Approach to the Patient With a Suboptimal Statin Response: Causes and Algorithm for Clinical Management

Abstract

Context

Statins are the lipid-lowering therapy of choice for the prevention of atherosclerotic cardiovascular disease (ASCVD) but their effectiveness in lowering low-density lipoprotein cholesterol (LDL-C) can substantially differ between individuals. In this mini-review, we describe the different causes for a suboptimal statin response and an algorithm for the diagnosis and clinical management of these patients.

Evidence Acquisition

A PubMed search using the terms “statin resistance,” “statin sensitivity,” “statin pharmacokinetics,” “cardiovascular disease,” and “lipid-lowering therapies” was performed. Published papers in the past 10 years that were relevant to the topic were examined to provide content for this mini-review.

Evidence Synthesis

Suboptimal lowering of LDL-C by statins is a major problem in the clinical management of patients and limits the value of this therapeutic approach. There are multiple causes of statin hyporesponsiveness with compliance being the most common explanation. Other causes, such as analytical issues with LDL-C measurement and the presence of common lipid disorders (familial hypercholesterolemia, elevated lipoprotein[a] and secondary dyslipidemias) should be excluded before considering primary statin resistance from rare genetic variants in lipoprotein-related or drug-metabolism genes. A wide variety of nonstatin lipid-lowering drugs are now available and can be added to statins to achieve more effective LDL-C lowering.

Conclusions

The evaluation of statin hyporesponsiveness is a multistep process that can lead to the optimization of lipid-lowering therapy for the prevention of ASCVD. It may also lead to the identification of distinct types of dyslipidemias that require specific therapies and/or the genetic screening of family members.

Statins for more than 3 decades have been the primary therapy for lowering plasma lipids to reduce atherosclerotic cardiovascular disease (ASCVD) events (1). Statins inhibit 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, an enzyme early in the cholesterol (C) biosynthetic pathway (Fig. 1). Hepatic C depletion after statin treatment induces a compensatory upregulation of the receptor for low-density lipoproteins (LDLs), which then increases the clearance of LDL from the circulation. Because LDL is one of the main drivers of atherosclerosis (2, 3), the lowering of LDL, which is most often monitored by its C content (LDL-C), reduces ASCVD risk (2, 4).

Figure 1.

Mechanism of action of lipid-lowering medications. The major pathways of lipoprotein metabolism are shown along with the site and mechanism of action of lipid-lowering drugs. Triglyceride (TG)-rich very low-density lipoproteins (VLDLs) are secreted by the liver and are transformed into LDL after lipolysis by LPL. LPL is inhibited by ANGPTL3/8 complex. LDL is cleared from the plasma by hepatic LDL receptors, which are downregulated by excess intracellular hepatic cholesterol and the binding to PCSK9 in the plasma. Excess LDL in plasma gets deposited in the wall of arterial vessels where it triggers atherosclerosis. (−) indicates downregulation of indicated pathway.

Several different types of statins have been developed (Table 1), and in general the newer ones are more effective (1). Based on the dose and type of statin, they are often categorized by their degree of LDL lowering. Rather than the use of “normal reference ranges” for LDL-C, the desired degree of LDL-C lowering depends on the level of risk (4). For all high-risk patients (diabetes, LDL-C ≥ 190 mg/dL, or a 10-year risk score ≥ 20%), it is recommended that LDL-C be lowered by more than 50% by the use of a high-intensity statin regimen (4). Moderate-intensity statins, which are mostly reserved for patients with intermediate ASCVD risk, typically result in a 30% to 50% reduction in LDL-C. For secondary prevention, when the goal is to lower LDL-C below 70 mg/dL or even lower (4, 5), high-dose statins are often used in conjunction with a nonstatin lipid-lowering medication (5). Sex and race do not change these recommendations but do affect the calculation of the 10-year risk score, which is used to further stratify intermediate-risk patients (4).

Table 1.

List of high, moderate, and low-intensity statin regimens

High-intensity statins		Moderate-intensity statins		Low-intensity statins
Name	Daily dose, mg	Name	Daily dose, mg	Name	Daily dose, mg
Atorvastastina (Lipitor)	40-80	Atorvastatina (Lipitor)	10-20	Simvastatina (Zocor)	10
Rosuvastatinb (Crestor, Ezallor)	20-40	Rosuvastatinb (Crestor, Ezallor)	5-10	Pravastatinb (Pravachol)	1-20
		Simvastatina (Zocor)	20-40	Lovastatina (Altoprev)	20
		Pravastatinb (Pravachol)	40-80	Fluvastatina (Lescol)	20-40
		Lovastatina (Altoprev)	40
		Fluvastatin XLa (Lescol XL)	80
		Fluvastatina (Lescol)	40 mg, 2×/d
		Pitavastatina (Livalo, Zypitamag)	2-4

From Grundy et al (4).

Lipophilic statins.

Hydrophilic statins.

The older low-intensity statins (see Table 1) usually result in less than a 30% reduction in LDL-C, which is now viewed as inadequate for primary prevention. These statins may still be used in statin-intolerant patients to avoid side effects but usually along with a second nonstatin lipid-lowering agent. Even on high-intensity statins, some patients will show less than a 30% reduction. A suboptimal statin response (statin hyporesponsiveness) can occur for a wide variety of reasons in as many as half of all patients (6, 7), but it is most often due to noncompliance (8). About 15% to 20% of patients, however, appear to be statin resistant, which is often defined as showing less than a 15% reduction in LDL-C on maximally tolerated statins (8-10). In the Jupiter trial, 42.8% of patients on a daily dose of 20 mg of rosuvastatin showed less than a 50% reduction of LDL-C and 10.8% showed either no reduction or a small increase in LDL-C (11). Patients exhibiting a statin hyporesponsive have been shown by intravascular ultrasound to undergo further plaque progression compared to adequately treated patients (9) and to develop significantly more ASCVD events (7, 10).

In this review, we first describe a typical case of statin hyporesponsiveness. Next, the various potential causes of statin hyporesponsiveness, and an algorithm for the clinical management of these patients, are discussed. Finally, we conclude with a brief description of new nonstatin lipid-lowering drugs that one should consider prescribing for a patient that does not achieve sufficient LDL-C lowering on a statin.

Clinical Case

A 60-year-old Hispanic woman with a past medical history of hypertension, and a 20-year history of type 2 diabetes mellitus presented with worsening chest pain on physical exertion. Her diabetes was under good control with a glycated hemoglobin A_1c of 5.7% when on treatment with metformin, empagliflozin, and liraglutide. She had no acute or chronic complications from her diabetes. The patient had been treated with statins, but she discontinued therapy because of a lack of response. She underwent cardiac catheterization and was found to have significant obstructive coronary artery disease. A 3-vessel coronary artery bypass graft surgery was performed, which relieved her angina.

Her initial lipid profile revealed a total C (TC) of 232 mg/dL (reference range, 50-199 mg/dL), high-density lipoprotein C (HDL-C) of 48 mg/dL(reference range, 40-60 mg/dL), non-HDL-C of 184 mg/dL (reference range, < 130 mg/dL), triglycerides (TGs) of 186 mg/dL (references range, 35-149 mg/dL), direct LDL-C of 158 mg/dL (reference range, 5-99 mg/dL), calculated LDL-C (Sampson-NIH) of 150 mg/dL, and an apolipoprotein B (apoB) of 121 mg/dL (reference range, 51-97 mg/dL). She was started on atorvastatin 40 mg daily plus ezetimibe 10 mg daily. Her follow-up lipid profile about 3 months later showed TC of 206 mg/dL, HDL-C of 50 mg/dL, non-HDL-C of 156, TGs of 129 mg/dL, direct LDL-C of 129 mg/dL, calculated LDL-C (Sampson-NIH (12)) of 133 mg/dL, and an apoB of 90 mg/dL. Because her LDL-C reduction after a combined statin therapy with ezetimibe was still suboptimal, she was referred to a lipid clinic for further evaluation.

After providing a detailed family history, the patient was found to have a first-degree relative with premature coronary disease and mixed dyslipidemia. In addition, both of her siblings had mixed dyslipidemia. Besides her history of diabetes, she had no other potential explanations for her dyslipidemia. On physical exam, there were no xanthomas, corneal arcus, xanthelasma, or palmar xanthomas. Adherence to the statin/ezetimibe treatment was considered adequate. Her Dutch Lipid Clinic Network Score was only 2 points, making the diagnosis of familial hypercholesterolemia (FH) unlikely. She was tested for lipoprotein(a) (Lp(a)) and found to have a marked increase level of 208 mg/dL. The patient was recommended to intensify her lifestyle factor changes and was started on evolocumab (140 mg subcutaneous every 2 weeks). She was already on low-dose aspirin. Recommendations were made to screen the extended family for elevated Lp(a).

Causes of Statin Hyporesponsiveness

As described in this case, there are several potential explanations one should consider when evaluating a statin-hyporesponsive patient, who shows less than a 30% reduction of LDL-C on a moderate- or high-intensity statin. We have broadly categorized the various possible explanations for statin hyporesponsiveness into the following 4 categories: analytical problems, compliance issues, common lipid disorders, and primary statin resistance. We also describe an algorithm for the clinical approach to the statin-hyporesponsive patient (Fig. 2).

Figure 2.

Flowchart for diagnostic evaluation of statin hyporesponsiveness. Statin-hyporesponsive patients, who are defined as having less than a 30% reduction in low-density lipoprotein cholesterol (LDL-C) on moderate- or high-intensity statins, should be further evaluated according to this algorithm. First, compliance and analytical problems related to the LDL-C measurement should be excluded. Next, the presence of 3 relatively common lipid disorders should be investigated. Primary statin resistance is largely a presumptive diagnosis after excluding other possibilities, but there are pharmacogenetic tests that can be considered. Regardless of the cause, nonstatin lipid-lowering therapy should be considered in all patients who do not achieve sufficient LDL-C lowering on a maximally tolerated dose of statins.

Analytical Problems

Most laboratories still calculate LDL-C by the Friedewald equation, which depends on the measurement of lipids in the standard lipid panel (TC, TGs, and HDL-C) (13). Because of biological and analytical variability, LDL-C values between 2 visits can randomly vary by as much as 15% to 20% in a patient with no change in therapy. Furthermore, most current guidelines now state that a nonfasting sample is suitable for the initial evaluation of a patient (4). If one specimen is collected in a fasting state and the other after a meal, the change in TGs can have a major effect in the accuracy of the calculation, particularly by the Friedewald equation, which has a negative bias for high TG samples (13, 14). There are newer, more accurate LDL-C equations, such as the Sampson-NIH or Martin-Hopkins equations, that are less affected by hypertriglyceridemia and ideally should be used instead of the Friedewald equation (15). There are homogenous assays for directly measuring LDL-C, but at times they can also detect C on other lipoprotein fractions besides LDL, particularly for dyslipidemic patients (16, 17). Although not common, patients with elevated lipoprotein X, an abnormal C-rich particle that accumulates in cholestasis, will often have falsely elevated LDL-C by all the different calculations and by some direct LDL-C methods (18).

When investigating a patient for possible statin hyporesponsiveness, it is therefore important to test for LDL-C on multiple occasions and to consider the method used to either estimate or measure LDL-C. It is also important to note whether the specimen used for the measurement was fasting, especially when comparing longitudinal values. Ideally, a patient should also be on a stable dose of statins for at least 4 to 6 weeks before testing the effect of a statin on LDL-C levels. To overcome some of the analytical problems related to LDL-C, one could consider measuring instead apoB, which is the main protein component of LDL. Statins tend to lower larger subspecies of LDL more than smaller ones and hence LDL-C typically decreases more than apoB on statins, which may explain why treating to apoB targets may be a more effective approach (19).

Compliance Issues

Concurrent with or after ruling out an analytical problem, one should also consider whether a patient is adherent to statin therapy. It is often difficult to motivate asymptomatic patients to take daily medications, particularly for preventing possible future disease outcomes. Several studies have shown that as many as 30% to 50% of all patients will discontinue their statins after only a few months. Even for those that continue with their statin therapy, on average they take only about 85% of the prescribed pills (20-25). One should also query the patient on how they take their statin and when. Some studies have shown that short-acting statins work best when taken before bedtime (26), whereas this is not observed for longer-acting statins like for atorvastatin and rosuvastatin (27-29). Asking about any changes in lifestyle factors, such as diet or exercise, can also be useful because of their potential independent effect on modulating LDL-C levels.

Some of the lack of compliance to statins is because of their side effects, the most common of which is myalgia (30). It usually manifests as a dull muscle ache or weakness of proximal limb muscles with normal plasma creatine kinase. In severe cases, it can progress to myositis, elevated creatine kinase, and even rhabdomyolysis. Usually, more potent statins, such as atorvastatin and rosuvastatin, are more likely to cause muscle-related side effects. In well-controlled clinical trials, the frequency of myalgia is usually less than 10% (31). In clinical practice, side effects from statins appear to be 2 to 3 times higher, but this may be due to a nocebo effect (32). Some recommendations have suggested that a patient should be offered at least 2 (preferably 3) different types of statins before considering other types of lipid-lowering therapies (5, 33). Strategies for the clinical management of patients with statin-associated muscle symptoms have been well described (30), and when followed the majority of patients initially complaining of myalgias can be placed back on a statin without symptoms (30). Statins can also cause a mild increase in liver transaminases, but this rarely results in clinically relevant liver injury (34).

For underserved parts of the population, who do not receive adequate medical care, there may be a large variety of socioeconomic reasons for poor compliance (35). For example, lack of trust in the medical field, transportation issues, and inability to afford copayments can create major barriers to the use of any medication. Besides directly asking patients about their statin use, indirect methods, such as questionnaires, pill counts, and electronic medication monitoring, are alternative approaches that show promise for improving adherence (36). Although infrequently done in routine clinical practice, there are therapeutic drug monitoring assays for measuring statins (37). Finally, education of physicians and other health care workers and the active participation of pharmacists can improve compliance (21, 38).

Common Lipid Disorders

After excluding analytical and compliance issues, one should consider 3 relatively common lipid disorders, namely FH, elevated Lp(a), and secondary dyslipidemias, as potential causes for statin hyporesponsiveness. Although all 3 of these diagnoses/conditions should have been considered during the initial evaluation of a patient before starting a statin, it is prudent to conduct a reevaluation when a patient does not show a good response to statins. Identification of any one of these types of common lipid disorders can have a major effect on clinical care.

Familial hypercholesterolemia

FH is a relatively common inherited disorder due to mutations in the LDLR gene, which encodes for the LDL receptor (LDLR) (39). These patients are at a marked increased risk of ASCVD because of their lifelong increase in LDL-C. They can present with tendon xanthomas, corneal arcus, and xanthelasmas from increased tissue deposition of C. Because of the importance of LDLR in the mechanism of action of statins, depending on the degree to which their LDLR mutation affects function, these patients often show a poor response to statin therapy (40).

According to the number of alleles involved, FH is classified into 2 major forms: heterozygous and homozygous FH. In more than 90% of cases, heterozygous FH is caused by loss-of-function mutations in the LDLR gene. Other less common causes are heterozygous mutations in APOB gene, which affects its LDLR-binding domain, or heterozygous gain-of-function mutations in proprotein convertase subtilisin/kexin type 9 (PCSK9) (41, 42), a protein that affects the recycling of the LDLR (see Fig. 1). Homozygous FH individuals have the same exact mutation in both alleles. Compound heterozygous patients have different type of mutations on both alleles, whereas combined heterozygous patients have different mutations on 2 different genes that affect LDL levels (43). Once a proband with FH is identified, it is important to conduct cascade screening in family members to possibly identify other individuals who may be at increased risk (44, 45).

It is estimated that the prevalence of heterozygous FH is about 1 in 200 to 1 in 500 individuals, depending on the population (42, 46), but it is as high as 1 in 31 among patients with ASCVD (47). Homozygous FH affects about 1 in 160 000 to 300 000 people (48). Despite its clinical importance, FH is frequently underdiagnosed (42), especially in Latin America (49). Usually LDL-C levels in heterozygous FH is 2 to 3 times higher than normal, with TC of 310 to 580 mg/dL. LDL-C levels in homozygous FH can be up to 10 times higher, with TC of 460 to 1160 mg/dL (42, 46). LDLR defective mutations often have 2% to 25% residual LDLR activity, but LDLR-negative or null mutations with less than 2% residual activity have been described and are often refractory to statin therapy (46, 50). The presence of other gene variants, such as genes encoding LDLRR adaptor protein 1 (LDLRAP1), apolipoprotein E (APOE), or adenosine triphosphate-binding cassette transporters G5 and G8 (ABCG5 and ABCG8), may have a positive or negative effect on LDL-C levels through gene-gene interactions.

The Dutch Lipid Clinic Network Score is a commonly used to screen for FH patients (51) and is based on a combination of family history, clinical history, physical findings, and laboratory testing. It is generally recommended than anyone with an LDL-C of 190 mg/dL or greater be evaluated for FH (4). Genetic testing is routinely available to confirm the diagnosis, and family members of confirmed probands should also be screened. As described later in the therapy section, FH patients often require statins plus multiple nonstatin lipid-lowering therapies and still often do not achieve good LDL-C lowering. Because of the complexity of treating these patients and their high risk, it is best that they are referred to a lipid specialist.

Elevated lipoprotein(a)

Nearly 30% of the population with Lp(a) levels greater than 30 mg/dL or 75 nmol/L are at increased risk of ASCVD (52, 53). Lp(a) is an LDL-like particle with the apolipoprotein(a) (apo(a)) protein covalently connected to apoB via a disulfide bond (52, 54). Apo(a) contains a kringle KIV domain, which exists in 10 subtypes. Among them, KIV₂ varies in individuals in its copy number from 1 to greater than 40, resulting in different size polymorphisms for apo(a). Those individuals with single-nucleotide variations that correspond to shorter apo(a) size variants typically have higher plasma Lp(a) levels and increased risk (52). Several studies have reported on the association between the apo(a) size isoform with Lp(a) levels and ASCVD risk, but the plasma level of Lp(a) appears to be sufficient for assessing ASCVD risk (55).

Unlike LDL-C, Lp(a) does not typically change much in response to diet or other lifestyle interventions. Typically, statins and most other lipid-lowering drugs do not lower Lp(a), and sometimes statins may actually cause a slight increase (56, 57). A recent meta-analysis found that the increase in Lp(a) on statins could vary between 8% to 24% (58). The mechanism for this increase is not well understood but may be due to increased apo(a) gene expression from the statin therapy (58).

When LDL-C is calculated, it includes C that is carried by Lp(a). For most individuals, the fraction of C on Lp(a) is less than 6% of the total amount of C on LDL (59). Lp(a), however, has a log-normal population distribution. Some patients therefore can contain a much larger fraction of C on Lp(a), especially when their LDL has been effectively lowered by statins. This makes patients with elevated Lp(a) like in our case report appear to be relatively resistant to statins because their calculated LDL-C or directly measured LDL-C does not drop as much after statin therapy. Alternative measures of LDL, such as apoB, do not correct for this problem because it is also present on Lp(a). Non–HDL-C, which includes C on all apoB-containing lipoproteins, also includes C on Lp(a). There are specific immunoassays for Lp(a), but these assays have limitations because of the kringle repeat domain number variation of apo(a), which can differentially affect the accuracy of the measurement (14, 60). In addition, the heterogeneity in the size and composition of Lp(a) makes it difficult to estimate from the immunoassay test result how much C is on the Lp(a). There are specific gel-based assays for directly measuring Lp(a)-C, but they are available only in specialty reference laboratories (61).

As shown in Fig. 2, we recommend that Lp(a) be measured, if not already known, in all patients who show a poor response to statin therapy. This is consistent with most current guidelines that recommend that Lp(a) be measured at least once in assessing ASCVD risk. If Lp(a) is elevated, a higher dose of a statin and/or a second nonstatin lipid-lowering agent may be needed to try to maximize “true” LDL-C lowering. Any other risk factors should also be carefully managed because of the high risk of ASCVD in these patients. The patient's family should also be screened for elevated Lp(a) because of its strong genetic association. New drug therapies for specifically lowering Lp(a) are being developed but are not yet approved (62). Treatment with low-dose aspirin, which reduces the risk of thrombosis in these patients, can be considered (63, 64).

Secondary dyslipidemias

A wide variety of conditions, such as hypothyroidism, nephrotic syndrome, cholestasis, inflammatory diseases, and obesity have been reported to be associated with decreased responsiveness to statins. Most of these are also considered secondary causes of dyslipidemia, which accounts for about a third of all patients with dyslipidemia. It is mostly due to acquired medical conditions or concomitant drug therapy (65).

The mechanism for why these secondary dyslipidemias can sometimes cause statin hyporesponsiveness is known in a few cases. For example, hypothyroid patients often present with elevated LDL-C and typically show a poor statin response. This is thought to occur because of a reduction of hepatic LDLR from decreased triiodothyronine, which is needed for optimum gene expression of the receptor (66). Nephrotic syndrome is another example of secondary dyslipidemia with relative statin resistance. In these patients, expression of PCSK9 is often increased, leading to increased degradation of LDLR. In addition, there is increased production of LDL in nephrotic syndrome (67). Except for patients with cholestatic liver disease and elevated lipoprotein X, most patients with liver disease have decreased plasma LDL-C levels and other lipids because of reduced hepatic production (68).

Many drugs can also contribute to statin resistance by a variety of mechanism. HIV antiretroviral therapies promote C biosynthesis and reduce statin blood levels (69). Amiodarone, presumably by causing hypothyroidism, has also been reported to cause statin resistance (70). As discussed in more detail in the following section, some statins are metabolized by cytochrome P450 (CYP450)-dependent pathways, so some drugs like phenytoin, rifampicin, and carbamazepine, and even some foods (71) that induce CYP450 pathways can cause statin resistance (72). Corticosteroids, anabolic steroids, retinoids, and cyclosporine A are other frequent drug causes of statin resistance (73).

A rare secondary cause of statin resistance is the formation of autoantibodies against the LDLR or apoB (74-77). These can spontaneously form in response to the formation of new epitopes from the oxidation of LDL (74, 75, 78). Paraproteins in patients with multiple myeloma can also form against LDL and block its plasma clearance (76, 77). These patients do not typically respond well to statins, but their LDL-C will often decrease after treatment of their multiple myeloma.

In general, most of the secondary causes of statin resistance are best addressed by treating the underlying condition. Doing so may resolve the dyslipidemia and or help the statin, if still needed, to work more effectively. In fact, this is the reason for the general recommendation that secondary dyslipidemias should always be addressed first before initiating statin therapy.

Primary Statin Resistance

Numerous gene variants are associated with statin effectiveness and can be considered as causes for primary statin resistance (79). A partial list of the more thoroughly investigated genes is shown in Table 2. They can be broadly classified into 2 groups. First are those that are directly related to lipoprotein metabolism and affect either LDL production or its catabolism. The second group are drug metabolism–related genes that affect statin pharmacokinetics.

Table 2.

Genes mutations associated with reduced statin response

Gene	Variant	Allele	Mechanism/Influence	Ref.
Lipid metabolism
APOA1	rs533556	—	APOA1 gene mutation causes decreased cholesterol reduction after statin therapy	(80)
APOA5	rs662799	−1131T > C	ApoA5 mutation associated with hypertriglyceridemia and poor statin response	(81)
APOE	rs429358	388T > C (E4)	E4 genotype is least responsive to statins compared with carriers of E2 (526C > T) or E3 (388T-526C)	(81-83)
CETP	rs1532624	C > A	CETP variant associated with poor statin response and increase of TC	(79, 80)
HMGCR	rs17244841 rs17238540	g.331648A > T, g.27506T > G	HMGCR variants, target of statins, with decreased lipid-lowering efficacy of statins	(81, 83)
LDLR	AvaII (rs5925), PvuII (rs256954), HincII (rs688)	T > C, A > G, C > T	LDLR genotypes with decreased lipid-lowering efficiency of statins	(79-83)
LPA	rs10455872	A > G	LPA variant with increased Lp(a) and decreased lipid-lowering efficacy of statins	(79, 113)
PCSK9	rs17111584	T > C	PCSK9 variant increases LDLR degradation and decreases statin response	(81)
Drug metabolism
ABCB1	rs1128503, rs1045642, rs2032582	C1236T, C3435T, G2677T	ABC transporters belong to efflux transporters of various substrates including statins	(81)
ABCC2	rs717620	−24C > T
CYP7A1	—	A-204C	CYP450 catalyzes the phase I reaction of drugs including statins, and both variants affect statin metabolism causing their lower efficacy in decreasing LDL-C	(79)
CYP3A5	—	3		(82)
CYP3A4	—	1G		(81)
SLCO1B1	rs2306283	388A > G	Functional defects in this transporter affect cellular statin uptake and its efficacy	(81)
	rs2900478 rs4149056	—	Polymorphisms associated with less LDL-C reduction in response to statin treatment	(113) (110)

Abbreviations: LDL-C, low-density lipoprotein cholesterol; LDLR, LDL receptor; Lp(a), lipoprotein a; TC, total cholesterol.

Besides LDLR and PCSK9, which have already been discussed, genetic variants in HMGCR, the gene that encodes for HMG-CoA reductase, have also been associated with statin resistance (84). It is not certain if it is due to a decrease ability of statins to inhibit the enzyme or from greater compensatory upregulation of the HMGCR gene. Other lipoprotein-related genes associated with statin resistance include those that encode for apoE, a ligand for the LDLR and other related lipoprotein receptors, and for cholesteryl ester transfer protein (CETP) (85), which transfers neutral lipids between lipoproteins. Transcriptomic analysis of immortalized lymphoblastoid cells from patients treated with statins have uncovered even more potential genetic loci that may affect statin responsiveness (86). In general, the effect size for many of these genes is relatively small and may vary depending on the population.

A wide variety of drug metabolism–related genes also contribute to statin resistance (see Table 2). In many cases, these same genes but different variants also account for increased statin sensitivity (33). Some like SLCO1B1 (solute carrier organic anion transporter family 1B1) are involved in the hepatic uptake of statins (87, 88), whereas other like the ATP-binding cassette (ABC) transporters are involved in the efflux of statins from cells (89). Perhaps the most clinically relevant are the CYP3A5 and CYP3A4 genes, which inactivate statins by oxidation (90). This pathway is more important for lipophilic statins (see Table 1) like atorvastatin and do not affect as much hydrophilic statins like pravastatin, which is sulfated and undergoes more renal elimination than other statins (28). Other cytochromes like CYP2C9 and other elimination pathways affect other statins (27), but the role of variations in these genes is not as well understood.

Although it is now becoming easier and less expensive to perform pharmacogenetic analysis of drug metabolism genes (91), there is no general recommendation to do so for assessing statin resistance. In fact, even though there is more evidence to support its possible clinical utility (92), there is currently no recommendation to conduct pharmacogenetic testing for managing statin sensitivity. The diagnosis of primary statin resistance can, therefore, be considered a presumptive diagnosis made after excluding other, more common causes (see Fig. 2). Given that some types of primary statin resistance are due to drug metabolism genes, it may be reasonable to empirically test both a high-intensity lipophilic and hydrophilic statin before adding a nonstatin lipid medication.

Nonstatin Lipid-Lowering Therapies

Several new therapies have emerged in the past few years (Table 3) that target complementary metabolic pathways to the HMG-CoA reductase inhibition by statins (see Fig. 1). This provides new drug options for patients when increasing the statin dose or changing the type of statin does not improve LDL-C lowering. Unless a patient has a mixed dyslipidemia with elevated TGs, the 2 nonstatin therapies that should be considered first, according to most guidelines, are ezetimibe and anti-PCSK9 therapy (4, 5, 30).

Table 3.

Nonstatin lipid-lowering drugs

Drug	Reduction of LDL-C, %	Mechanism	Ref.
Bempedoic acid	18-29	Inhibitor of ATP citrate lyase	(114)
Evinacumab	23-50	mAb inhibitor of angiopoietin-like 3, which inhibits lipoprotein and endothelial lipase	(115)
Ezetimibe	21-27	Targets NPC1L1 protein, which mediates intestinal absorption of C	(95)
Lomitapide	38-50	Inhibits MTP	(116)
Mipomersen	25-37	Antisense oligonucleotide preventing translation of apoB	(117)
PCSK9 inhibitors	50-60	siRNA, mAb therapies that block degradation of LDLR in hepatocytes	(104, 106)

Abbreviations: apoB, apolipoprotein B; ATP, adenosine triphosphate; C, cholesterol; LDL-C, low-density lipoprotein cholesterol; LDLR, LDL receptor; mAb, monoclonal antibody; MTP, microsomal triglyceride transfer protein; NPC1L1, Niemann-Pick C1-like 1; siRNA, small interfering RNA.

Ezetimibe is a relatively inexpensive drug that can either be taken alone or in a combined pill with a statin (93, 94). It inhibits the NPC1L1 protein in the intestine and reduces C absorption (see Fig. 1). When used in conjunction with statins, it lowers LDL-C by an additional 21% to 27% (95). C absorption is known to be upregulated in response to statin therapy (96), which may explain why statins plus ezetimibe is an effective combination. The IMPROVE-IT outcome study supports the recommendation that ezetimibe should be used as a second-line therapy when the desired therapeutic goal is not reached at the maximum tolerated dose of statins, or when statins cannot be prescribed (97).

Another relatively inexpensive oral agent that one could consider is bile acid sequestrants (see Fig. 1), which by reducing bile acid reabsorption deplete hepatic C levels and upregulate LDLR (98). The reductions in LDL-C, however, are relatively modest, and bile acid sequestrants can cause gastrointestinal side effects (99). Another possible consideration is bempedoic acid (see Fig. 1), which is a new drug that blocks C biosynthesis upstream from HMGCoA reductase (100), specifically the ATP-citrate-lyase (100, 101). It is available as a separate drug or in combination with a statin and lowers LDL-C on top of statins by approximately 15% to 25% and up to 50% when combined with ezetimibe (102). Depending on the result of the ongoing CLEAR Outcomes study (https://clinicaltrials.gov/ct2/show/NCT02993406), it may become an even more attractive option in the future.

Monoclonal antibody (mAb) PCSK9 inhibitors, such as evolocumab and alirocumab, are the second main option for a combination therapy. The neutralization of PCSK9 in plasma by these mAbs increases the half-life of the LDLR on hepatocytes, leading to lower LDL-C levels (see Fig. 1). In different studies, such as ODDYSEY (103) and FOURIER (104), LDL-C lowering has been as high as 50% to 60% when used on top of statins and they reduce ASCVD events by an additional 15%.

Inclisiran, a small interfering RNA (siRNA) targeted against PCSK9 messenger RNA, is a recently approved drug to be used in conjunction with a statin (105). It contains a triple-antennary N-acetylgalactosamine (GalNAc) oligosaccharide that targets the siRNA to the liver, where PCSK9 is produced by promoting its uptake by the asialoglycoprotein receptor (105). This siRNA has a long-acting effect (6 months) even though its plasma half-life is less than a day. This is mainly due its sustained interaction with the intracellular RNA-induced silencing complex. In the ORION clinical trials, inclisiran reduced LDL-C by approximately 40% to 60% (106). The ORION-4, a phase 3 study involving more than 15 000 patients, is testing the effect of inclisiran on ASCVD events in patients with existing CVD (https://clinicaltrials.gov/ct2/show/NCT03705234?cond=orion+4&draw=2&rank=1). This drug may be a more convenient option for patients compared to mAbs to PCSK9 because once a stable dose is established, it is administered only 2 times a year. In general, because of their expense, anti-PCSK9 therapies are mostly used in secondary prevention in patients who do not achieve sufficient LDL-C lowering on more conventional therapies like statins or statins plus ezetimibe (107).

Another promising treatment target is angiopoietin-like 3 (ANGPTL3). This protein inhibits lipoprotein lipase and endothelial lipase, 2 enzymes that hydrolyze TGs and lower TG-rich lipoproteins (see Fig. 1). ANGPTL3 forms a complex with ANGPTL8 and is induced after feeding and helps direct dietary fat to adipocytes for storage (108). Evinacumab is a fully humanized mAb against ANGPTL3 and has been shown to decrease TGs (−75%) and LDL-C (−23%) (109). Both subcutaneous (300-450 mg every 1 or 2 weeks) and intravenous (15 mg/kg every 4 weeks) administrations of evinacumab have been used in patients for whom the combination of PCSK9 antibodies and the maximum tolerated dose of statins did not achieve LDL-C treatment targets. After 16 weeks, subcutaneous evinacumab resulted in a 45% decrease in LDL-C, whereas intravenous evinacumab resulted in a 49.9% decrease. Evinacumab also had a significant impact in lowering LDL-C in homozygous FH and has received US Food and Drug Administration approval as an add-on treatment for homozygous FH individuals older than age 12 years.

New therapeutic approaches have also been developed to lower Lp(a) with antisense oligonucleotide or siRNAs. For example, the HORIZON trial, currently in phase 3, is evaluating the effect of pelacarsen, an antisense oligonucleotide against apo(a), on clinical events (110). In a phase 2 trial it was demonstrated to cause a dose-dependent reduction up to 80% in Lp(a) (111). In the future, drugs to lower Lp(a) could be a major advance in addressing the residual ASCVD risk that remains after effective LDL-C lowering.

Conclusions

As outlined in Fig. 2, it is useful to carefully investigate the cause of statin hyporesponsiveness for 2 main reasons. First, it may lead to the optimization of lipid-lowering therapy for a patient. Numerous studies have demonstrated that the benefit of LDL-C lowering for reducing ASCVD events is largely proportional to the degree it is lowered. Thus, it is important to identify what type of statin and dose or second lipid-lowering agent is needed for each patient to meet their primary or secondary LDL-C prevention goals. Second, uncovering the cause of suboptimal statin response may reveal a genetic or secondary cause of dyslipidemia that may require a specific therapy and the screening of family members.

As we describe in our case, making sure that a patient is adherent with the prescribed statin therapy and excluding an analytical problem in LDL-C testing will likely be the first steps in the evaluation of a statin-hyporesponsive patient. The exact order for excluding the 3 common diseases that may account for suboptimal statin response is not that important, but it should be done before considering a rare cause of primary statin resistance. At this time, primary statin resistance is probably best considered a presumptive diagnosis. The identification of the exact cause of primary statin resistance may be challenging and may not alter clinical management. In our case, the identification of a patient with high Lp(a) triggered the recommendation for cascade family screening. Additional genetic testing for the exact molecular cause for the high Lp(a) was not conducted in our case because it has not been found to be beneficial in the clinical management of patients (112).

In summary, the evaluation of statin hyporesponsiveness is a multistep process that depends on routine laboratory testing and clinical approaches already followed in the management of ASCVD risk. It can lead to the optimization of lipid-lowering therapy and to the identification of disorders that require specific therapy and/or the genetic screening of family members.

Abbreviations

ABC

ATP-binding cassette

ABCG5

adenosine triphosphate-binding cassette transporters G5

ABCG8

adenosine triphosphate-binding cassette transporters G8

ANGPTL3

angiopoietin-like 3

apoB

apolipoprotein B

APOE

apolipoprotein E

ASCVD

atherosclerotic cardiovascular disease

C

cholesterol

CETP

cholesteryl ester transfer protein

CYP450

cytochrome P450

FH

familial hypercholesterolemia

GalNAc

N-acetylgalactosamine

HDL-C

high-density lipoprotein cholesterol

HMG-CoA

3-hydroxy-3-methylglutaryl coenzyme A

LDL

low-density lipoprotein

LDL-C

low-density lipoprotein cholesterol

LDLR

LDL receptor

LDLRAP1

LDLR adaptor protein 1

Lp(a)

lipoprotein(a)

mAb

monoclonal antibody

PCSK9

proprotein convertase subtilisin/kexin type 9

siRNA

small interfering RNA

TC

total cholesterol

TGs

triglycerides

Funding

This work is supported by intramural DIR funds from the National Heart, Lung, and Blood Institute (to A.W., M.A., and A.T.R.).

Disclosures

The authors have nothing to disclose.

Data Availability

Data sharing is not applicable to this article because no data sets were generated or analyzed during the present study.