ORIGINAL E n d o c r i n e

ARTICLE C a r e

Familial Hypercholesterolemia in the Danish General Population: Prevalence, Coronary Artery Disease, and Cholesterol-Lowering Medication Marianne Benn, Gerald F. Watts, Anne Tybjaerg-Hansen, and Børge G. Nordestgaard Department of Clinical Biochemistry (M.B., B.G.N.) and The Copenhagen General Population Study (M.B., A.T.-H., B.G.N.), Herlev Hospital, Copenhagen University Hospital, DK-2730 Herlev, Denmark; School of Medicine and Pharmacology (G.F.W.), Lipid Disorders Clinic, Royal Perth Hospital, University of Western Australia, Perth WA6000, Australia; Department of Clinical Biochemistry (A.T.-H.), Rigshospitalet, Copenhagen University Hospital, DK-2100 Copenhagen, Denmark; and Faculty of Health Sciences (M.B., A.T.-H., B.G.N.), University of Copenhagen, DK-1455 Copenhagen, Denmark

Context: The diagnosis of familial hypercholesterolemia (FH) can be made using the Dutch Lipid Clinic Network criteria. This employs the personal and family history of premature coronary artery disease and hypercholesterolemia and the presence of a pathogenic mutation in the low-density lipoprotein receptor (LDLR) and apolipoprotein B (APOB) genes. Objective: We employed this tool to investigate the prevalence of FH and the associations between FH and coronary artery disease and cholesterol-lowering medication in the Copenhagen General Population Study. Setting: The study was of an unselected, community-based population comprising 69,016 participants. Main Outcome Measures: FH (definite/probable) was defined as a Dutch Lipid Clinic Network score higher than 5. Coronary artery disease was myocardial infarction or angina pectoris. Results: The prevalence of FH was 0.73% (one in 137). Of participants with FH, 20% had an LDLR or APOB mutation. The prevalence of coronary artery disease among FH participants was 33%. Only 48% of subjects with FH admitted to taking cholesterol-lowering medication. The odds ratio for coronary artery disease off cholesterol-lowering medication was 13.2 (10.0 –17.4) in definite/probable FH compared with non-FH subjects, after adjusting for age, gender, body mass index, hypertension, metabolic syndrome and diabetes, and smoking. The corresponding adjusted odds ratio for coronary artery disease in FH subjects on cholesterol-lowering medication was 10.3 (7.8 –13.8). Conclusion: The prevalence of FH appears to be higher than commonly perceived in a general population of white Danish individuals, with at least half of affected subjects not receiving cholesterol-lowering medication. The very high risk of coronary artery disease irrespective of use of medication reflects the extent of underdiagnosis and undertreatment of FH in the community and primary care. (J Clin Endocrinol Metab 97: 3956 –3964, 2012)

F

amilial hypercholesterolemia (FH) is a readily preventable cause of premature coronary artery disease (CAD) (1). FH is the most common and serious monogenic

disorder of lipid metabolism (OMIM number 143890), causing premature CAD due to accelerated atherosclerosis from birth (2). FH is an autosomal dominantly inherited

ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2012 by The Endocrine Society doi: 10.1210/jc.2012-1563 Received March 2, 2012. Accepted July 24, 2012. First Published Online August 14, 2012

Abbreviations: BMI, Body mass index; CAD, coronary artery disease; CI, confidence interval; DLCN, Dutch Lipid Clinic Network; FH, familial hypercholesterolemia; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

3956

jcem.endojournals.org

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

disorder caused primarily by mutations in the gene encoding the low-density lipoprotein (LDL) receptor, LDLR (3). Less frequent mutations in the APOB and PCSK9 genes have similar functional consequences (4). In FH, the classical defect in the LDL receptor pathway leads to decreased clearance of LDL-cholesterol from plasma, and this increases the plasma concentration of LDL and hence total cholesterol (4). The tools for diagnosing adult FH are the Make Early Diagnosis to Prevent Early Death (5), Simon Broome (6), and Dutch Lipid Clinic Network (DLCN) criteria (7). Although these can all predict FH mutations in adults (8), the DLCN criteria have been widely recommended recently for clinical use in Europe (9) and Australasia (10). This tool provides a numerical probability of having FH based on the personal and family history of premature CAD, the plasma level of LDL-cholesterol, and the presence of specific clinical stigmata (arcus cornealis and tendon xanthomata). It can also afford a sensitive approach upon which a decision to carry out a genetic test may be made (10, 11). Genetic testing for FH is particularly useful when there is little knowledge of the extended family. FH is genetically heterogeneous, and in Denmark the mutational spectrum is intermediate, with APOB and LDLR mutations accounting for the majority of mutations, and in particular three mutations in the LDLR gene accounting for 36% of known LDLR mutations in FH patients (12, 13). Despite routine screening of FH patients, disease-causing mutations in the PCSK9 gene have not been reported in Denmark so far. Excluding rare populations subject to a gene founder effect (in whom FH is particularly common), the community prevalence of FH is estimated to be one in 500 (3, 14, 15). Reports can vary, however, from one in 200 to one in 2000 (3, 14 –16). The available prevalence data may be inaccurate, because they are based on hospital patients, registry samples, or calculations employing the HardyWeinberg equation and the estimated frequency of homozygous FH (3, 14 –16). The prevalence of FH has not been hitherto investigated in an unselected communitybased population. This information is important for designing screening strategies for FH in primary care (10). The standardized mortality rate of CAD and risk of a coronary event are increased in people with untreated FH (2, 14). These data derive classically from prospective studies of FH subjects selected from registries or kinships (17, 18). The specific evidence for treating hypercholesterolemia in FH is based on selected observational studies showing that long-term statin medication decreases CAD events and mortality in FH to a level comparable to or approaching that of the background population. No data have been reported on the risk of CAD or the potential

jcem.endojournals.org

3957

effectiveness of statins in FH diagnosed in a large sample from a community-based population not subject to ascertainment bias. In the present study, we first estimated the prevalence of FH in 69,016 individuals from the Danish general population, the Copenhagen General Population Study, using phenotypic assessment based on the DLCN criteria and on the common mutations causative of FH in the Danish population (12, 13). Second, we examined the risk of CAD according to the diagnosis of FH and the potential impact of cholesterol-lowering medication.

Subjects and Methods The Copenhagen General Population Study The Copenhagen General Population Study is a study of a general population initiated in 2003 with ongoing enrollment (19, 20). Individuals were selected from the national Danish Civil Registration System to reflect the adult Danish population aged 20 –100 yr and were all whites of Danish descent. A total of 69,209 participants were included at the time of analyses; however, 193 individuals (two with a definite, three with a probable, and 20 with a possible diagnosis of FH and 168 unlikely to have FH) were excluded due to hypothyroidism defined as TSH higher than 5 mIU/liter and a total T4 below 50 nmol/liter or a total T3 below 0.9 nmol/liter, leaving 69,016 participants in the study. Data were obtained from a self-completed questionnaire that was reviewed together by a field investigator on the day of attendance, a brief physical examination, and nonfasting venous blood samples. The study was approved by Copenhagen University Hospital and by a Danish ethical committee (H-KF-01144/01) and conducted according to the Declaration of Helsinki. Informed written consent was obtained from participants.

Diagnostic criteria for fFH The diagnosis of FH was established using the following modified version of the DLCN criteria (7, 10, 22) and numerical score (Supplemental Table 1, published on The Endocrine Society’s Journals Online web site at http://jcem.endojournals.org): family history of a first-degree relative with known premature (⬍55 yr for men; ⬍60 yr for women) CAD or vascular disease and/or a first-degree relative with known hypercholesterolemia (1 point); personal history of premature CAD (ages as above, 2 points) or premature cerebral or peripheral vascular disease (ages as above, 1 point); LDL-cholesterol higher than 8.5 mmol/liter (⬎330 mg/ dl, 8 points), 6.5– 8.4 mmol/liter (250 –329 mg/dl, 5 points), 5.0 – 6.4 mmol/liter (190 –249 mg/dl, 3 points), or 4.0 – 4.9 mmol/liter (155–189 mg/dl, 1 point); presence of an LDLR W23X, W66G, or W556S (accounting for 36% of the LDLR mutation spectrum in the Danish population) (12, 13) or an APOB R3500Q mutation (8 points). A diagnosis of FH was considered definite if the total score was greater than 8, probable if the score was 6 – 8, possible if the score was 3–5, and unlikely if the score was below 3 points (10, 21, 22). We did not employ other criteria relating to LDL-cholesterol in children and family or personal details of tendon xanthomata or arcus cornealis to define FH because they were not recorded in the main study.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

3958

Benn et al.

Familial Hypercholesterolemia

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

Information on a diagnosis of CAD/ischemic heart disease (World Health Organization International Classification of Diseases, ICD8:410 – 414; ICD10:I20 –I25) was collected from January 1, 1977, through May 9, 2011, by reviewing all hospital admissions and diagnoses entered in the national Danish Patient Registry and all causes of death entered in the national Danish Causes of Death Registry (19, 20), and 5654 individuals had a diagnosis of CAD, with no individuals lost to follow-up.

Hg for those with diabetes), diastolic blood pressure at least 90 mm Hg (ⱖ85 mm Hg for those with diabetes), and/or use of antihypertensive medication (24). Metabolic syndrome was defined according to internationally agreed criteria (25). Diabetes mellitus was defined as self-reported diabetes, use of antidiabetic medication, a nonfasting plasma glucose over 11.0 mmol/liter (⬎198 mg/dl), and/or hospitalization due to diabetes (ICD8: 249 –250; ICD10:E10 –E11, E13–E14). Smokers were current smokers. Cholesterol-lowering medication was self-reported with over 96% being statin.

Laboratory analyses

Statistical analysis

Plasma concentrations of cholesterol, high-density lipoprotein (HDL)-cholesterol, triglycerides, and glucose were measured by standard enzymatic assays; apolipoprotein B and apolipoprotein A-I were measured immunochemically (Thermo Fisher Scientific/Konelab). LDL-cholesterol was calculated using the Friedewald equation when plasma triglycerides were less than or equal to 4.0 mmol/liter (ⱕ352 mg/dl) and measured directly in 2681 individuals (corresponding to 3.9% of the population) due to a plasma triglyceride higher than 4.0 mmol/liter (Thermo Fisher Scientific/Konelab, Espoo, Finland). For the diagnostic classification, plasma LDL-cholesterol concentrations were multiplied by 1.43 in individuals receiving cholesterol-lowering medication, corresponding to an estimated 30% reduction in LDL-cholesterol (23).

Data were analyzed using Stata/SE version 11.2. HardyWeinberg equilibrium of genotypes was evaluated by ␹2 tests. The point prevalence of each definition of FH was estimated as a percentage for all participants and according to age and gender. Frequency of nonlipid risk factors among the definitions of FH was compared by ␹2 tests. The odds ratio in predicting definite/ probable FH was estimated for each criterion in the DLCN criteria by comparing risk between individuals fulfilling vs. not fulfilling the specific criterion using logistic regression adjusted for age and gender and for LDL-cholesterol levels by comparing individuals with a level above and below each of the criterion thresholds. Positive and negative predictive values in predicting FH were also calculated for each criterion. One-way ANOVA was used to compare lipid, lipoprotein, and apolipoprotein concentrations among the diagnostic groupings of FH. Student’s t test was used for comparison between individuals, respectively, on and off cholesterol-lowering medication. Risk of CAD for individuals with a diagnosis of definite/probable and possible FH relative to those with unlikely FH was estimated by multivariable logistic regression, adjusting for gender, age, BMI, hypertension, metabolic syndrome, diabetes mellitus, and smoking.

Coronary artery disease

Genotyping LDLR W23X, W66G, and W556S and APOB R3500Q mutations were genotyped in 60,710 individuals by TaqMan (Applied Biosystems, Foster City, CA); these mutations are heterozygous. Genotypes were verified by sequencing of randomly selected samples of each variant. There was 100% agreement between TaqMan and sequencing results.

Results Other covariates Body mass index (BMI) was measured as weight (kilograms) divided by measured height (meters) squared. Hypertension was defined as systolic blood pressure at least 140 mm Hg (ⱖ135 mm

Prevalence of FH The prevalence of individuals classified with definite FH (DLCN criteria ⬎8 points) was 0.20% (one in 504),

TABLE 1. Characteristics of participants in the Copenhagen General Population Study according to the diagnostic probability of FH, based on the DLCN criteria Diagnostic probability of FH Number Prevalence (%) Gender (% women) Age (yr) BMI (kg/m2) Obesity, BMI ⱖ30 kg/m2 (%) Metabolic syndrome (%) Diabetes mellitus (%) Hypertension, all (%) Hypertension, non-diabetes mellitus (%) Hypertension, diabetes mellitus (%) Current smokers (%) Cholesterol-lowering medication (%)

Definite 137 0.20 54 58 (46 – 64) 25.7 (22.9 –28.9) 19 21 3.0 67 67

Probable 365 0.53 61 59 (52– 66) 27.5 (24.9 –30.8) 32 39 6.3 79 78

Possible 4,295 6.3 56 59 (52– 66) 30.0 (24.5–29.8) 23 35 4.0 77 77

Unlikely 64,219 93 55 59 (47– 67) 25.5 (23.2–28.4) 16 21 3.8 67 66

All 69,016 100 55 58 (47– 67) 25.6 (23.2–28.5) 17 22 3.8 67 66

75

91

90

91

91

16 57

31 45

25 21

20 9.4

20 11

Continuous values are summarized as median (interquartile range) and categorical variables as percentage of individuals in the diagnostic group.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

jcem.endojournals.org

3959

lesterol-lowering treatment compared with those with unlikely FH (Table 1). Diagnostic criteria for FH Table 2 shows the percentage of individuals with definite/probable, possible, and unlikely FH who met each diagnostic criterion employed to diagnose the condition. In those with definite/probable FH, the frequency with which the criteria met were (in decreasing order) as follows: LDL-cholesterol of 6.5– 8.4 mmol/liter (5 points, 53%); a first-degree relative with known elevated cholesterol levels (1 point, 50%); LDLFIG. 1. Prevalence of definite or probable and possible FH according to DLCN criteria in the cholesterol of 5.0 – 6.4 mmol/liter (3 Copenhagen General Population Study by 20-yr age groups and gender, based on 69,016 points, 32%); a patient with premature individuals. CAD (2 points, 28%); carrying LDLR W23X, W66G, or W556S or APOB probable FH (5– 8 points) was 0.53% (one in 189), definite or probable FH combined (⬎5 points) 0.73% (one in R3500Q mutation (8 points, 20%); a first-degree relative 137), possible FH 6.3% (one in 16) (3–5 points), and un- with premature cardiovascular disease (1 point, 9%); likely FH 93% (⬍3 points) (Table 1). The prevalence of LDL-cholesterol of 4.0 –5.9 mmol/liter (1 point, 9%); and definite, probable, and possible FH by 20-yr age groups LDL-cholesterol above 8.5 mmol/liter (8 points, 5%). Inand gender is shown in Fig. 1. For comparison, prevalence terestingly, of people with unlikely FH, 9 and 25% had a of probable FH by the Make Early Diagnosis to Prevent first-degree relative with premature CAD and elevated Early Death (5) criteria was 0.80% (one in 128), whereas cholesterol level, respectively. prevalence of definite or possible FH by Simon Broome (6) criteria was 4.1% (one in 25). Individuals with a diagnosis Prediction of FH Because the prevalence of the criteria included in the of FH, and in particular probable FH, more often were women; were obese; had metabolic syndrome, diabetes DLCN criteria differ, and because each criterion contribmellitus, and hypertension; and were smokers and on cho- utes differently to the overall diagnosis of FH (1– 8 points), TABLE 2. Proportion of individuals with selected DLCN criteria in diagnostic categories of FH in the Copenhagen General Population Study Diagnostic probability of FH

Family history First-degree relative with premature cardiovascular disease or First-degree relative with known elevated cholesterol levels Personal clinical history Patient with premature CAD Patient with premature cerebral or peripheral vascular disease Plasma cholesterol, LDL-cholesterol in mmol/litera ⱖ8.5 (330 mg/dl) 6.5– 8.4 (250 –329 mg/dl) 5.0 – 6.4 (190 –249 mg/dl) 4.0 – 4.9 (155–189 mg/dl) ⬍4.0 (⬍155 mg/dl) Genetic analysis LDLR (W23X, W66G, W556S) or APOB (R3500Q) mutations (%)

Definite/probable, n ⴝ 502

Possible, n ⴝ 4,295

Unlikely, n ⴝ 64,219

9 50

15 39

9 25

1

28 8

20 4

1 1

2 1

5 53 32 9 2

0 0 79 9 12

0 0 0 21 80

8 5 3 1

20

0

0

8

Score

Values are percentages of individuals within each diagnostic group fulfilling each criterion. a LDL-cholesterol was calculated using the Friedewald equation when plasma triglycerides were less than or equal to 4.0 mmol/liter (352 mg/dl) and otherwise measured directly (3.9% of the population). For individuals on lipid-lowering medication (11%), LDL-cholesterol in millimoles per liter was multiplied by 1.43 before categorizing each participant by LDL-cholesterol level.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

3960

Benn et al.

Familial Hypercholesterolemia

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

we estimated their relative contributions in predicting FH (Table 3). All individuals with plasma LDL-cholesterol of 8.5 mmol/liter or higher or a mutation had FH, because fulfilling one of these criteria alone is diagnostic for definite or probable FH. For the remaining criteria, odds ratios for FH were (in decreasing order) as follows: 1786 [95% confidence interval (CI) ⫽ 1349 –2366] for LDLcholesterol above 6.5 mmol/liter compared with below; 196 (144 –268) for LDL-cholesterol above 5.0 mmol/liter compared with below; 177 (91–342) for LDL-cholesterol above 4.0 mmol/liter compared with below; 3.0 (2.1– 4.2) for a patient with premature CAD vs. without; 1.9 (1.6 – 2.3) for a first degree relative with known elevated cholesterol levels vs. without; 1.7 (0.9 –3.1) for a patient with premature cerebral or peripheral disease vs. without; and 1.5 (1.1–1.9) for a first-degree relative with premature cardiovascular disease vs. without. Consistent with the above, the criteria with the highest positive predictive values for FH were LDL-cholesterol higher than 8.5 mmol/liter (100%), presence of a mutation (100%), and LDL-cholesterol higher than 6.5 mmol/liter (65%), with much lower positive predictive values for LDL-cholesterol higher than 5 mmol/liter (11%), LDLcholesterol higher than 4 mmol/liter (3%), and family history of premature cardiovascular disease or hypercholesterolemia (1%).

Lipid, lipoprotein, and apolipoprotein levels Lipid levels were measured in the nonfasting state. Median LDL-cholesterol levels as a function of time since last meal is shown in Supplemental Fig. 1. Range in median LDL-cholesterol levels due to time since last meal was 0.4 mmol/liter (0.3– 0.5 mmol/liter), with the lowest levels 0 –2 h after the last meal. Table 4 shows the plasma lipid, lipoprotein, and apolipoprotein concentrations according to FH diagnostic group and treatment with cholesterollowering medication. Individuals with a diagnosis of definite/probable FH not treated with cholesterol-lowering medication had 89% higher LDL-cholesterol levels (6.1 vs. 3.2 mmol/liter), 52% higher total cholesterol levels (8.5 vs. 5.6 mmol/liter), and 69% higher apolipoprotein B levels (186 vs. 110 mg/dl) compared with those unlikely to have FH. The corresponding values for individuals receiving cholesterol-lowering medication were 106% higher LDL-cholesterol levels (4.7 vs. 2.3 mmol/liter), 52% higher total cholesterol levels (7.1 vs. 4.7 mmol/liter), and 62% higher apolipoprotein B levels (155 vs. 95 mg/dl). Risk of CAD Among individuals not receiving cholesterol-lowering medication, the adjusted odds ratio for CAD was 13.2 (95% CI ⫽ 10.0 –17.4) for those with definite/probable FH and 4.8 (4.3–5.3) for those with possible FH relative

TABLE 3. Contribution of each criterion of the DLCN score to the diagnosis of definite or probable FH in the Copenhagen General Population Study FH Odds ratio (95% CI) Family history First-degree relative with premature cardiovascular 1.5 (1.1–1.9)a disease or First-degree relative with known elevated 1.9 (1.6 –2.3)a cholesterol levels Personal clinical history Premature CAD 3.0 (2.1– 4.2)a Premature cerebral or peripheral vascular disease 1.7 (0.9 –3.1)a Plasma cholesterol, LDL-cholesterol in mmol/liter ⱖ8.5 (330 mg/dl) All had FH ⱖ6.5 (250 mg/dl) 1786 (1349 –2366)b ⱖ5.0 (190 mg/dl) 196 (144 –268)b ⱖ4.0 (155 mg/dl) 177 (91–342)b ⬍4.0 (155 mg/dl) 0.006 (0.002– 0.01)c Genetic analysis LDLR (W23X, W66G, W556S) or APOB All had FH (R3500Q) mutation

P value

Positive predictive Negative predictive value (%) value (%)

0.003

1

99

⬍0.001

1

99

⬍0.001 0.11

2 1

99 99

⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001

100 65 11 3 0

99 99 99 99 97

100

99

Individuals fulfilling the DLCN criteria for a definite or probable diagnosis of FH are considered to have FH, whereas individuals categorized as possible or unlikely by the same criteria are not. a

Odds ratio for each criterion is risk in individuals fulfilling the specific criteria vs. those not fulfilling the same criteria used as reference group.

b

Odds ratio in groups by LDL-cholesterol levels is risk in individuals with an LDL-cholesterol level above the threshold vs. those below.

c

Risk in individuals is relative to individuals with an LDL-cholesterol level above 4.0 mmol/liter.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

jcem.endojournals.org

3961

TABLE 4. Lipid and apolipoprotein levels as a function of the DLCN criteria for a diagnosis of FH in individuals on and off cholesterol-lowering medication from the general population Diagnostic category of FH Cholesterol-lowering medication, off/on

Definite/probable Off

On

Possible Off

Unlikely On

Off

All On

Off

On

Number 260 242 3,380 915 58,158 6,061 61,798 7,218 LDL-cholesterol ( mmol/liter) 6.5 (5.1– 6.8)a 4.7 (3.8 –5.1)a 5.2 (5.0 –5.5)a 3.6 (2.8 –3.9)a 3.2 (2.7–3.8)a 2.3 (1.8 –2.7)a 3.3 (2.7–3.9)a 2.4 (1.9 –2.9)a LDL-cholesterol (mmol/liter); 6.5 (5.3– 6.9) 4.8 (3.9 –5.2) 5.2 (5.0 –5.5)a 3.5 (2.6 –3.8)a 3.5 (2.9 – 4.0)a 2.2 (1.8 –2.7)a 3.6 (3– 4.2)a 2.3 (1.8 –2.9)a metabolic syndrome LDL-cholesterol (mmol/liter); 5.6 (4.2– 6.5)a 4.7 (4.4 –5.2)a 5.1 (3.7–5.3)a 3.1 (2.1–3.7)a 3.0 (2.4 –3.5)a 1.9 (1.5–2.4)a 3.0 (2.4 –3.6)a 2.0 (1.6- (2.5)a diabetes mellitus Total cholesterol (mmol/liter) 8.6 (7.5–9.4)a 7.0 (6.3–7.6)a 7.6 (7.1– 8.1)a 5.9 (5.0 – 6.4)a 5.6 (4.9 – 6.3)a 4.7 (4.1–5.2)a 5.7 (5.0 – 6.4)a 4.8 (4.2–5.4)a Apolipoprotein B (mg/dl) 182 (158 –208)a 148 (127–173)a 159 (142–181)a 122 (104 –140)a 106 (88 –127)a 91 (76 –109)a 108 (89 –131)a 95 (79 –117)a Triglycerides (mmol/liter) 1.9 (1.3–2.7) 1.8 (1.2–2.6) 1.8 (1.3–2.5) 1.8 (1.3–2.6) 1.4 (1.0 –2.0)a 1.6 (1.1–2.3)a 1.4 (1.0 –2.1)a 1.6 (1.1–2.3)a HDL-cholesterol (mmol/liter) 1.4 (1.2–1.7) 1.4 (1.2–1.7) 1.4 (1.2–1.8) 1.4 (1.1–1.7) 1.6 (1.3–1.9)a 1.5 (1.2–1.9)a 1.6 (1.3–1.9)a 1.5 (1.2–1.8)a Apolipoprotein A1 (mg/dl) 158 (141–178) 163 (142–182) 161 (144 –179) 157 (140 –178) 157 (140 –178) 157 (139 –178) 157 (140 –178) 158 (139 –178)

Continuous values are median (interquartile range). To convert cholesterol values to milligrams per deciliter, multiply values in millimoles per liter by 38.6. To convert triglyceride values to milligrams per deciliter, multiply values in millimoles per liter by 89. a

P value ⬍0.001 for difference between individuals off and on cholesterol-lowering medication within diagnostic category.

to individuals with unlikely FH (Fig. 2). Results were similar in women and men. Among individuals receiving cholesterol-lowering medication, the corresponding odds ratios were 10.3 (7.8 –13.8) and 14.8 (12.9 –17.1). Excluding individuals with plasma triglyceride levels above 4 mmol/liter or with metabolic syndrome gave similar results (Supplemental Fig. 2).

Discussion This is the first study to assess the prevalence of FH in a large community-based population. The findings suggest that FH may be encountered in approximately one in 137 people in the community. This is more frequent than the commonly reported prevalence of one in 500 for heterozygous FH (3, 14, 15). Previous estimates of the frequency of FH have been derived from clinical registry populations or from the Hardy-Weinberg equation and the estimated prevalence of homozygous FH (3, 14 –16). Neither of these approaches truly reflects the status in the general community, where the phenotypic expression of FH may be less severe than in clinically established FH (26). The lower prevalence of FH in our younger subjects reflects their lower DLCN criteria scores owing to a lower prevalence of CAD (2) and fewer self-reported family histories of CAD and hypercholesterolemia. The higher prevalence of FH in women compared with men above the age of 79 yr clearly reflects the later onset of CAD in women in general and with FH (2, 14). The higher prevalence of smoking in participants with probable FH may reflect that this group has a high prevalence of participants fulfilling the DLCN criteria of premature coronary artery disease and cerebral or peripheral vascular disease and that smoking is a significant driver for these diseases.

We found that the prevalence of FH based on common LDLR and APOB mutations alone (12) was one in 650, which agrees with conventional estimates (3, 14, 15). Our genotypic data, based on common but a limited number of mutations (11, 12), are likely to underestimate the true genetically defined prevalence of FH. This is a limitation of the study; however, the true prevalence of genetically defined FH would only be found if all 69,016 individuals had a complete screen for mutations in all known candidate genes for FH. Pathogenic mutations may also not be identified in up to 30% of individuals with definite phenotypic FH (10, 27). Our prevalence of CAD in FH was high at 28% and is consistent with other reports in partially treated individuals (2, 17, 18, 28, 29). In the present study, the range in median LDL-cholesterol levels due to time since last meal was a modest 0.4 mmol/liter (0.3– 0.5 mmol/liter), with the lowest levels 0 –2 h after the last meal (30). Therefore, the postprandial state may result in a small but significant depression in plasma LDL-cholesterol levels (31), and this might have resulted in underestimation of the DLCN criteria score and hence the prevalence of FH. We concede that 11% of participants were taking cholesterol-lowering medication. However, we adjusted for this cholesterol-lowering effect when estimating the pretreatment LDL-cholesterol levels using a factor of 1.43 corresponding to an estimated 30% reduction in LDL-cholesterol for a median daily statin dose (23). This adjustment is an approximation and may have introduced a bias; however, in Denmark in the period from 1995–2007, 96% of individuals on lipid-lowering medication received statin, and of those, 68% received simvastatin with a 30% estimated reduction in LDL-cholesterol level for the median prescribed daily dose, 14% atorvastatin (35% median reduction in LDL-cholesterol),

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

3962

Benn et al.

Familial Hypercholesterolemia

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

That less than half of the individuals with FH were on cholesterollowering medication concurs with the shortfall in detection and management of people with FH in other populations (16, 33). Also, compared with non-FH subjects, individuals with FH had 13-fold and 10-fold risk of CAD in relation to being off and on cholesterol-lowering medication, respectively. The findings of increased prevalence of CAD in subjects with unlikely FH and possible FH on cholesterol-lowering medications compared with those not taking cholesterol-lowering medications are probably due to the cross-sectional design of the study; that is, those with CAD are more likely to be treated with cholesterollowering medications than those without CAD. Therefore, in the present study, one could also use the phrase relative prevalence instead of odds ratio for risk of CAD. The most striking data from this study show that despite having CAD in 84 subjects with definite or probable FH and in 604 with possible FH, they were not being treated with cholesterol-lowering medications. This clearly reveals undertreatment. The risk of CAD with respect to cholesterol-lowering medication is higher than previously reported in selected cohorts of clinic or registry patients with FH (28, 29). The cross-sectional study design, use of a community sample, bias in self-reported medication, and short FIG. 2. Risk of CAD as a function of the DLCN criteria for a diagnosis of FH in individuals on and off lipid-lowering medication from the general population. Upper panel, risk in all duration of treatment for a disorder participants; middle panel, risk in women; lower panel, risk in men. Odds ratios were with life-long risk of CAD could exmultivariable adjusted for gender, age, hypertension, diabetes mellitus, BMI, metabolic plain our less favorable findings relasyndrome, and smoking. tive to selected cohort studies (28, 29). European and Danish community stud9% pravastatin (25% median reduction in LDL-choles- ies have demonstrated that a significant proportion of terol), 4% lovastatin (30% median reduction in LDL-cho- people at risk of CAD are not being treated to recomlesterol), and 3% fluvastatin (25% median reduction in mended plasma levels of LDL-cholesterol, even if they LDL-cholesterol). are on a statin (34 –36). Among participants on cholesFH accelerates the rate of atherosclerosis, particularly terol-lowering medication, those categorized with posin coronary arteries and the proximal aorta (2). The high sible FH had a higher risk of CAD compared with those prevalence of CAD in our subjects considered to have FH with definite/probable FH. This could be due to less compared with those without FH concurs with other re- aggressive treatment in those with possible FH, both ports (2, 14, 17, 18). As suggested elsewhere, the preva- with lipid-lowering medication and of other CAD risk lence of subclinical atherosclerosis and asymptomatic factors; however, we do not have data to support this CAD is likely to be high in FH (2, 32). suggestion.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

The potentially high prevalence of FH in this Danish community is likely to reflect other populations not subject to a founder effect, although the mutation spectrum may differ from country to country. Our findings point to the need to actively seek index cases in the community and identify new cases in families using cascade screening employing both genetic and phenotypic criteria (6, 10). We emphasize that early detection of FH in children and the young is a particular challenge; recent guidelines suggest that all 9- to 11-yr-old children be routinely tested for hypercholesterolemia (37). A more practical approach would be to target individuals in the community with a family history of hypercholesterolemia and/or premature CAD (6, 9, 10). This could be based on an initial nonfasting lipid profile followed if required by a fasting lipid profile and a genetic test for FH (10, 31). Finally, our results also underscore a significant shortfall in the treatment of high-risk subjects with hypercholesterolemia (16, 33–35). This points to the need to institute early therapy with lifestyle modifications and statins to reduce their subsequent risk of CAD (6, 9, 10). Development of national models of care and health policy that integrate medical care between primary care physicians and specialists in tertiary hospitals is likely to achieve the best outcome for individuals with FH and other atherogenic dyslipidemias in our community. The diverse approaches to grappling with this issue among countries remain a major challenge for preventive medicine (21). In conclusion, the prevalence of FH may be higher than commonly perceived in a general population study of white Danish individuals. Also, only half of those with FH were on cholesterol-lowering medication. The very high risk of CAD irrespective of use of medication reflects the extent of underdiagnosis and undertreatment of FH in the community. This requires implementation of universal and targeted screening for FH, increased public awareness of the importance of hypercholesterolemia, and more education and training of primary care physicians on treating FH.

Acknowledgments We are indebted to the staff and participants of the Copenhagen General Population Study. We thank technicians Hanne Damm, Mette Refstrup, and Dorthe Uldall Andersen for expert technical assistance. Address all correspondence and requests for reprints to: Børge G. Nordestgaard, Professor, Chief Physician, M.D., D.M.Sc., Department of Clinical Biochemistry, Herlev Hospital, Copenhagen University Hospital, Herlev Ringvej 75, DK-2730 Herlev, Denmark. E-mail: [email protected].

jcem.endojournals.org

3963

This work was supported by The Danish Medical Research Council, Herlev Hospital, Copenhagen University Hospital, Copenhagen County Foundation, and Chief Physician Johan Boserup and Lise Boserup’s Fund, Denmark. Disclosure Summary: G.F.W. has received lecture and/or consultancy honoraria from AstraZeneca, Merck, Pfizer, Abbott, Sanofi-Aventis, Roche, Amgen, Novartis, Boehringer-Ingelheim, GlaxoWellcome, and Genfit; B.G.N. has received lecture and/or consultancy honoraria from AstraZeneca, Merck, Pfizer, Karo Bio, Omthera, Abbott, Sanofi-Aventis, and Regeneron. The other authors have no conflict of interest.

References 1. Watts GF, Lewis B, Sullivan DR 2007 Familial hypercholesterolemia: a missed opportunity in preventive medicine. Nat Clin Pract Cardiovasc Med 4:404 – 405 2. Austin MA, Hutter CM, Zimmern RL, Humphries SE 2004 Familial hypercholesterolemia and coronary heart disease: a HuGE association review. Am J Epidemiol 160:421– 429 3. Austin MA, Hutter CM, Zimmern RL, Humphries SE 2004 Genetic causes of monogenic heterozygous familial hypercholesterolemia: a HuGE prevalence review. Am J Epidemiol 160:407– 420 4. Soutar AK, Naoumova RP 2007 Mechanisms of disease: genetic causes of familial hypercholesterolemia. Nat Clin Pract Cardiovasc Med 4:214 –225 5. Williams RR, Hunt SC, Schumacher MC, Hegele RA, Leppert MF, Ludwig EH, Hopkins PN 1993 Diagnosing heterozygous familial hypercholesterolemia using new practical criteria validated by molecular genetics. Am J Cardiol 72:171–176 6. National Collaborating Centre for Primary Care 2008 Identification and management of familial hypercholesterolaemia. NICE clinical guideline 71. London: National Institute for Health and Clinical Excellence 7. World Health Organization 1999 Familial hypercholesterolaemia (FH). Report of a second WHO consultation. Geneva: World Health Organization 8. Hadfield SG, Humphries SE 2005 Implementation of cascade testing for the detection of familial hypercholesterolaemia. Curr Opin Lipidol 16:428 – 433 9. Catapano AL, Reiner Z, De BG, Graham I, Taskinen MR, Wiklund O, Agewall S, Alegria E, Chapman MJ, Durrington P, Erdine S, Halcox J, Hobbs R, Kjekshus J, Perrone FP, Riccardi G, Storey RF, Wood D 2011 ESC/EAS Guidelines for the management of dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Atheroscler 217(Suppl 1):S1–S44 10. Watts GF, Sullivan DR, Poplawski N, van Bockxmeer F, HamiltonCraig I, Clifton PM, O’Brien R, Bishop W, George P, Barter PJ, Bates T, Burnett JR, Coakley J, Davidson P, Emery J, Martin A, Farid W, Freeman L, Geelhoed E, Juniper A, Kidd A, Kostner K, Krass I, Livingston M, Maxwell S, et al. 2011 Familial hypercholesterolaemia: a model of care for Australasia. Atheroscler Suppl 12:221–263 11. Walma EP, Visseren FL, Jukema JW, Kastelein JJ, Hoes AW, Stalenhoef AF 2006 [The practice guideline ‘Diagnosis and treatment of familial hypercholesterolaemia’ of the Dutch Health Care Insurance Board]. Ned Tijdschr Geneeskd 150:18 –23 (Dutch) 12. Jensen HK, Jensen LG, Meinertz H, Hansen PS, Gregersen N, Faergeman O 1999 Spectrum of LDL receptor gene mutations in Denmark: implications for molecular diagnostic strategy in heterozygous familial hypercholesterolemia. Atherosclerosis 146: 337–344 13. Damgaard D, Larsen ML, Nissen PH, Jensen JM, Jensen HK, Soerensen VR, Jensen LG, Faergeman O 2005 The relationship of

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.

3964

14.

15.

16.

17. 18.

19.

20.

21.

22.

23.

24.

25.

Benn et al.

Familial Hypercholesterolemia

molecular genetic to clinical diagnosis of familial hypercholesterolemia in a Danish population. Atherosclerosis 180:155–160 Marks D, Thorogood M, Neil HA, Humphries SE 2003 A review on the diagnosis, natural history, and treatment of familial hypercholesterolaemia. Atherosclerosis 168:1–14 Hopkins PN, Toth PP, Ballantyne CM, Rader DJ 2011 Familial hypercholesterolemias: prevalence, genetics, diagnosis and screening recommendations from the National Lipid Association Expert Panel on Familial Hypercholesterolemia. J Clin Lipidol 5:S9 –S17 Neil HA, Hammond T, Huxley R, Matthews DR, Humphries SE 2000 Extent of underdiagnosis of familial hypercholesterolaemia in routine practice: prospective registry study. BMJ 321:148 Scientific Steering Committee SBRG 1991 Risk of fatal coronary heart disease in familial hypercholesterolaemia. BMJ 303:893– 896 Sijbrands EJ, Westendorp RG, Defesche JC, de Meier PH, Smelt AH, Kastelein JJ 2001 Mortality over two centuries in large pedigree with familial hypercholesterolaemia: family tree mortality study. BMJ 322:1019 –1023 Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A 2007 Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease, and death in men and women. JAMA 298:299 –308 Zacho J, Tybjaerg-Hansen A, Jensen JS, Grande P, Sillesen H, Nordestgaard BG 2008 Genetically elevated C-reactive protein and ischemic vascular disease. N Engl J Med 359:1897–1908 Aarden E, Van Hoyweghen I, Horstman K 2011 The paradox of public health genomics: definition and diagnosis of familial hypercholesterolaemia in three European countries. Scand J Public Health 39:634 – 639 Reiner Z, Catapano AL, De Backer G, Graham I, Taskinen MR, Wiklund O, Agewall S, Alegria E, Chapman MJ, Durrington P, Erdine S, Halcox J, Hobbs R, Kjekshus J, Filardi PP, Riccardi G, Storey RF, Wood D, Bax J, Vahanian A, Auricchio A, Baumgartner H, Ceconi C, Dean V, Deaton C, et al. 2011 ESC/EAS Guidelines for the management of dyslipidaemias. Eur Heart J 32:1769 –1818 Jones PH, Davidson MH, Stein EA, Bays HE, McKenney JM, Miller E, Cain VA, Blasetto JW 2003 Comparison of the efficacy and safety of rosuvastatin versus atorvastatin, simvastatin, and pravastatin across doses (STELLAR* Trial). Am J Cardiol 92:152–160 Whitworth JA; World Health Organization; International Society of Hypertension Writing Group 2003 2003 World Health Organization (WHO)/International Society of Hypertension (ISH) statement on management of hypertension. J Hypertens 21:1983–1992 Alberti KG, Eckel RH, Grundy SM, Zimmet PZ, Cleeman JI, Donato KA, Fruchart JC, James WP, Loria CM, Smith Jr SC; International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; International Association for the Study of Obesity 2009 Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology

J Clin Endocrinol Metab, November 2012, 97(11):3956 –3964

26.

27.

28.

29.

30.

31. 32.

33.

34.

35.

36.

37.

and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 120:1640 –1645 Tybjaerg-Hansen A, Jensen HK, Benn M, Steffensen R, Jensen G, Nordestgaard BG 2005 Phenotype of heterozygotes for low-density lipoprotein receptor mutations identified in different background populations. Arterioscler Thromb Vasc Biol 25:211–215 Humphries SE, Norbury G, Leigh S, Hadfield SG, Nair D 2008 What is the clinical utility of DNA testing in patients with familial hypercholesterolaemia? Curr Opin Lipidol 19:362–368 Versmissen J, Oosterveer DM, Yazdanpanah M, Defesche JC, Basart DC, Liem AH, Heeringa J, Witteman JC, Lansberg PJ, Kastelein JJ, Sijbrands EJ 2008 Efficacy of statins in familial hypercholesterolaemia: a long term cohort study. BMJ 337:a2423 Neil A, Cooper J, Betteridge J, Capps N, McDowell I, Durrington P, Seed M, Humphries SE 2008 Reductions in all-cause, cancer, and coronary mortality in statin-treated patients with heterozygous familial hypercholesterolaemia: a prospective registry study. Eur Heart J 29:2625–2633 Langsted A, Freiberg JJ, Nordestgaard BG 2008 Fasting and nonfasting lipid levels: influence of normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk prediction. Circulation 118:2047–2056 Watts GF, Cohn JS 2011 Whither the lipid profile: feast, famine, or no free lunch? Clin Chem 57:363–365 Descamps OS, de Meester A, Cheron P, Kastelein JJ, Heller FR 2003 Silent ischaemia in familial hypercholesterolemia. Atheroscler Suppl 4:7– 8 Pijlman AH, Huijgen R, Verhagen SN, Imholz BP, Liem AH, Kastelein JJ, Abbink EJ, Stalenhoef AF, Visseren FL 2010 Evaluation of cholesterol lowering treatment of patients with familial hypercholesterolemia: a large cross-sectional study in The Netherlands. Atherosclerosis 209:189 –194 Kotseva K, Wood D, De Backer G, De Bacquer D, Pyo¨ra¨la¨ K, Reiner Z, Keil U 2010 EUROASPIRE III Management of cardiovascular risk factors in asymptomatic high-risk patients in general practice: cross-sectional survey in 12 European countries. Eur J Cardiovasc Prev Rehabil 17:530 –540 Kotseva K, Wood D, De Backer G, De Bacquer D, Pyo¨ra¨la¨ K, Keil U 2009 EUROASPIRE III: a survey on the lifestyle, risk factors and use of cardioprotective drug therapies in coronary patients from 22 European countries. Eur J Cardiovasc Prev Rehabil 16:121–137 Siggaard-Andersen N, Freiberg JJ, Nordestgaard BG 2012 Only a fraction of patients with ischaemic diseases or diabetes are treated to recommended target values for plasma lipids. Dan Med J 59:A4470 Daniels SR, Gidding SS, de Ferranti SD 2011 Pediatric aspects of familial hypercholesterolemias: recommendations from the National Lipid Association Expert Panel on Familial Hypercholesterolemia. J Clin Lipidol 5:S30 –S37

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 18 January 2017. at 09:55 For personal use only. No other uses without permission. . All rights reserved.