Statins increase Lp(a) plasma level: is this clinically relevant?

Abstract

Lipoprotein(a) [Lp(a)] is a ‘compound’ lipoprotein in which apolipoprotein B-100 of an LDL particle is covalently bound to apolipoprotein(a) [apo(a)]. Plasma levels of Lp(a) are genetically determined to a large extent. High circulating levels of Lp(a) (usually defined as >50 mg/dL) represent an independent cardiovascular (CV) risk factor for CV disease and contribute to the residual CV risk observed in statin-treated subjects. Genome-wide association and Mendelian randomization studies have clearly established a causal association between high Lp(a) levels and vascular atherosclerosis, acute myocardial infarction, and atherosclerotic aortic valve stenosis. However, whether Lp(a) is still a risk factor in patients with CV events and on statin therapy is uncertain. In fact, whereas some studies have suggested a role for Lp(a) in predicting CV risk even in subjects with well-controlled LDL cholesterol (LDL-C) levels, other studies did not. While no approved specific Lp(a)-lowering drugs exist, LDL-C-lowering drugs have shown different abilities to affect Lp(a) levels. To date, ezetimibe in monotherapy was shown to reduce plasma levels of Lp(a) in patients with primary hypercholesterolaemia by 7%. Proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibition reduces Lp(a) by 20–30%, likewise nicotinic acid. Data on the effect of statins on Lp(a) levels are conflicting: some studies suggest a neutral role, while others report a statin-induced Lp(a) increase. In their study in this issue of the European Heart Journal, Tsimikas et al. performed a subject-level meta-analysis to evaluate the effect of statin therapy on Lp(a) levels. Using individual data from six randomized clinical trials with 5256 participants, and using the same assay, they found that Lp(a) levels were significantly increased in statin-treated patients, with a mean percentage change from baseline ranging from 8.5% to 19.6%, compared with 0.4% to –2.3% in the placebo group, an effect independent of the baseline Lp(a) levels. This result is in agreement with some previous reports indicating that statins increase Lp(a) levels by 10–20% and, given the individual data analysis of the study, makes the finding very robust. Of note, comparability with other studies is not apt, as the measurements of Lp(a) are not fully standardized owing to the heterogeneity of the apo(a) size. Furthermore, statin treatment changes the average size and composition of LDL particles; therefore, the possibility of a change in the apo(a) immunoreactivity may not be discarded a priori. The question is then: is this statin-induced Lp(a) increase clinically relevant? Percentages may be misleading and we should consider absolute Lp(a) levels: in the presence of low baseline Lp(a) levels, as occurs in the vast majority of the population, the 10–20% increase induced by statins should not affect the overall CV benefit of this therapy; on the other hand, the presence of high baseline Lp(a) levels might result in a significant absolute Lp(a) increase which, in turn, might modulate the beneficial effect of LDL-C lowering. A recent meta-analysis of patient-level data from seven statin trials reported that hazard ratios for high Lp(a) levels at both baseline and on statin were comparable, suggesting that statins are unlikely to affect Lp(a)related CV risk, although the association of Lp(a) levels with CV risk was stronger in patients on statins than in those taking placebo, indicating the possibility that the Lp(a)-attributable risk was better appreciated after LDL-C reduction. A possible explanation for this finding is that in subjects with better controlled LDL-C levels, Lp(a) levels become a more important determinant of CV risk, and this may be particularly true in patients with high levels of Lp(a) (>50 mg/dL). To date, no randomized trials have been performed to assess the effect of a statin-mediated increase of Lp(a) levels on cardiovascular risk; however, we can leverage on the findings from genetic studies and clinical trials reporting the impact of Lp(a) lowering on CV risk. Two Mendelian randomization analyses (which analysed 43 and 27 single nucleotide polymorphisms, respectively) calculated the magnitude of plasma Lp(a) level change required to achieve a clinical benefit and found that a large absolute decrease in Lp(a) levels (100 mg/dL and 65.7 mg/dL, respectively) is needed to result in a CV risk reduction comparable with that obtained with an 39 mg/dL (1 mmol/L)