Prediction of response to lipid-lowering treatments

KEYWORDS:

• Ezetimibe
• PCSK9 inhibitors
• statins

1. Introduction

Reduction of low-density lipoprotein (LDL) cholesterol with statin therapy is a well-established strategy to reduce cardiovascular events, and statins have become one of the most widely prescribed groups of drugs worldwide. Newer drugs that reduce LDL cholesterol, including ezetimibe and the monoclonal antibodies binding to proprotein convertase subtilisin kexin type 9 (PCSK9), have also been shown to reduce cardiovascular events in large clinical trials. Guidelines are being continually updated to recommend more intensive statin treatment to reduce LDL cholesterol and to recommend the appropriate role for the newer therapies. However, the identification of high-risk patients is largely based on clinical phenotypes and LDL cholesterol levels, and currently genetic factors do have a major role in determining which patients to treat or what the most appropriate therapy would be.

2. Statin tolerability and efficacy

The reduction in LDL cholesterol with statins, or with ezetimibe and the PCSK9 inhibitors, shows a wide variation between individuals. It is generally assumed that a greater reduction in LDL cholesterol will be associated with a greater reduction in cardiovascular events, and the major predictor of the absolute reduction in LDL cholesterol is the baseline level. Whether a precision medicine approach to statin therapy might improve patient selection to maximize benefits and reduce the risk of adverse effects was discussed in a recent review in this journal [1].

Many studies have examined the pharmacogenetics of statin responses, largely related to the reduction in LDL cholesterol rather than the clinical outcomes. Candidate gene studies examined common single-nucleotide polymorphisms (SNPs) in genes for the drug-metabolizing enzymes and drug transporters involved in statin pharmacokinetic pathways along with obvious targets in the pharmacodynamic pathways, such as the genes encoding HMG-CoA reductase (HMGCR) and the LDL receptor (LDLR).

Variants in the drug transporter genes have emerged as being more important in statin tolerability than in the LDL cholesterol response. The rs4149056 (c.521T>C, p.V174A) SNP in SLCO1B1, encoding the organic anion transporting polypeptide 1B1 (OATP1B1) transporter, reduces hepatic uptake of all statins except fluvastatin [2]. This polymorphism was identified in the genome-wide association study (GWAS) for myopathy related to high-dose simvastatin and was also associated with slightly less reduction in LDL cholesterol with simvastatin. It is the only variant recommended for genotyping by the clinical pharmacogenomics implementation consortium to predict a response to statins, which is for the increased risk of myopathy rather than reduced efficacy with simvastatin [3].

The other polymorphism having a significant effect on statin pharmacokinetics and response is in the gene for the ATP-binding cassette (ABC) transporter, ABCG2, which normally limits intestinal absorption and enhances biliary excretion of most statins. The loss-of-function c.421C>A (rs2231142) polymorphism in ABCG2 is associated with higher plasma concentrations and increased LDL cholesterol lowering with rosuvastatin, which has shown in candidate gene studies [4] and a GWAS from the JUPITER (Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin) trial [5]. Plasma concentrations of rosuvastatin were also increased with the c.521T>C polymorphism in SLCO1B1 [4,6], whereas systemic exposure to atorvastatin was associated with the c.521T>C and the c.388A>G (rs2306283) polymorphisms in SLCO1B1 and a marker of CYP3A activity as well as some phenotypic factors [6]. Therefore, genotyping for these transporter SNPs would be useful to predict the maximum safe dose of these statins and should be more accurate than simply basing the maximum dose on ethnicity [1].

Most of the large placebo-controlled statin trials have been subjected to GWAS analysis, and a pharmacogenetic meta-analysis of these GWAS to identify genetic signals related to the LDL cholesterol response to statins confirmed associations found in individual GWAS with APOE and LPA and also identified loci in SLCO1B1 and the rs646776 SNP in SORT1, the gene which encodes sortilin-1, another protein involved in lipoprotein regulation [7]. The APOE gene has three apolipoprotein (apo) E alleles which influence baseline lipid concentrations, and carriers of the E2 allele show a greater reduction in LDL cholesterol with most statins compared to carriers of the other APOE alleles. Apo(a), encoded by LPA, is the key protein defining lipoprotein(a) [Lp(a)] particles, and SNPs in LPA influence the plasma concentration of Lp(a), which in turn influences the LDL cholesterol reduction with statins and the overall cardiovascular risk.

Polymorphisms in PCSK9 also influence the response to statins in some GWAS [5], and the many gain-of-function and loss-of-function mutations in PCSK9 have a significant impact on circulating LDL cholesterol levels [8]. Some of these polymorphisms in PCSK9 were related to the risk of cardiovascular disease in Mendelian randomization studies and are likely to influence the response to the anti-PCSK9 monoclonal antibodies as well as to statins, but genetic analyses from the PCSK9 inhibitor studies are not available yet. Likewise, functional variants in the Niemann–Pick C1 Like 1 protein gene (NPC1L1) influence baseline LDL cholesterol levels, and a Mendelian randomization study found that lower LDL cholesterol mediated by NPC1L1 polymorphisms was associated with a lower risk of coronary heart disease (CHD) similar to the effects of HMGCR polymorphisms [9].

Whether NPC1L1 polymorphisms influenced the outcome of IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial) is not yet known, but the benefit from the addition of ezetimibe to simvastatin after acute coronary syndrome was greater in patients with diabetes mellitus and in those with a high-risk score without diabetes mellitus [10]. In the FOURIER trial (Further Cardiovascular Outcomes Research With PCSK9 Inhibition in Patients With Elevated Risk) with the PCSK9 inhibitor evolocumab, the absolute risk reductions were greatest in patients with higher-baseline high-sensitivity C-reactive protein (hsCRP) and event rates were lowest in patients who achieved the lowest hsCRP and LDL cholesterol levels [11]. Considering the high cost of the PCSK9 inhibitors, having a genetic test to help predict the clinical benefit might well be cost-effective, whereas for statins and ezetimibe, which are generic and inexpensive in most countries, cost-effectiveness of genetic tests would be limited although it may still be reasonable considering the lifelong nature of the therapy.

3. Imprecision medicine

The benefit in reduction in cardiovascular events with statins and other lipid-lowering treatments is currently mainly predicted by the overall cardiovascular risk and the baseline level of LDL cholesterol or other risk markers. Schork considered that treatment with statins was an example of imprecision medicine, even when the number needed to treat (NNT) to prevent one combined event over 5 years was 20, as in the JUPITER study with rosuvastatin in subjects with relatively low LDL cholesterol (<130 mg/dL) and elevated hsCRP (≥2 mg/L) [12]. The NNT value of 20 is lower than seen with many widely accepted medical interventions for primary or secondary cardiovascular prevention but still means that 19 out of each 20 patients treated will have no benefit in preventing cardiovascular events over the five years of the study, although the reduction in LDL cholesterol may confer benefits later.

4. Improving prediction

The monogenic condition of familial hypercholesterolemia provides an obvious example of when genetic testing can help to define the level of cardiovascular risk [13]. Because of the lifetime exposure to high levels of LDL cholesterol, the CHD risk is much greater than would be predicted from the usual risk assessment algorithms. However, the need for and intensity of treatment are still influenced by the phenotype, in particular the LDL cholesterol level and the presence of other cardiovascular risk factors.

The more common polygenic variants contributing to cardiovascular risk factors such as plasma LDL cholesterol levels also contribute to overall CHD risk, and by combining these together, a genetic risk score (GRS) can be obtained. One study using this approach used a GRS based on 27 genetic variants and combined data from a cohort study and four randomized controlled trials with statin therapy in both primary and secondary prevention [14]. The GRS identified individuals at increased risk for both incident and recurrent CHD events, and those subjects with the highest GRS values derived the largest relative and absolute clinical benefit from statin therapy with a roughly threefold decrease in the NNT to prevent one CHD event between those with low and high genetic risk in the primary prevention trials.

Another study applied a GRS derived from 57 common SNPs in the primary prevention WOSCOPS (West of Scotland Coronary Prevention Study) statin trial and found that the relative risk reduction with statin therapy was 44% in the top quintile of polygenic risk score compared with 24% in all the other groups despite similar reductions in LDL cholesterol [15]. Likewise, the NNT to prevent nonfatal myocardial infarction or death from CHD attributed to statin in the high genetic risk group was only 13 compared to 38 in the lower-risk groups. The GRS was also associated with a greater extent of subclinical atherosclerosis determined by coronary artery calcification and carotid artery plaque burden.

More accurate assessment of subclinical atherosclerosis with noninvasive imaging and identification of novel biomarkers to detect the presence of vulnerable plaque could further refine the assessment of cardiovascular risk. Combining these with the collective effect of multiple common genetic variants could improve the prediction of cardiovascular risk and reduce the NNT even further. Further pharmacogenetic analyses from the studies with ezetimibe and PCSK9 inhibitors should help to establish which patients would benefit most from these new therapies. The precision medicine approach to lipid-lowering therapy appears to be nearing fruition.

Declaration of interest

B Tomlinson has received research funding from Amgen, Merck Sharp and Dohme, Pfizer, and Roche. He has also acted as a consultant, advisor or speaker for Amgen, Merck Serono, and Sanofi. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This manuscript was not funded.