Familial hypercholesterolemia (FH) is an autosomal disorder characterized by increased levels of total cholesterol and low density lipoprotein cholesterol. The FH clinical phenotype has been shown to be associated with increased coronary heart disease and premature death. Mutations in the low density lipoprotein receptor gene (LDLR) can result in the FH phenotype, and there is evidence that receptor-negative mutations result in a more severe phenotype than do receptor-defective mutations. Mutations in the apolipoprotein B-100 gene (APOB) can result in a phenotype that is clinically indistinguishable from familial hypercholesterolemia, and mutations in this gene have also been shown to be associated with coronary heart disease. Preliminary research indicates that the FH phenotype is influenced by other genetic and environmental factors; however, it is not clear if these are synergistic interactions or simply additive effects.

The genetic causes of heterozygous familial hypercholesterolemia (FH) have been a subject of study since the early 1900s (1), and they have already been reviewed for the Human Genome Epidemiology Network (2). Briefly, the frequency of FH is reported as 1/500 for Caucasian populations (3). FH is characterized by autosomal inheritance of increased total cholesterol and low density lipoprotein (LDL) cholesterol, primarily attributable to mutations in the low density lipoprotein receptor gene (LDLR)
(4,5). LDLR is located on chromosome 19 at 19p13.1-p13.3 (6), and over 700 mutations have been identified in this gene (7,8). Mutations in two other genes also cause the clinical FH phenotype. One of these is the apolipoprotein B-100 gene (APOB), located on chromosome 2p23-24 (9,10), that codes for the protein component of LDL particles (11). In contrast to LDLR, only a small number of functional mutations have been identified in APOB. The third gene, proprotein convertase subtilisin/kexin type 9 (PCSK9), was recently identified on chromosome 1p32 (12). To date, no epidemiologic research has investigated mutations in PCSK9.

LDLR mutations can be classified according to the effect they have on LDL receptor protein function
(13). The LDL receptor protein is a cell surface receptor that removes LDL particles from the plasma by way of receptor-mediated endocytosis. In class 1 mutations, the LDL receptor protein is not synthesized; in class 2 mutations, the LDL receptor is not transported to the Golgi; in class 3 mutations, the LDL receptor does not properly bind with the LDL particles; in class 4 mutations, bound surface receptors are not internalized; and in class 5 mutations, the internalized LDL particles are not released in the endosome. The majority of mutations identified to date are class 2 or class 3 mutations occurring in the ligand-binding and epithelial growth factor precursor regions of the gene (14). Class 1 mutations are alternatively referred to as "null" or "receptor-negative" mutations in the literature, whereas mutations from classes 2-5 are termed "receptor defective." For this review, we will use the terms "receptor negative" and "receptor defective."

Coronary heart disease and its clinical manifestation of myocardial infarction are widely recognized to be a multifactorial disorder, with contributions from both environmental and genetic factors. The development of hypertension and diabetes, both of which also have environmental and genetic components, is strongly associated with increased coronary heart disease risk (15). Increasing age and male gender are strongly associated with coronary heart disease risk, with men typically developing clinically important disease 10-15 years before women, who in general are protected to a degree until after menopause (16). Of the environmental factors, smoking is the major contributor and is associated with a roughly twofold higher lifetime risk (17). Lack of exercise and the associated adiposity, as well as a high intake of saturated fats and a low intake of certain vitamins, are also associated with increased risk (16). The mechanism of action of these factors is thought at least in part to be through determining differences in the plasma levels of lipids and lipoproteins that are atherogenic. High levels of LDL cholesterol and low levels of high density lipoprotein cholesterol have consistently been shown to be associated with coronary heart disease risk (18). Evidence of the strong genetic component for coronary heart disease risk is supported by the consistent association between a reported family history of early coronary heart disease and a personal increased risk (19), with risk associated with family history being in the order of 1.7-fold higher, even after adjusting for other classical risk factors (20). Although such analyses do not distinguish between familial aggregation of environmental or lifestyle risk factors and inherited factors, these data strongly support the role of inherited factors in the mechanisms of coronary heart disease.

The specific genes involved in these processes and their variants in the general population are the subject of much research but are beyond the scope of this review. With an estimated heterozygous frequency of 1/500 (3), FH accounts for only a small fraction of the familial cases of FH. Another gene that has been well examined in the context of coronary heart disease is the gene coding for the apolipoprotein E (APOE). This association has previously been reviewed for the Human Genome Epidemiology Network (21). There are three common variants of this gene called E3, E2, and E4 , with the E4 allele being associated with higher and the E2 allele being associated with lower levels of plasma apolipoprotein B-containing proteins such as LDL (22). As would be expected from the known risk associated with LDL cholesterol levels, carriers of the E4 allele tend to have higher and carriers of the E2 allele tend to have lower coronary heart disease risk (23), such that these common variants explain from 2 percent to 3 percent of the population variance in coronary heart disease risk.

The first stage of the development of the atherosclerotic lesion is thought to be dysfunction of the vessel wall endothelium, which in healthy vessels maintains vascular tone and blood pressure. Endothelial dysfunction can be detected in the peripheral vessels of subjects with high coronary heart disease risk factors (such as FH) as early as in the first decade of life (24). Animal models suggest that endothelial dysfunction is caused by a wide range of insults, including inflammatory processes (25) as a result of infectious agents, smoking, or elevated levels of lipoproteins such as LDL. When LDL enters the vessel wall through a dysfunctional endothelial barrier, the LDL particles are oxidized and recruit monocytes from the blood. These cells take up the LDL and may then exit the site of the lesion, allowing the damage to be limited and healed. However, in subjects with high plasma levels of LDL, this process is overwhelmed, and the monocytes, differentiated into macrophages, become lipid laden and "foamy" in appearance under the microscope (26). These macrophage-foam cells are the hallmark of the developing atherosclerotic lesion. In later stages, the burden of toxic lipids results in cellular death and the deposition of cholesterol as crystals in the expanding atherosclerotic plaque (27).

Although the plaque itself may occupy an increasing proportion of the lumen and thus restrict blood flow, this is not associated with clinical symptoms until stenosis approaches 70 percent or greater
(27). At this stage, ischemia may develop, especially upon exercise, and is seen as the chest pains of angina. The clinically more serious event of a myocardial infarction occurs if the plaque ruptures. The resulting thrombus may completely occlude the already narrowed vessel and, downstream, ischemia may cause permanent damage to the myocardial tissue. If the affected area of the heart is extensive or localized in a critical region, the result may be fatal.

Rupture occurs due to the degradation of the vessel wall matrix by metalloproteinases (28). Much research interest is currently focused on the cellular and tissue control of expression of these enzymes and their natural inhibitors, as well as on the role of common genetic variants and environmental mediators, such as inflammation and smoking. However, one of the major determinants of both the initiation of vessel damage and the rate of development of the atherosclerotic lesion seems to be plasma levels of key lipoprotein particles, including LDL levels. Further research on FH and on the impact of treatment in FH patients will continue to enhance our understanding of the relations between LDL levels and coronary heart disease.

We identified studies of FH and coronary heart disease through two methods. First, to identify classic papers in the early literature, we performed hand searches of papers in our collections and the reference lists of extensive review articles (3,29). Second, we searched MEDLINE and PubMed using combinations of the terms "familial hypercholesterolemia," " LDLR ," " APOB," "apolipoprotein B," "coronary heart disease," and "cardiovascular disease," in addition to "genetics" and "epidemiology." Studies that examined noncoronary vascular disease outcomes, such as stroke (30), peripheral vascular disease (31), and intima-medial wall thickness (32), are reviewed separately for the Human Genome Epidemiology Network (33).

Studies detailing the magnitude of the association between clinical FH and coronary heart disease are listed by geographic location in Table 1 and Web table 1. (This information is described in the first of three supplementary tables; each is referred to as "Web table" in the text and is posted on the website of the Human Genome Epidemiology Network ( http://www.cdc.gov/genomics/hugenet/reviews.htm ) as well as on the Journal's website.) Table 1 is limited to studies that provide a risk estimate for the association, whereas Web table 1 lists all studies with a control group and includes additional information on the prevalence and cumulative probability of coronary heart disease in FH and control populations. The following discussion of the association studies is organized by study design to highlight the primary findings and to identify methodological limitations.

click to view

Table 1  Association studies of clinical familial hypercholesterolemia and coronary  heart disease by geographic location

Studies estimating cumulative probability of coronary heart disease
In one of the earliest association studies, Slack (34) compared 104 patients who had type II hyperbetalipoproteinemia (clinical FH) with 41 patients who had type III, type IV, or type V hyperlipoproteinemia (hypertriglyceridemia) in the United Kingdom. Medical records and resting electrocardiogram results were obtained, and index patients were followed for 1-10 years. The cumulative probability of a first attack of ischemic heart disease by age 60 was higher for patients with clinical FH (85.4 percent for males, 57.5 percent for females) than for patients with type III, type IV, or type V hyperlipoproteinemia (53.3 percent for males, 25 percent for females). Soon afterward in the United States, Stone et al. (35) examined 1,023 first- and second-degree relatives of 116 FH index patients diagnosed at the National Institutes of Health from 1964 to 1970. The risk of fatal or nonfatal coronary heart disease by age 60 years was 52 percent for male and 31.8 percent for female relatives with FH compared with 12.7 percent and 9.1 percent for relatives without FH. Additional studies determined a similarly high risk of premature coronary heart disease among patients with clinical FH in Japan (cumulative probability of myocardial infarction by age 60 years: ~35 percent for males, ~20 percent for females) (36), in Quebec (mean age of onset of ischemic heart disease: ~40 years for males, ~50 years for females) (37), in France (mean age of onset of ischemic heart disease: 44.2 years for males, 53.1 years for females) (38), and in Norway (cumulative probability of coronary heart disease symptoms by age 60 years: 83 percent for males, 70 percent for females) (39). Because the latter four studies did not include a control group, the magnitude of association could not be estimated, and they are not included in Table 1 or Web table 1.

Although recent studies have also found an association between clinical heterozygous FH and coronary heart disease (40,41), the early studies described above were performed before the widespread use of statins for treatment of hypercholesterolemia and, thus, give a more accurate reflection of the natural history of the disease. However, the early studies have two primary limitations. First, the life-table analyses for some studies included deceased relatives (35,36). Second, index patients were selected from hospitals (34,35,38) or lipid clinics (36,37), and some had a prior history of coronary heart disease or xanthomatosis. As a result, these study subjects may have had a more severe form of FH or may have other genetic/environmental predisposing factors for coronary heart disease compared with a population-based sample. Both of these biases could spuriously inflate the observed association.

Taken together, however, the results of these studies uniformly demonstrate an increased burden of premature coronary heart disease and death associated with the presence of FH. In general, the onset of disease appears to be delayed approximately 10 years for women compared with men (37,38) and is lower in Japan (36) than in Western countries.

Cohort studies of standardized mortality ratios for coronary heart disease and all-cause mortality
Jensen et al. (42) prospectively followed 331 individuals (181 with FH, 150 normocholesterolemic) in 11 Danish families from 1944 to 1964. Again, a high prevalence of early coronary heart disease was seen in FH patients (45.1 percent for males by age 50 years). An increased number of deaths for FH-affected relatives was observed compared with the general Danish population. An indirect standardization was performed. The national death rate for each 10-year age bracket, by sex, for 1943-1964 was reported by the Copenhagen statistics department, and these rates were used to calculate the expected number of deaths for the study population's size in each age and sex bracket. The resulting standardized mortality ratio of observed to expected deaths was elevated for both sexes (males: standardized mortality ratio (SMR) = 2.88, 95 percent confidence interval (CI): 1.73, 4.46; females: SMR = 1.71, 95 percent CI: 0.912, 2.93). This increase was not observed for the unaffected relatives (SMR = 1.03, 95 percent CI: 0.562, 2.01). A Japanese study (43) examined 527 heterozygotes over 10 years and observed 41 deaths. Thirty patients died from coronary heart disease, a number 10.9 times higher than the proportional mortality of cardiovascular deaths in the general Japanese population. In addition, the mean age of death from a cardiac event was significantly younger for males (54 years) than females (68 years).

The Simon Broome Familial Hypercholesterolemia Register Group has been recruiting patients from lipid clinics in the United Kingdom since 1980. The first publication of findings from this large prospective cohort study presented data on 526 patients for a total of 2,234 person-years from 1980 to 1989 (44). The second publication expanded the size of the cohort to a total 1,185 patients followed for 8,770 person-years from 1980 to 1995 (45). The observed number of deaths was compared with the number expected on the bases of age, sex, and calendar period death rates for the general population of England and Wales . Again, indirect standardization showed an increase in the number of deaths due to coronary heart disease for both males (SMR = 2.6, 95 percent CI: 1.7, 3.8) and females (SMR = 3.7, 95 percent CI: 2.3, 5.8) (45). The large size of this study allowed determination of age-specific mortality rates. The absolute risk of coronary heart disease increased with age (45) and, notably, a large relative risk of fatal coronary heart disease was seen in young adults. For the 1980-1995 follow-up, an over 100-fold increase in risk for females (SMR = 125.00, 95 percent CI: 15.1, 451.3) and almost 50-fold increase for males (SMR = 48.4, 95 percent CI: 17.8, 105.5) were reported (45). In contrast, the relative risk in the same time period for persons aged 60-75 years was only 2.6 (95 percent CI: 1.3, 4.5) for females and 1.1 (95 percent CI: 0.5, 2.3) for males (45). This demonstrates a decrease in the relative risk of fatal coronary heart disease with increasing age.

Further, the time frame of the Simon Broome study overlaps with the introduction of statins in the early 1990s, allowing a comparison of standard mortality rates before and after the widespread use of these medications. The relative risk of coronary mortality in patients aged 20-59 years was higher from 1980 to 1991 (SMR = 8.0, 95 percent CI: 4.8, 12.6) than from 1992 to 1995 (SMR = 3.7, 95 percent CI: 1.6, 7.2) (45), suggesting that treatment is effective in lowering the risk of death from coronary disease in patients with clinical FH.

Family studies comparing all-cause standardized mortality ratios
Two studies by Sijbrands et al. (46,47) examined all-cause mortality standardized mortality ratios for relatives of FH individuals, indirectly standardized to the age-, sex-, and calendar period-specific mortality rates of the general Dutch population. The pool of relatives studied combined approximately 50 percent affected and 50 percent unaffected individuals, so the reported standardized mortality ratio is an underestimate of the effect of the FH mutations. The first study traced 855 first-degree relatives of 113 unrelated patients (47). The authors observed an increased risk for all-cause mortality (SMR = 1.34, 95 percent CI: 1.16, 1.55). Similar to the Simon Broome Register, no excess mortality was found in older FH individuals (patients aged 80-103 years: SMR = 0.96, 95 percent CI: 0.60, 1.46). The design of the study allowed a comparison of the all-cause mortality standardized mortality ratio among families where FH was in part ascertained by premature onset of coronary heart disease and among families without coronary heart disease. An increased relative risk of death was observed in the premature coronary heart disease families (SMR = 1.46, 95 percent CI: 1.09, 1.94). This finding motivated the second study (46) in which Dutch records were used to trace the ancestry of three selected probands with the same mutation. A common ancestor pair living in 1830 was identified, and all living descendants of that pair were screened for the V408M mutation. A total of 412 individuals were found over eight generations of the transmission lines of the mutation, 250 of whom lived for at least 20 years. The overall relative risk for the 250 individuals (50 percent affected) was 1.32 (95 percent CI: 1.03, 1.67). Since these 250 identified individuals were not selected on the basis of clinical manifestations of FH, they better represent the natural course of FH, free from selection from cardiovascular disease. The level of excess mortality varied over time. There was no excess mortality in the 19th and early 20th centuries. The standardized mortality ratio then reached a peak between 1935 and 1964 (SMR = 1.78, 95 percent CI: 1.13, 2.76) and declined in the latter half of the 20th century. Taken together, these family studies show that selecting patients based on only clinical manifestations may overestimate mortality risk and that strong environmental factors may influence the increased mortality in patients with FH.

These results also complement earlier family studies of FH. A series of analyses based on survivors of myocardial infarction and their families in the United States in the 1970s by Goldstein et al. (48) and Hazzard et al. (49) found a high prevalence (3 percent) of heterozygous FH among survivors of myocardial infarction. Similar estimates (5 percent) were provided from the United Kingdom in 1972 by Patterson and Slack (50). In 1986, Williams et al. (51) screened 77 heterozygous FH members of four Utah pedigrees. They found that all 26 males born in the last two generations surveyed (after 1900) had coronary disease. In contrast, only one of five males born in the 19th century had coronary heart disease before the age of 60 years, providing additional evidence that environmental factors influence the association between clinical FH and coronary heart disease.

Allele-specific associations
Web table 2 and Web table 3 summarize studies that have examined the association between phenotypic outcomes and specific mutations in LDLR and APOB, respectively. The studies of LDLR mutations and coronary heart disease are among subjects with FH and/or their relatives, while the APOB studies are population-based case-control studies. It is worth noting that these studies differ in terms of their definition and methods of ascertainment of coronary heart disease. It may be the case that some mutations are associated with specific clinical manifestations of coronary heart disease, and variation in the coronary heart disease severity across studies limits the generalizability of the results.

Because of the large number of allelic variants, LDLR mutations are classified into two groups: 1) receptor-negative alleles and 2) receptor-defective alleles. However, there can be variation within these groups. For example, mutations in repeat 5 of the binding domain, which are coded by exon 4, have been shown to be associated with a more severe phenotype than other receptor-defective mutations (52). When possible, we have included comparisons of lipid levels with control (53,54), unaffected relative (41), or normolipidemic (55) subjects in Web table 2, because LDL cholesterol levels and coronary heart disease prevalence in the general population vary by geographic location.

The effects of specific mutations can most easily be compared within founder populations in which a small number of alleles are responsible for the clinical FH phenotype. For example, a study in the Afrikaner populations shows a more severe phenotype, in terms of both lipid levels and coronary heart disease outcomes, for V408M , a receptor-negative mutation, than for D206E , a receptor-defective mutation (56). Similarly, studies of French Canadians are able to compare FH subjects who have primarily a 15-kilobase, receptor-negative deletion with those who have primarily the W66G receptor-defective mutation (54,55). However, because of their common origin, the FH heterozygotes in these populations may share other genetic or environmental factors, and these potential confounding factors may spuriously inflate the observed association (52).

Observational studies of small numbers of families or individuals have also noted additional LDLR alleles with atypically mild (57-62) or atypically severe (63,64) phenotypes; however, the results of these studies may not be applicable to the general population.

To have large enough samples for phenotypic comparisons, studies in nonfounder populations have grouped LDLR mutations. Typically, mutations are grouped by either class, as was done in England and Wales (52), Italy (65), Norway (66), and Spain (67), or mutation type, as was done in Northern Ireland (68). As in the founder populations, these studies typically find LDL cholesterol levels and risk of cardiac death to be higher in individuals with receptor-negative alleles than in those with receptor-defective alleles (Web table 2). These results are supported by additional studies that have reported increased cholesterol levels for null alleles, but they did not ascertain coronary heart disease risk (69, 70). It has been noted (47) that the patient populations for many of these studies were selected from lipid clinics and, therefore, may have other environmental or familial factors that predisposed them to coronary heart disease. To address this issue, the Dutch national screening program for FH identified first-degree relatives of FH patients who were carriers of LDLR mutations, many of whom did not have clinical signs of FH. This program further identified relatives of these carriers who also had LDLR mutations. The index cases were then excluded from analysis to minimize the number of FH subjects ascertained through clinical manifestations of coronary heart disease (41). As in the other studies, receptor-negative alleles were found to be associated with a more severe phenotype than were receptor-defective alleles. This difference was due to primarily a particularly mild phenotype in the receptor-defective allele N534H/2393del9 (41).

Case-control studies in Denmark (71) and the United States (72) have shown the APOB R3500Q allele to be associated with coronary heart disease, while a study in France (73) had inconclusive results (Web table 3). In the latter study, two of 622 cases and one of 639 controls were found to carry the R3500Q mutation; however, further analysis showed the control to have a history of coronary heart disease also. There is no evidence for an association between the R3531C allele and either coronary heart disease (71,72) or hyperlipidemia (71,72,74). Association studies of the R3500W allele have not been practical because of the low frequency of the mutation in FH heterozygotes in Western populations (75-77). The R3500Q mutation results, on average, in a phenotype that is slightly more mild than that caused by mutations in LDLR (78-80). The clinical phenotype associated with APOB mutations is termed "familial defective apolipoprotein B-100" or "FDB." Recent work indicates that LDL cholesterol plasma levels may be lower in subjects with APOB mutations than in subjects with LDLR mutations, because of decreased intermediate-density lipoprotein to LDL transfer (81). However, there is a large overlap in cholesterol distributions for individuals with LDLR mutations compared with individuals with APOB mutations. As a result, familial defective apolipoprotein B-100 is generally considered to be clinically indistinguishable from FH (82).

The phenotypic expression of heterozygous FH is quite variable, and at least part of this variation is due to the underlying molecular heterogeneity of the disease. Some studies demonstrate that age of onset of coronary heart disease clusters within families (39); however, phenotypic variation is still observed in families or populations sharing the same LDLR or APOB mutation (56,83,84), indicating that the clinical FH phenotype is influenced by additional environmental and/or genetic risk factors as well.

Gene-environment interactions
As noted above, multigenerational family studies demonstrate that the association of heterozygous FH with excess cardiovascular mortality varies over time. Specifically, studies by Sijbrands et al. (46) and Williams et al. (51) both noted a later onset of coronary heart disease mortality for FH heterozygotes in the 1900s compared with their 20th century descendants. Both studies propose that this mortality change is most likely due to changes in the environment, specifically an increase in dietary fat and sedentary lifestyle. Smaller studies have also demonstrated intrafamilial variability (53,84) among first-degree relatives sharing the same LDLR mutations.

This interaction with environmental factors is also illustrated by comparing the phenotypic expression of heterozygous FH geographically (85). For example, total and LDL cholesterol levels in FH heterozygotes of similar genetic background vary in different parts of the world (86-88), even after controlling for differences in the underlying mutation. Pimstone et al. (89) matched Chinese FH subjects in Canada to FH heterozygotes in China with similar LDLR mutations. The subjects residing in Canada had higher concentrations of LDL cholesterol and an increased prevalence of tendinous xanthomata and coronary heart disease. Pereira et al. (90) examined FH heterozygotes in three Cuban families of Spanish descent in which one third of family members carried the LDLR V408M mutation common in the Afrikaner population. Although all the subjects had elevated LDL cholesterol, cardiovascular complications were rarely observed in the Cuban subjects compared with Afrikaners.

Several standard coronary risk factors have been shown to be associated with increased coronary risk in FH heterozygotes (69,91-93). As in non-FH patients, sex and age are strong predictors of risk (3,92) as are obesity (94,95), diabetes (92,96,97), lipid levels (91,98,99), and smoking (91,92,96). However, these studies have examined only FH individuals and have not included non-FH individuals as controls; therefore, it is not clear that there is a gene-environment interaction for any of these risk factors. Instead, they may just be additive effects, with the increase in risk for individuals with FH being equivalent to the increased risk observed in the general population.

The association between clinical FH and coronary heart disease is well established, and there is evidence that receptor-negative mutations result in a more severe phenotype than do receptor-defective mutations. Further research in this field should focus on clarifying the genotype-phenotype relation and on understanding the impact of statins and other forms of treatment on reducing cardiovascular disease risk in individuals with LDLR or APOB mutations.

This project was supported under a cooperative agreement from the Centers for Disease Control and Prevention through the Association of Schools of Public Health (grants U36/CCU300430-20 and U36/CCU300430-22). This research was also supported by National Institutes of Health grant HL-49513 from the National Heart, Lung, and Blood Institute. S. E. H. is supported by the British Heart Foundation and in part by a grant from the Department of Health and Departments of Trade and Industry to the London Genetics Knowledge Park . R. L. Z. is supported by funding from the Departments of Health and of Trade and Industry, through the Cambridge Genetics Knowledge Park , and by the Wellcome Trust.

The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the Centers for Disease Control and Prevention or the Association of Schools of Public Health.

Correspondence to Dr. Melissa A. Austin, F 363 Health Sciences Building, Box 357236, School of Public Health and Community Medicine, University of Washington, 1959 NE Pacific Avenue, Seattle, WA 98195 (e-mail: maustin@u.washington.edu).

Table

• Table 1 - Association studies of clinical familial hypercholesterolemia and coronary  heart disease by geographic location

Web Tables

• Web Table 1 - Association studies of clinical familial hypercholesterolemia (FH) and coronary heart disease (CHD) by geographic location
• Web Table 2 - Association studies of specific low-density lipoprotein receptor gene (LDLR) alleles and coronary heart disease (CHD) by geographic location
• Web Table 3 - Association studies of specific apolipoprotein B (APOB) alleles and coronary heart disease (CHD) by geographic location

List of References

• MedPed Coordinating Center at the University of Utah (last accessed 2/2008): More about familia hypercholesterolemia
• Center for Genetic Education: Familial Hypercholesterolaemia and Cardiovascular Disease
• Mortality from Familial Hypercholesterolaemia (last accessed 2/2008)
  Online presentation by Eric Sijbrands, Erasmus Univeristy Medical Center , Rotterdam , Denmark
• GeneCards: LDLR (last accessed 2/2008)
• GeneCards: APOB (last accessed 2/2008)