Association of Elevated Lipoprotein(a) levels with vulnerable plaque features assessed by multimodality coronary imaging in patients with coronary artery disease: A systematic review and Meta-analysis

Abstract

Background

Coronary artery plaques are the pathological basis of coronary artery disease (CAD), and the rupture of vulnerable plaques and thrombosis are important causes of major adverse cardiovascular events (MACE). Lipoprotein(a) is a lipid particle which has been proven to be related to the risk of MACE, but the association between its level and the specific imaging phenotype of vulnerable coronary artery plaques has not been clearly established.

Methods

We searched PubMed, Embase, and the Cochrane Library databases from inception to June 19, 2026, for observational studies evaluating the association between elevated serum Lp(a) levels and coronary plaque characteristics. For OCT-derived features, we compared the relationship between elevated serum Lp(a) levels and the formation of thin-cap fibroatheroma (TCFA), plaque rupture, maximum lipid arc, and minimum lumen area (MLA) in patients with coronary artery disease. Similarly, for CCTA-derived features, we compared the relationship between elevated serum Lp(a) levels and plaque burden, positive remodeling, spotty calcification, napkin ring sign, and low attenuation plaques (LAPs) in patients with CAD.

Results

This meta-analysis included 13 studies involving 6671 patients (7 studies used optical coherence tomography (OCT) and 6 studies used coronary computed tomography angiography). For the primary outcomes, a meta-analysis of 7 OCT studies (n = 1773 patients) showed that elevated Lp(a) was associated with a significantly increased risk of TCFA (OR 2.57; 95% CI 1.54–4.30; I² = 61%; P = 0.0003). Conversely, a meta-analysis of 5 CCTA studies (n = 3581 patients) revealed no statistically significant difference in plaque burden between the high and normal Lp(a) groups (SMD 0.27; 95% CI -0.01 to 0.56; I² = 90%; P = 0.06). For secondary outcomes, elevated Lp(a) was significantly associated with a higher risk of OCT-derived plaque rupture (4 studies; OR 1.71; 95% CI 1.25–2.34; I² = 0%; P = 0.0008) and a larger maximum lipid arc (3 studies; SMD 0.84; 95% CI 0.31–1.38; I² = 79%; P = 0.002). For CCTA-derived features, elevated Lp(a) was associated with an increased risk of spotty calcification (3 studies, OR 1.78; 95% CI 1.19–2.68; I² = 14%; P = 0.005), napkin-ring sign (3 studies, OR 1.46; 95% CI 1.05–2.03; I² = 0%; P = 0.02), and LAPs (3 studies, OR 1.52; 95% CI 1.06–2.17; I² = 69%; P = 0.02).

Conclusion

This meta-analysis suggests that high Lp(a) levels are associated with several imaging-defined high-risk coronary plaque features, with the most robust evidence observed for OCT-derived TCFA and plaque rupture. By integrating evidence from both OCT and CCTA, our findings provide a comprehensive imaging-based characterization of the vulnerable plaque phenotype associated with elevated Lp(a), although several CCTA-derived findings should be interpreted cautiously. Further research is needed to determine whether Lp(a)-lowering therapy can modify high-risk coronary plaque features and reduce subsequent cardiovascular events.

Graphical abstract

1. Introduction

Coronary artery disease (CAD) remains a leading cause of global morbidity and mortality, frequently presenting with acute and severe clinical presentations such as myocardial infarction and cardiac death [1], [2], [3]. Although traditional cardiovascular risk factors such as smoking, diabetes, and hypercholesterolemia are well established, CAD may also occur in individuals without standard modifiable risk factors (SMuRFs), highlighting the need for additional markers of atherosclerotic risk [4], [5]. Advances in coronary imaging, particularly coronary computed tomography angiography (CCTA) and optical coherence tomography (OCT), have enabled detailed assessment of plaque burden and vulnerability-related features, such as thin-cap fibroatheroma (TCFA), low-attenuation plaque (LAPs), positive remodeling, and the napkin-ring sign [6], [7]. These high-risk plaque features are considered imaging markers of plaque vulnerability and are closely related to plaque rupture and subsequent thrombosis [8].

Lipoprotein(a) [Lp(a)] is a large, LDL-like particle containing apolipoprotein(a) and has pro-atherogenic and pro-thrombotic properties through its involvement in lipid deposition, inflammation, coagulation, and fibrinolysis [9], [10]. Unlike conventional lipid parameters, Lp(a) levels are largely genetically determined and remain relatively stable throughout life, making it a clinically attractive biomarker for cardiovascular risk stratification [11]. Growing evidence from epidemiological studies and Mendelian randomization analyses has established elevated Lp(a) as a causal risk factor for myocardial infarction, ischemic stroke, and calcific aortic valve stenosis [2], [12], [13], [14], [15]. However, the mechanisms linking Lp(a) to adverse cardiovascular outcomes, particularly its relationship with the morphology and composition of coronary plaques, remain incompletely understood.

Several observational studies have examined the association between Lp(a) and coronary plaque features assessed using CCTA or OCT, but their findings remain inconsistent [16], [17]. Existing evidence remains fragmented because most studies focused on selected plaque features within a single imaging modality, used heterogeneous Lp(a) thresholds, and enrolled different clinical populations. Differences in study populations, sample size, Lp(a) assays, and imaging definitions may partly explain these discrepancies and limit definitive conclusions. Therefore, we conducted this systematic review and meta-analysis to quantitatively synthesize available evidence on the relationship between Lp(a) levels and coronary plaque characteristics, including plaque burden, high-risk plaque features, and plaque composition. By integrating OCT-derived microstructural plaque features and CCTA-derived high-risk plaque characteristics, this study aimed to clarify the broader imaging-defined vulnerable plaque phenotype associated with elevated Lp(a) and to inform future Lp(a)-based cardiovascular risk stratification.

2. Methods

This review was conducted in accordance with the recommendations from the Cochrane Collaboration (Cochrane Guidelines) and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines [18]. The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO; registration ID: CRD420251232659).

2.1. Eligibility criteria

Clinical human studies were considered eligible irrespective of publication language. Studies were deemed eligible if they met all of the following inclusion criteria: (1) The research population consisted of adults with coronary artery disease. (2) Studies that conducted invasive assessment of coronary plaque characteristics (CPC) using OCT, or non-invasive assessment using CCTA. (3) Studies analyzing the relationship between Lp(a) levels or elevated Lp(a) and coronary plaque characteristics. (4) Studies describing the outcomes of coronary plaque characteristics, including: thin-cap fibroatheroma, minimum lumen area, maximum lipid arc, spotty calcification, LAPs, positive remodeling, napkin-ring sign, plaque burden, and other relevant features. (5) Eligible study designs included observational studies, including cross-sectional studies, prospective cohort studies, and retrospective cohort studies. (6) Full-text articles and conference abstracts with sufficient information for eligibility assessment and outcome extraction were eligible.

2.2. Search strategy and data extraction

We systematically searched PubMed (MEDLINE), Embase, and Cochrane Central Register of Controlled Trials to identify reports meeting the inclusion criteria published from database inception to 19 June 2026. In addition to the database search, reference lists of relevant studies were manually screened to identify additional relevant reports. The search terms used included lipoprotein(a), coronary artery disease, and coronary imaging modalities, including OCT and CCTA. The complete search strategy is provided in Table S2. After removing duplicates, two authors (Z.J.Y. and C.Y·Y.) screened titles and abstracts and assessed potentially eligible reports using the same prespecified eligibility criteria, irrespective of publication language. An additional search of the grey literature was performed using OpenGrey (http://www.opengrey.eu/), Greynet (http://greynet.org/) and ProQuest Dissertations & Theses with the same terms.

Moreover, reference list screening was performed to identify additional relevant reports. Two authors (Z.J.Y. and C.Y.Y.) independently extracted baseline characteristics, and extracted outcome data, both following a standardized data form. In the event of disagreements, differences were resolved through discussion. The data extracted from each study included several elements, among them (1) Author, Year (2) Study design (3) Study, Population size (4) Disease (5) Cut-off for elevated Lp(a) (6) Age (7) Sex (8) Race (9) Reported HRP.

2.3. Endpoints

Our study focused on comparing coronary plaque characteristics (CPCs) between patients with elevated Lp(a) levels and those with normal or lower Lp(a) levels. Elevated Lp(a) was defined as any level above the study-specific laboratory cut-off, given the lack of a universally accepted threshold in the literature and the absence of patient-level data [19]. According to the data format of each included study, Lp(a) was analyzed as a categorical variable, and participants were divided into elevated-Lp(a) and normal-Lp(a) groups. For categorical outcomes (e.g., presence or absence of plaque features), events were defined as the presence of specific high-risk plaque characteristics rather than clinical cardiovascular events. We recorded the numbers of events and non-events in the elevated-Lp(a) and normal-Lp(a) groups, respectively. For continuous outcomes (e.g., quantitative plaque parameters), we recorded mean values and standard deviations for each group. For OCT-derived features, the association between lipoprotein(a) levels and the incidence of thin-cap fibroatheroma (TCFA) in patients with coronary artery disease is the primary endpoint. The secondary endpoint is the association between Lp(a) levels and plaque rupture, maximum lipid arc, and minimum lumen area. For CCTA-derived features, the association between lipoprotein(a) levels and plaque burden in patients with coronary artery disease is the primary endpoint. The secondary endpoint is the association between Lp(a) levels and positive remodeling, spotty calcification, napkin ring sign, and LAPs.

The CPCs of interest, as defined according to the respective imaging modalities, included TCFA, high plaque burden, low minimum lumen area (MLA), LAP, spotty calcification, the napkin ring sign, positive remodeling, and maximum lipid arc [20]. Of note, for patients presenting with acute myocardial infarction, plaque rupture was assessed specifically at culprit lesions as defined in the original studies.

2.4. Lp(a) measurement

Lp(a) concentrations were measured using immunoassay-based methods in the original studies, including immunoturbidimetric assays and enzyme-linked immunosorbent assays (ELISA). The reporting units varied, with some studies presenting Lp(a) in mg/dL and others in nmol/L. Because the relationship between mg/dL and nmol/L is highly dependent on apo(a) isoform size, a standardized conversion could not be applied [21]. Therefore, Lp(a) values were extracted and analyzed in their originally reported units.

Cut-off values defining elevated Lp(a) ranged from 30 to 70 mg/dL or 60–125 nmol/L and were based on study-specific laboratory reference standards. High and normal Lp(a) groups were categorized according to the definitions provided in the original studies [22]. To address the heterogeneity in reporting units, patients were stratified into elevated or normal Lp(a) groups based on the study-specific clinical cut-off values defined by the respective original studies, allowing for a categorical meta-analysis without the need for mathematical unit conversions.

2.5. Definition of imaging characteristics

The standard definitions and diagnostic thresholds of imaging-derived plaque characteristics (including TCFA, plaque rupture, maximum lipid arc, MLA, plaque burden, positive remodeling, spotty calcification, napkin-ring sign, and LAP) utilized in the CCTA and OCT literature were extracted from each original study. Detailed diagnostic criteria and thresholds for these imaging phenotypes are provided in Appendix S1.

2.6. Quality assessment

The quality assessment was conducted independently by two authors (Z.J.Y and C.Y.Y.) using the Newcastle-Ottawa Scale (NOS) to evaluate the included cohort studies, and the Agency for Healthcare Research and Quality (AHRQ) to assess the included cross-sectional studies, as well as the risk of bias in observational studies [23]. Inconsistencies were resolved by consulting the third author (S.Y.).

The quality assessment of the cohort study was performed using the Newcastle-Ottawa Quality Assessment Scale for Case-observational studies (NOS) [23]. The NOS comprises 8 items, with a total score exceeding 6 indicating high quality, while a score ranging from four to six indicates moderate quality. Cross-sectional study utilized evaluation criteria established by the Agency for Healthcare Research and Quality (AHRQ) [24]. The AHRQ comprises 11 items; when an answer is ‘yes,’ one point is assigned, while no points are assigned if the answer is ‘no’ or ‘not reported.’ High quality was defined as a score 8–11, whereas moderate quality was defined as a score 4–7. The results of the quality assessment were used to evaluate the reliability of the included evidence.

The certainty of evidence for each outcome was further assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. Detailed NOS/AHRQ scoring and GRADE evidence profiles are presented in the Table S4 and Table S5 [25]. Disagreements were resolved by consensus.

2.7. Statistical analysis

In this meta-analysis, we conducted a pooled statistical analysis to determine the relationship between high Lp(a) levels and the incidence of coronary artery plaque characteristics, and performed subgroup analysis for the primary outcome TCFA.

For binary outcomes, the number of events and non-events was extracted from the elevated-Lp(a) and normal-Lp(a) groups of each study. Study-specific log odds ratios (lnORs) and their standard errors (SEs) were initially calculated from the raw 2 × 2 contingency tables utilizing Stata 17.0 (StataCorp, College Station, TX, USA). The resulting lnORs and SE values were then entered into Review Manager (RevMan) version 5.4, which converted lnORs back to odds ratios (ORs) and performed the meta-analysis using an inverse-variance random-effects model. For continuous outcomes, we extracted the group-specific means and standard deviations (SDs) as reported in the original studies and calculated standardized mean differences (SMDs) with their 95% confidence intervals (CIs) using Review Manager (RevMan) version 5.4. All continuous outcomes were synthesized using random-effects models. The data extracted and pooled in this meta-analysis were primarily raw, unadjusted event counts and continuous metrics (univariate data) from the original studies. Multi-variable adjusted odds ratios (ORs) were not pooled due to the high heterogeneity in the selection of adjustment covariates across the trials; however, the impact of these potential confounding factors is discussed qualitatively.

Considering the anticipated between-study heterogeneity, a random-effects model was preferred for all endpoints. Heterogeneity was assessed using Cochran's Q test and I² statistics. A P-value less than 0.10 or I² statistics exceeding 50% were deemed indicative of significant heterogeneity, prompting additional sensitivity analyses when feasible. In these cases, we performed leave-one-out sensitivity analyses to ensure that results were not dependent on a single study.

To explore potential sources of heterogeneity, prespecified subgroup analyses were performed for the TCFA endpoint by stratifying studies based on geographical region and clinical presentation. Subgroups were analyzed utilizing the same random-effects model as in the primary analysis, and differences between subgroups were assessed through interaction testing. All subgroup analyses were exploratory.

To assess the robustness of the pooled results, a leave-one-out sensitivity analysis was performed using Stata 17.0 (StataCorp LLC, Texas, USA). This analysis iteratively removed each included study to evaluate the impact of individual studies on the overall effect estimate.

Publication bias was assessed visually through funnel plots and statistically utilizing Egger's test for outcomes comprising at least 10 studies.

All statistical tests were two-tailed, and a P-value <0.05 was considered statistically significant. All statistical analyses were conducted using Review Manager (RevMan) version 5.4, developed by the Cochrane Collaboration.

3. Results

3.1. Study selection and baseline characteristics

The study selection process and baseline characteristics of this research are detailed in the PRISMA flow chart (Fig. 1). A total of 1911 relevant studies were initially identified through systematic retrieval. Following the removal of 152 duplicate documents, 1683 of the remaining 1759 studies were excluded based on titles and abstracts because they did not meet the predefined inclusion criteria. Ultimately, 13 studies were included in the systematic review. Among these, 33 studies were excluded for not reporting outcomes related to coronary artery plaque characteristics, while another 15 studies were letters, editorials, or meta-analyses that did not meet the inclusion criteria. The basic characteristics of all excluded studies are detailed in Table S3.

Fig. 1.

Flow chart of the study selection process for a meta-analysis of the relationship between lipoprotein (a) levels and the imaging manifestations of HRP characteristics in patients with coronary artery disease.

Additionally, eight documents were acquired through grey literature retrieval. After a thorough screening process, one of these documents met the criteria and were included in this meta-analysis.

In total, 13 studies were included in this meta-analysis, encompassing 6671 patients (Table 1). Regarding racial and geographical distribution, five studies were conducted in European regions, while eight studies were from Asian regions. All studies were observational in design, including 4 cross-sectional studies and 9 cohort studies (Table 1).

Table 1.

Characteristics of the included studies stratified by coronary imaging modality.

Author, Year	Study design	Study, Population size	⁎#Disease	Cut-off for elevated Lp(a)	Age, years Mean (SD)	Men (%)	Race	Reported HRP
OCT
Di Muro, 2025 [27]	Retrospective cohort study	Single-center study 202	ACS	300 mg/L	67.3 (11.3)	162 (80.2)	Italy	TCFA, MLA, Plaque rupture
Kato, 2022 [29]	Cross-sectional study	Single-center study 185	ACS	30 mg/dL	70.0 (8.5)	153 (82.7)	Japan	TCFA, MLA, Max lipid arc
Niccoli, 2016 [16]	Cross-sectional study	OCT cohort 51	ACS	30 mg/dL	65.0 (11.0)	29 (56.9)	Italy	TCFA, MLA, Maximum lipid arc, plaque rupture
Wang, 2025 [30]	Retrospective cohort study	NCT02140801 177	ACS	30 mg/dL	68.3 (10.4)	134 (75.7)	China	TCFA, PAV
Yuan, 2023 [31]	Retrospective cohort study	Single-center study 119	Stable CAD	30 mg/dL	63.9 (9.0)	96 (80.7)	China	TCFA, MLA, Maximum lipid arc
Fang, 2025 [28]	Retrospective cohort study	Abstract 1005	ACS	75 nmol/L	56.6 (11.5)	754 (75.0)	China	TCFA, Plaque rupture
Chandra, 2022 [26]	Cross-sectional study	Single-center study 34	ACS	75 nmol/L	/	<35	India	TCFA
CCTA
Breitbart, 2025 [36]	Cross-sectional study	COPRODUCTION 1317	Stable CAD	125 nmol/L	63.3 (11.1) #	913 (69.3)	Germany	PR, NRS, SC, LAP
Dai 2022 [32]	Retrospective cohort study	NCT05323227 377	Stable CAD	60 mg/dL	63.6 (9.3)	264 (70.0)	China	PR, NRS, SC, LAP
Kaiser, 2022 [33]	Prospective cohort study	DIAMOND 191	Stable CAD	70 mg/dL	65.9 (8.3)	152 (80.0)	The UK	Plaque burden
Lan, 2025 [34]	Prospective cohort study	Single-center study 1694	Stable CAD	30 mg/dL	62.0 (7.0)	1031 (60.9)	China	PR, NRS, SC, LAP, Plaque burden, PAV
Nurmohamed, 2024 [17]	Prospective cohort study	Single-center study 267	Stable CAD	125 nmol/L	57.1 (7.3)	153 (57.0)	Netherlands	Plaque burden
Yang, 2026 [35]	Retrospective cohort study	WYX: NO.2023–226 1052	Stable CAD	30 mg/dL	61.0 (10.0)	550 (52.3)	China	MLA

Lp(a): Lipoprotein(a); HRP: High-Risk Plaque features; OCT: Optical Coherence Tomography; ACS: Acute Coronary Syndrome; TCFA: Thin-Cap Fibroatheroma; MLA: Minimum Lumen Area; CAD: Coronary Artery Disease; PAV: Percent Atheroma Volume; ISNA: In-Stent Neoatherosclerosis; AMI: Acute Myocardial Infarction; CCTA: Coronary Computed Tomography Angiography; PR: Positive Remodeling; NRS: Napkin-Ring Sign; SC: Spotty Calcification; LAP: Low-Attenuation Plaque; PCI: Percutaneous Coronary Intervention.

^⁎Patients were categorized as having ACS if the original study enrolled patients with ACS or AMI; otherwise, patients with CAD, stable CAD, coronary artery stenosis, non-obstructive CAD, atherosclerotic CAD with type 2 diabetes, or in-stent neoatherosclerosis were classified as having stable coronary artery disease.

^#Since the average age of the coronary heart disease patients could not be accurately obtained from the text, we used the average age of all the baseline populations in the text as a substitute.

3.2. Study characteristics

Thirteen studies were published between 2016 and 2026, including four cross - sectional studies, six retrospective cohort studies, and three prospective cohort studies. In total, 13 studies were included in this meta-analysis, encompassing 6671 patients. Among studies with available age and sex data, the weighted mean age was 61.9 years, and males accounted for 66.2% of the population. The included populations comprised patients with stable coronary artery disease or acute coronary syndrome. Of the 13 included studies, seven enrolled patients with stable coronary artery disease and six enrolled patients with acute coronary syndrome. The detection thresholds of Lp(a) varied across studies, ranging from 30 to 70 mg/dL or 60 to 125 nmol/L. All studies reported the characteristics of HRP evaluated by CCTA or OCT, mainly including positive remodeling, LAP, napkin-ring sign, plaque burden, minimum lumen area, and TCFA. The characteristics of the studies are presented in Table 1.

3.3. Primary endpoint

A total of 7 studies were included, involving 1773 patients, to explore the relationship between elevated Lp(a) and TCFA assessed by OCT [16], [26], [27], [28], [29], [30], [31]. Compared to patients with normal Lp(a), the incidence of TCFA in the Lp(a)-elevated group showed a relative increase of 157% (OR, 2.57; 95% CI, 1.54–4.30; p = 0.0003) (Fig. 2A).

Fig. 2.

Forest plots of the association between elevated Lp(a) levels and OCT- derived TCFA (A) and CCTA-derived Plaque Burden (B). Panels show study-specific and pooled effect estimates (random-effects, inverse-variance method) with 95% confidence intervals (CI). (A) OCT- derived TCFA: A meta-analysis of data from 7 studies revealed that the probability of TCFA was significantly elevated by 157% in patients with high Lp(a) levels relative to those with normal Lp(a) levels (OR = 2.57; 95% CI: 1.54–4.30; P = 0.0003). Heterogeneity: I² = 61%.(B) CCTA-derived Plaque Burden (continuous variable): A meta-analysis of data from 6 studies demonstrated a moderate difference in the plaque burden between patients with high Lp(a) levels and those with normal Lp(a) levels (SMD 0.27; 95% CI -0.01 - 0.56; P = 0.06). Heterogeneity: I² = 90%.

A total of 5 out of the 6 CCTA studies were included, involving 3581 patients, to explore the relationship between elevated Lp(a) and plaque burden assessed by CCTA [16], [32], [33], [34], [35]. There was no statistically significant difference in plaque burden between patients with elevated Lp(a) and those with normal Lp(a). (SMD 0.27; 95% CI -0.01 - 0.56; I² = 90%; p = 0.06) (Fig. 2B).

3.4. Secondary endpoints

3.4.1. Other plaque characteristics evaluated by OCT

Regarding the secondary endpoints evaluated using OCT, compared with patients in the normal Lp(a) group, patients in the elevated Lp(a) group exhibited a significantly increased risk of plaque rupture by 71% (OR 1.71; 95% CI 1.25–2.34; I² = 0%; P = 0.0008) (Fig. 3A) [16], [27], [28], [31]. Regarding continuous variables, elevated Lp(a) levels were significantly associated with a larger maximum lipid arc when compared to patients in the normal Lp(a) group (SMD 0.84; 95% CI 0.31–1.38; I² = 79%; P = 0.002) (Fig. 3B) [16], [29], [31]. However, no statistically significant association was found with the minimum lumen area (SMD -0.06; 95% CI -0.32 - 0.19; I² = 69%; P = 0.63) (Fig. 3C) [16], [27], [29], [31], [35].

Fig. 3.

Forest plots of the association between elevated Lp(a) levels and OCT-derived HRP Features. Panels show study-specific and pooled effect estimates (random-effects, inverse-variance method) with 95% confidence intervals (CI).

3.4.2. The characteristics of plaques evaluated by CCTA

Among the plaque characteristics evaluated based on CCTA, compared with patients in the Lp(a) normal group, patients in the Lp(a) elevated group demonstrated a significantly relative increase risk of developing spotty calcification (OR 1.78; 95% CI 1.19–2.68; I² = 14%; P = 0.005) (Fig. 4B) [32], [34], [36]. Additionally, they demonstrated a significantly relative increase risk of presenting with a napkin-ring sign (OR 1.46; 95% CI 1.05–2.03; I² = 0%; P = 0.02) (Fig. 4C) [32], [34], [36], and a significantly relative increase risk of developing LAPs (OR 1.53; 95% CI 1.06–2.17; I² = 69%; P = 0.02) (Fig. 4D) [32], [34], [36]. However, no statistically significant difference was observed in the association with positive remodeling (OR 1.42; 95% CI 0.84–2.41; I² = 82%; P = 0.20) (Fig. 4A) [32], [34], [36].

Fig. 4.

Forest plots of the association between elevated Lp(a) levels and CCTA-derived HRP Features. Panels show study-specific and pooled effect estimates (random-effects, inverse-variance method) with 95% confidence intervals (CI). (A) Positive remodeling: A meta-analysis of data from 3 studies demonstrated that there was no significant difference in the probability of positive remodeling between patients with high Lp(a) levels and those with normal Lp(a) levels (OR: 1.42; 95% CI: 0.84–2.41; P = 0.20). Heterogeneity: I² = 82%. (B) Spotty calcification: A meta-analysis of data from 3 studies revealed that the probability of spotty calcification was significantly elevated by 78% in patients with high Lp(a) levels relative to those with normal Lp(a) levels (OR = 1.78; 95%CI: 1.19–2.68; P = 0.005). Heterogeneity: I² = 14%. (C) Napkin-ring sign: A meta-analysis of data from 3 studies revealed that the probability of napkin-ring sign was significantly elevated by 46% in patients with high Lp(a) levels relative to those with normal Lp(a) levels (OR = 1.46; 95%CI: 1.05–2.03; P = 0.02). Heterogeneity: I² = 0%. (D) Low attenuating plaque: A meta-analysis of data from 3 studies revealed that the probability of low attenuating plaque was significantly elevated by 53% in patients with high Lp(a) levels relative to those with normal Lp(a) levels (OR = 1.52; 95%CI: 1.06–2.17; P = 0.01). Heterogeneity: I² = 69%.

3.4.3. Sensitivity analyses

A leave-one-out sensitivity analysis was performed for each outcome to assess the robustness of the pooled estimates. The association between elevated Lp(a) and OCT-derived TCFA remained statistically significant after sequential exclusion of each individual study, indicating a robust primary finding. Notably, exclusion of Yuan et al. (2023) reduced the heterogeneity of the TCFA analysis from moderate to negligible while preserving the significant association, suggesting that this study was the main contributor to heterogeneity [31]. The association with OCT-derived plaque rupture was also robust, with the pooled effect remaining significant after omission of any single study. For maximum lipid arc, the association remained significant in all leave-one-out analyses, although heterogeneity was substantially reduced after excluding Yuan et al. (2023) [31]. In contrast, the CCTA-derived plaque burden result was not fully robust: the overall association was borderline non-significant and highly heterogeneous, and statistical significance changed after omission of individual studies. Similarly, the associations with spotty calcification, napkin-ring sign, and low-attenuation plaque were sensitive to the exclusion of specific studies, likely reflecting the limited number of studies available for these endpoints. The null findings for minimum lumen area and positive remodeling remained unchanged in leave-one-out analyses (Fig. S1).

3.4.4. Subgroup analysis

To investigate the source of the moderate heterogeneity (I² = 61%) observed in the pooled analysis of the primary outcome TCFA, subgroup analyses were conducted based on patient region and clinical disease stage. The forest plot of the subgroup analysis is shown in Fig. S3 and Fig. 5.

Fig. 5.

Subgroup Analysis of the Association Between Elevated Lp(a) and TCFA by Clinical Presentation Estimates were calculated using a random-effects model. Subgroup analyses evaluating the association between elevated Lp(a) levels and the presence of TCFA stratified by clinical presentation. Elevated Lp(a) was significantly associated with TCFA in patients with ACS (OR 2.27; 95% CI 1.51–3.43; P < 0.0001), in those with acute myocardial infarction treated with PCI (OR 1.80; 95% CI 1.27–2.57; P = 0.001), and showed the strongest association in patients with in-stent neoatherosclerosis (ISNA) (OR 13.74; 95% CI 4.32–43.66; P < 0.0001).

When stratified by region, elevated Lp(a) was associated with a significantly increased risk of TCFA in both Asian cohorts (OR 2.61; 95% CI 1.41–4.82; P = 0.002; 5 studies) and European cohorts (OR 2.75; 95% CI 1.25–6.06; P = 0.01; 2 studies), with no meaningful difference in effect magnitude between the two regions (Fig. S3).

When stratified by disease type, the association was generally positive across clinical settings, but the magnitude of association varied significantly among different clinical phenotypes (group difference P = 0.004, I² = 81.6%). In patients with ACS, Lp(a) elevation was significantly associated with an increased risk of TCFA (OR, 2.27; 95% CI, 1.51–3.43; P < 0.0001; 5 studies), and there was no heterogeneity within this subgroup (I² = 0%). In patients with acute myocardial infarction (AMI) treated with percutaneous coronary intervention (PCI), a significant positive association was also observed (OR, 1.80; 95% CI, 1.27–2.57; P = 0.001; 1 study). Notably, the strength of this association was most significant in In-Stent Neoatherosclerosis (ISNA) patients (OR, 13.74; 95% CI, 4.32–43.66; P < 0.00001; 1 study) (Fig. 5).

3.5. Quality assessment

The cross-sectional studies were evaluated using the AHRQ scale, while the bias risk of the cohort studies was assessed using the NOS scale. Detailed results are presented in Table S5 of the online supplementary materials.

Overall, among the 13 studies, 10 (76.9%) were classified as high quality and three (23.1%) as moderate quality. Although the proportion of high-quality studies was substantial, all included studies were observational in design, and some outcomes exhibited instability during sensitivity analysis, limiting the overall certainty of evidence in this meta-analysis.

3.6. Publication bias

The funnel plot for the primary endpoint displays a symmetrical distribution of each study's weight, converging toward the combined effect size as the weight increases (Fig. S2). The number of studies included in some outcome measures was insufficient (<10), preventing the conduction of Egger's and Begg's tests.

4. Discussion

4.1. Major findings

In this systematic review and meta-analysis involving 13 observational studies and 6671 patients, elevated Lp(a) levels were associated with several imaging-defined vulnerable plaque features assessed by both OCT and CCTA. The most robust evidence was observed for OCT-derived TCFA and plaque rupture, whereas several CCTA-derived high-risk plaque features showed directionally consistent but less stable associations in leave-one-out sensitivity analyses.

Although several individual studies have reported associations between Lp(a) and coronary plaque features, the available evidence has remained fragmented across different imaging modalities, plaque endpoints, Lp(a) thresholds, and clinical settings.

The novelty of the present study lies in systematically and quantitatively integrating complementary OCT and CCTA evidence to define the vulnerable plaque phenotype associated with elevated Lp(a).

This approach moves beyond isolated single-feature associations and clarifies which imaging markers of plaque vulnerability are most consistently related to elevated Lp(a).

Among all endpoints, the association between elevated Lp(a) and OCT-defined TCFA was the most robust. Patients with high Lp(a) levels demonstrated more than a two-fold increase in TCFA risk compared with those with normal Lp(a) levels. TCFA represents the quintessential vulnerable plaque phenotype and is characterized by a thin fibrous cap overlying a lipid-rich necrotic core. Its pathophysiological relevance is well established, as disruption of the fibrous cap directly precipitates acute coronary thrombotic event [37]. In addition to TCFA, elevated Lp(a) levels were associated with other OCT-derived markers of vulnerability, including plaque rupture and significantly larger maximum lipid arc, suggesting more extensive lipid infiltration and heightened susceptibility to mechanical stress [38]. Collectively, these findings highlight a consistent microstructural pattern in which elevated Lp(a) is associated with thinner caps, larger lipid pools, and a morphology more prone to rupture. The associations observed using CCTA were directionally consistent with the OCT findings. For CCTA-derived features, elevated Lp(a) showed significant associations with low-attenuation plaque, the napkin-ring sign, and spotty calcification, whereas the association with plaque burden was borderline and did not reach statistical significance. Moreover, leave-one-out analyses indicated that several CCTA-derived findings were sensitive to individual studies, and therefore should be interpreted cautiously [39].

Most included studies were classified as high quality. However, all were observational in design. Several associations—particularly those derived from CCTA endpoints—exhibited heterogeneity and sensitivity to individual studies. Particularly, a single study focusing on ISNA contributed disproportionately to heterogeneity in the TCFA analysis. These considerations indicate that although the evidence base is directionally consistent, the overall certainty remains limited and highlights the need for large, standardized, imaging-based prospective studies.

4.2. Comparison with previous studies

Several studies have investigated the association between Lp(a) and coronary artery calcification (CAC). Most of these studies were conducted in individuals without established cardiovascular disease. Previous research has found that the level of Lp(a) is positively correlated with the incidence of CAC, and a high level of Lp(a) increases the probability of CAC > 0. These data support the notion that Lp(a) has atherogenic properties; however, these studies primarily focus on primary prevention populations and exhibit substantial heterogeneity [40].

Our research focused on secondary prevention of coronary artery disease. It further indicates that among patients with coronary artery disease, those with high Lp(a) levels often exhibit a higher TCFA ratio, a larger lipid core, and more plaque rupture. These changes are highly correlated with spontaneous or perioperative acute coronary events. Similarly, the CCTA study also shows that elevated Lp(a) is positively correlated with non-calcified plaque burden, LAPs, and high-risk imaging patterns (such as the napkin ring sign) [41].

However, due to the limited sample size and inconsistent definitions in previous original studies, their results showed certain differences. Compared with these individual studies, our meta-analysis provides a broader synthesis by jointly evaluating OCT-derived microstructural features and CCTA-derived non-invasive high-risk plaque characteristics.

Our findings suggest that elevated Lp(a) is more consistently associated with vulnerability-related features, including TCFA, plaque rupture, maximum lipid arc, low-attenuation plaque, napkin-ring sign, and spotty calcification, than with plaque size alone. By systematically integrating data from multiple studies, our meta-analysis provides more structured evidence supporting an association between elevated Lp(a) and an imaging-defined vulnerable plaque phenotype, while also highlighting that several CCTA-derived findings require cautious interpretation.

4.3. Potential mechanisms

The instability of coronary artery HRPs is primarily driven by local and systemic inflammation, altered shear stress, and upregulated matrix metalloproteinase activity [42]. Among high-risk plaque phenotypes, TCFA is a key rupture-prone morphology associated with plaque rupture and subsequent coronary events. Plaque progression is characterized by vascular smooth muscle cell (VSMC) apoptosis, extracellular matrix degradation, neovascularization, arterial remodeling, fibrous-cap disruption with thrombosis, and progressive necrosis and calcification. Plaque vulnerability is a dynamic rather than static process that is continually shaped by biological and biomechanical factors. Moreover, Lp(a) has been implicated in promoting the development of high-risk coronary plaques through its pro-atherogenic, pro-inflammatory, and pro-thrombotic effects, thereby contributing to plaque instability. Similar to LDL particles, the structural properties of Lp(a) facilitate its binding to lysine residues on apolipoproteins, allowing subintimal accumulation, lipid-streak formation, and initiation of early atherogenesis [22], [43]. Moreover, as the principal carrier of oxidized phospholipids (OxPLs), Lp(a)-bound OxPLs have been shown to activate endothelial cells, promote monocyte transendothelial migration, stimulate immune-cell activation, enhance chemotaxis, and amplify inflammatory signaling pathways [44], [45], [46]. In addition, apolipoprotein(a) shares structural homology with plasminogen, enabling competitive inhibition of plasminogen activation, which reduces fibrinolytic capacity and facilitates the formation of persistent thrombus after plaque disruption [47], [48], [49]. Through these convergent mechanisms, elevated Lp(a) ultimately drives the evolution of plaque morphology toward a more rupture-prone phenotype.

The findings from our subgroup analyses further support these mechanistic insights. In the subgroup analysis stratified by patient region, no material difference in the pooled effect size between Asian and European cohorts was observed (P for subgroup difference = 0.92). However, substantial heterogeneity was detected within the Asian subgroup (I² = 65%), suggesting the presence of additional underlying modifiers. Further subgroup analysis by clinical presentation demonstrated a highly consistent association among patients with ACS (I² = 0%), whereas the markedly elevated effect reported in Yuan et al. (2023) (OR = 13.74; 95% CI 4.32–43.66) was identified as the dominant contributor to overall heterogeneity [31].

Importantly, the markedly elevated effect observed in Yuan et al. (2023) does not reflect conventional native-vessel TCFA, but rather represents the vulnerability characteristics of ISNA, a distinct pathological entity. ISNA is biologically predisposed to forming highly unstable, lipid-rich neoatherosclerotic plaques, which provides a mechanistic explanation for the substantially larger effect size [50]. Accordingly, this effect should not be directly compared with studies evaluating native-vessel TCFA in ACS or stable CAD, as the underlying disease substrates differ fundamentally. This pattern aligns with current pathophysiological concepts in which Lp(a) accelerates the development of unstable, lipid-rich neoatherosclerotic plaques within the stent through enhanced atherogenesis, inflammation, and antifibrinolytic activity [51]. Therefore, when the research outcome is precisely targeted at this direct mechanism, observing a strong effect is reasonable.

Additional mechanistic data also support the biological plausibility of our findings. Experimental studies have demonstrated that elevated Lp(a) levels (>50 mg/dL) promote calcification in primary human smooth-muscle cells and valve interstitial cells, potentially through inflammation-mediated pathways, an effect that can be attenuated by the E06 antibody [52]. Single-cell extracellular-vesicle microarray analyses have shown that Lp(a) stimulation alters vesicular composition and enhances the release of pro-calcific vesicles [52]. These findings support a direct mechanistic link between elevated Lp(a) concentrations and plaque instability. These clinical research results indicate that elevated Lp(a) levels are closely related to the formation and instability of coronary artery HRPs.

4.4. Lp(a), plaque vulnerability, and clinical outcomes

The findings of this study suggest that the clinical relevance of Lp(a) should be emphasized in individualized cardiovascular risk assessment. Elevated Lp(a) reflects genetically determined long-term exposure to atherogenic risk and is associated with multiple imaging-defined HRP features, contributing to greater plaque instability even when LDL-C is adequately controlled. In the included studies, Di Muro et al. (2025) reported that elevated Lp(a) levels were associated with a more rupture-prone HRP phenotype, and that patients continued to exhibit higher residual cardiovascular risk even after adjustment for other risk factors [27]. Niccoli et al. (2016) reported that in ACS, patients with non-ST-elevation myocardial infarction (NSTEMI) patients exhibited higher Lp(a) concentrations, suggesting that Lp(a) is more strongly linked to plaque instability than to stable ischemia [16]. Moreover, when elevated Lp(a) coexists with multiple HRP features, the risk of MACE appears to be greatest, with a graded increase in risk corresponding to higher Lp(a) levels and a greater burden of unstable plaque characteristics [32]. Another study further confirmed that in patients with non-obstructive CAD, the combined assessment of Lp(a) and pericoronary adipose tissue CT attenuation (PCATa) significantly improves the predictive ability of MACE [35].

Furthermore, beyond anatomic morphology, functional evaluation of lesions provides crucial prognostic insights. Recent evidence has underscored the clinical impact of myocardial perfusion evaluated with CT in identifying hemodynamically significant stenosis and predicting adverse cardiovascular events [53]. Integrating anatomical high-risk plaque characteristics derived from CCTA with functional CT perfusion measurements represents a highly promising clinical direction. This comprehensive approach could help clinicians determine whether Lp(a)-driven high-risk plaques not only pose a structural risk of rupture but also cause functional myocardial ischemia, thereby further refining precision risk stratification.

With the development of new Lp(a)-targeting therapies, such as antisense oligonucleotides and siRNA, identifying the imaging phenotypes associated with elevated Lp(a) may help identify patients most likely to benefit from early intervention [54]. At the same time, combining Lp(a) testing with advanced coronary imaging is expected to further improve the accuracy of existing risk prediction models and promote residual risk management based on pathological mechanisms. However, due to the differences in follow-up endpoints, case composition, and follow-up time among different studies, prospective large-sample imaging-prognosis studies are still needed for verification.

4.5. Strengths and limitations

A major strength of this study is that it systematically integrates multimodality coronary imaging evidence from both OCT and CCTA to evaluate the association between elevated Lp(a) and a broad spectrum of vulnerable plaque features.

This design allowed us to synthesize invasive microstructural plaque information and non-invasive high-risk plaque phenotypes within a unified imaging framework. Secondly, we conducted subgroup analysis of the primary outcome TCFA to explore potential sources of heterogeneity and found that clinical presentation, particularly the inclusion of ISNA, may have contributed to the observed heterogeneity. Thirdly, we conducted a systematic evaluation of the quality of research methods and imaging definitions to support the reliability of the results.

However, this study also has limitations. All included studies were observational, and despite quality assessment, residual confounding cannot be excluded. Large-scale prospective imaging studies and interventional studies are needed to validate these findings and determine whether Lp(a)-lowering therapy can modify vulnerable plaque features. Second, a major limitation of this meta-analysis is the substantial heterogeneity in the thresholds and assay methodologies used to define elevated Lp(a) levels across the included studies (commonly used thresholds included 30 mg/dL [300 mg/L] and 125 nmol/L). Due to the lack of a standardized universal threshold in the clinical literature and the absence of patient-level raw data, patients were grouped using study-specific reference standards, which may introduce classification bias and complicate the identification of a single clinical cutoff. Third, one included record was available only as a conference abstract rather than a full-text article, and the lack of complete methodological and statistical details may have introduced potential bias into data extraction and risk-of-bias assessment. Fourth, the sample sizes and adjustment variables of the studies were inconsistent. Fifth, several secondary CCTA-derived endpoints were based on only three studies, and leave-one-out sensitivity analyses showed that the statistical significance of some associations changed after exclusion of individual studies. Therefore, these endpoint-specific findings should be regarded as suggestive rather than definitive. In this study, binary variables were mainly recalculated based on the original data, and the effect sizes were unadjusted. Due to the small number of included studies, formal publication bias detection could not be conducted, which may affect the reliability of pooled estimates.

5. Conclusion

Elevated Lp(a) is associated with multiple imaging features of coronary plaque vulnerability, with the most robust association observed for OCT-derived TCFA and plaque rupture. These findings support the potential role of Lp(a) in the development of unstable atherosclerotic lesions and highlight its potential importance in improving risk stratification beyond traditional lipid parameters. Nonetheless, given the observational nature of the included studies and the threshold heterogeneity of Lp(a), standardized prospective imaging-outcome trials are warranted to validate these findings. As targeted Lp(a)-lowering therapies continue to advance, identifying the imaging phenotypes associated with elevated Lp(a) may help guide treatment allocation and inform future precision-prevention strategies.

CRediT authorship contribution statement

Jiayu Zhang: Writing – review & editing, Formal analysis, Data curation. Yang Tang: Writing – review & editing. Yuyang Cao: Writing – review & editing. Queyun Sun: Project administration. Zhao Ma: Methodology. Sida Jia: Data curation. Na Xu: Software. Jinqing Yuan: Project administration, Funding acquisition. Ying Song: Investigation, Funding acquisition.

Fundings

the Capital's Funds for Health Improvement and Research (2026–2-4036); the National High Level Hospital Clinical Research Funding (2024-GSP-TJ-8); the Special Project of “Science & Technology Deputy Director” in Changping District (T2025-ZX057), the National Clinical Research Center for Cardiovascular Diseases, Fuwai Hospital, Chinese Academy of Medical Sciences [NCRC2022003].

(A) Plaque rupture: A meta-analysis of data from 4 studies revealed that the probability of plaque rupture was significantly elevated by 71% in patients with high Lp(a) levels relative to those with normal Lp(a) levels (OR = 1.71; 95% CI: 1.25–2.34; P = 0.0008). Heterogeneity: I² = 0%. (B) Maximum lipid arc (continuous variable): A meta-analysis of data from 4 studies demonstrated a significant difference in the maximum lipid arc between patients with high Lp(a) levels and those with normal Lp(a) levels (SMD = 0.84; 95% CI: 0.31–1.38; P = 0.002). Heterogeneity: I² = 79%. (C) Minimum lumen area (continuous variable): A meta-analysis of data from 5 studies demonstrated no significant difference in the minimum lumen area between patients with high Lp(a) levels and those with normal Lp(a) levels (SMD = −0.06; 95% CI: −0.32 - 0.19; P = 0.63). Heterogeneity: I² = 61%.

Declaration of competing interest

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.

Appendix A. Supplementary data

Supplementary material 1 Supplementary Tables, Figures, and Appendices