Discordant response of low-density lipoprotein cholesterol and lipoprotein(a) levels to monoclonal antibodies targeting proprotein convertase subtilisin/kexin type 9

Abstract

Clinical trials testing proprotein convertase subtilisin/kexin type 9 inhibitors (PCSK9i) have demonstrated an unanticipated but significant lipoprotein (a) (Lp(a))-lowering effect, on the order of 25% to 30%. Although the 50% to 60% reduction in low-density lipoprotein (LDL)-cholesterol (LDL-C) achieved by PCSK9i is mediated through its effect on LDL receptor (LDLR) preservation, the mechanism for Lp(a) lowering is unknown. We sought to characterize the degree of concordance between LDL-C and Lp(a) lowering because of PCSK9i in a standard of care patient cohort. Participants were selected from our Center for Preventive Cardiology, an outpatient referral center in a tertiary academic medical center. Subjects were included in this study if they had (1) at least 1 measurement of LDL-C and Lp(a) before and after initiation of the PCSK9i; (2) baseline Lp(a)>10mg/dL; and (3) continued adherence to PCSK9i therapy. They were excluded if (1)they were undergoing LDL apheresis; (2) pre- or post-PCSK9i LDL-C or Lp(a) laboratory values were censored; or (3) subjects discontinued other lipid-modifying therapies. In total, 103 subjects were identified as taking a PCSK9i and 26 metall inclusion and exclusion criteria. Concordant response to therapy was defined as an LDL-C reduction >35% and an Lp(a) reduction >10%. The cohort consisted of 26 subjects (15 females, 11 males, mean age 63±12years). Baseline mean LDL-C and median Lp(a) levels were 167.4±72mg/dL and 81mg/dL (interquartile range 38-136mg/dL), respectively. The average percent reductions in LDL-C and Lp(a) were 52.8% (47.0-58.6) and 20.2% (12.2-28.1). The correlation between %LDL and %Lp(a) reduction was moderate, with a Spearman's correlation of 0.56 (P<.01). All subjects except for 1 had a protocol-appropriate LDL-C response to therapy. However, only 16 of the 26 (62%; 95% confidence interval 41%-82%) subjects had a protocol-concordant Lp(a) response. Although some subjects demonstrated negligible Lp(a) reduction associated with PCSK9i, there were some whose Lp(a) decreased as much as 60%. In this standard-of-care setting, we demonstrate moderate correlation but large discordance (∼40%) in these 2 lipid fractions in response to PCSK9i. The results suggest that pathways beyond the LDLR are responsible for Lp(a) lowering and indicate that PCSK9i have the potential to significantly lower Lp(a) in select patients, although confirmation in larger multicenter studies is required.