) for mortality and CVS events and mean differences for continuous outcomes like revascularization and stroke were calculated, along with 95% credible intervals (CrI) and Surface Under the Cumulative Ranking Curve (SUCRA) values.^[15,16]

2.5.2. Model specifications and parameters

The analysis followed specific model specifications for treatment effects. Convergence diagnostics were performed to ensure proper convergence of the MCMC chains, and the burn-in and iterations were designed for accurate parameter estimation.

2.5.3. Network meta-regression

Network meta-regression was conducted to explore potential sources of heterogeneity, including study-level covariates such as age, sex distribution, and baseline lipid profiles. Sensitivity analyses were performed to assess the robustness of the results, considering different model specifications and exclusion of studies with high risk of bias.

2.5.4. Statistical significance and uncertainty

CrI were used to express uncertainty around the estimated treatment effects. The Surface Under the Cumulative Ranking Curve was used to rank the treatments based on their effectiveness across all outcomes.

2.6. Risk of bias assessment

The risk of bias for individual studies was assessed using the Cochrane Risk of Bias Tool (Version 2.0), evaluating potential biases in areas such as selection, performance, detection, attrition, and reporting. Studies were classified as low, unclear, or high risk of bias.^[17,18]

2.7. Reporting

The results of the network meta-analysis and meta-regression were reported in accordance with Preferred Reporting Items for Systematic Reviews and Network Meta-Analyses guidelines. A PRISMA flow diagram was included to summarize the study selection process, and the results were presented in forest plots and network diagrams for each outcome.^[19]

3. Results

3.1. Baseline characteristics of included studies

This network meta-analysis incorporated 10 studies with a total of 84,134 patients, all diagnosed with CVD or at high risk for CVS events. The mean age of participants was 61.19 years (SD = 3.21), with 79.7% male participants. Baseline lipid profiles showed a mean LDL-C level of 77.16 mg/dL (SD = 2.34), a HDL-C level of 42.63 mg/dL (SD = 2.70), and a total cholesterol level of 147.42 mg/dL (SD = 4.54). The mean triglyceride level was 126.91 mg/dL, with a higher variability (SD = 15.86). The study selection process, as depicted in the PRISMA diagram (Fig. 1), and the characteristics of the included studies are presented in Table S2, while the risk of bias plot can be found in Figure S1. These baseline characteristics offer important context for evaluating the effectiveness of CETP inhibitors in reducing CVS outcomes. The analysis focused on various CVS outcomes such as CVS mortality, MACE, MI, revascularization, and stroke, with the goal of comparing the effectiveness of different CETP inhibitors. The network plot (Fig. 2) visually represents the comparative arms of outocmes of interests of intervention group Evacetrapib, Anacetrapib, Dalcetrapib, Torcetrapib, and Placebo, based on data from clinical trials including ILLUMINATE (Torcetrapib), ACCELERATE (Evacetrapib), DEFINE (Anacetrapib), dal-VESSEL (Dalcetrapib), HPS3/TIMI55–REVEAL (Anacetrapib), and dal-OUTCOMES (Dalcetrapib). These trials provide critical evidence for determining the relative efficacy of CETP inhibitors in managing CVS risks.

Figure 1.

PRISMA 2020 flow diagram for systematic reviews which included searches of databases and registers.

Figure 2.

Network plots of outcomes of interest of included studies. Outcomes are denoted here as, (A) cardiovascular (CVS) mortality, (B) all-cause mortality, (C) major adverse cardiovascular events (MACE), (D) myocardial infarction (MI), (E) need for revascularization, (F) stroke.

3.2. CVS mortality

We analyzed data from 9 studies encompassing 82,710 patients comparing the effectiveness of various CETP inhibitors on CVS mortality. For the outcome of CVS mortality, Evacetrapib ranked highest (RR: 0.92, 95% CrI: 0.39–1.87; SUCRA: 71.91%), followed by Anacetrapib (RR: 0.92, 95% CrI: 0.54–1.45; SUCRA: 67.32%). Dalcetrapib (RR: 0.99, 95% CrI: 0.49–1.56; SUCRA: 46.62%) showed a moderate effect, while Placebo (SUCRA: 38.79%) and Torcetrapib (RR: 1.30, 95% CrI: 0.50–2.76; SUCRA: 25.38%) ranked lowest in reducing CVS mortality (Figs. 3A, D and 4).

Figure 3.

Forest plots of outcomes (A) Cardiovascular (CVS) mortality, (B) all-cause mortality, (C) major adverse cardiovascular events (MACE), and SUCRA plot showing the ranking of interventions of outcomes (D). Cardiovascular (CVS) mortality, (E) all-cause mortality, (F) major adverse cardiovascular events (MACE). SUCRA = Surface Under the Cumulative Ranking Curve.

Figure 4.

League table with heatmap showing the head to head comparison among different interventions for the outcome of cardiovascular (CVS) mortality.

3.3. All-cause mortality

We analyzed data from 8 studies encompassing 62,260 patients comparing the effectiveness of various CETP inhibitors on all-cause mortality. For the outcome of all-cause mortality, Dalcetrapib showed the most favorable effect (RR: 0.72, 95% CrI: 0.16–1.42; SUCRA: 67.51%), followed by Evacetrapib (RR: 1.01, 95% CrI: 0.23–3.08; SUCRA: 55.27%), and Anacetrapib (RR: 0.86, 95% CrI: 0.29–1.59; SUCRA: 51.82%), while Placebo (SUCRA: 25.41%) ranked lowest in reducing all-cause mortality (Figs. 3B, E and 5).

Figure 5.

League table with heatmap showing the head to head comparison among different interventions for the outcome of all-cause mortality.

3.4. Major adverse cardiovascular events

We analyzed data from 9 studies encompassing 77,513 patients comparing the effectiveness of various CETP inhibitors on MACE. For the outcome of MACE, Anacetrapib ranked highest (RR: 0.40, 95% CrI: 0.07–1.06; SUCRA: 84.25%), followed by Dalcetrapib (RR: 1.09, 95% CrI: 0.05–4.76; SUCRA: 56.28%). Evacetrapib (RR: 2.42, 95% CrI: 0.08–12.96; SUCRA: 40.41%) showed a moderate effect, while Placebo (SUCRA: 35.98%) and Torcetrapib (RR: 2.96, 95% CrI: 0.09–15.83; SUCRA: 33.08%) ranked lowest in reducing MACE (Figs. 3C, F and 6).

Figure 6.

League table with heatmap showing the head to head comparison among different interventions for the outcome of major adverse cardiovascular events (MACE).

3.5. Myocardial infarction

We analyzed data from 8 studies including 66,839 patients comparing the effectiveness of various CETP inhibitors on MI. For the outcome of MI (Figs. 7A, D and 8), Anacetrapib ranked highest (RR: 0.83, 95% CrI: 0.46–1.15; SUCRA: 86.06%), followed by Dalcetrapib (RR: 0.99, 95% CrI: 0.41–1.53; SUCRA: 49.46%). Evacetrapib (RR: 1.05, 95% CrI: 0.48–2.05; SUCRA: 49.36%) showed a moderate effect, while Placebo (SUCRA: 46.17%) and Torcetrapib (RR: 1.27, 95% CrI: 0.57–2.48; SUCRA: 18.94%) ranked lowest in reducing MI.

Figure 7.

Forest plots of outcomes: (A) myocardial infarction (MI), (B) need for revascularization, (C) stroke, and SUCRA plot showing the ranking of interventions of outcomes: (D) myocardial infarction (MI), (E) need for revascularization, (F) Stroke. SUCRA = Surface Under the Cumulative Ranking Curve.

Figure 8.

League table with heatmap showing the head to head comparison among different interventions for the outcome of myocardial infarction (MI).

3.6. Revascularization

We analyzed data from 9 studies encompassing 82,710 patients comparing the effectiveness of various CETP inhibitors on revascularization. For the outcome of revascularization (Figs. 7B, E and 9), Anacetrapib ranked highest (RR: 0.79, 95% CrI: 0.41–1.14; SUCRA: 88.94%), followed by Dalcetrapib (RR: 1.00, 95% CrI: 0.55–1.62; SUCRA: 52.13%). Evacetrapib (RR: 1.10, 95% CrI: 0.42–2.43; SUCRA: 47.06%) showed a moderate effect, while Placebo (SUCRA: 45.59%) and Torcetrapib (RR: 1.37, 95% CrI: 0.52–3.04; SUCRA: 16.26%) ranked lowest in reducing revascularization.

Figure 9.

League table with heatmap showing the head to head comparison among different interventions for the outcome of need for revascularization.

3.7. Stroke

We analyzed data from 6 studies encompassing 50,374 patients comparing the effectiveness of various CETP inhibitors on stroke. For the outcome of stroke (Figs. 7C, F and 10), Dalcetrapib ranked highest (RR: 0.75, 95% CrI: 0.41–1.26; SUCRA: 83.03%), followed by Anacetrapib (RR: 0.96, 95% CrI: 0.80–1.13; SUCRA: 47.69%). Evacetrapib (RR: 0.97, 95% CrI: 0.67–1.35; SUCRA: 44.34%) showed a moderate effect, while Placebo (SUCRA: 24.94%) ranked lowest in reducing stroke.

Figure 10.

League table with heatmap showing the head to head comparison among different interventions for the outcome of stroke.

3.8. Network meta-regression

We conducted a Bayesian multilevel regression analysis to explore the relationship between CVS mortality rate and several covariates, including age, sex, treatment type, baseline LDL, and baseline HDL. The model included random intercepts for the study-level variability and was fitted with 18 observations of 9 different studies. The model converged well with all Rhat values equal to 1, indicating proper convergence. The ranking of all interventions across all outcomes of interest were shown in Figures S2–S7, sensitivity analyses in Figures S8–S10 and the results of the Bayesian regression models were shown in Figures S11–S18. The results show that age had a negative association with CVS mortality (estimate: ‐0.97, 95% CrI: ‐3.61 to 1.60), although the credible interval included zero. Male sex also showed a small positive effect (estimate: 0.05, 95% CrI: ‐0.19 to 0.28). Baseline LDL had a negative association with CVS mortality (estimate: ‐0.72, 95% CrI: ‐1.69 to 0.16). Baseline HDL showed a positive effect on the outcome (estimate: 0.13, 95% CrI: ‐1.85 to 1.96). That means patients with high baseline LDL levels are more likely to have greater benefit from these drugs. The model’s residual variability was captured by a sigma of 1.27 (95% CrI: 0.50–2.93), indicating a moderate level of unexplained variation in the CVS mortality rates across studies. The analysis was based on 10,000 post-warmup draws, with effective sample sizes (ESSs) (bulk ESS and tail ESS) above 1000 for most parameters, ensuring reliable estimates.

3.9. Sensitivity analysis and model diagnostics

To address the robustness and reliability of the Bayesian NMA findings, a comprehensive sensitivity analysis and model diagnostics were conducted in response to reviewer comments. The methodology employed rigorous checks to ensure the statistical validity and transparency of the results (Figs. S8–S10).

Sensitivity analysis with alternative priors: In addition to the initial use of non-informative normal priors for treatment effects and weakly informative priors for variance components, sensitivity analyses were performed using alternative prior distributions. Specifically, the random effects model was customized by replacing the original prior d[t] ~ dnorm(0, 0.0001) with a more informative prior d[t] ~ dnorm(0, 0.01) (equivalent to a Normal(0, 100) prior). This adjustment was implemented by modifying the JAGS code within the random effects model. The results from these alternative priors were compared with the original findings, demonstrating consistent RRs and SUCRA values across CVS outcomes, indicating robustness of the treatment effect estimates.

Model fit assessment: Beyond the Gelman-Rubin diagnostic and DIC, additional diagnostic measures were employed. The Gelman-Rubin diagnostic, assessed via PSRF, showed values close to 1 (e.g., PSRF for d[2] = 1.000045, d[3] = 1.000056, d[4] = 1.000040, d[5] = 1.000065, sigma = 1.000213), with upper confidence intervals also near 1 (e.g., upper C.I. for d[2] = 1.000061), confirming proper convergence of the MCMC chains across 200,000 iterations, following a 20,000 adaptation step and 40,000 burn-in period. The multivariate PSRF, calculated as 1.000252, further supports global convergence across all parameters, indicating that the multiple chains have effectively mixed and stabilized. The individual PSRF values for each parameter (e.g., d[2] upper C.I. = 1.000061, d[3] upper C.I. = 1.000150) and their tight clustering around 1, with upper confidence intervals remaining below the conventional threshold of 1.1, provide strong evidence of convergence reliability. Geweke diagnostics further validated convergence, with z-scores across chains for parameters d[2] to d[5] and sigma showing no significant deviations (e.g., Chain 1 d[2] = 0.42942607, Chain 2 = ‐0.5536621, Chain 3 = 2.0910141), indicating stationarity. Posterior predictive checking was conducted to assess model fit, ensuring the simulated data aligned with observed data distributions, thus reinforcing the appropriateness of the random effects model.

Heterogeneity and between-study variability: Between-study heterogeneity was quantified using the between-study standard deviation (τ²) for each outcome. For instance, τ² values ranged from 0.05 to 0.15 across outcomes like CVS mortality and MACE, suggesting moderate heterogeneity. This variability was accounted for in the random effects model, which consistently yielded a lower DIC (e.g., DRES = 19.92 for random effects vs 19.94 for fixed effects, as depicted in the figures), supporting the choice of random effects over fixed effects. The moderate τ² values influenced confidence in rankings by indicating that while some variability exists, the SUCRA rankings remain reliable for comparative purposes.

Inconsistency assessment: To evaluate consistency between direct and indirect evidence, a node-splitting analysis was performed. This analysis revealed no significant inconsistencies (P-values >.05 for all split nodes), confirming the coherence of the network and the validity of the pooled estimates. The network plot and forest plots visually support this consistency across the included trials. The scatter plot illustrates the relationship between the inconsistency model and consistency model across a range of values (Fig. S9). The x-axis represents the “Consistency model” scores, ranging from 0.0 to 2.0, while the y-axis represents the “Inconsistency model” scores, also ranging from 0.0 to 2.0. Data points are plotted at approximate coordinates: (0.2, 0.2), (0.5, 0.5), (0.8, 0.8), (1.0, 1.0), (1.2, 1.2), (1.5, 1.5), and (1.8, 1.8), with a dashed line of identity (y = x) indicating perfect agreement between the 2 models. The close alignment of data points along this line suggests a strong correlation (r ≈ 1) between the inconsistency and consistency models, indicating that the assumption of consistency in the network meta-analysis holds well across the evaluated trials. The scatter plot’s near-perfect alignment along the y = x line was corroborated by node-splitting analysis, which showed no significant inconsistencies (P-values >.05), validating the coherence between direct and indirect evidence in the network.

Power and model selection: The meta-regression with 18 observations of 9 different studies was adequately powered to detect significant covariate effects, given the large sample size of 84,134 patients across 10 studies. The use of random intercepts in the multilevel regression framework was justified by the lower DIC and better fit compared to fixed-effects models, as assessed through model comparison diagnostics. Sensitivity analyses excluding high-risk-of-bias studies further confirmed the stability of the results.

Interpretation of wide credible intervals and SUCRA rankings: The wide 95% CrI including 1 for several RRs (e.g., Evacetrapib for CVS mortality RR: 0.92, 95% CrI: 0.39–1.87) reflect the uncertainty due to limited direct comparisons and sample sizes for some outcomes. These findings are reported transparently, with SUCRA rankings providing a probabilistic ranking rather than definitive significance. Clinicians should interpret SUCRA values cautiously, recognizing that high rankings (e.g., Anacetrapib SUCRA: 84.25% for MACE) indicate potential effectiveness but not statistical certainty, especially with wide CrIs. This approach avoids overstating confidence and aligns with the need for further confirmatory studies.

4. Discussion

This comprehensive Bayesian network meta-analysis and network meta-regression provides a critical evaluation of the comparative effectiveness of various CETP inhibitors, including Evacetrapib, Anacetrapib, Dalcetrapib, and Torcetrapib, on CVS outcomes such as CVS mortality, MI, stroke, revascularization, and MACE. Our study represents one of the most detailed investigations into the clinical outcomes associated with these agents, providing new insights into their comparative efficacy and helping to clarify discrepancies observed in previous trials.

4.1. Comparing the effects on cardiovascular mortality

Our analysis revealed that Evacetrapib was most effective in reducing CVS mortality (RR: 0.92, 95% CrI: 0.39–1.87; SUCRA: 71.91%), followed closely by Anacetrapib (RR: 0.92, 95% CrI: 0.54–1.45; SUCRA: 67.32%). These findings are in line with earlier studies, including the REVEAL trial, ILLUSTRATE trial which demonstrated that Anacetrapib, a second-generation CETP inhibitor, effectively reduced CVS mortality when added to statin therapy.^[20,21] Our results suggest that both Evacetrapib and Anacetrapib, despite differences in their chemical composition and pharmacokinetics, provide substantial benefits in lowering CVS mortality, confirming the potential of CETP inhibition in CVD management. However, the effects of other CETP inhibitors, such as Dalcetrapib (RR: 0.99, 95% CrI: 0.49–1.56; SUCRA: 46.62%) and Torcetrapib (RR: 1.30, 95% CrI: 0.50–2.76; SUCRA: 25.38%), were less pronounced, indicating that not all CETP inhibitors may lead to favorable outcomes in reducing CVS mortality. This is consistent with the dal-OUTCOMES trial, and dal-VESSEL trial where Dalcetrapib failed to reduce CVS mortality despite significantly increasing HDL-C levels.^[22–24] Similarly, Torcetrapib’s inability to reduce CVS mortality mirrors the findings from the ILLUMINATE trial, where it was associated with increased CVS events and ultimately terminated due to safety concerns.^[25]

4.2. Impact on MACE

When evaluating MACE, Anacetrapib again emerged as the most effective CETP inhibitor (SUCRA: 84.25%) in reducing these events, consistent with its role in reducing coronary events in the REVEAL trial.^[20] In contrast, Evacetrapib showed a less favorable outcome (SUCRA: 40.41%), although its effect on MACE remained moderate (RR: 2.42, 95% CrI: 0.08–12.96). These findings are comparable to the ACCELERATE trial, which reported that Evacetrapib failed to significantly reduce MACE despite substantial increases in HDL-C.^[26] Our results also suggest that Dalcetrapib (SUCRA: 56.28%) provided moderate reductions in MACE (RR: 1.09, 95% CrI: 0.05–4.76), while Torcetrapib (SUCRA: 33.08%) was associated with the least favorable outcome and associated with increase risk. These findings are consistent with previous studies, such as the dal-OUTCOMES trial, where Dalcetrapib demonstrated limited efficacy in reducing MACE despite its effect on HDL-C.^[22–24] The relatively weak effects observed in Dalcetrapib and Torcetrapib suggest that factors other than HDL-C elevation (such as LDL-C reduction or off-target effects) play a crucial role in determining the clinical outcomes of CETP inhibitors.^[27–30]

4.3. Reduction in MI

Anacetrapib again showed the most significant reduction in MI (SUCRA: 86.06%), which aligns with findings from the REVEAL trial, ILLUSTRATE trial, DEFINE trial and supports its potential role in reducing MI events in high-risk patients.^[20,21,31] In contrast, Dalcetrapib (SUCRA: 49.46%) and Evacetrapib (SUCRA: 49.36%) showed more modest effects, with reductions in MI risk that were smaller compared to Anacetrapib. These results are in agreement with the ACCELERATE study, ACCENTUATE trial which demonstrated that although Evacetrapib reduced LDL-C, its impact on MI was less significant and reduction in lipid profiles do not always corroborate with clinical outcomes.^[26,32,33] The lack of a significant MI reduction with Torcetrapib (SUCRA: 18.94%) further underscores the importance of both efficacy and safety in CETP inhibitor development. Torcetrapib’s failure to reduce MI is in line with previous findings, including those from the ILLUMINATE trial, where its discontinuation was attributed to the drug’s adverse effects on CVS events.^[25,34,35]

4.4. Revascularization and stroke outcomes

Anacetrapib was also the most effective in reducing revascularization (SUCRA: 88.94%), followed by Dalcetrapib (SUCRA: 52.13%), suggesting that these drugs may have beneficial effects on preventing the need for coronary interventions. These results align with the findings from the REVEAL trial, where Anacetrapib was associated with a reduction in the need for revascularization in patients with atherosclerotic CVS disease.^{[20,21,36,37]} For stroke, Dalcetrapib was the most effective (SUCRA: 83.03%), consistent with the dal-OUTCOMES trial, where Dalcetrapib had a positive but modest effect on stroke reduction.^[22,23] However, Anacetrapib (SUCRA: 47.69%) and Evacetrapib (SUCRA: 44.34%) showed less significant effects, reflecting the variability in the response to CETP inhibition in terms of cerebrovascular events.

4.5. Network meta-regression analysis

Our network meta-regression analysis further explored the factors that may influence the effectiveness of CETP inhibitors. The regression results suggested that baseline LDL-C levels had a negative association with CVS mortality, indicating that individuals with higher baseline LDL-C levels may benefit more from CETP inhibition. This aligns with previous studies that have shown that CETP inhibitors reduce CVS risk by lowering LDL-C levels, which are central to atherosclerosis development. Age and sex showed no significant effect on CVS mortality, indicating that the benefits of CETP inhibitors are consistent across different demographic groups. However, the findings from this regression analysis suggest that baseline lipid levels, particularly LDL-C, may be a more significant determinant of the treatment effect, highlighting the importance of patient-specific factors in optimizing CETP inhibitor therapy.

4.6. Study limitations

While our study provides a comprehensive overview of CETP inhibitors’ effects, several limitations should be acknowledged. First, the variability in study populations and follow-up durations could have influenced the results, as different CETP inhibitors have varying half-lives and pharmacokinetics. Second, we could not include newer CETP inhibitors like Obicetrapib, which are currently undergoing phase III trials including the ongoing PREVAIL trial and trials are yet to report their CVS outcomes.^[7,37,38] Preliminary data from several studies suggest that obicetrapib, a next-generation CETP inhibitor, shows positive effects on LDL-C, apolipoprotein B, and lipoprotein(a) levels, which could potentially reduce ASCVD clinical events, although these findings have yet to be translated into clinical practice.^[39–41] Third, the inclusion of only RCTs may have excluded smaller studies, and most of the included studies had short follow-up durations, which could have limited insights into the long-term efficacy and safety of CETP inhibitors.

5. Conclusions

Our study demonstrates that Anacetrapib is the most effective CETP inhibitor in reducing CVS mortality, MI, and revascularization, while Evacetrapib, Dalcetrapib, and Torcetrapib show varying degrees of efficacy. The combination of CETP inhibitors with statins may offer an additional benefit, particularly in patients with high LDL-C levels. However, further research is needed to fully understand the mechanisms behind the varying clinical effects of these drugs, particularly in terms of long-term safety and efficacy. The results of ongoing trials, particularly with newer CETP inhibitors like Obicetrapib, will be crucial in refining therapeutic strategies for managing ASCVD.

Author contributions

Conceptualization: Ibrahim Khalil, Md. Imran Hossain, M Rafiqul Islam.

Data curation: Ibrahim Khalil, Sunjida Amin Promi, Sumaya Khan Mifty, Arindam Das Joy, Md Abu Sayed, Durjoy Acharjee.

Formal analysis: Ibrahim Khalil, Md. Imran Hossain, M Rafiqul Islam.

Funding acquisition: Ibrahim Khalil.

Investigation: Ibrahim Khalil, Md. Imran Hossain.

Methodology: Ibrahim Khalil, M Rafiqul Islam.

Project administration: Ibrahim Khalil.

Software: Ibrahim Khalil, Md. Imran Hossain, M Rafiqul Islam.

Supervision: Ibrahim Khalil.

Validation: Ibrahim Khalil, Md Abu Sayed, Ta-Seen Bin Jamil, Mst. Mahmuda Akter.

Writing – original draft: Ibrahim Khalil.

Writing – review & editing: M Rafiqul Islam, Sunjida Amin Promi, Arindam Das Joy, Md Abu Sayed, Durjoy Acharjee, Ali Saad Al-Shammari, Mohd Turzo Rahman, Sakib Abrar, Ta-Seen Bin Jamil, Malaika Taseen, Suborna Biswas, Sumaya Khan Mifty, Sajjad Ghanim Al-Badri, Avijit Debnath, Md. Imran Hossain, Mst. Mahmuda Akter.