Statins Increase Lp(a) Plasma Level: Is This Clinically Relevant?

Angela Pirillo; Alberico Luigi Catapano


Eur Heart J. 2020;41(24):2285-2287. 

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.[1] 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.[1] 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.[1,2] However, whether Lp(a) is still a risk factor in patients with CV events and on statin therapy is uncertain.[3] 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.[2,4] 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%.[5] Proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibition reduces Lp(a) by 20–30%,[6] 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.[3]

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.[7] 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%[8] and, given the individual data analysis of the study, makes the finding very robust.[2] 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.[9] 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).[10]

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) LDL-C level reduction.[11,12] Furthermore, a 10 mg/dL genetically determined reduction of Lp(a) level was associated with a 5.8% lower risk of coronary heart disease, compared with the 14.5% coronary heart disease risk reduction observed with a 10 mg/dL genetically determined reduction of LDL-C.[11] These findings may help explain, at least in part, why randomized trials which evaluated the effect of hypocholesterolaemic drugs able to reduce Lp(a) levels by 20–35% (such as niacin, cholesterol ester transfer protein inhibitors, and PCSK9 inhibitors) do not show a direct impact of Lp(a) reduction on CV risk over and above the LDL-C-lowering effect,[13–16] as these trials were largely underpowered to show an effect. Of note, in these trials, median baseline Lp(a) levels were low (<20 mg/dL), and thus a 30% decrease would result in a limited absolute reduction in circulating Lp(a), which may not translate into a clinically measurable benefit.[6] Agents that specifically and potently lower Lp(a) levels by inhibiting apo(a) synthesis are under investigation and are intended for use in patients with established CV disease and elevated baseline Lp(a) levels;[17] the cut-off point of Lp(a) plasma levels should be carefully considered among the inclusion criteria. We should remember that ~20% of the general population has an Lp(a) plasma level >50 mg/dL.[1]

Another crucial point that the authors addressed is the mechanism by which atorvastatin increases Lp(a). Tsimikas et al. evaluated the in vitro effect of atorvastatin in HepG2 cells.[7] A dose- and time-dependent increase of the expression of LPA mRNA as well as an increase of synthesis and secretion of apo(a) was observed, which is proposed be responsible for the increase of Lp(a).[7] The mechanisms underlying this effect have not been investigated; we suggest that they may relate to the reduction of intracellular cholesterol on the LXR–FXR (liver X receptor–farnesoid X receptor) axis. An FXR-responsive element is located within the promoter of the LPA gene coding for apo(a)[18] and we suggest that it may be involved in the modulation of apo(a) synthesis by statins[19,20] (Take home figure). The representativeness of these models for humans, however, requires careful assessment. We must acknowledge, in fact, that HepG2 cells may not represent the best model for bile acid production, a pathway intimately linked to the LXR–FXR axis, raising the question of whether the stimulatory effect on Lp(a) observed in vitro can be translated to humans. Furthermore, in a kinetic study, no effects of atorvastatin on Lp(a) levels were reported, with no changes in Lp(a) fractional catabolic rate or in its production rate.[21] In the same study, evolocumab, a PCSK9 inhibitor, significantly reduced Lp(a)plasma levels by reducing its production rate, but, when administered together with atorvastatin, the Lp(a) reduction derived from an increase in the fractional catabolic rate.[21] These results were not confirmed in another study,[22] reflecting, perhaps, the complexity of Lp(a) metabolism, the possible involvement of multiple receptors in its catabolism, as well as the metabolic heterogeneity of patients enrolled in the different studies. Finally, all reported data are related to atorvastatin, and whether the effect can be translated to other statins remains unclear. These observations, together with the conflicting results obtained in statin trials, suggest the need for further studies to better address the role of lipid-lowering drugs in Lp(a) metabolism.

Take home figure.

Apo(a) expression is regulated by both positive and negative factors, among which farnesoid X receptor (FXR) plays a key role. Following the activation by bile acids, FXR binds to a negative DR-1 control element located in the promoter of LPA and represses apo(a) expression (A). Statins, by reducing cholesterol and, as a consequence, bile acid levels, may negatively affect FXR activation, thus resulting in an increased synthesis of apo(a) (B).

In summary, the clinical relevance of the Lp(a) increase caused by atorvastatin is not yet demonstrated, and randomized clinical studies are needed. Ideally, these studies should be performed in subjects/patients with high Lp(a) plasma levels. Until then, statins continue to represent the cornerstone for the treatment of hypercholesterolaemia.