Efficacy and safety of novel lipoprotein(a)‐targeted therapies: A systematic review and meta‐analysis of randomised controlled trials

Elevated lipoprotein(a) [Lp(a)] is a genetically determined, causal risk factor for atherosclerotic cardiovascular disease (ASCVD), yet it remains underrecognized in high-risk populations.1 This risk is particularly relevant in patients with metabolic syndrome, where current therapies like PCSK9 inhibitors offer only modest Lp(a) reductions.2 This therapeutic gap has driven the development of novel agents, including antisense oligonucleotides (ASOs), small interfering RNAs (siRNAs), and oral small-molecule inhibitors, which have shown promise in recent trials.3 However, a pooled quantitative assessment of this new class of drugs is lacking. Therefore, the objective of this systematic review and meta-analysis was to quantify the efficacy of emerging Lp(a)-targeted therapies in reducing Lp(a) and related lipids, while assessing their early safety profile. This systematic review and meta-analysis was registered on PROSPERO (CRD420251109408) and is reported in accordance with the PRISMA guidelines. We systematically searched PubMed, Embase, and Cochrane CENTRAL through 9 November 2025. Our updated and expanded search strategy included both specific drug names (e.g., pelacarsen, muvalaplin) and broad mechanistic terms (e.g., “Lp(a) inhibitor,” “LPA siRNA”). The full, detailed search string for all databases is available in the Supplementary Appendix 1. Two authors independently screened all records. We included RCTs comparing Lp(a)-targeted therapies versus placebo in patients with Lp(a) ≥125 nmol/L and extracted data from the highest-dose arms to assess maximal therapeutic efficacy. The primary efficacy outcomes were the mean differences (MDs) in Lp(a), LDL-C, and ApoB. Key secondary outcomes included C-reactive protein (hsCRP), and safety was assessed by serious adverse events and injection site reactions. A pre-specified subgroup analysis by drug class (siRNA, ASO, oral inhibitor) was performed. We pooled MDs and odds ratios (ORs) using a random-effects model with R software. Statistical heterogeneity was assessed with the I2 statistic. Study quality was assessed using the Cochrane RoB 2 tool and the GRADE approach. A detailed description of our complete methodology—including eligibility criteria, data extraction procedures, handling of missing data, quality and bias assessment, and the full statistical analysis plan with the statistical formulas used—is provided in the Supplementary Appendix 2. Our search identified 1282 records, from which 8 RCTs (reported across 9 publications) met the inclusion criteria (Supplementary Figure S1).4-12 The 8 RCTs included 899 patients (569 receiving therapy; 330 placebo). The pooled population was 63.0% male, with a mean age ranging from 43.7 to 64.8 years, and a mean baseline Lp(a) ranging from 203 to 302 nmol/L. Therapeutic approaches included siRNA (5 trials), ASO (2 trials), and an oral inhibitor (1 trial). Detailed baseline characteristics are presented in Supplementary Table S1. To address the high overall heterogeneity between therapies, our primary efficacy assessment was the pre-specified subgroup analysis (Figure 1). A significant difference was found between therapeutic classes (p for subgroup differences <0.001). SiRNA agents demonstrated the largest mean reduction (MD −93.73%; 95% CI −100.48 to −86.98; p <0.001; I2 = 86.4%), followed by the oral inhibitor (MD −85.80%; 95% CI −88.00 to −83.60; p <0.001) and ASOs (MD −67.67%; 95% CI −76.64 to −58.71; p <0.001; I2 = 22.9%). Notably, substantial heterogeneity persisted within the siRNA subgroup, suggesting true potency differences between individual agents. In contrast to Lp(a), the effects on other atherogenic lipoproteins were consistent across drug classes (Figure 2). Treatment significantly lowered pooled LDL-C levels (MD −21.64%; 95% CI −26.34 to −16.93; p <0.001; I2 = 31.1%; Figure 2A), with low overall heterogeneity and no difference between subgroups (p = 0.238). A similar and consistent reduction was observed for ApoB (MD −14.37%; 95% CI −16.40 to −12.35; p <0.001; I2 = 0.0%; Figure 2B), with no difference between classes (p = 0.807). No significant effects were found in triglyceride levels (MD −0.94%; 95% CI −7.72 to 5.84; p = 0.793; I2 = 45.3%; Supplementary Figure S2A) or HDL-C levels (MD −0.40%; 95% CI −6.38 to 5.58; p = 0.901; I2 = 87.3%; Supplementary Figure S2B). Similarly, no significant change was observed in the systemic inflammatory marker hsCRP (MD −0.04 mg/L; 95% CI −0.23 to 0.15; p = 0.713; I2 = 0.0%; Supplementary Figure S2C). The overall safety profile was favourable. The incidence of total adverse events (OR 1.50; 95% CI 0.82–2.75; p = 0.184; I2 = 31.6%; Supplementary Figure S3A), as was the incidence of serious adverse events was similar between groups (OR 0.96; 95% CI 0.43–2.18; p = 0.932; I2 = 27.1%; Supplementary Figure S3B). The one consistent safety signal was a significantly higher incidence of injection site reactions (OR 10.26; 95% CI 2.23–47.20; p <0.001; I2 = 68.7%; Supplementary Figure S3C). Analyses of other biomarkers showed significant reductions in total cholesterol and oxidised phospholipids (oxApoB, oxApoA), which are detailed in the appendix. The overall quality of the included trials was high, with six of the eight studies at low risk of bias (RoB 2). The certainty of evidence (GRADE) was rated high for ApoB and LDL-C reductions, but moderate for the primary Lp(a) outcome (downgraded for high heterogeneity). Visual inspection of the funnel plots suggested potential small-study effects for the Lp(a) outcome, though this was likely confounded by the high inter-class heterogeneity; plots for LDL-C and ApoB showed no clear evidence of bias. Furthermore, a comprehensive leave-one-out sensitivity analysis confirmed the robustness of the primary findings, showing that no single study unduly influenced the pooled estimates for Lp(a), LDL-C, or ApoB. All detailed figures and tables—including the PRISMA diagram, baseline characteristics, full quality/bias assessments (funnel plots, RoB2, GRADE tables), and all supplementary forest plots—are available in the Supplementary Material. The marked potency of these novel therapies positions them as a potential solution to a long-standing therapeutic challenge in lipid management. Elevated Lp(a) is a “treatment-resistant” risk factor, as it is largely unaffected by lifestyle interventions –such as intermittent fasting,13 dietary supplements like green coffee extract,14 or specific dietary components like animal protein15 that have been shown to modulate other cardiometabolic or inflammatory outcomes—and is ineffectively lowered by statins. Non-targeted therapies like PCSK9 inhibitors offer only modest reductions (~25%), even in meta-analyses of diabetic or metabolic syndrome populations.2 Epidemiological modelling suggests a substantial Lp(a) reduction is required to meaningfully lower cardiovascular risk.16 Our analysis provides the first pooled evidence that this new generation of drugs, particularly siRNAs, can achieve a level of reduction that meets or exceeds this proposed therapeutic threshold. A primary finding of our analysis is that the high overall heterogeneity (I2 = 94.5%) observed in a simple pooled analysis is clearly explained by pre-specified differences between therapeutic classes (p for subgroup differences <0.001), as shown in Figure 1. SiRNA agents were the most potent (MD −93.73%), followed by the oral inhibitor (−85.80%) and ASOs (−67.67%). Furthermore, the high heterogeneity within the siRNA subgroup (I2 = 86.4%) suggests that a uniform class effect is an oversimplification, pointing to true potency differences between individual agents (e.g., Olpasiran vs. Zerlasiran). Our interpretation is strongly supported by a recent network meta-analysis, which also compared siRNA agents and identified Olpasiran as the most potent.17 We also identified two key ancillary findings. First, treatment consistently lowered LDL-C by approximately 22% and ApoB by 14% across all drug classes (Figure 2). The LDL-C reduction must be contextualized, as standard assays co-measure cholesterol from Lp(a) particles (Lp(a)-C)—a known measurement artefact.1 Thus, a portion of this effect is an inherent result of potent Lp(a) lowering. Second, we observed a clear decoupling of inflammation: despite potent Lp(a) reduction, there was no corresponding change in systemic hsCRP. This finding supports a model where Lp(a)'s pathogenic actions are confined to the local arterial wall. The safety profile from these Phase 2 trials appears favourable, with no significant increase in serious adverse events. The one consistent safety signal was the higher rate of injection-site reactions, a known characteristic of oligonucleotide therapeutics, as also noted in the network meta-analysis by Liu et al.17 The development of effective oral agents, such as muvalaplin (included in our analysis), may circumvent this issue entirely.12 This study has limitations. First, while our subgroup analysis explained the primary source of heterogeneity, the estimates for the non-siRNA arms were based on a small number of trials (n = 2 for ASO, n = 1 for oral). Second, this meta-analysis relies on Phase 2 trials using surrogate endpoints and hard cardiovascular outcomes are not yet available. Third, most included trials were industry-funded, a potential source of bias. Finally, future studies must ensure equitable inclusion of diverse racial and ethnic groups, as Lp(a) levels and associated risk vary significantly across populations.1 The clinical benefit of these agents is currently being evaluated in large-scale cardiovascular outcome trials (e.g., Lp(a) HORIZON, OCEAN(a)-Outcomes), which will be essential to establish their definitive role in clinical practice.18 In this meta-analysis, emerging Lp(a)-targeted therapies achieved a marked mean reduction in Lp(a) concentrations, with siRNAs appearing most potent. Significant ancillary reductions in LDL-C and ApoB were also observed. While associated with a higher rate of injection site reactions, the overall early safety profile was favourable. These findings on surrogate endpoints provide a strong rationale for the ongoing cardiovascular outcome trials needed to establish definitive clinical benefit. Anisio Uchoa Leite Santana: Conceptualization, Methodology, Data Curation, Formal Analysis, Investigation, Writing—Original Draft, Project Administration. Muhammad Owais Yusufzai: Data Curation, Investigation, Writing—Review and Editing. Arthur Pinheiro Silveira Freitas: Formal Analysis, Validation, Writing—Review and Editing. Vitor Batalha de Camargo: Data Curation, Validation, Investigation. Lucas de Queiroz Menezes: Investigation, Writing—Review and Editing. Miguel Chaves Lenzi: Methodology, Validation, Writing—Review and Editing. Clara Uchoa Leite Santana: Investigation, Writing—Review and Editing. Debora de Queiroz Menezes: Investigation, Writing—Review and Editing. Marcilio de Oliveira Filho: Investigation, Methodology. Andre Jorge Nogues de Almeida: Methodology, Writing—Review and Editing. Alfredo Rogerio Dias Dalcomuni: Methodology, Writing—Review and Editing. Jose Eduardo Viana de Rezende Meira: Writing—Review and Editing, Validation. Eduardo Bello Martins: Conceptualization, Supervision, Writing—Review and Editing. Carlos Vicente Serrano Junior: Conceptualization, Supervision, Writing—Review and Editing. We would like to thank the participants and investigators of all included trials for their contributions. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.70349. The data supporting the findings of this systematic review and meta-analysis are derived from the included randomized controlled trials, which are fully cited in the manuscript. The data extraction tables and the final dataset generated for the statistical analysis are available from the corresponding author upon reasonable request. Data S1 Supplementary Materials Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.