PCSK9 as predictor for recurrent cardiovascular disease in familial hypercholesterolemia

Familial hypercholesterolemia (FH) is a common (1:250) autosomal dominant genetic disorder characterized by high plasma levels of low-density lipoprotein cholesterol (LDL-C) and increased risk for (premature) cardiovascular disease (CVD). This risk ranges from a 2.2- up to 25.8-fold increased CVD risk compared to non-FH subjects with LDL-C below 3.25 mmol/L. ¹ FH can be diagnosed clinically, using scoring systems such as the Dutch Lipid Clinic Network (DLCN) criteria, and/or by identification of a functional variant in one of three culprit FH genes. The CVD risk in FH patients is not totally mitigated by the use of statins, and is still approximately two-fold increased compared to non-affected family members. ² Both risk stratification and additional therapy are deemed crucial to prevent future events. Proprotein convertase subtilisin/kexin type 9 (PCSK9) lowering agents have been shown to lower both LDL-C levels and CVD risk in large randomized controlled trials, and are, therefore, of particular importance in the treatment of FH. ³

PCSK9, however, may also serve as a CVD risk indicator and in the current issue of this journal, Cao et al. investigated whether plasma PCSK9 levels hold predictive value in patients with FH. They showed that PCSK9 concentrations are positively associated with coronary artery calcification (CAC) and recurrent cardiovascular events (CVEs) during a 3.6-year follow-up in 249 FH patients after their first CVE (defined as myocardial infarction, stroke, unstable angina, percutaneous coronary intervention, coronary artery bypass grafting, or peripheral arterial revascularization). After adjustment for traditional risk factors, PCSK9 concentrations were positively associated with CAC scores (β: 0.51, 95% confidence interval (CI): 0.32–0.70, p < 0.001). Furthermore, patients in the highest tertile of PCSK9 levels had a 5.90 hazard ratio (95% CI: 2.20–11.34, p < 0.01) for recurrent CVE compared to patients in the lowest tertile. These results are in line with earlier reported associations between plasma PCSK9 levels and CAC ⁴ and cardiovascular events. ⁵

A strength of the current study is the well-defined FH population with a DLCN score >6 (probable FH and definite FH) or a molecular diagnosis of FH. A large proportion (57%) of subjects were heterozygous carriers of a FH causing variants in low-density lipoprotein receptor or apolipoprotein B (ApoB). Additionally, only three patients were lost to follow-up. Unique to this cohort is that the lipid-lowering therapy was standardized (20 mg rosuvastatin + 10 mg ezetimibe daily) after the admission for the initial CVE.

Several interesting observations from this study warrant further discussion. The aim of this study was to assess the predictive value of PCSK9 levels for recurrent CVEs in FH patients. However, the baseline characteristics in Table 1 show that LDL-C and ApoB levels increased per PCSK9 tertile. Additionally, more genetically confirmed “true” FH patients are present in the highest PCSK9 group. Despite correction for LDL-C levels and mutation status in the multivariate analysis, these two observations might confound the association between PCSK9 and cardiovascular events, since a “true” FH diagnosis and high LDL-C might reflect more severe (lifelong) hypercholesterolemia and, thus, a higher risk for recurrent events. As shown by Khera et al., FH mutation carriers face a 2–3-fold increased CVD risk compared to non-FH mutation carriers with similar LDL-C levels. ¹ One could, therefore, question whether mutation status may outperform PCSK9 levels in predicting recurrent CVE.

Given that the association between PCSK9 level and CVD in the current hypercholesterolemic study population persisted even after adjustment for LDL-C, suggests that PCSK9 promotes atherosclerosis beyond its involvement in lipid metabolism or is a biomarker for other atherosclerotic processes. Despite some evidence that PCSK9 plasma concentration is associated with thrombocyte aggregation in patients with atrial fibrillation, ⁶ there is little and inconclusive evidence that PCSK9 is relevantly involved in other biological processes beyond lipid metabolism. ⁷ In fact, both the randomized clinical trials where PCSK9 inhibition were evaluated as well as the mendelian randomization studies on LDL-C-lowering variants in the PCSK9 gene show that the CVD risk reduction is directly proportional to the LDL-C level, and completely in line with the effect of statin therapy and variants in the HMGCR gene.^3,8

The predictive role of PCSK9 levels for CVE in FH should be viewed in a broader future perspective. Many attempts are currently being taken to generate a personalized medicine approach. For example, polygenic risk scores are used to capture the aggregated burden of small effects of common variants on CVD risk, and proteomic panels can combine the predictive capacity of multiple biomarkers. The question arises whether a single PCSK9 measurement will evoke more predictive value compared to a broad-scale genomic and proteomic assessment, which will incorporate PCSK9 genetic scores and plasma PCSK9 levels.

Cao et al. furthermore show that PCSK9 levels may serve as a predictor of CAC severity.⁹ CAC measurements can be confounded by statin use in the investigated population, since these medications induce coronary calcification, while lowering the risk of CVEs through plaque stabilization. ¹⁰ In the current study, this confounder appears to not play an important role, since 88% of patients were on statin therapy at the time of inclusion, and this percentage did not significantly vary between PCSK9 tertiles. It is reassuring that PCSK9 not only predicts future CVEs but also correlates with CAC severity scores. From a clinical perspective, however, there appears little relevance for this correlation, since CAC is already established as a good proxy for atherosclerosis. CAC scoring in itself, however, can play a large role in risk prediction in asymptomatic FH patients. Results from Cao et al. indirectly lend credence to the observation made by Miname et al. that a CAC score of 0 precluded the occurrence of atherosclerotic cardiovascular disease events in a cohort of asymptomatic FH patients, although this CAC-free subset was younger and predominantly female.^9,11

Cao et al. propose that measuring PCSK9 in FH patients could help provide tailored intensified lipid-lowering therapy to those patients who are at increased risk.⁹ It is subject of debate to what extent absolute PCSK9 plasma levels could be used for tailoring specific PCSK9 inhibition therapy, since it was shown that baseline PCSK9 levels in four phase-three randomized controlled trials with the PCSK9 antibody evolocumab were not related to the achieved LDL-C reduction. ¹² Whether baseline PCSK9 levels are also independent of CVE endpoints in trials is unknown. Moreover, the study by Cao et al. might lead to the spurious conclusion that PCSK9 levels can be used to identify patients that require intensified treatment because of a high risk for recurrent events.⁹ The ultimate goal is to lower CVD risk in FH patients with adequate treatment. Since FH is a genetic disease with lifelong LDL-C accumulation in the vessel wall, treatment should be initiated as early as possible. Therefore, risk stratification in a middle-aged FH patient with established CVD, with the goal of intensifying treatment is probably “too little, too late”. In a perfect world, a FH patient is diagnosed as early as possible and would from that point on receive aggressive LDL-C lowering treatment, which nowadays comprises high-intensity statin in combination with ezetimibe and (if available) PCSK9 inhibition. After all, the cumulative lifetime achieved LDL-C level remains the major determinant for cardiovascular outcomes in FH patients.