Editorial comment: Frequent detection of familial hypercholesterolemia mutations in familial combined hyperlipidemia

Familial combined hyperlipidemia (FCHL) is the most common genetic cause of hyperlipidemia, affecting approximately 1% of the population. It was first described in the Seattle Myocardial Infarction Study in 1973 (1). This condition is characterized by variable lipid phenotypes (increased levels of triglycerides, TG, or of cholesterol, or of both lipids) in the proband and in relatives that may vary within an individual from time to time. FCHL contributes to ≥ 20% of CAD in males under the age of 60 years (2, 3).

The diagnosis of FCHL is difficult, requiring data from family members who may not be available. Percentile cutoffs, often greater than the 90th or 95^th percentile for cholesterol, LDL-C or triglyceride elevations, are utilized to establish the diagnosis, As lipid levels vary over time, both a negative and a positive diagnosis can be unreliable. An increase in apoB (4, 5), together with elevated numbers of small-dense LDL particles appears to be a more consistent phenotype for FCHL than TC and TG, whose levels vary over time (6).

Though originally described as autosomal dominant, more recent data indicate a more complex inheritance. Several loci with major and minor impacts on risk of FCHL, as well as environmental factors, have been demonstrated.

Variants in upstream transcription factor 1 (USF1) were linked in a Finnish FCHL cohort (7) and associated in Dutch (8) and Mexican (9) patients. Paradoxically, a dyslipidemia haplotype was associated with lower CVD risk (10). The USF1 dyslipidemia haplotype has been associated with lower IL6 and CRP levels increasing with age, suggesting that the reduced risk may not involve regulation of lipid genes (11). However, this locus only accounts for some FCHL families, suggesting other factors remain undetected. Other loci implicated include lipoprotein lipase (12) and the apolipoprotein A1/C3/A4/A5 gene cluster (13, 14). Additionally, a large numbers of chromosomal regions have been implicated by linkage in FCHL families. Clearly, FCHL is genetically heterogeneous.

Heterozygous familial hypercholesterolemia (FH) is diagnosed based on high LDL cholesterol levels (usually >250 mg/dl) in untreated adults and frequent tendinous xanthomas. FH is autosomal dominant and associated with early vascular disease. FH can be confirmed by molecular demonstration of a mutant LDL receptor (LDLR) gene. However, molecular studies are rarely done clinically, as the results do not influence treatment. A defect in the ligand for the LDL receptor, apoB, is a less common cause of FH as are missense mutations of PCSK9 (15). FH occurs in about 1 in 500 individuals in the population, and accounts for about 4% of premature myocardial infarctions (1). Hypertriglyceridemia can independently occur in FH causing mixed hyperlipidemia leading to confusion with FCHL.

Against this background and potential uncertainties about the diagnosis of FCHL, Civiera et al searched for classical FH mutations in FCHL using probes for 203 LDL receptor mutations and 4 apoB mutations (16). Their novel microarray techniques (Lipochip) (17) recognized 88% of 230 different LDLR mutations from Spain. Methodology to find rearrangements in the LDLR gene was also utilized (18).143 unrelated middle aged men and women with FCHL were selected from the clinical records of two lipid clinics (16). Inclusion required LDL cholesterol levels of >170 mg/dl or non-HDL cholesterol of >220 mg/dl when triglyceride levels were over 400 mg/dl. At least one first degree relative had to have hyperlipidemia which was defined as total cholesterol and/or triglyceride levels >90th percentile. Functional LDLR mutations were found in 28 (20%) patients. No apoB mutations were detected. Patients with LDLR mutations had higher levels of cholesterol LDL cholesterol and apoB levels as compared with patients without these mutations and their triglyceride levels were somewhat higher in those without LDLR mutation. Diabetes was found in 26 (23%) of patients without LDLR mutations and was absent in all LDLR mutation carriers. Considering only non-diabetes patients with cholesterol levels of >335 mg/dl and apoB levels of more than 185 mg/dl, 48 (or 42% of study subjects) had aLDLR mutations. The detected LDLR mutations were likely to be pathogenic, since most of the LDR abnormalities of this study had been observed in other Spanish FH patients (17).

While it is not surprising that “weaker” mutations in LDLR might result in diagnostic uncertainty in the absence of molecular diagnosis, the high frequency of FH mutations in non-diabetics with a presumptive clinical diagnosis of FCHL is unexpected. While part of this excess can be ascribed to selection or referral bias of more complicated patients to a referral center, and to the atypically high LDL levels in this cohort, these data indicate potential difficulties with the clinical diagnoses of FH and FCHL. Subjects with LDLR mutations lacked tendon xanthomas, which would have suggested FH but frequently is absent. Furthermore, patients with LDLR mutations in this study had lower lipid levels than those typical for FH. The lack of individuals with isolated elevated TG in a family with presumptive FCHL should suggest the possibility of FH rather than FCHL.

Similar studies on FCHL using molecular search for LDLR mutations are strongly suggested. What should be the role of molecular testing for LDLR mutations in FCHL? Currently, LDLR molecular studies are rarely done clinically. This is both due to the absence of an impact on patient management and the ability to test for affected relatives using lipid determination rather than molecular studies which are not generally available. While the presence of a LDLR mutation allows definite diagnosis of FH, studies of lipid levels alone will give informative diagnostic results in affected and non-affected family members in both FH and FCHL.

Should molecular testing be more widely utilized, some consideration must be given to the optimal test. Available options include coverage of the whole gene with current methodology (19) or a microarray including individual known mutations such as used by the current (16) study and the Tejedor et al study (17). The content of a microarray mutation panel for FH detection must of course reflect the various mutations in a given population which varies. Note that at least 800 different FH mutations exist. With falling cost of comprehensive DNA sequencing, an approach covering the entire LDLR gene including promoters is likely to detect all LDLR mutations and will ultimately be possible.

A separate issue is the utility of APOE testing in subjects with a presumptive diagnosis of FCHL. Again, it is useful to consider whether or not knowledge of the results of the diagnostic test will change management of the patient. For APOE there are special considerations because the APOE ε4 allele is a risk factor for late onset Alzheimer’s disease (AD) and is found in approximately 25% of the population in heterozygotes (ε2ε4 and ε3ε4) and in less than 1% of homozygotes (ε4ε4) (20), depending on race. APOE testing therefore requires a tailored approach. In the Civeria study 4 of 115 FCHL patients carried the APOE ε2ε2 genotype (16). These patients have type III hyperlipidemia (dysbetalipoproteinemia or remnant removal disease) which can be diagnosed by examining VLDL composition (21) or with APOE phenotyping by gradient gel electrophoresis (21). Less than 1% of most populations are APOE ε2ε2. In view of the higher risk of AD (2–3 times higher for ε4 heterozygotes and 15 times higher for ε4ε4 homozygotes (22) compared with non-ε4 genotypes), and the fact that ε2 and ε4 are separately testable polymorphisms, many physicians request only APOE ε2 testing or do not inform patients of APOE ε4 status and its significance for AD. Ideally, before any testing that may disclose ε4 status, all patients should be advised about potential risks for AD and offered clinical genetics consultation.