Learning From Patients With Ultrarare Conditions

Cholesterol Hoof Beats

Corresponding Author

• Download figure
• Open in new tab
• Download powerpoint

• atherosclerosis
• genetics
• hyperlipoproteinemia
• low-density lipoprotein
• rare diseases
• registries

“When you hear hoof beats, think of horses not zebras,” Theodore Woodward advised his interns 70 years ago. The practical value of this aphorism is upheld daily. Nonetheless, we have learned much from “zebras,” that is, patients with a rare diagnosis underlying their clinical presentation. Because these conditions are so rare, it can be challenging to round up sufficient numbers to characterize clinical, biochemical, and molecular features, let alone execute adequately powered randomized outcomes trials. Creative approaches are needed to gather patients with rare genetic disorders to yield adequate numbers for testing a scientific hypothesis.

A particularly instructive, rare genetic condition is homozygous familial hypercholesterolemia (HoFH) (1). Although definitions and diagnostic criteria vary, HoFH patients generally have untreated levels of total and low-density lipoprotein (LDL) cholesterol exceeding 520 mg/dl (13 mmol/l) and 390 mg/dl (10 mmol/l), respectively (2). Patients frequently have skin and tendinous xanthomas; some affected children are initially referred to dermatologists. Untreated HoFH patients are markedly predisposed to severe early atherosclerotic cardiovascular disease (ASCVD) (1); coronary deaths in the first decade of life are well-described. Aortic root calcification and valvular stenosis often develop later in life (3).

Although precise numbers are elusive, there are possibly 1,500 to 2,000 HoFH patients in North America. Patients typically carry 2 mutated alleles of the LDLR gene encoding the LDL receptor or less commonly 2 mutated alleles of the APOB or PCSK9 genes encoding, respectively, the receptor ligand apolipoprotein B or the receptor degrading zymogen proprotein convertase subtilisin kexin 9 (PCSK9) (1). Because of these various molecular lesions, HoFH patients cannot clear LDL particles and require methods to directly remove them, such as weekly or biweekly plasmapheresis or LDL apheresis (4). HoFH patients respond poorly to statins, bile acid resins, and ezetimibe, all of which require functional LDL receptors to mediate their pharmacologic effects. In contrast, these agents are effective in obligate heterozygote FH (HeFH) parents of HoFH children, since each heterozygote parent has 1 normal allele that can be up-regulated. Newer therapies such as mipomersen, lomitapide, evolocumab, and evinacumab and gene therapy appear to have improved LDL-lowering efficacy (5,6), and some of these have already been approved for treatment of HoFH.

In contrast to HoFH, HeFH has a worldwide prevalence of ∼1 in 250 people (7), and most cardiologists have likely encountered such patients. HeFH patients have LDL cholesterol concentrations of >190 mg/dl (>5.0 mmol/l); some have tendinous xanthomas, premature corneal arcus, and xanthelasmas (8). ASCVD predisposition in HeFH is very strong (8) but not nearly as severe as in HoFH patients who carry a “double dose” of the defective gene (1). The rarity of HoFH follows mathematically from the: 1) small chance that 2 HeFH parents will randomly mate; and 2) laws of genetics, whereby one-fourth of their children inherit both defective alleles, leading to clinical HoFH (1). Sometimes, HeFH is diagnosed retrospectively in parents once an affected child has a diagnosis of HoFH (1). Technically, HoFH is usually an autosomal codominant condition, meaning that HeFH parents carrying 1 broken allele express an abnormal phenotype, which is additive to create a very severe phenotype in the affected child with biallelic mutations (1).

Clinical heterogeneity among HoFH patients has long been appreciated. For instance, some Sardinian children with a phenotype resembling HoFH had parents and close relatives with normal lipid levels (9). This suggested a true “autosomal recessive” inheritance pattern, which by definition stipulates normal phenotypes in heterozygous carriers. Genetic mapping studies coordinated by Helen Hobbs and Jonathan Cohen identified a new causative gene in these families, namely LDLRAP1, which encodes an adaptor protein that chaperones the LDL receptor through its intracellular itinerary (10).

The discovery that LDLRAP1 mutations caused autosomal recessive hypercholesterolemia (ARH) was a major scientific advance, revealing new aspects of LDL receptor biology (10). Concurrently, questions arose about whether the underlying molecular heterogeneity of HoFH was associated with clinical heterogeneity. ARH was like a golden zebra among black-and-white zebras; but were there any clinical implications of having a different underlying molecular cause? Because 1 mutant LDLRAP1 allele had no obvious clinical consequences, would 2 mutant alleles result in less severe clinical features than in HoFH individuals with 2 mutant LDLR alleles? Indeed, earlier case reports and small anecdotal series of ARH patients suggested that ARH was less severe clinically than classical HoFH (11).

Our knowledge is now advanced by an observational study reported in this issue of the Journal in which D'Erasmo et al. (12) searched published articles and contacted lipid clinics performing apheresis to collect a total of 52 patients (28 females) with molecularly proven ARH from 6 countries, the largest cohort with this condition ever assembled. Despite the heterogeneity of health records and management protocols, the authors nonetheless harmonized baseline demographic and clinical data to obtain fairly reliable information for longitudinal treatments, LDL cholesterol response, and incident atherosclerotic cardiovascular disease (CVD) and aortic events, deriving Kaplan-Meier plots for these important outcomes (12).

Most ARH patients received high-intensity statin therapy, with or without ezetimibe, with or without LDL apheresis. Six patients took lomitapide, and none took a PCSK9 inhibitor. Over a mean follow-up of 14.1 years, the mean nadir LDL cholesterol concentration was reduced by 67% and 88% in the groups not taking and taking lomitapide, respectively. Twenty-three percent of ARH patients achieved LDL-C of <100 mg/dl (<2.6 mmol/l); 26.9% had incident new ASCVD diagnoses (1.9% per year), and 11.5% had developed newly diagnosed aortic valve stenosis (0.9% per year). Using the general Italian population as the reference standard, the risk of disease endpoints was markedly increased in both sexes. Perhaps most importantly, these percentages were virtually identical to treatment efficacy and rates of ASCVD and aortic disease progression reported in observational studies of patients with typical HoFH (13,14). The authors concluded that ARH is not a “less severe” form of HoFH but, instead, that ARH patients express important disease readouts with severity that appears indistinguishable from typical HoFH.

The report by D'Erasmo et al. (12) provides a great example of how careful longitudinal observational studies can yield important clinical lessons for ultrarare diseases. The relatively high number of ARH subjects, obtained by methodically contacting clinics, yielded new findings that have changed prior thinking about this condition. Because disease progression appears identical to typical HoFH, the same types and intensities of therapies need to be considered. Whether ARH patients show a distinctive response to new medications for severe hypercholesterolemia needs to be evaluated (5). Going forward, we need to streamline the collection of such information for other rare disorders (i.e., to make the herding of zebras more efficient), perhaps through global linkages among specialty clinics, disease-specific registries, and other clinical databases (15,16).

• 2018 American College of Cardiology Foundation