Serum and Plasma Osteoprotegerin Levels in Aortic Stenosis: A Stratified Meta-Analysis Comparing Patients and Controls

Background: Aortic stenosis (AS) is a common and life-threatening valvular disease in older adults, characterized by progressive calcification of the aortic valve. Osteoprotegerin (OPG), a cytokine in the RANK/RANKL/OPG bone regulatory pathway, has been implicated in vascular calcification and might serve as a biomarker in AS. However, studies measuring circulating OPG in AS have reported inconsistent results, possibly due to differences in sample matrix (serum vs. plasma) and measurement units (pg/mL vs. pmol/L). Objective: We performed a comprehensive meta-analysis to compare OPG levels in AS patients versus control subjects, stratified by sample type and reporting units, and to assess heterogeneity and moderators of the observed differences. Methods: We searched databases for studies reporting OPG concentrations in AS patients and a non-AS control group. Following PRISMA guidelines, 24 studies (with a total of 1693 patients) were included after screening and eligibility assessment. Data were pooled in four subgroup meta-analyses defined by sample matrix (plasma or serum) and OPG units (pg/mL or pmol/L). Random-effects models were used to estimate mean differences between AS and control populations. We report pooled differences, 95% confidence intervals (CI), p -values, between-study heterogeneity statistics (tau², I²), 95% prediction intervals, and meta-regression moderator tests for the effect of disease status (AS vs control). Results: In plasma (pg/mL) measurements (8 study cohorts), the pooled OPG levels did not differ significantly between AS and controls (mean difference ≈ 192 pg/mL, p = 0.93). Heterogeneity was extreme (tau² ≈ 8.34×10^6, I² = 99.9%), and the 95% prediction interval ranged from approximately 7200 to +6800 pg/mL, essentially encompassing no effect as well as large differences. In plasma (pmol/L) studies (4 cohorts), AS patients had significantly higher OPG by about +3.54 pmol/L (95% CI: +2.29 to +4.79 pmol/L, p < 0.0001) than controls. Heterogeneity was moderate-to-high (tau² = 0.21, I² = 78.5%), but the AS vs control status explained ~93% of between-study variance (R² = 93%). The 95% prediction interval remained positive (approximately +2.0 to +5.1 pmol/L), suggesting a consistent elevation of OPG in AS across these plasma-pmol studies. For serum (pg/mL) studies (7 cohorts), AS patients also showed markedly higher OPG levels than controls (mean difference ≈ +348 pg/mL, 95% CI: +152 to +544, p = 0.0005). Between-study heterogeneity was very high (tau² ≈ 7729, I² = 99.1%), though about two-thirds of this variance was accounted for by AS vs control grouping (R² ~67%). The predicted range of true differences was wide (+87 to +609 pg/mL) but entirely above zero, indicating a robust increase in serum OPG in AS despite study variability. In contrast, in serum (pmol/L) studies (10 cohorts), there was no significant OPG difference between AS and controls (mean difference +1.17 pmol/L, 95% CI: 1.77 to +4.10, p = 0.44). Heterogeneity again approached 100% (tau² = 5.43, I² = 99.6%) with essentially 0% explained by disease status (no detectable effect). The prediction interval spanned from 4.3 to +6.6 pmol/L, including the possibility of lower or higher OPG in AS. In meta-regression terms, the moderator test for population (AS vs control) was significant in the plasma-pmol and serum-pg analyses (QM p < 0.001), but non-significant in plasma-pg and serum-pmol (p = 0.93 and p = 0.44, respectively). No other moderators were significant in exploratory analyses. There was no clear evidence of publication bias, though formal tests were limited by the small number of studies per subgroup. Conclusions: Circulating OPG levels in AS patients versus controls show a mixed picture. In two subsets (plasma OPG in pmol/L and serum OPG in pg/mL), we found a significant elevation of OPG in AS, whereas in the other two subsets (plasma OPG in pg/mL and serum OPG in pmol/L) there was no detectable difference. All analyses were characterized by substantial heterogeneity and broad prediction intervals, reflecting divergent results across studies. These findings suggest that OPG may be higher in AS patients in certain contexts, but methodological differences (sample type, assay calibration) and population variations likely influence the results. OPG’s utility as a general biomarker for AS is therefore uncertain. Standardization of measurement and further research are needed to clarify OPG’s role in AS.