Confounding by Indication in a Propensity-Score-Matched Comparison of PCSK9 Inhibitors Versus GLP-1 Receptor Agonists in Elevated Lipoprotein(a): A Cautionary Real-World Analysis

Background Elevated lipoprotein(a) is a well-established cardiovascular risk factor with limited therapeutic options. Both PCSK9 inhibitors and GLP-1 receptor agonists are used in patients with elevated Lp(a), yet any observational comparison between these two drug classes is highly vulnerable to confounding by indication. Methods We used the TriNetX U.S. Collaborative Network to identify adults with elevated Lp(a) (>125 nmol/L or >50 mg/dL) initiated on either a PCSK9 inhibitor (evolocumab, alirocumab, or inclisiran) or a GLP-1 receptor agonist within one year of diagnosis. After 1:1 propensity score matching, 1,594 patients were included in each cohort. One-year cardiovascular outcomes were compared using Cox proportional hazards regression. Results After matching, MACE-3 occurred in 11.3% versus 9.1% of patients (HR 1.25, 95% CI 1.01–1.56). Numerically higher rates of MACE-4, incident ASCVD, and peripheral artery disease were also observed in the PCSK9 inhibitor arm. These associations ran opposite to randomized-trial evidence and occurred in the presence of residual imbalance on body-mass index (standardized difference 0.647), HbA1c, and total cholesterol. Conclusions A naïve reading of these results would suggest GLP-1 receptor agonists outperform PCSK9 inhibitors. This interpretation is almost certainly an artifact of confounding by indication. Residual imbalance in body-mass index, the coherent excess of all atherosclerotic endpoints in the PCSK9 inhibitor arm, and lower statin initiation in that arm indicate that matching failed to equalize baseline cardiovascular severity. This analysis serves as a cautionary case: propensity-score matching on measured covariates cannot support causal comparison between drug classes with such divergent prescribing indications.