Author + information
- Received March 2, 2016
- Revision received March 22, 2016
- Accepted March 22, 2016
- Published online June 7, 2016.
- Amit V. Khera, MDa,b,
- Hong-Hee Won, PhDc,
- Gina M. Peloso, PhDb,d,
- Kim S. Lawson, MSe,
- Traci M. Bartz, MSf,
- Xuan Deng, MScd,
- Elisabeth M. van Leeuweng,
- Pradeep Natarajan, MD, MMSca,b,
- Connor A. Emdin, HBScb,
- Alexander G. Bick, PhDb,
- Alanna C. Morrison, PhDe,
- Jennifer A. Brody, BAh,
- Namrata Gupta, PhDb,
- Akihiro Nomura, MDb,i,
- Thorsten Kessler, MDj,
- Stefano Duga, PhDk,
- Joshua C. Bis, PhDh,
- Cornelia M. van Duijn, PhDg,
- L. Adrienne Cupples, PhDd,
- Bruce Psaty, MD, PhDh,l,
- Daniel J. Rader, MDm,
- John Danesh, DPhiln,
- Heribert Schunkert, MDj,
- Ruth McPherson, MDo,
- Martin Farrall, MDp,
- Hugh Watkins, MD, PhDp,
- Eric Lander, PhDb,
- James G. Wilson, MDq,
- Adolfo Correa, MD, PhDr,
- Eric Boerwinkle, PhDe,
- Piera Angelica Merlini, MDs,
- Diego Ardissino, MDt,
- Danish Saleheen, MBBS, PhDu,
- Stacey Gabriel, PhDb and
- Sekar Kathiresan, MDa,b,∗ ()
- aCenter for Human Genetic Research, Cardiovascular Research Center and Cardiology Division, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts
- bProgram in Medical and Population Genetics, Broad Institute, Cambridge, Massachusetts
- cSamsung Advanced Institute for Health Sciences and Technology, Sungkyunkwan University, Samsung Medical Center, Seoul, Republic of Korea
- dDepartment of Biostatistics, Boston University School of Public Health, Boston, Massachusetts
- eHuman Genetics Center and Institute of Molecular Medicine, University of Texas-Houston Health Science Center, Houston, Texas
- fDepartment of Biostatistics, University of Washington, Seattle, Washington
- gDepartment of Epidemiology, Erasmus Medical Center, Rotterdam, the Netherlands
- hCardiovascular Health Research Unit, University of Washington, Seattle, Washington
- iDivision of Cardiovascular Medicine, Kanazawa University Graduate School of Medical Science, Kanazawa, Japan
- jDeutsches Herzzentrum München, Technische Universität München, Deutsches Zentrum für Herz-Kreislauf-Forschung, München, Germany, and Munich Heart Alliance, München, Germany
- kDepartment of Biomedical Sciences, Humanitas University, Rozzano, Milan, Italy, and Humanitas Clinical and Research Center, Rozzano, Milan, Italy
- lDepartments of Medicine, Epidemiology, and Health Services, University of Washington, Seattle, Washington
- mDepartments of Genetics, University of Pennsylvania, Philadelphia, Pennsylvania
- nPublic Health and Primary Care, University of Cambridge, Cambridge, Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge and National Institute for Health Research Blood and Transplant Research Unit in Donor Health and Genomics, Department of Public Health and Primary Care, University of Cambridge, Cambridge, United Kingdom
- oUniversity of Ottawa Heart Institute, Ottawa, Canada
- pDivision of Cardiovascular Medicine, Radcliffe Department of Medicine and the Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, United Kingdom
- qDepartment of Physiology and Biophysics, University of Mississippi Medical Center, Jackson, Mississippi
- rJackson Heart Study, Department of Medicine, University of Mississippi Medical Center, Jackson, Mississippi
- sOspedale Niguarda, Milano, Italy
- tDivision of Cardiology, Azienda Ospedaliero-Universitaria di Parma, University of Parma, Parma, Italy, and ASTC: Associazione per lo Studio Della Trombosi in Cardiologia, Pavia, Italy
- uBiostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania
- ↵∗Reprint requests and correspondence:
Dr. Sekar Kathiresan, Cardiovascular Research Center & Center for Human Genetics, Massachusetts General Hospital, 185 Cambridge Street, CPZN 5.252, Boston, Massachusetts 02114.
Background Approximately 7% of American adults have severe hypercholesterolemia (untreated low-density lipoprotein [LDL] cholesterol ≥190 mg/dl), which may be due to familial hypercholesterolemia (FH). Lifelong LDL cholesterol elevations in FH mutation carriers may confer coronary artery disease (CAD) risk beyond that captured by a single LDL cholesterol measurement.
Objectives This study assessed the prevalence of an FH mutation among those with severe hypercholesterolemia and determined whether CAD risk varies according to mutation status beyond the observed LDL cholesterol level.
Methods Three genes causative for FH (LDLR, APOB, and PCSK9) were sequenced in 26,025 participants from 7 case-control studies (5,540 CAD case subjects, 8,577 CAD-free control subjects) and 5 prospective cohort studies (11,908 participants). FH mutations included loss-of-function variants in LDLR, missense mutations in LDLR predicted to be damaging, and variants linked to FH in ClinVar, a clinical genetics database.
Results Among 20,485 CAD-free control and prospective cohort participants, 1,386 (6.7%) had LDL cholesterol ≥190 mg/dl; of these, only 24 (1.7%) carried an FH mutation. Within any stratum of observed LDL cholesterol, risk of CAD was higher among FH mutation carriers than noncarriers. Compared with a reference group with LDL cholesterol <130 mg/dl and no mutation, participants with LDL cholesterol ≥190 mg/dl and no FH mutation had a 6-fold higher risk for CAD (odds ratio: 6.0; 95% confidence interval: 5.2 to 6.9), whereas those with both LDL cholesterol ≥190 mg/dl and an FH mutation demonstrated a 22-fold increased risk (odds ratio: 22.3; 95% confidence interval: 10.7 to 53.2). In an analysis of participants with serial lipid measurements over many years, FH mutation carriers had higher cumulative exposure to LDL cholesterol than noncarriers.
Conclusions Among participants with LDL cholesterol ≥190 mg/dl, gene sequencing identified an FH mutation in <2%. However, for any observed LDL cholesterol, FH mutation carriers had substantially increased risk for CAD.
Severe hypercholesterolemia, defined as having a low-density lipoprotein (LDL) cholesterol level ≥190 mg/dl, is a treatable risk factor for coronary artery disease (CAD) (1,2). Current treatment guidelines place particular emphasis on intensive life-style and pharmacological therapy in this population (3). One cause of severely elevated LDL cholesterol is familial hypercholesterolemia (FH), an autosomal dominant monogenic disorder linked to impaired hepatic clearance of LDL cholesterol particles (4). Patients with LDL cholesterol ≥190 mg/dl are often assumed to have FH, but this may not be the case. Large-scale gene sequencing provides an opportunity to clarify the diagnostic yield and clinical impact of identifying an FH mutation in severely hypercholesterolemic patients.
Previous studies of the diagnostic yield of genetic testing in severe hypercholesterolemia have focused on subjects with clinically suspected FH and reported FH mutation prevalence has ranged from 20% to 80% (5–16). This variability is likely caused by differing ascertainment schemes utilizing family history, physical examination features, elevated LDL cholesterol at a young age, or referral to specialized clinics, each of which may enrich for monogenic causes. In contrast, if ascertainment from the general population is solely on the basis of elevated LDL cholesterol, the extent to which FH mutations contribute to severe hypercholesterolemia is unknown. Such knowledge may inform the design and effectiveness of universal FH screening proposals (17,18).
Knowledge of FH mutation status could also provide CAD risk information beyond that from a single LDL cholesterol measurement (4,18). An FH mutation could lead to higher cumulative exposure to LDL cholesterol levels over a lifetime; as such, within any stratum of LDL cholesterol, the risk of CAD might be greater if the LDL elevation is due to a monogenic mutation versus other causes. The extent to which the presence of a causal FH mutation modulates CAD risk is uncertain.
We analyzed gene sequences of 3 FH genes (low-density lipoprotein receptor [LDLR], apolipoprotein B [APOB], and proprotein convertase subtilisin/kexin type 9 [PCSK9]) in 12 distinct cohorts, including 26,025 participants, to determine: 1) the diagnostic yield of gene sequencing to identify an FH mutation in severely hypercholesterolemic participants; and 2) the clinical impact of an FH mutation on CAD risk within any given stratum of LDL cholesterol levels.
Whole-exome sequencing was performed in 7 previously described CAD case-control cohorts of the Myocardial Infarction Genetics Consortium (Online Table 1), including the Italian Atherosclerosis, Thrombosis, and Vascular Biology study (19), the ESP-EOMI (Exome Sequencing Project Early–Onset Myocardial Infarction) study (20), a nested case-control of the JHS (Jackson Heart Study) (21), the Munich Myocardial Infarction study (22), the Ottawa Heart Study (23), the PROCARDIS (Precocious Coronary Artery Disease) study (24), and PROMIS (Pakistan Risk of Myocardial Infarction Study) (25). The effect of lipid-lowering therapy in those reporting use at the time of lipid measurement was taken into account by dividing the measured total cholesterol and LDL cholesterol by 0.8 and 0.7, respectively, as implemented previously (26–28). Primary, severe LDL cholesterol elevation was defined as ≥190 mg/dl, in accordance with current cholesterol treatment guidelines (3).
The prevalence of an FH mutation in participants with LDL cholesterol >190 mg/dl was additionally determined in 11,908 participants from 5 prospective cohort studies derived from the CHARGE (Cohorts for Heart and Aging Research in Genomic Epidemiology) Consortium (29), ARIC (Atherosclerosis Risk in Communities), Cardiovascular Health Study, FHS (Framingham Heart Study), Rotterdam Baseline Study, and Erasmus Rucphen Family Study (Online Table 2).
To determine the cumulative exposure to LDL cholesterol according to FH mutation status, publically available data from the National Center for Biotechnology Information dbGAP database were analyzed, which included 5,727 ARIC cohort participants and 2,714 FHS Offspring Study participants.
CAD case-control whole-exome sequencing was performed at the Broad Institute (Cambridge, Massachusetts) as described previously (20). Population-based cohort sequencing was performed at the Baylor College of Medicine (Houston, Texas) for the ARIC, CHS, and FHS cohorts and at Erasmus Medical Center (Rotterdam, Netherlands) for the Rotterdam Baseline Study and Erasmus Rucphen Family Study cohorts. Additional sequencing methodology details are available in the Online Appendix.
Genetic variant annotation
Three classes of genetic variants were aggregated with respect to association with FH: 1) loss-of-function variants in LDLR: single-base changes that introduce a stop codon, leading to premature truncation of a protein (nonsense), insertions or deletions (indels) of deoxyribonucleic acid (DNA) that scramble protein translation beyond the variant site (frameshift), or point mutations at sites of pre-messenger ribonucleic acid splicing that alter the splicing process (splice-site); 2) missense variants in LDLR predicted to be deleterious by each of 5 in silico prediction algorithms (LRT score, MutationTaster, PolyPhen-2 HumDiv, PolyPhen-2 HumVar, and Sorting Intolerant From Tolerant [SIFT]), as described previously (20,30); and 3) variants in LDLR, APOB, or PCSK9 annotated as “pathogenic” or “likely pathogenic” in ClinVar, a publicly available archive of genetic variations associated with clinical phenotypes (31). Additional sensitivity analyses aggregated all rare (allele frequency <0.01) missense mutations in LDLR; exon 26 of APOB, which encodes key components of apolipoprotein B binding to the LDL receptor and harbors the majority of APOB variants linked to FH (32); and those that produce a change in PCSK9 at an amino acid associated with FH in ClinVar. Rare synonymous variants at these same locations were included as a negative control. Software used to annotate observed variants included Variant Effect Predictor (version 77) (33) and the associated LOFTEE plugin (34), as well as the dbNSFP database (version 3.0b1) (35).
Longitudinal analysis of LDL cholesterol exposure
Individuals with an FH mutation and LDL cholesterol ≥130 mg/dl were identified in the ARIC and FHS Offspring Study cohorts. LDL cholesterol values were adjusted in those who reported lipid-lowering therapy by dividing measured values by 0.7. Mean LDL cholesterol exposure was calculated as cumulative exposure, determined via area under the curve analysis, divided by length of follow-up. Twenty-seven FH mutation carriers met the inclusion criteria described previously and were matched 1:1 to a mutation-negative control according to age (within 10 years), sex, statin use at time of last visit, and similar LDL cholesterol at last visit (within 10 mg/dl). A match was available in 25 of 27 participants (93%). Mean LDL cholesterol exposure was compared among those with and without FH mutation using a paired Student t test.
The impact of aggregations of genetic variants on levels of LDL cholesterol was assessed with linear regression, with adjustments for age, age squared, sex, cohort, and the first 5 principal components of ancestry. Odds ratios (ORs) for CAD were calculated by use of logistic regression with adjustment for sex, cohort, and the first 5 principal components of ancestry. In analyses conducted on ordinal strata of LDL cholesterol, participants with LDL cholesterol <130 mg/dl and no mutation linked to FH served as the reference group.
Analyses were performed with R version 3.2.2 software (R Project for Statistical Computing, Vienna, Austria). Figures were created with the ggplot2 package within R (36).
Within the Myocardial Infarction Genetics Consortium CAD case-control cohorts, a total of 14,117 participants with both LDL cholesterol level and sequence data for FH genes were available for analysis: 8,577 CAD-free control subjects and 5,540 CAD case subjects (Online Table 3). The study population included 10,421 men (74% of participants) with a mean age of 53 ± 14 years. Proportions of self-identified race were 47%, 46%, and 7% for white, South Asian, and black, respectively. Forty-seven percent of study participants had a history of hypertension, 27% had a history of diabetes mellitus, 34% were current smokers, and 14% were taking lipid-lowering medications.
Sequencing identified 86 variants linked to FH because they led to loss of function in LDLR, were missense mutations in LDLR predicted to be damaging by each of 5 computer prediction algorithms, or were a variant in LDLR, APOB, or PCSK9 previously linked to FH in the ClinVar genetics database. These included 13 premature stop (“nonsense”) codons, 6 splice acceptor/donor variants, 4 frameshift mutations, and 63 missense mutations (Online Table 4).
Mutations linked to FH were found in 164 participants, including 48 CAD-free control subjects (OR: 0.6%; 95% confidence interval [CI]: 0.4% to 0.7%) and 116 CAD case subjects (OR: 2.1%; 95% CI: 1.7% to 2.5%) (Online Table 5). The mutation was located in LDLR for 141 participants (86%), in APOB for 22 (13%), and in PCSK9 for 1 (0.6%) (Online Table 4). Only 1 homozygote (or compound heterozygote) participant was identified; a 33-year-old patient with LDL cholesterol of 539 mg/dl and CAD was homozygous for a p.Q33* premature stop codon in LDLR.
Diagnostic yield of gene sequencing in severe hypercholesterolemia
Among 8,577 CAD-free control participants from the Myocardial Infarction Genetics Consortium cohorts, LDL cholesterol approximated a normal distribution (Online Figure 1). The prevalence of an FH mutation increased across categories of LDL cholesterol levels (p < 0.001) (Online Figure 2). Of 8,577 control participants, 430 (5% of control sample) had LDL cholesterol ≥190 mg/dl, and only 8 of these carried an FH mutation (OR: 1.9%; 95% CI: 0.9% to 3.8%) (Table 1, Central Illustration).
This prevalence estimate was replicated in 11,908 participants from 5 prospective cohort studies of the CHARGE consortium: 956 (8%) had LDL cholesterol >190 mg/dl, and of these, 16 (OR: 1.7%; 95% CI: 1.0% to 2.8%) harbored an FH mutation. Across the 12 studies (n = 20,485), 1,386 participants (7%) displayed LDL cholesterol ≥190 mg/dl, of whom 24 (1.7%) carried an FH mutation (Table 1).
Clinical impact of FH mutation identification on CAD risk
In the Myocardial Infarction Genetics Consortium case-control studies, the presence of an FH mutation was associated with a 50 mg/dl (95% CI: 44 to 57 mg/dl) increase in LDL cholesterol and a 3.8-fold (95% CI: 2.6 to 5.4) increase in odds of CAD. These effects were most pronounced in those with loss-of-function mutations in LDLR (Figure 1). Average LDL cholesterol was 190 mg/dl in those with an FH mutation, and 73 of 164 mutation carriers (45%) had LDL cholesterol ≥190 mg/dl. However, despite the observed large effect on average levels, a wide spectrum of circulating LDL cholesterol concentrations was noted in those who were mutation positive (Figure 2). Forty-four of 164 (27%) mutation carriers had an observed LDL cholesterol level <130 mg/dl, which reflects incomplete penetrance. An aggregation of all rare missense mutations had a modest impact on both LDL cholesterol and CAD risk. As expected, synonymous mutations, which do not change the amino acid sequence, had no effect on either parameter (Figure 1). FH mutations were also associated with a nominally significant reduction in high-density lipoprotein cholesterol (−1.9 mg/dl; 95% CI: −3.7 to −0.1; p = 0.04) but not with circulating triglycerides (p = 0.36).
Within the Myocardial Infarction Genetics Consortium case-control cohort populations, those with an FH mutation were at higher risk of CAD than those without a mutation (Table 2) (p value for difference = 0.001). For participants with both LDL cholesterol ≥190 mg/dl and an FH mutation, the odds of CAD were increased 22-fold (OR: 22.3; 95% CI: 10.7 to 53.2) compared with those with LDL cholesterol <130 mg/dl and no mutation. For participants with LDL cholesterol ≥190 mg/dl and no FH mutation, odds of CAD were increased 6-fold (OR: 6.0; 95% CI: 5.2 to 6.9) compared with the same reference group. This difference persisted after additional adjustment for measured LDL cholesterol level (p = 0.02).
Separation of the population into clinically relevant categories of LDL cholesterol levels demonstrated trends toward higher risk in those with an FH mutation within each stratum (Central Illustration, Online Table 6). The impact of an FH mutation was similar across strata of LDL cholesterol levels (p value for interaction = 0.51). Within the subgroup of participants with LDL cholesterol in the ≥190 to 220 mg/dl range, those with a mutation had 17-fold increased CAD risk versus 5-fold for those without a mutation. This was despite similar observed LDL cholesterol levels in this stratum (mean LDL cholesterol 205 mg/dl in those with an FH mutation versus 203 mg/dl in those without; p value for difference = 0.22).
Cumulative LDL cholesterol exposure according to FH mutation status
For any given observed LDL cholesterol level, those harboring a mutation might have a higher average lifetime LDL cholesterol exposure than those who do not harbor a mutation; this could explain the higher CAD risk among mutation carriers. We tested this hypothesis using 2 prospective cohort studies, ARIC and the FHS Offspring Study, in which sequencing data and serial measurements of LDL cholesterol were available. We identified 25 participants with an FH mutation and LDL cholesterol ≥130 mg/dl. Mean LDL cholesterol at time of last study visit was 185 mg/dl. Compared with matched noncarriers with similar LDL cholesterol at the last visit, participants with an FH mutation had a 17 mg/dl (95% CI: 5 to 29 mg/dl; p = 0.007) higher average LDL cholesterol exposure in the years preceding the last visit (Figure 3, Online Table 7).
Among 20,485 multiethnic participants from 12 studies, we found that 1,386 (7%) had severe hypercholesterolemia (LDL cholesterol ≥190 mg/dl), and only a small fraction (<2%) of those also carried an FH mutation. However, within any stratum of LDL cholesterol, those who carried an FH mutation were at substantially higher risk for CAD than those who did not. This increased CAD risk among mutation carriers was explained at least in part by a greater cumulative lifetime exposure to LDL cholesterol.
These results permit several conclusions. First, FH mutations explain only a small fraction of severe hypercholesterolemia in the population. Previous reports noted a substantially higher rate of mutation detection in those with clinically suspected FH, ascertained on the basis of features (e.g., family history, physical examination, or severe hypercholesterolemia at a young age) that enrich for a monogenic origin (5–16). Here, we address a scientific question (what fraction of severely hypercholesterolemic subjects carry a mutation in any of 3 genes causal for FH?) that is distinct from these earlier seminal reports. When participants were ascertained solely on the basis of a single elevated LDL cholesterol level, we identified an FH mutation in fewer than 2% of severely hypercholesterolemic subjects. These sequencing results are broadly consistent with those of a recent study of 98,098 subjects from the Copenhagen General Population Study in which genotyping of the 4 most common FH mutations was used to extrapolate overall FH mutation prevalence. In that Danish study, of 5,332 subjects with LDL cholesterol ≥5 mmol/l (193 mg/dl), fewer than 5% were predicted to harbor an FH mutation (28).
If not a monogenic mutation in the 3 FH genes, what might be the cause of elevated LDL cholesterol in the remaining >95% of participants with severe hypercholesterolemia? Possibilities include polygenic hypercholesterolemia, life-style factors, or a combination of these. For example, subjects in the top quartile of a polygenic LDL cholesterol gene score composed of 95 common variants were 13-fold more likely to have high LDL cholesterol (37). Similarly, subjects in the top decile of a LDL cholesterol gene score composed of 12 common variants were 4.2-fold more likely to have LDL ≥190 mg/dl in the U.K. Whitehall II study (38). Future genetic studies might identify additional causal variants, genes beyond those considered in this study, or large-effect regulatory variants that underlie severe hypercholesterolemia. Other nongenetic explanations for severe LDL cholesterol elevations include secondary causes (e.g., hypothyroidism or nephrotic syndrome), life-style factors such as dietary fat, or some combination of these.
Second, within any stratum of a single observed LDL cholesterol level, CAD risk was higher in those with an FH mutation than in those without. This novel finding reinforces the potential utility of genetic testing to provide risk information beyond the LDL cholesterol level. We analyzed 25 matched pairs of participants with similarly elevated LDL cholesterol levels at the time of ascertainment and found a higher cumulative exposure to LDL cholesterol in those with an FH mutation. These data support the hypothesis that an FH mutation, present since birth, increases CAD risk via lifelong exposure to high LDL cholesterol (39). By contrast, an isolated elevation in LDL cholesterol in those without a genetic predisposition might reflect a time-limited exposure related to a current environmental perturbation or a value that is more likely to regress toward the mean in the future. Future studies might identify additional metabolic parameters, such as increased lipoprotein(a) levels (40), that also contribute to the excess CAD risk noted in those with an FH mutation.
Finally, these data contribute to ongoing discussion regarding how to define FH. Classically, FH refers to elevated LDL cholesterol caused by a single mutation in any of several genes segregating in an autosomal dominant manner. Alternate approaches to 2 features, LDL cholesterol threshold and mutation definition, affect FH prevalence estimates (Table 3). An approach that includes all participants with untreated LDL cholesterol ≥190 mg/dl (i.e., without an FH mutation requirement) would combine nongenetic and genetic causes and classify approximately 7% of the U.S. adult population as having FH. An alternative possibility is to withhold an LDL cholesterol threshold and require only a stringent mutation definition; in such an analysis of 20,485 participants, we identified an FH mutation in 97 participants (1 in 211). This estimate is nearly identical to a population-based analysis in the Copenhagen General Population Study (1 in 217) (28). However, if one additionally requires that an FH mutation is accompanied by an elevated LDL cholesterol level, FH prevalence in our study declines (1 in 301 with an LDL threshold ≥130 mg/dl and 1 in 853 with an LDL threshold ≥190 mg/dl).
With regard to defining an FH mutation, all schemata agree on the inclusion of loss-of-function alleles in LDLR, but they differ on how to handle missense mutations. For missense mutations, we applied a rigorous threshold, requiring that the mutation be designated as damaging by each of 5 computer prediction algorithms or be previously annotated as pathogenic in the ClinVar clinical genetics database. A key advantage of this approach is that it ensures that classification is both fully reproducible and generalizable to genes beyond those related to FH.
When routine genetic testing is not available, clinical scoring systems, such as the Dutch Lipid Clinical Network, Simon Broome, and MEDPED criteria, have been developed to approximate FH status (4). Ongoing collaborative efforts on how to optimally incorporate population-based genetic sequencing data into existing frameworks for the clinical diagnosis of FH will be critically important.
First, our data did not permit us to stratify participants by family history or physical examination features, as suggested by the Dutch Lipid Clinic Network and Simon Broome criteria (41,42). Second, we accounted for an estimated 30% reduction in LDL cholesterol in those undergoing lipid-lowering therapy, as previously implemented (26–28). This approach might imperfectly estimate untreated LDL cholesterol, given heterogeneity in drug selection, dosing, and response and variability across baseline LDL cholesterol levels or mutation status. However, a sensitivity analysis limited to Myocardial Infarction Genetics Consortium cohort participants not undergoing lipid-lowering therapy similarly noted a pronounced difference in risk among severely hypercholesterolemic participants stratified by mutation status (Online Table 8). Third, current exome-sequencing techniques inadequately capture structural and copy-number genetic variation, and as such, some FH mutations might have been missed. Fourth, our approach to annotating missense variants using prediction algorithms and the ClinVar database might have led to misclassification in some cases. Additional studies that implement large-scale functional screens of identified variants or that pool phenotypes across additional studies could provide additional refinement of pathogenicity annotations. Lastly, FH mutation prevalence was determined in CAD-free control subjects and population-based cohorts. These participants survived to middle age, and few had clinically manifest CAD, which raises the possibility of survivorship or selection bias. Our case-control population was enriched for participants with premature CAD; effect estimates of mutations on coronary risk might be different in patients with later disease onset.
Genetic sequencing identified an FH mutation in only a small proportion of severely hypercholesterolemic participants; however, for any given observed LDL cholesterol level, risk for CAD was substantially higher in FH mutation carriers than in noncarriers, which was likely related in large part to higher lifelong exposure to atherogenic LDL particles. A primary goal of precision medicine is to use molecular diagnostics to identify a small subset of the population at increased disease risk in which to deliver an intervention. Systematic efforts to identify and treat severely hypercholesterolemic patients who carry an FH mutation could represent one such opportunity.
COMPETENCY IN MEDICAL KNOWLEDGE: For any given observed LDL cholesterol level, carriers of a familial hypercholesterolemia mutation are at substantially increased risk of coronary disease compared with noncarriers, which is likely related to increased lifelong exposure to LDL cholesterol.
TRANSLATIONAL OUTLOOK: Additional research is needed to understand whether genetic testing can prove clinically useful in guiding the treatment of people with severe hypercholesterolemia to reduce risk of CAD.
For an expanded Methods section including references as well as supplemental tables and figures, please see the online version of this article.
Field work, genotyping, and standard clinical chemistry assays in PROMIS were principally supported by grants awarded to the University of Cambridge from the British Heart Foundation, U.K. Medical Research Council, Wellcome Trust, EU Framework 6–funded Bloodomics Integrated Project, Pfizer, Novartis, and Merck. Additional support for PROMIS was provided by the U.K. Medical Research Council (MR/L003120/1), British Heart Foundation (RG/13/13/30194), U.K. National Institute for Health Research Cambridge Biomedical Research Centre, European Research Council (268834), and European Commission Framework Programme 7 (HEALTH-F2-2012-279233). The Jackson Heart Study is supported by contracts HHSN268201300046C, HHSN268201300047C, HHSN268201300048C, HHSN268201300049C, and HHSN268201300050C from the National Heart, Lung, and Blood Institute (NHLBI) and the National Institute on Minority Health and Health Disparities. The Munich Study is supported by the German Federal Ministry of Education and Research (BMBF) in the context of the e:Med program (e:AtheroSysMed) and the FP7 European Union project CHS-NHLBI.org. The Italian Atherosclerosis, Thrombosis, and Vascular Biology (ATVB) Study was supported by a grant from RFPS-2007-3-644382 and Programma di ricerca Regione-Università 2010-2012 Area 1–Strategic Programmes–Regione Emilia-Romagna. Funding for the exome-sequencing project (ESP) was provided by RC2 HL103010 (HeartGO), RC2 HL102923 (LungGO), and RC2 HL102924 (WHISP). Exome sequencing was performed through RC2 HL102925 (BroadGO) and RC2 HL102926 (SeattleGO). Exome sequencing in ATVB, PROCARDIS, Ottawa, PROMIS, Munich Study, and the Jackson Heart Study was supported by 5U54HG003067 (to Drs. Lander and Gabriel). The views expressed in this manuscript are those of the authors and do not necessarily represent the views of the NHLBI; the National Institutes of Health; or the U.S. Department of Health and Human Services. The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.(261123). Additional grants were received from the Fondation Leducq (CADgenomics: Understanding Coronary Artery Disease Genes, 12CVD02). This study was also supported through the Deutsche Forschungsgemeinschaft cluster of excellence “Inflammation at Interfaces” and SFB 1123. Funding support for “Building on GWAS for NHLBI-diseases: the U.S. CHARGE Consortium” was provided by the National Institutes of Health through the American Recovery and Reinvestment Act of 2009 (5RC2HL102419). Data for “Building on GWAS for NHLBI-diseases: the U.S. CHARGE Consortium” were provided by Eric Boerwinkle on behalf of the Atherosclerosis Risk in Communities (ARIC) Study, L. Adrienne Cupples, principal investigator for the Framingham Heart Study (FHS), and Bruce Psaty, principal investigator for the Cardiovascular Health Study (CHS). Sequencing was performed at the Baylor Genome Center (U54 HG003273). The ARIC Study is conducted as a collaborative study supported by NHLBI contracts (HHSN268201100005C, HHSN268201100006C, HHSN268201100007C, HHSN268201100008C, HHSN268201100009C, HHSN268201100010C, HHSN268201100011C, and HHSN268201100012C). The FHS is conducted and supported by the NHLBI in collaboration with Boston University (contract No. N01-HC-25195), and its contract with Affymetrix, Inc., for genome-wide genotyping services (contract N02-HL-6-4278), for quality control by FHS investigators using genotypes in the SNP Health Association Resource (SHARe) project. A portion of this research was conducted using the Linux Cluster for Genetic Analysis (LinGA-II) funded by the Robert Dawson Evans Endowment of the Department of Medicine at Boston University School of Medicine and Boston Medical Center. This CHS research was supported by NHLBI contracts HHSN268201200036C, HHSN268200800007C, N01HC55222, N01HC85079, N01HC85080, N01HC85081, N01HC85082, N01HC85083, and N01HC85086 and NHLBI grants U01HL080295, R01HL087652, R01HL105756, R01HL103612, and R01HL120393, with additional contribution from the National Institute of Neurological Disorders and Stroke. Additional support was provided through R01AG023629 from the National Institute on Aging. A full list of principal CHS investigators and institutions can be found at
Dr. Khera is supported by an American College of Cardiology/Merck Fellowship award and has received consulting fees from Merck and Amarin Corporation. Dr. Peloso is supported by National Heart, Lung, and Blood Institute award K01HL125751. Dr. Kessler is supported by a Deutsches Zentrum für Herz-Kreislauf-Forschung rotation grant. Dr. Psaty has served on a data safety and monitoring board for a clinical trial funded by Zoll LifeCor; and on a steering committee for the Yale Open Data Access project, funded by Johnson & Johnson. Dr. Rader has received consulting fees from Aegerion Pharmaceuticals, Alnylam Pharmaceuticals, Eli Lilly and Company, Pfizer, Sanofi, and Novartis; is an inventor on a patent related to lomitapide that is owned by the University of Pennsylvania and licensed to Aegerion Pharmaceuticals; and is a cofounder of Vascular Strategies and Staten Biotechnology. Dr. Ardissino has received speaker fees from AstraZeneca, Boehringer Ingelheim, Johnson & Johnson, Bayer, Daiichi-Sankyo, GlaxoSmithKline, Eli Lilly and Company, Boston Scientific, Bristol-Myers Squibb, Menarini Group, Novartis, and Sanofi; and research grants from GlaxoSmithKline, Eli Lilly and Company, Pfizer, and Novartis. Dr. Saleheen has received grants from Pfizer and the National Institutes of Health. Dr. Kathiresan is supported by a research scholar award from the Massachusetts General Hospital, the Donovan Family Foundation, and R01 HL127564; has received grants from Bayer Healthcare, Aegerion Pharmaceuticals, and Regeneron Pharmaceuticals; consulting fees from Merck, Novartis, Sanofi, AstraZeneca, Alnylam Pharmaceuticals, Leerink Partners, Noble Insights, Quest Diagnostics, Genomics PLC, and Eli Lilly and Company; and holds equity in San Therapeutics and Catabasis Pharmaceuticals. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Khera, Won, and Peloso contributed equally to this work.
- Abbreviations and Acronyms
- apolipoprotein B
- coronary artery disease
- confidence interval
- familial hypercholesterolemia
- low-density lipoprotein
- low-density lipoprotein receptor
- odds ratio
- proprotein convertase subtilisin/kexin type 9
- Received March 2, 2016.
- Revision received March 22, 2016.
- Accepted March 22, 2016.
- American College of Cardiology Foundation
- Stone N.J.,
- Robinson J.G.,
- Lichtenstein A.H.,
- et al.
- Gidding S.S.,
- Champagne M.A.,
- de Ferranti S.D.,
- et al.,
- American Heart Association Atherosclerosis, Hypertension, and Obesity in Young Committee of the Council on Cardiovascular Disease in Young,
- Council on Cardiovascular and Stroke Nursing,
- Council on Functional Genomics and Translational Biology, and Council on Lifestyle and Cardiometabolic Health
- Humphries S.E.,
- Whittall R.A.,
- Hubbart C.S.,
- et al.
- van der Graaf A.,
- Avis H.J.,
- Kusters D.M.,
- et al.
- Ahmad Z.,
- Adams-Huet B.,
- Chen C.,
- Garg A.
- Klančar G.,
- Grošelj U.,
- Kovač J.,
- et al.
- Nordestgaard B.G.,
- Chapman M.J.,
- Humphries S.E.,
- et al.,
- European Atherosclerosis Society Consensus Panel
- Atherosclerosis, Thrombosis, and Vascular Biology Italian Study Group
- Taylor H.A. Jr..
- McPherson R.,
- Pertsemlidis A.,
- Kavaslar N.,
- et al.
- Cholesterol Treatment Trialists' (CTT) Collaborators
- Benn M.,
- Watts G.F.,
- Tybjærg-Hansen A.,
- Nordestgaard B.G.
- Psaty B.M.,
- O'Donnell C.J.,
- Gudnason V.,
- et al.,
- CHARGE Consortium
- Landrum M.J.,
- Lee J.M.,
- Riley G.R.,
- et al.
- Borén J.,
- Ekström U.,
- Agren B.,
- Nilsson-Ehle P.,
- Innerarity T.L.
- McLaren W.,
- Pritchard B.,
- Rios D.,
- Chen Y.,
- Flicek P.,
- Cunningham F.
- ↵Karczewski K. J. LOFTEE (Loss-Of-Function Transcript Effect Estimator). 2016. Available at: https://github.com/konradjk/loftee. Accessed March 31, 2016.
- Wikham H.
- Brown M.S.,
- Goldstein J.L.
- ↵Alonso R, Andres E, Mata N, et al., SAFEHEART Investigators. Lipoprotein(a) levels in familial hypercholesterolemia: an important predictor of cardiovascular disease independent of the type of LDL receptor mutation. J Am Coll Cardiol 2014;63:1982–9.
- World Health Organization Human Genetics Programme
- National Collaborating Centre for Primary Care