Lipoprotein(a) and calcific aortic valve disease initiation and progression: a systematic review and meta-analysis

Abstract

Although evidence indicates the association of lipoprotein(a) [Lp(a)] with atherosclerosis, the link with calcific aortic valve disease (CAVD) is unclear. This systematic review and meta-analysis explores the connection between Lp(a) and aortic valve calcification and stenosis (AVS). We included all relevant studies, indexed in eight databases, up to February 2023. A total of 44 studies (163 139 subjects) were included, with 16 of them being further meta-analysed. Despite considerable heterogeneity, most studies support the relationship between Lp(a) and CAVD, especially in younger populations, with evidence of early aortic valve micro-calcification in elevated-Lp(a) populations. The quantitative synthesis showed higher Lp(a) levels, by 22.63 nmol/L (95% CI: 9.98-35.27), for patients with AVS, while meta-regressing the data revealed smaller Lp(a) differences for older populations with a higher proportion of females. The meta-analysis of eight studies providing genetic data, revealed that the minor alleles of both rs10455872 and rs3798220 LPA gene loci were associated with higher risk for AVS (pooled odds ratio 1.42; 95% CI: 1.34-1.50 and 1.27; 95% CI: 1.09-1.48, respectively). Importantly, high-Lp(a) individuals displayed not only faster AVS progression, by a mean difference of 0.09 m/s/year (95% CI: 0.09-0.09), but also a higher risk of serious adverse outcomes, including death (pooled hazard ratio 1.39; 95% CI: 1.01-1.90). These summary findings highlight the effect of Lp(a) on CAVD initiation, progression and outcomes, and support the early onset of Lp(a)-related subclinical lesions before clinical evidence.

Keywords: Aortic valve stenosis; Calcific aortic valve disease; Lipoprotein(a); Lp(a); Meta-analysis.

Graphical Abstract

Figure 1

PRISMA flowchart for study selection.

Figure 2

(A–C) Forest plots showing the pooled (A) standardised mean difference (MD) in lipoprotein(a) between patients with aortic valve stenosis and those without, (B) MD only for studies reporting in nmol/L (only the subcohort of individuals with measurements in nmol/L was used from the study by Wodaje et al.) and (C) standardised MD for patients with severe against those with milder stenosis. Random effects model was applied with the size of each marker corresponding to its relative study weight. AVS, Aortic valve stenosis; Lp(a), Lipoprotein(a); SD, Standard deviation; CI, Confidence interval; MD, Mean difference (SMD, Standardised MD); I², Higgins’ and Thompson’s I² statistic.

Figure 3

(A and B) Forest plots showing the pooled odds ratio of AVS for two LPA SNPs: (A) rs10455872 minor allele G and (B) rs3798220 minor allele C. Random effects model was used with the size of each marker corresponding to its relative study weight. OR, Odds ratio; CI, Confidence interval; N, Number of participants; I², Higgins’ and Thompson’s I² statistic.

Figure 4

(A and B) Forest plots showing the (A) pooled mean difference (MD) in annualised peak aortic velocity change (measured in m/s/year), and (B) the pooled hazard ratio for serious adverse events (death, aortic valve replacement or stenosis-related hospitalisation), between patients with low and those with high lipoprotein(a). Random effects model was applied, with the size of each marker corresponding to its relative study weight. Lp(a), Lipoprotein(a); SD, Standard deviation; CI, Confidence interval; MD, Mean difference; HR, Hazard ratio; N, Number of participants; I², Higgins’ and Thompson’s I² statistic.