A Comparative Analysis of Phenotypic Predictors of Mutations in Familial Hypercholesterolemia

The Journal of Clinical Endocrinology & Metabolism, Apr 2018

The gold standard for diagnosing familial hypercholesterolemia (FH) is identification of a causative pathogenic mutation. However, genetic testing is expensive and not widely available.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://academic.oup.com/jcem/article-pdf/103/4/1704/24565765/jc.2017-02622.pdf

A Comparative Analysis of Phenotypic Predictors of Mutations in Familial Hypercholesterolemia

Abstract Context The gold standard for diagnosing familial hypercholesterolemia (FH) is identification of a causative pathogenic mutation. However, genetic testing is expensive and not widely available. Objective To compare the validity of the Dutch Lipid Clinic Network (DLCN), Simon Broome (SB), Make Early Diagnosis to Prevent Early Deaths (MEDPED), and American Heart Association (AHA) criteria in predicting an FH-causing mutation. Design, Setting, and Patients An adult cohort of unrelated patients referred to a lipid clinic for genetic testing. Main Outcome Measures Odds ratio (OR), area under the curve (AUC), sensitivity, and specificity. Results A pathogenic FH-causing mutation was detected in 30% of 885 patients tested. Elevated low-density lipoprotein (LDL) cholesterol and personal or family history of tendon xanthomata were independent predictors of a mutation (OR range 5.3 to 16.1, P < 0.001). Prediction of a mutation for the DLCN and SB definite and MEDPED criteria (ORs 9.4, 11.7, and 10.5, respectively) was higher than with the AHA criteria (OR 4.67). The balance of sensitivity and specificity was in decreasing order DLCN definite (Youden Index 0.487), MEDPED (0.457), SB definite (0.274), and AHA criteria (0.253), AUC being significantly higher with DLCN definite and MEDPED than other criteria (P < 0.05). Pretreatment LDL cholesterol and tendon xanthomata had the highest AUC in predicting a mutation. Conclusions The DLCN, SB, and MEDPED criteria are valid predictors of an FH-causing mutation in patients referred to a lipid clinic, but concordance between these phenotypic criteria is only moderate. Use of pretreatment LDL cholesterol and tendon xanthomata alone may be particularly useful for deciding who should be genetically tested for FH. Familial hypercholesterolemia (FH) is a codominantly inherited disorder resulting in elevated low-density lipoprotein (LDL) cholesterol concentration and increased risk of premature atherosclerotic cardiovascular disease (ASCVD) (1). FH is primarily caused by mutations in the LDL receptor (LDLR), and less commonly by mutations in apolipoprotein B (APOB) and proprotein convertase subtilisin/kexin type 9 (PCSK9) genes (2). The prevalence of heterozygous FH is estimated to be 1 in 250 in unselected community populations (3–5). Although genetic testing is the reference standard for diagnosing FH and has great value in risk assessment and cascade screening, it is expensive and not widely available (6). Deciding whether to undertake a genetic test in patients with phenotypic FH referred to a clinic is a key consideration. The clinical diagnosis of FH is based on identifying a personal and family history of premature coronary artery disease (CAD), tendon xanthomata, corneal arcus, and elevated LDL cholesterol (7–9). However, there is no standardized phenotypic diagnosis of FH (10). The most commonly recommended diagnostic tools include the Dutch Lipid Clinic Network (DLCN), the Simon Broome (SB), and Make Early Diagnosis to Prevent Early Deaths (MEDPED) criteria (7–9). However, a detailed family history and physical stigmata required by the DLCN and SB criteria may be difficult to elicit. A simplified diagnosis of FH, based on LDL cholesterol level and family history of elevated LDL cholesterol or premature CAD, was recently proposed by the American Heart Association (AHA) (11). We previously reported on the characteristics of patients with FH in a service for lipid disorders in Western Australia (6, 12). We now compare the diagnostic validity of the aforementioned clinical tools in discriminating between the presence and absence of a pathogenic gene variant causative of FH. We thereby aimed to identify the optimal phenotype for predicting a mutation and how best to use genetic testing for FH in the clinic. Materials and Methods Study design A cross-sectional comparison was undertaken of unrelated patients aged ≥18 years with or without a recognized mutation causative of FH. The care of these patients has been previously described (6, 12). In brief, adult patients with suspected FH were recruited via referral from general practice, coronary care, or other specialists. Reasons for suspecting FH were generally an elevated plasma LDL cholesterol with a family or personal history of CAD. Patients were evaluated by specialist physicians at the Lipid Disorders Clinic at Royal Perth Hospital (Perth, Western Australia); if the diagnosis was at least possible FH, they were genetically tested after appropriate counseling and consent. The study was approved by the Royal Perth Hospital Human Research Ethics Committee. Clinical data Age, gender, history of coronary and other vascular disease, diabetes, hypertension, smoking status, examination findings (weight, height, blood pressure), and medications were recorded. The definitions of CAD, diabetes, hypertension, and obesity were previously described (6, 12). Phenotypic diagnosis of FH Details were obtained of family and personal history of hypercholesterolemia, premature coronary or vascular disease, tendon xanthomata (bilateral, subcutaneous nodules on Achilles tendons or at ligamentous insertions), and corneal arcus; pretreatment plasma LDL cholesterol concentration was recorded. Where a pretreatment plasma lipid profile was not available (25% of the cohort), a correction that accounted for type, dose, and frequency of statins and use of ezetimibe (Supplemental Table 1) was used to estimate pretreatment LDL cholesterol concentrations (13). Patients were classified phenotypically, using the same individual variable, according to the DLCN, SB, MEDPED, and AHA diagnostic criteria (Supplemental Table 2) (7-11). Because ∼30% of Lp(a) mass is cholesterol (14), in subsidiary analyses LDL cholesterol was also adjusted for the cholesterol content of Lp(a) by subtracting 30% of total Lp(a) mass from the measured LDL cholesterol concentration. Genetic diagnosis of FH The diagnosis of FH was based on the presence of a pathogenic gene variant. DNA extraction, multiplex ligation-dependent probe amplification, and Sanger sequencing of the 18 exons of the LDLR, exon 7 of PCSK9, and part of APOB exons 26 and 29 were performed, as previously described (6). In 10% of patients, genetic testing was based on next-generation sequencing (NGS), performed by Ion Torrent sequencing using a TargetSeq (Life Technologies, Waltham, MA) custom capture panel of lipid genes and polymorphisms, derived from LipidSEquation (15). Mutation detection rates with NGS (29%) and Sanger sequencing (30%) were not significantly different. Multiplex ligation-dependent probe amplification was performed in all patients with DLCN probable or definite FH in whom Sanger sequencing or NGS did not identify a mutation. Laboratory analyses Venous blood samples were collected during clinic visits. Plasma cholesterol, high-density lipoprotein (HDL) cholesterol, and triglyceride concentrations were determined by standard enzymatic methods. LDL cholesterol was calculated by the Friedewald Equation (16), but in patients with plasma triglyceride >4.5 mmol/L, a direct LDL cholesterol assay was used. Total apoB and Lp(a) [Quantia Lp(a) assay and standards] were determined by immunoassay (Abbott Laboratories, Abbott Park, IL), with interassay coefficient of variation of <5%. Statistical analyses All data were analyzed using SPSS (version 21, Chicago, IL). Differences in clinical characteristics between FH index cases with and without a pathogenic mutation (LDLR, APOB, and PCSK9) were tested with χ2 or independent t tests. Skewed variables, including plasma triglyceride and Lp(a) concentrations, were log-transformed. Logistic regression was used to assess prediction of a mutation. The discriminant value of plasma LDL cholesterol concentration and the DLCN score in predicting a mutation were depicted using receiver operator characteristic (ROC) area under the curve (AUC). Cohen’s κ coefficient was used to measure the phenotypic agreement between DLCN, SB, MEDPED, and AHA criteria. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and Youden index were used to compare the diagnostic value of the clinical tools and their components in predicting a mutation. Improvement in prediction of a mutation with individual variable relative to the DLCN and SB definitions was estimated as differences between the AUCs and compared using C-statistics. Statistical significance was defined at the 5% level. Results A total of 931 adult patients with at least possible FH according to the DLCN criterion (score ≥3) was genetically screened for a causative mutation for FH from March 2007 to September 2017. Patients with homozygous FH, compound or double heterozygous FH (n = 6), LDLR or APOB variants of uncertain significance (n = 30), pathogenic or uncertain ABCG5/8 variants (n = 5), or those with frank secondary causes of hypercholesterolemia (n = 5) were excluded from the analysis. Of the remaining 885 index cases, 267 (30%) of the screened cases had a pathogenic FH-causing mutation: 237 were found to have mutation in the LDLR gene (60% missense, 20% nonsense/frameshift, 11% splice, 8% large deletion/duplication, and 1% promoter variants), whereas others were heterozygous for APOB (n = 25), PCSK9 (n = 3), and APOE p.Leu167del (n = 2) mutations. The characteristics of the 885 patients with (M+) and without (M−) a mutation are shown in Table 1. M+ individuals were older, and had a higher frequency of non-Caucasian origin, tendon xanthomata, and corneal arcus (P < 0.01 for all) than M− patients. There were no significant differences in the proportion of men, smokers, or patients with CAD or cerebral/peripheral vascular disease, type 2 diabetes or obesity, and use of antihypertensive agents or aspirin between the M+ and M− groups. Compared with M− individuals, M+ individuals had higher total cholesterol, non-HDL cholesterol, LDL cholesterol (treated and pretreatment), and apoB levels, and more frequent use of lipid-lowering medication than M− individuals (P < 0.01 for all). Plasma HDL cholesterol and Lp(a) levels were not significantly different between the two groups (P > 0.05 for both), but plasma triglyceride levels were significantly higher in the M− group. M+ individuals had a higher DLCN score (10.6 ± 4.0 vs 6.2 ± 2.5) and prevalence of DLCN definite FH (65% vs 17%), SB definite FH (31% vs 4%), MEDPED-defined FH (79% vs 26%), and AHA-defined FH (89% vs 64%) than M− individuals (P < 0.001 for all). Table 1. Demographic, Clinical, and Biochemical Characteristics of the Patients With FH in Relation to Presence (Mutation+) and Absence (Mutation−) of an FH-Causing Mutation  Total Mutation+ Mutation− P No. of subjects 885 267 618  Age at diagnosis, y 51 ± 13 48 ± 14 52 ± 13 <0.001 Male, % 46 41 47 0.091 Ethnicity    0.004  Caucasian, % 89 84 91   Asian, % 6 10 5   Other (Mixed/Aboriginal/African/Middle Eastern), % 5 6 4  Early CAD, % 44 48 43 0.141 Early cerebral or peripheral vascular disease, % 4 5 4 0.719 Tendon xanthomata, % 12 31 4 <0.001 Corneal arcus, % 27 36 24 0.001 Smoking (ever), %) 42 41 42 0.767 Hypertension, % 29 23 32 0.012 Type 2 diabetes, % 7 5 7 0.375 Obesity (body mass index >30 kg/m2), % 30 28 32 0.386 On lipid-lowering medication, % 66 75 63 0.001 On antihypertensive medication, % 29 27 30 0.422 On aspirin, % 28 31 27 0.223 Total cholesterol, mmol/L 5.7 ± 1.8 6.0 ± 2.0 5.6 ± 1.7 0.018 Triglyceride, mmol/L 1.8 ± 1.1 1.4 ± 0.7 1.9 ± 1.2 <0.001 HDL cholesterol, mmol/L 1.4 ± 0.6 1.3 ± 0.4 1.4 ± 0.6 0.168 Non-HDL cholesterol, mmol/L 4.4 ± 1.8 4.6 ± 2.0 4.3 ± 1.6 0.007 LDL cholesterol, mmol/L 3.6 ± 1.6 3.9 ± 1.9 3.4 ± 1.5 <0.001 Pretreated LDL cholesterol, mmol/L 6.6 ± 1.7 8.1 ± 1.8 6.0 ± 1.1 <0.001 Lipoprotein(a), g/L 0.30 (0.28–0.32) 0.27 (0.24–0.31) 0.31 (0.29–0.35) 0.073 ApoB, g/L 1.1 ± 0.4 1.2 ± 0.4 1.1 ± 0.3 0.001 DLCN criteria scorea 7.6 ± 3.6 10.6 ± 4.0 6.2 ± 2.5 <0.001 DLCN categoryb      Definite FH, % 31 65 17 <0.001  Probable FH, % 36 27 40 <0.001  Possible FH, % 33 8 43 <0.001 SBb      Definite FH, % 12 31 4 <0.001  Possible FH, % 63 63 63 0.685 MEDPED probable heterozygous FH, % 42 79 26 <0.001 AHA heterozygous FH, % 71 89 64 <0.001   Total Mutation+ Mutation− P No. of subjects 885 267 618  Age at diagnosis, y 51 ± 13 48 ± 14 52 ± 13 <0.001 Male, % 46 41 47 0.091 Ethnicity    0.004  Caucasian, % 89 84 91   Asian, % 6 10 5   Other (Mixed/Aboriginal/African/Middle Eastern), % 5 6 4  Early CAD, % 44 48 43 0.141 Early cerebral or peripheral vascular disease, % 4 5 4 0.719 Tendon xanthomata, % 12 31 4 <0.001 Corneal arcus, % 27 36 24 0.001 Smoking (ever), %) 42 41 42 0.767 Hypertension, % 29 23 32 0.012 Type 2 diabetes, % 7 5 7 0.375 Obesity (body mass index >30 kg/m2), % 30 28 32 0.386 On lipid-lowering medication, % 66 75 63 0.001 On antihypertensive medication, % 29 27 30 0.422 On aspirin, % 28 31 27 0.223 Total cholesterol, mmol/L 5.7 ± 1.8 6.0 ± 2.0 5.6 ± 1.7 0.018 Triglyceride, mmol/L 1.8 ± 1.1 1.4 ± 0.7 1.9 ± 1.2 <0.001 HDL cholesterol, mmol/L 1.4 ± 0.6 1.3 ± 0.4 1.4 ± 0.6 0.168 Non-HDL cholesterol, mmol/L 4.4 ± 1.8 4.6 ± 2.0 4.3 ± 1.6 0.007 LDL cholesterol, mmol/L 3.6 ± 1.6 3.9 ± 1.9 3.4 ± 1.5 <0.001 Pretreated LDL cholesterol, mmol/L 6.6 ± 1.7 8.1 ± 1.8 6.0 ± 1.1 <0.001 Lipoprotein(a), g/L 0.30 (0.28–0.32) 0.27 (0.24–0.31) 0.31 (0.29–0.35) 0.073 ApoB, g/L 1.1 ± 0.4 1.2 ± 0.4 1.1 ± 0.3 0.001 DLCN criteria scorea 7.6 ± 3.6 10.6 ± 4.0 6.2 ± 2.5 <0.001 DLCN categoryb      Definite FH, % 31 65 17 <0.001  Probable FH, % 36 27 40 <0.001  Possible FH, % 33 8 43 <0.001 SBb      Definite FH, % 12 31 4 <0.001  Possible FH, % 63 63 63 0.685 MEDPED probable heterozygous FH, % 42 79 26 <0.001 AHA heterozygous FH, % 71 89 64 <0.001  Values are expressed as mean ± standard deviation or geometric mean (95% CI); P value for Mutation+ vs Mutation−. a DLCN score refers to the score without the genetic analysis component [i.e., family history, clinical history, physical examination (tendon xanthomata and corneal arcus), and LDL cholesterol level only]; tendon xanthomata were diagnosed with reference to the Achilles tendons. b DLCN and SB FH categories do include the genetic analysis component into the classification. View Large Table 1. Demographic, Clinical, and Biochemical Characteristics of the Patients With FH in Relation to Presence (Mutation+) and Absence (Mutation−) of an FH-Causing Mutation  Total Mutation+ Mutation− P No. of subjects 885 267 618  Age at diagnosis, y 51 ± 13 48 ± 14 52 ± 13 <0.001 Male, % 46 41 47 0.091 Ethnicity    0.004  Caucasian, % 89 84 91   Asian, % 6 10 5   Other (Mixed/Aboriginal/African/Middle Eastern), % 5 6 4  Early CAD, % 44 48 43 0.141 Early cerebral or peripheral vascular disease, % 4 5 4 0.719 Tendon xanthomata, % 12 31 4 <0.001 Corneal arcus, % 27 36 24 0.001 Smoking (ever), %) 42 41 42 0.767 Hypertension, % 29 23 32 0.012 Type 2 diabetes, % 7 5 7 0.375 Obesity (body mass index >30 kg/m2), % 30 28 32 0.386 On lipid-lowering medication, % 66 75 63 0.001 On antihypertensive medication, % 29 27 30 0.422 On aspirin, % 28 31 27 0.223 Total cholesterol, mmol/L 5.7 ± 1.8 6.0 ± 2.0 5.6 ± 1.7 0.018 Triglyceride, mmol/L 1.8 ± 1.1 1.4 ± 0.7 1.9 ± 1.2 <0.001 HDL cholesterol, mmol/L 1.4 ± 0.6 1.3 ± 0.4 1.4 ± 0.6 0.168 Non-HDL cholesterol, mmol/L 4.4 ± 1.8 4.6 ± 2.0 4.3 ± 1.6 0.007 LDL cholesterol, mmol/L 3.6 ± 1.6 3.9 ± 1.9 3.4 ± 1.5 <0.001 Pretreated LDL cholesterol, mmol/L 6.6 ± 1.7 8.1 ± 1.8 6.0 ± 1.1 <0.001 Lipoprotein(a), g/L 0.30 (0.28–0.32) 0.27 (0.24–0.31) 0.31 (0.29–0.35) 0.073 ApoB, g/L 1.1 ± 0.4 1.2 ± 0.4 1.1 ± 0.3 0.001 DLCN criteria scorea 7.6 ± 3.6 10.6 ± 4.0 6.2 ± 2.5 <0.001 DLCN categoryb      Definite FH, % 31 65 17 <0.001  Probable FH, % 36 27 40 <0.001  Possible FH, % 33 8 43 <0.001 SBb      Definite FH, % 12 31 4 <0.001  Possible FH, % 63 63 63 0.685 MEDPED probable heterozygous FH, % 42 79 26 <0.001 AHA heterozygous FH, % 71 89 64 <0.001   Total Mutation+ Mutation− P No. of subjects 885 267 618  Age at diagnosis, y 51 ± 13 48 ± 14 52 ± 13 <0.001 Male, % 46 41 47 0.091 Ethnicity    0.004  Caucasian, % 89 84 91   Asian, % 6 10 5   Other (Mixed/Aboriginal/African/Middle Eastern), % 5 6 4  Early CAD, % 44 48 43 0.141 Early cerebral or peripheral vascular disease, % 4 5 4 0.719 Tendon xanthomata, % 12 31 4 <0.001 Corneal arcus, % 27 36 24 0.001 Smoking (ever), %) 42 41 42 0.767 Hypertension, % 29 23 32 0.012 Type 2 diabetes, % 7 5 7 0.375 Obesity (body mass index >30 kg/m2), % 30 28 32 0.386 On lipid-lowering medication, % 66 75 63 0.001 On antihypertensive medication, % 29 27 30 0.422 On aspirin, % 28 31 27 0.223 Total cholesterol, mmol/L 5.7 ± 1.8 6.0 ± 2.0 5.6 ± 1.7 0.018 Triglyceride, mmol/L 1.8 ± 1.1 1.4 ± 0.7 1.9 ± 1.2 <0.001 HDL cholesterol, mmol/L 1.4 ± 0.6 1.3 ± 0.4 1.4 ± 0.6 0.168 Non-HDL cholesterol, mmol/L 4.4 ± 1.8 4.6 ± 2.0 4.3 ± 1.6 0.007 LDL cholesterol, mmol/L 3.6 ± 1.6 3.9 ± 1.9 3.4 ± 1.5 <0.001 Pretreated LDL cholesterol, mmol/L 6.6 ± 1.7 8.1 ± 1.8 6.0 ± 1.1 <0.001 Lipoprotein(a), g/L 0.30 (0.28–0.32) 0.27 (0.24–0.31) 0.31 (0.29–0.35) 0.073 ApoB, g/L 1.1 ± 0.4 1.2 ± 0.4 1.1 ± 0.3 0.001 DLCN criteria scorea 7.6 ± 3.6 10.6 ± 4.0 6.2 ± 2.5 <0.001 DLCN categoryb      Definite FH, % 31 65 17 <0.001  Probable FH, % 36 27 40 <0.001  Possible FH, % 33 8 43 <0.001 SBb      Definite FH, % 12 31 4 <0.001  Possible FH, % 63 63 63 0.685 MEDPED probable heterozygous FH, % 42 79 26 <0.001 AHA heterozygous FH, % 71 89 64 <0.001  Values are expressed as mean ± standard deviation or geometric mean (95% CI); P value for Mutation+ vs Mutation−. a DLCN score refers to the score without the genetic analysis component [i.e., family history, clinical history, physical examination (tendon xanthomata and corneal arcus), and LDL cholesterol level only]; tendon xanthomata were diagnosed with reference to the Achilles tendons. b DLCN and SB FH categories do include the genetic analysis component into the classification. View Large In univariate regression analysis (Table 2), family history of ASCVD (premature CAD and/or cerebral/peripheral vascular disease) or elevated LDL cholesterol and presence of tendon xanthomata, corneal arcus, and elevated LDL cholesterol were all significant predictors of a mutation [odds ratio (OR) range 3.00 to 27.0, P < 0.001 for all]. Using the DLCN criteria, the ORs for an FH-causing mutation were 9.47 [95% confidence interval (CI) 6.81 to 13.2] and 8.58 (5.40 to 13.6) for definite and for probable/definite FH, respectively. The ORs for SB definite, possible/definite FH, MEDPED, and AHA criteria were 11.7 (7.15 to 19.1), 7.79 (4.57 to 13.3), 10.5 (7.48 to 14.9), and 4.67 (3.07 to 7.09), respectively. Age and ethnicity were also significant predictors of a mutation (OR 0.977, 95% CI 0.966 to 0.987 and 2.02, 95% CI 1.32 to 3.09, respectively); i.e., mutations were more likely in younger and non-Caucasian patients. As seen in Table 2, family history, physical stigmata, and elevated LDL cholesterol remained independent predictors of the presence of a mutation in multivariate regression models. In stepwise regression analysis, including tendon xanthomata, family history of physical stigmata, and the DLCN definite criteria, elevated LDL cholesterol (≥8.5 mmol/L) was the best predictor of the presence of a mutation (β-coefficient 2.22, standard error 0.351). The β-coefficients for tendon xanthomata, family history of physical stigmata, and DLCN definite criteria were 0.924 (standard error 0.319), 2.01 (0.278), and 1.01 (0.235), respectively. After adjusting for Lp(a) cholesterol, elevated LDL cholesterol remained an independent predictor of the presence of a mutation in the multivariate regression models. Inclusion of age, gender, and ethnicity as independent variables also did not alter the significance of the aforementioned variables in detecting a mutation. Table 2. Predictors of the Presence of an FH-Causing Mutation According to Individual Phenotypic Criteria and Composite Diagnostic Tools for FH  Univariate Logistic Regression Multivariate Logistic Regression  OR 95% CI P OR a 95% CI P Individual phenotypic criteria       Family history        Premature CAD and/or vascular disease or elevated LDL cholesterol 3.00 1.92–4.67 <0.001 3.24 1.86–5.64 <0.001  Tendon xanthomata and/or corneal arcus or children with elevated LDL cholesterol 12.0 7.39–19.4 <0.001 8.65 4.98–15.03 <0.001 Clinical history        Presence of premature CAD 1.24 0.93–1.66 0.139     Presence of premature cerebral or PVD 0.82 0.39–1.71 0.597    Physical examination        Tendinous xanthomata 10.7 6.64–17.2 <0.001 5.32 2.94–9.61 <0.001  Corneal arcus 1.75 1.29–2.40 <0.001 1.55 1.04–2.31 0.033 Pretreated LDL cholesterol, mmol/L        LDL cholesterolb 2.81 2.41–3.27 <0.001     LDL cholesterol ≥8.5 27.0 14.8–49.3 <0.001 16.1 8.42–30.6 <0.001  LDL cholesterol ≥6.5 8.77 6.20–12.4 <0.001 7.54 5.11–11.2 <0.001c  LDL cholesterol ≥5.0 15.2 4.77–48.5 <0.001 14.9 4.48–49.3 <0.001c Composite diagnostic tools       DLCN scoreb 1.49 1.41–1.58 <0.001    DLCN definite FHd 9.47 6.81–13.2 <0.001    DLCN probable/definite FHd 8.58 5.40–13.6 <0.001    SB definite FHd 11.7 7.15–19.1 <0.001    SB possible/definite FH 7.79 4.57–13.3 <0.001    MEDPED probable heterozygous FH 10.5 7.48–14.9 <0.001    AHA heterozygous FH 4.67 3.07–7.09 <0.001      Univariate Logistic Regression Multivariate Logistic Regression  OR 95% CI P OR a 95% CI P Individual phenotypic criteria       Family history        Premature CAD and/or vascular disease or elevated LDL cholesterol 3.00 1.92–4.67 <0.001 3.24 1.86–5.64 <0.001  Tendon xanthomata and/or corneal arcus or children with elevated LDL cholesterol 12.0 7.39–19.4 <0.001 8.65 4.98–15.03 <0.001 Clinical history        Presence of premature CAD 1.24 0.93–1.66 0.139     Presence of premature cerebral or PVD 0.82 0.39–1.71 0.597    Physical examination        Tendinous xanthomata 10.7 6.64–17.2 <0.001 5.32 2.94–9.61 <0.001  Corneal arcus 1.75 1.29–2.40 <0.001 1.55 1.04–2.31 0.033 Pretreated LDL cholesterol, mmol/L        LDL cholesterolb 2.81 2.41–3.27 <0.001     LDL cholesterol ≥8.5 27.0 14.8–49.3 <0.001 16.1 8.42–30.6 <0.001  LDL cholesterol ≥6.5 8.77 6.20–12.4 <0.001 7.54 5.11–11.2 <0.001c  LDL cholesterol ≥5.0 15.2 4.77–48.5 <0.001 14.9 4.48–49.3 <0.001c Composite diagnostic tools       DLCN scoreb 1.49 1.41–1.58 <0.001    DLCN definite FHd 9.47 6.81–13.2 <0.001    DLCN probable/definite FHd 8.58 5.40–13.6 <0.001    SB definite FHd 11.7 7.15–19.1 <0.001    SB possible/definite FH 7.79 4.57–13.3 <0.001    MEDPED probable heterozygous FH 10.5 7.48–14.9 <0.001    AHA heterozygous FH 4.67 3.07–7.09 <0.001     a Multivariate logistic regression (variables enter on model: family history of premature CAD and/or vascular disease or elevated LDL cholesterol, family history of tendon xanthomata and/or corneal arcus, tendon xanthomata, corneal arcus, and pretreated LDL cholesterol ≥8.5 mmol/L). b Expressed as a continuous variable. c Replacing LDL cholesterol ≥8.5 mmol/L with lower LDL cholesterol cutoff. d Refers to the classification without the genetic analysis component. View Large Table 2. Predictors of the Presence of an FH-Causing Mutation According to Individual Phenotypic Criteria and Composite Diagnostic Tools for FH  Univariate Logistic Regression Multivariate Logistic Regression  OR 95% CI P OR a 95% CI P Individual phenotypic criteria       Family history        Premature CAD and/or vascular disease or elevated LDL cholesterol 3.00 1.92–4.67 <0.001 3.24 1.86–5.64 <0.001  Tendon xanthomata and/or corneal arcus or children with elevated LDL cholesterol 12.0 7.39–19.4 <0.001 8.65 4.98–15.03 <0.001 Clinical history        Presence of premature CAD 1.24 0.93–1.66 0.139     Presence of premature cerebral or PVD 0.82 0.39–1.71 0.597    Physical examination        Tendinous xanthomata 10.7 6.64–17.2 <0.001 5.32 2.94–9.61 <0.001  Corneal arcus 1.75 1.29–2.40 <0.001 1.55 1.04–2.31 0.033 Pretreated LDL cholesterol, mmol/L        LDL cholesterolb 2.81 2.41–3.27 <0.001     LDL cholesterol ≥8.5 27.0 14.8–49.3 <0.001 16.1 8.42–30.6 <0.001  LDL cholesterol ≥6.5 8.77 6.20–12.4 <0.001 7.54 5.11–11.2 <0.001c  LDL cholesterol ≥5.0 15.2 4.77–48.5 <0.001 14.9 4.48–49.3 <0.001c Composite diagnostic tools       DLCN scoreb 1.49 1.41–1.58 <0.001    DLCN definite FHd 9.47 6.81–13.2 <0.001    DLCN probable/definite FHd 8.58 5.40–13.6 <0.001    SB definite FHd 11.7 7.15–19.1 <0.001    SB possible/definite FH 7.79 4.57–13.3 <0.001    MEDPED probable heterozygous FH 10.5 7.48–14.9 <0.001    AHA heterozygous FH 4.67 3.07–7.09 <0.001      Univariate Logistic Regression Multivariate Logistic Regression  OR 95% CI P OR a 95% CI P Individual phenotypic criteria       Family history        Premature CAD and/or vascular disease or elevated LDL cholesterol 3.00 1.92–4.67 <0.001 3.24 1.86–5.64 <0.001  Tendon xanthomata and/or corneal arcus or children with elevated LDL cholesterol 12.0 7.39–19.4 <0.001 8.65 4.98–15.03 <0.001 Clinical history        Presence of premature CAD 1.24 0.93–1.66 0.139     Presence of premature cerebral or PVD 0.82 0.39–1.71 0.597    Physical examination        Tendinous xanthomata 10.7 6.64–17.2 <0.001 5.32 2.94–9.61 <0.001  Corneal arcus 1.75 1.29–2.40 <0.001 1.55 1.04–2.31 0.033 Pretreated LDL cholesterol, mmol/L        LDL cholesterolb 2.81 2.41–3.27 <0.001     LDL cholesterol ≥8.5 27.0 14.8–49.3 <0.001 16.1 8.42–30.6 <0.001  LDL cholesterol ≥6.5 8.77 6.20–12.4 <0.001 7.54 5.11–11.2 <0.001c  LDL cholesterol ≥5.0 15.2 4.77–48.5 <0.001 14.9 4.48–49.3 <0.001c Composite diagnostic tools       DLCN scoreb 1.49 1.41–1.58 <0.001    DLCN definite FHd 9.47 6.81–13.2 <0.001    DLCN probable/definite FHd 8.58 5.40–13.6 <0.001    SB definite FHd 11.7 7.15–19.1 <0.001    SB possible/definite FH 7.79 4.57–13.3 <0.001    MEDPED probable heterozygous FH 10.5 7.48–14.9 <0.001    AHA heterozygous FH 4.67 3.07–7.09 <0.001     a Multivariate logistic regression (variables enter on model: family history of premature CAD and/or vascular disease or elevated LDL cholesterol, family history of tendon xanthomata and/or corneal arcus, tendon xanthomata, corneal arcus, and pretreated LDL cholesterol ≥8.5 mmol/L). b Expressed as a continuous variable. c Replacing LDL cholesterol ≥8.5 mmol/L with lower LDL cholesterol cutoff. d Refers to the classification without the genetic analysis component. View Large Table 3 shows the concordance and discordance rates between the four clinical diagnostic tools. There was a moderate agreement between DLCN and SB definite (ϰ=0.455), with an overall discordance rate of 19%. However, there was a poor agreement between DLCN probable/definite and MEDPED criteria (ϰ=0.120, concordance rate 67%, discordance rate 33%), between SB possible/definite and MEDPED criteria (ϰ=0.185, concordance rate 56%, discordance rate 44%), and between the AHA criteria and DLCN definite (ϰ=0.120, concordance rate 49%, discordance rate 51%), SB definite FH (ϰ=0.028, concordance rate 36%, discordance rate 64%), and MEDPED criteria (ϰ=0.176, concordance rate 56%, discordance rate 44%). The κ coefficient between the DLCN probable/definite and the SB possible/definite and AHA criteria was 0.378 and 0.338, respectively, and between SB definite and MEDPED criteria 0.205, reflecting fair agreement between them. There was a very good agreement between the SB possible/definite and AHA criteria (ϰ=0.908, concordance rate 99%, discordance rate 1%). Table 3. Concordance and Discordance Rates for Diagnostic Outcomes According to DLCN, SB, MEDPED, and AHA Phenotypic Tools  Concordance Rate Discordance Rate κ Coefficient* DLCN definite FH vs SB definite FH     For total group (n = 885) 81% 19% 0.455  For positives (n = 276)b 38% 62%   For negatives (n = 609)b 98% 2%  DLCN probable/definite FH vs SB possible/definite FH     For total group (n = 885) 74% 26% 0.378  For positives (n = 594)b 87% 13%   For negatives (n = 291)b 49% 51%  DLCN definite FH vs MEDPED FH     For total group (n = 885) 73% 27% 0.372  For positives (n = 276)b 73% 27%   For negatives (n = 609)b 72% 28%  DLCN probable/definite FH vs MEDPED FH     For total group (n = 885) 67% 33% 0.120  For positives (n = 276)b 56% 44%   For negatives (n = 609)b 88% 12%  DLCN definite FH vs AHA FH     For total group (n = 885) 49% 51% 0.120  For positives (n = 276)b 83% 17%   For negatives (n = 609)b 34% 66%  DLCN probable/definite FH vs AHA FH     For total group (n = 885) 71% 29% 0.338  For positives (n = 594)b 82% 18%   For negatives (n = 291)b 49% 51%  SB definite FH vs MEDPED FH     For total group (n = 885) 65% 35% 0.205  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 63% 37%  SB definite FH vs AHA FH     For total group (n = 885) 36% 64% 0.028  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 30% 70%  SB possible/definite FH vs MEDPED FH     For total group (n = 885) 56% 44% 0.185  For positives (n = 664)b 48% 52%   For negatives (n = 221)b 78% 22%  SB possible/definite FH vs AHA FH     For total group (n = 885) 99% 1% 0.908  For positives (n = 664)b 95% 5%   For negatives (n = 221)b 100% 0%  MEDPED FH vs AHA FH     For total group (n = 885) 56% 44% 0.176  For positives (n = 370)b 48% 52%   For negatives (n = 515)b 75% 25%    Concordance Rate Discordance Rate κ Coefficient* DLCN definite FH vs SB definite FH     For total group (n = 885) 81% 19% 0.455  For positives (n = 276)b 38% 62%   For negatives (n = 609)b 98% 2%  DLCN probable/definite FH vs SB possible/definite FH     For total group (n = 885) 74% 26% 0.378  For positives (n = 594)b 87% 13%   For negatives (n = 291)b 49% 51%  DLCN definite FH vs MEDPED FH     For total group (n = 885) 73% 27% 0.372  For positives (n = 276)b 73% 27%   For negatives (n = 609)b 72% 28%  DLCN probable/definite FH vs MEDPED FH     For total group (n = 885) 67% 33% 0.120  For positives (n = 276)b 56% 44%   For negatives (n = 609)b 88% 12%  DLCN definite FH vs AHA FH     For total group (n = 885) 49% 51% 0.120  For positives (n = 276)b 83% 17%   For negatives (n = 609)b 34% 66%  DLCN probable/definite FH vs AHA FH     For total group (n = 885) 71% 29% 0.338  For positives (n = 594)b 82% 18%   For negatives (n = 291)b 49% 51%  SB definite FH vs MEDPED FH     For total group (n = 885) 65% 35% 0.205  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 63% 37%  SB definite FH vs AHA FH     For total group (n = 885) 36% 64% 0.028  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 30% 70%  SB possible/definite FH vs MEDPED FH     For total group (n = 885) 56% 44% 0.185  For positives (n = 664)b 48% 52%   For negatives (n = 221)b 78% 22%  SB possible/definite FH vs AHA FH     For total group (n = 885) 99% 1% 0.908  For positives (n = 664)b 95% 5%   For negatives (n = 221)b 100% 0%  MEDPED FH vs AHA FH     For total group (n = 885) 56% 44% 0.176  For positives (n = 370)b 48% 52%   For negatives (n = 515)b 75% 25%   a Poor agreement <0.20, fair agreement 0.20 to 0.40, moderate agreement 0.40 to 0.60, good agreement 0.60 to 0.80, and very good agreement 0.80 to 1.00. b Positives and negatives refer to outcomes by the former phenotypic diagnostic category. View Large Table 3. Concordance and Discordance Rates for Diagnostic Outcomes According to DLCN, SB, MEDPED, and AHA Phenotypic Tools  Concordance Rate Discordance Rate κ Coefficient* DLCN definite FH vs SB definite FH     For total group (n = 885) 81% 19% 0.455  For positives (n = 276)b 38% 62%   For negatives (n = 609)b 98% 2%  DLCN probable/definite FH vs SB possible/definite FH     For total group (n = 885) 74% 26% 0.378  For positives (n = 594)b 87% 13%   For negatives (n = 291)b 49% 51%  DLCN definite FH vs MEDPED FH     For total group (n = 885) 73% 27% 0.372  For positives (n = 276)b 73% 27%   For negatives (n = 609)b 72% 28%  DLCN probable/definite FH vs MEDPED FH     For total group (n = 885) 67% 33% 0.120  For positives (n = 276)b 56% 44%   For negatives (n = 609)b 88% 12%  DLCN definite FH vs AHA FH     For total group (n = 885) 49% 51% 0.120  For positives (n = 276)b 83% 17%   For negatives (n = 609)b 34% 66%  DLCN probable/definite FH vs AHA FH     For total group (n = 885) 71% 29% 0.338  For positives (n = 594)b 82% 18%   For negatives (n = 291)b 49% 51%  SB definite FH vs MEDPED FH     For total group (n = 885) 65% 35% 0.205  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 63% 37%  SB definite FH vs AHA FH     For total group (n = 885) 36% 64% 0.028  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 30% 70%  SB possible/definite FH vs MEDPED FH     For total group (n = 885) 56% 44% 0.185  For positives (n = 664)b 48% 52%   For negatives (n = 221)b 78% 22%  SB possible/definite FH vs AHA FH     For total group (n = 885) 99% 1% 0.908  For positives (n = 664)b 95% 5%   For negatives (n = 221)b 100% 0%  MEDPED FH vs AHA FH     For total group (n = 885) 56% 44% 0.176  For positives (n = 370)b 48% 52%   For negatives (n = 515)b 75% 25%    Concordance Rate Discordance Rate κ Coefficient* DLCN definite FH vs SB definite FH     For total group (n = 885) 81% 19% 0.455  For positives (n = 276)b 38% 62%   For negatives (n = 609)b 98% 2%  DLCN probable/definite FH vs SB possible/definite FH     For total group (n = 885) 74% 26% 0.378  For positives (n = 594)b 87% 13%   For negatives (n = 291)b 49% 51%  DLCN definite FH vs MEDPED FH     For total group (n = 885) 73% 27% 0.372  For positives (n = 276)b 73% 27%   For negatives (n = 609)b 72% 28%  DLCN probable/definite FH vs MEDPED FH     For total group (n = 885) 67% 33% 0.120  For positives (n = 276)b 56% 44%   For negatives (n = 609)b 88% 12%  DLCN definite FH vs AHA FH     For total group (n = 885) 49% 51% 0.120  For positives (n = 276)b 83% 17%   For negatives (n = 609)b 34% 66%  DLCN probable/definite FH vs AHA FH     For total group (n = 885) 71% 29% 0.338  For positives (n = 594)b 82% 18%   For negatives (n = 291)b 49% 51%  SB definite FH vs MEDPED FH     For total group (n = 885) 65% 35% 0.205  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 63% 37%  SB definite FH vs AHA FH     For total group (n = 885) 36% 64% 0.028  For positives (n = 106)b 79% 21%   For negatives (n = 779)b 30% 70%  SB possible/definite FH vs MEDPED FH     For total group (n = 885) 56% 44% 0.185  For positives (n = 664)b 48% 52%   For negatives (n = 221)b 78% 22%  SB possible/definite FH vs AHA FH     For total group (n = 885) 99% 1% 0.908  For positives (n = 664)b 95% 5%   For negatives (n = 221)b 100% 0%  MEDPED FH vs AHA FH     For total group (n = 885) 56% 44% 0.176  For positives (n = 370)b 48% 52%   For negatives (n = 515)b 75% 25%   a Poor agreement <0.20, fair agreement 0.20 to 0.40, moderate agreement 0.40 to 0.60, good agreement 0.60 to 0.80, and very good agreement 0.80 to 1.00. b Positives and negatives refer to outcomes by the former phenotypic diagnostic category. View Large The ROC curve for LDL cholesterol and DLCN score in predicting an FH-causing mutation is shown in Supplemental Fig. 1. The higher the LDL cholesterol levels and DLCN score, the greater the specificity and the lower the sensitivity in detecting a mutation. The ROC curve for LDL cholesterol (AUC 0.835; 95% CI 0.806 to 0.865) and for DLCN score (AUC 0.816; 95% CI 0.784 to 0.847) was comparable in detecting a mutation (difference between ROC areas 0.020; 95% CI −0.001 to 0.048; P = 0.164). The AUC for LDL cholesterol, after adjusting for Lp(a) cholesterol, in detecting a mutation was 0.841 (95% CI 0.812 to 0.870), which was also not significantly different from that for DLCN score as a continuous variable (difference between ROC areas 0.011; 95% CI −0.017 to 0.040; P = 0.453). Table 4 shows the sensitivity, specificity, PPV, NPV, Youden index, and AUC for different diagnostic criteria in detecting an FH-causing mutation. A family history of physical stigmata, the presence of tendon xanthomata, and a LDL cholesterol ≥8.5 mmol/L had high specificities (>95% for all), but relatively low sensitivities (range 31.1% to 36.7%) in detecting a mutation. A family history of premature CAD or hypercholesterolemia had high sensitivity (90.2%), but low specificity (24.4%) in detecting a mutation. Lower plasma LDL cholesterol concentrations were more sensitive (from 36.7% to 98.9%) but less specific (from 97.9% to 14.7%) in detecting a mutation. The highest Youden index among the individual diagnostic components was 0.485 with a LDL cholesterol ≥6.5 mmol/L. Table 4. Sensitivity, Specificity, PPV and NPV, Youden Index, and AUC of Diagnostic Test for FH Criteria Sensitivity % (95% CI) Specificity % (95% CI) PPV % (95% CI) NPV % (95% CI) Youden Index AUC (95% CI) Change in AUC, Relative to, DLCN Definite FH a (95% CI) Change in AUC, Relative to, SB Definite, FH a (95% CI) Change in AUC, Relative to, DLCN Definite/Probable FH a (95% CI) Change in AUC, Relative to, SB Definite/Possible FH a (95% CI) Individual diagnostic criteria            Family history of premature ASCVD or elevated LDL cholesterol 90.2 (85.9, 93.4) 24.4 (21.1, 28.1) 34.0 (30.5, 37.7) 85.3 (79.0, 90.0) 0.146 0.573 (0.534, 0.613) −0.170b (−0.211, −0.128) −0.063b (−0.103, −0.024) −0.103b (−0.136, −0.070) −0.062b (−0.082, −0.043)  Family history of tendon xanthomata and/or corneal arcus 32.5 (27.1, 38.6) 96.1 (94.2, 97.4) 78.4 (69.4, 85.4) 76.7 (73.5, 79.6) 0.286 0.644 (0.601, 0.686) −0.100b (−0.138, −0.061) 0.007 (−0.030, 0.044) −0.033 (−0.071, 0.005) 0.008 (−0.029, 0.044)  Presence of premature CAD 47.9 (41.8, 54.1) 57.4 (53.4, 61.4) 32.7 (28.2, 37.7) 71.9 (67.6, 75.7) 0.053 0.527 (0.485, 0.568) −0.216b (−0.260, −0.172) −0.110b (−0.156, −0.064) −0.150b (−0.188, −0.111) −0.109b (−0.153, −0.065)  Presence of premature cerebral or PVD 3.74 (1.91, 6.99) 95.7 (94.1, 96.9) 26.3 (14.0, 43.4) 69.7 (66.4, 72.7) 0.001 0.496 (0.455, 0.537) −0.239b (−0.275, −0.204) −0.133b (−0.166, −0.100) −0.173b (−0.202, −0.143) −0.132b (−0.160, −0.104)  Tendon xanthomata 31.1 (25.7, 37.1) 96.0 (94.0–97.3) 76.9 (67.6, 84.2) 76.3 (73.1, 79.2) 0.271 0.635 (0.592, 0.678) −0.108b (−0.140, −0.077) −0.002 (−0.004, 0.001) −0.041b (−0.077, −0.006) −0.001 (−0.036, 0.034)  Corneal arcus 35.6 (29.9, 41.7) 76.1 (72.4, 79.3) 39.1 (33.0, 45.4) 73.2 (69.6, 76.6) 0.117 0.558 (0.516, 0.600) −0.185b (−0.219, −0.151) −0.079b (−0.119, −0.038) −0.118b (−0.154, −0.083) −0.078b (−0.120, −0.036)  LDL cholesterol ≥8.5 mmol/L 36.7 (31.0, 42.8) 97.9 (96.3, 98.8) 88.3 (80.5, 93.4) 78.1 (75.0, 81.0) 0.346 0.673 (0.631, 0.715) −0.070b (−0.101, −0.039) 0.036b (0.001, −0.071) −0.003 (−0.039, 0.032) 0.037b (0.001, 0.073)  LDL cholesterol ≥6.5 mmol/L 80.5 (75.2, 85.0) 68.0 (64.1, 71.6) 52.1 (47.1, 57.0) 89.0 (85.7, 91.6) 0.485 0.742 (0.707, 0.778) −0.001 (−0.037, 0.035) 0.106b (0.067, 0.144) 0.066b (0.037, 0.095) 0.107b (0.071, 0.143)  LDL cholesterol ≥5.0 mmol/L 98.9 (96.5, 99.7) 14.7 (12.1, 17.8) 33.4 (30.1, 36.8) 96.8 (90.3, 99.2) 0.136 0.568 (0.529, 0.607) −0.175b (−0.209, −0.141) −0.069b (−0.101, −0.037) −0.108b (−0.136, −0.081) −0.068b (−0.089, −0.046) Composite diagnostic tools            DLCN definite FH 65.2 (59.1, 70.8) 83.5 (80.3, 86.3) 63.0 (57.0, 68.7) 84.7 (81.6, 87.4) 0.487 0.743 (0.705, 0.781) Reference 0.106b (0.075, 0.138) 0.067b (0.034, 0.100) 0.107b (0.070, 0.145)  DLCN probable/definite FH 91.8 (87.6, 94.6) 43.5 (39.6, 47.5) 41.2 (37.2, 45.3) 92.4 (88.6, 95.1) 0.353 0.676 (0.641, 0.712) −0.067b (−0.100, −0.034) 0.040b (0.004, 0.075) Reference 0.041b (0.011, 0.070)  SB definite FH 31.1 (25.7, 37.1) 96.3 (94.4, 97.6) 78.3 (69.0, 85.5) 76.4 (73.2, 79.3) 0.274 0.637 (0.594, 0.680) −0.106b (−0.138, −0.075) Reference −0.040b (−0.075, −0.004) 0.001 (−0.034, 0.036)  SB possible/definite FH 94.0 (90.3, 96.4) 33.2 (29.5, 37.1) 37.8 (34.1, 41.6) 92.8 (88.3, 95.7) 0.272 0.636 (0.599, 0.673) −0.107b (−0.145, −0.070) −0.001 (−0.036, 0.034) −0.041b (−0.070, −0.011) Reference MEDPED probable FH 56.8 (51.5, 61.8) 88.9 (85.8, 91.4) 78.7 (73.1, 83.3) 74.1 (70.4, 77.5) 0.457 0.764 (0.729, 0.799) 0.021 (−0.009, 0.051) 0.088b (0.058, 0.118) 0.127b (0.097, −0.157) 0.128b (0.098, 0.158) AHA FH 89.1 (84.6, 92.5) 36.2 (32.5, 40.2) 37.7 (33.9, 41.6) 88.5 (83.8, 92.1) 0.253 0.627 (0.589, 0.665) −0.116b (−0.158, −0.075) −0.010 (−0.051, 0.031) −0.050b (−0.082, −0.017) −0.009 (−0.024, 0.006)  Criteria Sensitivity % (95% CI) Specificity % (95% CI) PPV % (95% CI) NPV % (95% CI) Youden Index AUC (95% CI) Change in AUC, Relative to, DLCN Definite FH a (95% CI) Change in AUC, Relative to, SB Definite, FH a (95% CI) Change in AUC, Relative to, DLCN Definite/Probable FH a (95% CI) Change in AUC, Relative to, SB Definite/Possible FH a (95% CI) Individual diagnostic criteria            Family history of premature ASCVD or elevated LDL cholesterol 90.2 (85.9, 93.4) 24.4 (21.1, 28.1) 34.0 (30.5, 37.7) 85.3 (79.0, 90.0) 0.146 0.573 (0.534, 0.613) −0.170b (−0.211, −0.128) −0.063b (−0.103, −0.024) −0.103b (−0.136, −0.070) −0.062b (−0.082, −0.043)  Family history of tendon xanthomata and/or corneal arcus 32.5 (27.1, 38.6) 96.1 (94.2, 97.4) 78.4 (69.4, 85.4) 76.7 (73.5, 79.6) 0.286 0.644 (0.601, 0.686) −0.100b (−0.138, −0.061) 0.007 (−0.030, 0.044) −0.033 (−0.071, 0.005) 0.008 (−0.029, 0.044)  Presence of premature CAD 47.9 (41.8, 54.1) 57.4 (53.4, 61.4) 32.7 (28.2, 37.7) 71.9 (67.6, 75.7) 0.053 0.527 (0.485, 0.568) −0.216b (−0.260, −0.172) −0.110b (−0.156, −0.064) −0.150b (−0.188, −0.111) −0.109b (−0.153, −0.065)  Presence of premature cerebral or PVD 3.74 (1.91, 6.99) 95.7 (94.1, 96.9) 26.3 (14.0, 43.4) 69.7 (66.4, 72.7) 0.001 0.496 (0.455, 0.537) −0.239b (−0.275, −0.204) −0.133b (−0.166, −0.100) −0.173b (−0.202, −0.143) −0.132b (−0.160, −0.104)  Tendon xanthomata 31.1 (25.7, 37.1) 96.0 (94.0–97.3) 76.9 (67.6, 84.2) 76.3 (73.1, 79.2) 0.271 0.635 (0.592, 0.678) −0.108b (−0.140, −0.077) −0.002 (−0.004, 0.001) −0.041b (−0.077, −0.006) −0.001 (−0.036, 0.034)  Corneal arcus 35.6 (29.9, 41.7) 76.1 (72.4, 79.3) 39.1 (33.0, 45.4) 73.2 (69.6, 76.6) 0.117 0.558 (0.516, 0.600) −0.185b (−0.219, −0.151) −0.079b (−0.119, −0.038) −0.118b (−0.154, −0.083) −0.078b (−0.120, −0.036)  LDL cholesterol ≥8.5 mmol/L 36.7 (31.0, 42.8) 97.9 (96.3, 98.8) 88.3 (80.5, 93.4) 78.1 (75.0, 81.0) 0.346 0.673 (0.631, 0.715) −0.070b (−0.101, −0.039) 0.036b (0.001, −0.071) −0.003 (−0.039, 0.032) 0.037b (0.001, 0.073)  LDL cholesterol ≥6.5 mmol/L 80.5 (75.2, 85.0) 68.0 (64.1, 71.6) 52.1 (47.1, 57.0) 89.0 (85.7, 91.6) 0.485 0.742 (0.707, 0.778) −0.001 (−0.037, 0.035) 0.106b (0.067, 0.144) 0.066b (0.037, 0.095) 0.107b (0.071, 0.143)  LDL cholesterol ≥5.0 mmol/L 98.9 (96.5, 99.7) 14.7 (12.1, 17.8) 33.4 (30.1, 36.8) 96.8 (90.3, 99.2) 0.136 0.568 (0.529, 0.607) −0.175b (−0.209, −0.141) −0.069b (−0.101, −0.037) −0.108b (−0.136, −0.081) −0.068b (−0.089, −0.046) Composite diagnostic tools            DLCN definite FH 65.2 (59.1, 70.8) 83.5 (80.3, 86.3) 63.0 (57.0, 68.7) 84.7 (81.6, 87.4) 0.487 0.743 (0.705, 0.781) Reference 0.106b (0.075, 0.138) 0.067b (0.034, 0.100) 0.107b (0.070, 0.145)  DLCN probable/definite FH 91.8 (87.6, 94.6) 43.5 (39.6, 47.5) 41.2 (37.2, 45.3) 92.4 (88.6, 95.1) 0.353 0.676 (0.641, 0.712) −0.067b (−0.100, −0.034) 0.040b (0.004, 0.075) Reference 0.041b (0.011, 0.070)  SB definite FH 31.1 (25.7, 37.1) 96.3 (94.4, 97.6) 78.3 (69.0, 85.5) 76.4 (73.2, 79.3) 0.274 0.637 (0.594, 0.680) −0.106b (−0.138, −0.075) Reference −0.040b (−0.075, −0.004) 0.001 (−0.034, 0.036)  SB possible/definite FH 94.0 (90.3, 96.4) 33.2 (29.5, 37.1) 37.8 (34.1, 41.6) 92.8 (88.3, 95.7) 0.272 0.636 (0.599, 0.673) −0.107b (−0.145, −0.070) −0.001 (−0.036, 0.034) −0.041b (−0.070, −0.011) Reference MEDPED probable FH 56.8 (51.5, 61.8) 88.9 (85.8, 91.4) 78.7 (73.1, 83.3) 74.1 (70.4, 77.5) 0.457 0.764 (0.729, 0.799) 0.021 (−0.009, 0.051) 0.088b (0.058, 0.118) 0.127b (0.097, −0.157) 0.128b (0.098, 0.158) AHA FH 89.1 (84.6, 92.5) 36.2 (32.5, 40.2) 37.7 (33.9, 41.6) 88.5 (83.8, 92.1) 0.253 0.627 (0.589, 0.665) −0.116b (−0.158, −0.075) −0.010 (−0.051, 0.031) −0.050b (−0.082, −0.017) −0.009 (−0.024, 0.006)  a A positive change (δ) in AUC value indicates an increased discriminant value in detecting an FH mutation compared with the reference test. A negative change (δ) in AUC value indicates a decreased discriminant value in detecting an FH mutation compared with the reference test. b Indicates significant differences in δ AUC compared with the reference using C-statistics; P < 0.05. View Large Table 4. Sensitivity, Specificity, PPV and NPV, Youden Index, and AUC of Diagnostic Test for FH Criteria Sensitivity % (95% CI) Specificity % (95% CI) PPV % (95% CI) NPV % (95% CI) Youden Index AUC (95% CI) Change in AUC, Relative to, DLCN Definite FH a (95% CI) Change in AUC, Relative to, SB Definite, FH a (95% CI) Change in AUC, Relative to, DLCN Definite/Probable FH a (95% CI) Change in AUC, Relative to, SB Definite/Possible FH a (95% CI) Individual diagnostic criteria            Family history of premature ASCVD or elevated LDL cholesterol 90.2 (85.9, 93.4) 24.4 (21.1, 28.1) 34.0 (30.5, 37.7) 85.3 (79.0, 90.0) 0.146 0.573 (0.534, 0.613) −0.170b (−0.211, −0.128) −0.063b (−0.103, −0.024) −0.103b (−0.136, −0.070) −0.062b (−0.082, −0.043)  Family history of tendon xanthomata and/or corneal arcus 32.5 (27.1, 38.6) 96.1 (94.2, 97.4) 78.4 (69.4, 85.4) 76.7 (73.5, 79.6) 0.286 0.644 (0.601, 0.686) −0.100b (−0.138, −0.061) 0.007 (−0.030, 0.044) −0.033 (−0.071, 0.005) 0.008 (−0.029, 0.044)  Presence of premature CAD 47.9 (41.8, 54.1) 57.4 (53.4, 61.4) 32.7 (28.2, 37.7) 71.9 (67.6, 75.7) 0.053 0.527 (0.485, 0.568) −0.216b (−0.260, −0.172) −0.110b (−0.156, −0.064) −0.150b (−0.188, −0.111) −0.109b (−0.153, −0.065)  Presence of premature cerebral or PVD 3.74 (1.91, 6.99) 95.7 (94.1, 96.9) 26.3 (14.0, 43.4) 69.7 (66.4, 72.7) 0.001 0.496 (0.455, 0.537) −0.239b (−0.275, −0.204) −0.133b (−0.166, −0.100) −0.173b (−0.202, −0.143) −0.132b (−0.160, −0.104)  Tendon xanthomata 31.1 (25.7, 37.1) 96.0 (94.0–97.3) 76.9 (67.6, 84.2) 76.3 (73.1, 79.2) 0.271 0.635 (0.592, 0.678) −0.108b (−0.140, −0.077) −0.002 (−0.004, 0.001) −0.041b (−0.077, −0.006) −0.001 (−0.036, 0.034)  Corneal arcus 35.6 (29.9, 41.7) 76.1 (72.4, 79.3) 39.1 (33.0, 45.4) 73.2 (69.6, 76.6) 0.117 0.558 (0.516, 0.600) −0.185b (−0.219, −0.151) −0.079b (−0.119, −0.038) −0.118b (−0.154, −0.083) −0.078b (−0.120, −0.036)  LDL cholesterol ≥8.5 mmol/L 36.7 (31.0, 42.8) 97.9 (96.3, 98.8) 88.3 (80.5, 93.4) 78.1 (75.0, 81.0) 0.346 0.673 (0.631, 0.715) −0.070b (−0.101, −0.039) 0.036b (0.001, −0.071) −0.003 (−0.039, 0.032) 0.037b (0.001, 0.073)  LDL cholesterol ≥6.5 mmol/L 80.5 (75.2, 85.0) 68.0 (64.1, 71.6) 52.1 (47.1, 57.0) 89.0 (85.7, 91.6) 0.485 0.742 (0.707, 0.778) −0.001 (−0.037, 0.035) 0.106b (0.067, 0.144) 0.066b (0.037, 0.095) 0.107b (0.071, 0.143)  LDL cholesterol ≥5.0 mmol/L 98.9 (96.5, 99.7) 14.7 (12.1, 17.8) 33.4 (30.1, 36.8) 96.8 (90.3, 99.2) 0.136 0.568 (0.529, 0.607) −0.175b (−0.209, −0.141) −0.069b (−0.101, −0.037) −0.108b (−0.136, −0.081) −0.068b (−0.089, −0.046) Composite diagnostic tools            DLCN definite FH 65.2 (59.1, 70.8) 83.5 (80.3, 86.3) 63.0 (57.0, 68.7) 84.7 (81.6, 87.4) 0.487 0.743 (0.705, 0.781) Reference 0.106b (0.075, 0.138) 0.067b (0.034, 0.100) 0.107b (0.070, 0.145)  DLCN probable/definite FH 91.8 (87.6, 94.6) 43.5 (39.6, 47.5) 41.2 (37.2, 45.3) 92.4 (88.6, 95.1) 0.353 0.676 (0.641, 0.712) −0.067b (−0.100, −0.034) 0.040b (0.004, 0.075) Reference 0.041b (0.011, 0.070)  SB definite FH 31.1 (25.7, 37.1) 96.3 (94.4, 97.6) 78.3 (69.0, 85.5) 76.4 (73.2, 79.3) 0.274 0.637 (0.594, 0.680) −0.106b (−0.138, −0.075) Reference −0.040b (−0.075, −0.004) 0.001 (−0.034, 0.036)  SB possible/definite FH 94.0 (90.3, 96.4) 33.2 (29.5, 37.1) 37.8 (34.1, 41.6) 92.8 (88.3, 95.7) 0.272 0.636 (0.599, 0.673) −0.107b (−0.145, −0.070) −0.001 (−0.036, 0.034) −0.041b (−0.070, −0.011) Reference MEDPED probable FH 56.8 (51.5, 61.8) 88.9 (85.8, 91.4) 78.7 (73.1, 83.3) 74.1 (70.4, 77.5) 0.457 0.764 (0.729, 0.799) 0.021 (−0.009, 0.051) 0.088b (0.058, 0.118) 0.127b (0.097, −0.157) 0.128b (0.098, 0.158) AHA FH 89.1 (84.6, 92.5) 36.2 (32.5, 40.2) 37.7 (33.9, 41.6) 88.5 (83.8, 92.1) 0.253 0.627 (0.589, 0.665) −0.116b (−0.158, −0.075) −0.010 (−0.051, 0.031) −0.050b (−0.082, −0.017) −0.009 (−0.024, 0.006)  Criteria Sensitivity % (95% CI) Specificity % (95% CI) PPV % (95% CI) NPV % (95% CI) Youden Index AUC (95% CI) Change in AUC, Relative to, DLCN Definite FH a (95% CI) Change in AUC, Relative to, SB Definite, FH a (95% CI) Change in AUC, Relative to, DLCN Definite/Probable FH a (95% CI) Change in AUC, Relative to, SB Definite/Possible FH a (95% CI) Individual diagnostic criteria            Family history of premature ASCVD or elevated LDL cholesterol 90.2 (85.9, 93.4) 24.4 (21.1, 28.1) 34.0 (30.5, 37.7) 85.3 (79.0, 90.0) 0.146 0.573 (0.534, 0.613) −0.170b (−0.211, −0.128) −0.063b (−0.103, −0.024) −0.103b (−0.136, −0.070) −0.062b (−0.082, −0.043)  Family history of tendon xanthomata and/or corneal arcus 32.5 (27.1, 38.6) 96.1 (94.2, 97.4) 78.4 (69.4, 85.4) 76.7 (73.5, 79.6) 0.286 0.644 (0.601, 0.686) −0.100b (−0.138, −0.061) 0.007 (−0.030, 0.044) −0.033 (−0.071, 0.005) 0.008 (−0.029, 0.044)  Presence of premature CAD 47.9 (41.8, 54.1) 57.4 (53.4, 61.4) 32.7 (28.2, 37.7) 71.9 (67.6, 75.7) 0.053 0.527 (0.485, 0.568) −0.216b (−0.260, −0.172) −0.110b (−0.156, −0.064) −0.150b (−0.188, −0.111) −0.109b (−0.153, −0.065)  Presence of premature cerebral or PVD 3.74 (1.91, 6.99) 95.7 (94.1, 96.9) 26.3 (14.0, 43.4) 69.7 (66.4, 72.7) 0.001 0.496 (0.455, 0.537) −0.239b (−0.275, −0.204) −0.133b (−0.166, −0.100) −0.173b (−0.202, −0.143) −0.132b (−0.160, −0.104)  Tendon xanthomata 31.1 (25.7, 37.1) 96.0 (94.0–97.3) 76.9 (67.6, 84.2) 76.3 (73.1, 79.2) 0.271 0.635 (0.592, 0.678) −0.108b (−0.140, −0.077) −0.002 (−0.004, 0.001) −0.041b (−0.077, −0.006) −0.001 (−0.036, 0.034)  Corneal arcus 35.6 (29.9, 41.7) 76.1 (72.4, 79.3) 39.1 (33.0, 45.4) 73.2 (69.6, 76.6) 0.117 0.558 (0.516, 0.600) −0.185b (−0.219, −0.151) −0.079b (−0.119, −0.038) −0.118b (−0.154, −0.083) −0.078b (−0.120, −0.036)  LDL cholesterol ≥8.5 mmol/L 36.7 (31.0, 42.8) 97.9 (96.3, 98.8) 88.3 (80.5, 93.4) 78.1 (75.0, 81.0) 0.346 0.673 (0.631, 0.715) −0.070b (−0.101, −0.039) 0.036b (0.001, −0.071) −0.003 (−0.039, 0.032) 0.037b (0.001, 0.073)  LDL cholesterol ≥6.5 mmol/L 80.5 (75.2, 85.0) 68.0 (64.1, 71.6) 52.1 (47.1, 57.0) 89.0 (85.7, 91.6) 0.485 0.742 (0.707, 0.778) −0.001 (−0.037, 0.035) 0.106b (0.067, 0.144) 0.066b (0.037, 0.095) 0.107b (0.071, 0.143)  LDL cholesterol ≥5.0 mmol/L 98.9 (96.5, 99.7) 14.7 (12.1, 17.8) 33.4 (30.1, 36.8) 96.8 (90.3, 99.2) 0.136 0.568 (0.529, 0.607) −0.175b (−0.209, −0.141) −0.069b (−0.101, −0.037) −0.108b (−0.136, −0.081) −0.068b (−0.089, −0.046) Composite diagnostic tools            DLCN definite FH 65.2 (59.1, 70.8) 83.5 (80.3, 86.3) 63.0 (57.0, 68.7) 84.7 (81.6, 87.4) 0.487 0.743 (0.705, 0.781) Reference 0.106b (0.075, 0.138) 0.067b (0.034, 0.100) 0.107b (0.070, 0.145)  DLCN probable/definite FH 91.8 (87.6, 94.6) 43.5 (39.6, 47.5) 41.2 (37.2, 45.3) 92.4 (88.6, 95.1) 0.353 0.676 (0.641, 0.712) −0.067b (−0.100, −0.034) 0.040b (0.004, 0.075) Reference 0.041b (0.011, 0.070)  SB definite FH 31.1 (25.7, 37.1) 96.3 (94.4, 97.6) 78.3 (69.0, 85.5) 76.4 (73.2, 79.3) 0.274 0.637 (0.594, 0.680) −0.106b (−0.138, −0.075) Reference −0.040b (−0.075, −0.004) 0.001 (−0.034, 0.036)  SB possible/definite FH 94.0 (90.3, 96.4) 33.2 (29.5, 37.1) 37.8 (34.1, 41.6) 92.8 (88.3, 95.7) 0.272 0.636 (0.599, 0.673) −0.107b (−0.145, −0.070) −0.001 (−0.036, 0.034) −0.041b (−0.070, −0.011) Reference MEDPED probable FH 56.8 (51.5, 61.8) 88.9 (85.8, 91.4) 78.7 (73.1, 83.3) 74.1 (70.4, 77.5) 0.457 0.764 (0.729, 0.799) 0.021 (−0.009, 0.051) 0.088b (0.058, 0.118) 0.127b (0.097, −0.157) 0.128b (0.098, 0.158) AHA FH 89.1 (84.6, 92.5) 36.2 (32.5, 40.2) 37.7 (33.9, 41.6) 88.5 (83.8, 92.1) 0.253 0.627 (0.589, 0.665) −0.116b (−0.158, −0.075) −0.010 (−0.051, 0.031) −0.050b (−0.082, −0.017) −0.009 (−0.024, 0.006)  a A positive change (δ) in AUC value indicates an increased discriminant value in detecting an FH mutation compared with the reference test. A negative change (δ) in AUC value indicates a decreased discriminant value in detecting an FH mutation compared with the reference test. b Indicates significant differences in δ AUC compared with the reference using C-statistics; P < 0.05. View Large With the composite diagnostic criteria, Table 4 also shows that SB definite FH had a higher specificity (96.3%) but lower sensitivity (31.1%) in detecting a mutation than DLCN definite FH (specificity 83.5%, sensitivity 65.2%) and MEDPED criteria (specificity 88.9%, sensitivity 56.8%). However, the DLCN definite FH (Youden index 0.487) and MEDPED criteria (Youden index 0.457) had a better balance between sensitivity and specificity than the SB definite (Youden index 0.274) in detecting a mutation. Although the AHA criteria had a high sensitivity (89.1%), its specificity (36.2%) was lower compared with DLCN definite, SB definite, or MEDPED criteria. The Youden indices for the composite were in decreasing order: DLCN definite, MEDPED criteria, DLCN probable/definite, SB definite, SB possible/definite, and AHA criteria. With reference to a DLCN definite diagnosis, the AUCs were significantly less (P < 0.05 for all) with family history of ASCVD (or elevated LDL cholesterol), family and personal history of physical stigmata, personal history of ASCVD, elevated LDL cholesterol (≥5.0 mmol/L or ≥8.5 mmol/L), DLCN probable/definite FH, SB possible/definite FH, and AHA criteria (AUC range −0.067 to −0.239). With reference to a SB definite diagnosis, the AUCs were significantly less (P < 0.05 for all) with family history of ASCVD (or elevated LDL cholesterol), personal history of ASCVD, corneal arcus, and elevated LDL cholesterol ≥5.0 mmol/L (AUC range −0.069 to −0.133). The AUC for a LDL cholesterol ≥6.5 mmol/L was significantly higher than a SB definite diagnosis (0.106, 95% CI 0.067 to 0.144), but not different from a DLCN definite diagnosis (−0.001, 95% CI −0.037 to 0.035). With reference to a DLCN probable/definite or SB possible/definite diagnosis, the AUCs were significantly less with family history of ASCVD (or elevated LDL cholesterol), personal history of ASCVD and corneal arcus, and LDL cholesterol ≥5.0 mmol/L (AUC range −0.062 to −0.173). The AUC for the MEDPED criteria was similar to the DLCN definite FH (0.487 vs 0.457, P > 0.05), but significantly higher compared with the SB definite, DLCN probable/definite, or SB possible/definite diagnosis (P < 0.05 for all). Although family history of physical stigmata showed no significant difference in AUCs, a LDL cholesterol ≥6.5 mmol/L had significantly higher AUCs when compared with the DLCN probable/definite or SB possible/definite diagnosis (0.066 and 0.107, respectively, P < 0.001). Discussion In a large sample of patients referred to a specialist clinic, we demonstrated that the DLCN, SB, and MEDPED phenotypes were valid predictors of an FH-causing mutation, the AHA criteria having the lowest diagnostic specificity. Among individual phenotypic components, the pretreatment LDL cholesterol and presence of tendon xanthomata afforded the best discriminant value in predicting a mutation that overall matched the DLCN definite, SB definite, and MEDPED criteria. The concordance between these phenotypic tools for FH was, however, only moderate. Several studies have examined the relationship between clinical diagnostic tools and genetically defined FH. Damgaard et al. (17) reported that the DLCN and SB criteria were comparably effective in identifying most carriers of an FH-causing mutation, with specificities >85%, but sensitivities of <45%; the discriminant value of individual diagnostic components was not investigated. The diagnostic yield of the DLCN and SB criteria in predicting the common FH-causing mutations was later confirmed in a larger community population, but there was limited information on family history and no records of physical stigmata of FH (4). Haralambos et al. (13) reported on the value of a modified DLCN criteria in predicting FH-causing mutations in Wales, but the findings were limited by lack of genetic data in individuals with lower DLCN score. Silva et al. (18) found that the DLCN score was a significant predictor of a mutation, but SB diagnostic criteria were not examined. Civeira et al. (19) also reported that the MEDPED criteria had high sensitivity and specificity in predicting a mutation in Spanish patients, but tendon xanthomata and age-adjusted LDL cholesterol were the most significant predictors of a mutation. We extend previous reports by comparing the discriminant value of different phenotypic criteria in predicting an FH-causing mutation among referrals to a specialist clinic. That a higher proportion of M+ than M− patients in our study fulfilled a phenotypic diagnosis of FH according to the DLCN definite, SB definite, MEDPED, and AHA criteria was anticipated. Tendon xanthomata and markedly elevated LDL cholesterol are classical hallmarks of FH (7–9, 11). Consistent with previous studies (18, 19, 21), the presence of tendon xanthomata and/or corneal arcus in the potential index cases was an independent predictor of the presence of an FH-causing mutation (Table 2). Likewise, a pretreatment LDL cholesterol >4.9 mmol/L was confirmed as an independent predictor of a mutation. Given that tendon xanthomata and elevated LDL cholesterol are major criteria for the DLCN and SB diagnostic tools (7, 9), it was anticipated that these phenotypic components were highly predictive of a mutation. Although the AHA tool was also predictive of a mutation, its specificity and PPV were significantly less than DLCN and SB definite and MEDPED criteria. In agreement with Civeira et al. (19), we found that a history of premature CAD was not a significant predictor of a mutation. This accords with recent understanding that the major causes of premature CAD are due to multigenic and adverse lifestyle factors. As suggested elsewhere, familial combined hyperlipidemia and common hypercholesterolemia with elevated Lp(a) are frequent phenocopies of FH (22). Although we studied referrals to a specialist clinic, ∼30% of patients with a definite phenotypic FH by DLCN and SB criteria did not have an FH-causing mutation. This points to polygenic hypercholesterolemia (attributed to common, small-effect LDL cholesterol-raising alleles), or unrecognized genetic causes of autosomal dominant hypercholesterolemia (23, 24). Conversely, >15% of patients with a mutation were not identified by a DLCN or SB definite diagnosis. This suggests less severe FH phenotypes associated with pathogenic LDLR variants causing less LDL-receptor dysfunction, reduced genetic penetrance, and/or coinheritance of protective genes (25). Earlier detection and treatment with statins can also modify phenotypic FH (26). The phenotypic diagnosis of FH appears more dependent on the pretreatment LDL cholesterol than on other clinical features of FH, consistent with the earlier MEDPED recommendations (8), although clinical stigmata allow greater refinement of the diagnosis (8, 27). The strength of the study was the use of well-standardized, comprehensive criteria for characterizing patients. However, there were limitations. The sample was predominantly Caucasian, but reflected the ethnic distribution of the Australian population. That the likelihood of an FH-causing mutation was increased in non-Caucasians reflects higher specificity in this group of applying Caucasian criteria. The diagnosis of FH in children based on LDL cholesterol was not examined, but is reviewed elsewhere (28). Because of ascertainment bias, our results may have limited applicability to the community diagnosis of FH, but our conclusions may be valuable when applied to targeted screening of electronic health records for FH in primary care (29). Approximately 25% of our patients were on cholesterol-lowering medication, and pretreatment LDL cholesterol was derived using adjustments that might not have fully reflected true values (13). Testing for cholesterol allelic gene scores might have helped identify patients with isolated polygenic hypercholesterolemia (24). Implications The reference diagnosis of FH is the identification of a pathogenic mutation that affects the clearance of LDL via the LDL-receptor pathway (1, 2). Genetic testing may be used to predict ASCVD risk (19), but is currently expensive and not widely available. Moreover, not all patients consent to genetic testing owing to personal preferences and potential discrimination by employers and insurers (3, 5, 11, 25). Knowledge of the phenotypic predictors of a mutation may be useful for improving the accuracy of diagnosis where genetic testing is not available, or not consented to, and for rationing genetic testing where resources are limited. With rising awareness of FH in primary care, an increasing number of patients will be referred to specialist clinics for confirmation of the diagnosis. Because of limited resources, it may be necessary to phenotypically identify patients who are very likely to have FH in whom genetic confirmation may be less valuable (unless combined with cascade screening), patients with an intermediate clinical likelihood in whom genetic testing may be particularly useful, and patients with low likelihood in whom genetic testing could be wasteful. Cascade screening for FH using LDL cholesterol and genetic testing is potentially cost effective (27, 30, 31). Phenotypic cascade testing from index cases with a definite clinical diagnosis is also intuitively more cost effective than from index cases with lesser probabilities of FH. Hence, genetic testing of index cases with an intermediate probability of FH may offer an effective use of laboratory resources in the context of cascade testing. Based on our results, Fig. 1 shows the application of the four phenotypic diagnostic tools to the detection of a mutation and to enable decision regarding genetic testing in patients referred to clinics. In this scheme, genetic testing could be recommended in patients with an intermediate probability of having a mutation and could be considered in patients with a low phenotypic probability in whom the clinical suspicion of FH is high. A caveat is that although prediction of a mutation appears statistically comparable among DLCN, SB, and MEDPED criteria, the phenotypic classification of patients can be discordant (Table 3). Figure 1. View largeDownload slide Flowcharts showing the use of the phenotypic (A) DLCN, (B) SB, (C) MEDPED, and (D) AHA diagnostic tools in predicting the likelihood of an FH-causing mutation and assisting decisions on whether to undertake genetic testing. Genetic testing is arbitrarily not routinely recommended where probability of a mutation is <15%, is recommended when the probability of a mutation is 15% to 60%, and is optional when the probability of a mutation >60%. Figure 1. View largeDownload slide Flowcharts showing the use of the phenotypic (A) DLCN, (B) SB, (C) MEDPED, and (D) AHA diagnostic tools in predicting the likelihood of an FH-causing mutation and assisting decisions on whether to undertake genetic testing. Genetic testing is arbitrarily not routinely recommended where probability of a mutation is <15%, is recommended when the probability of a mutation is 15% to 60%, and is optional when the probability of a mutation >60%. Given that information on individual family members is not always available, a simpler scheme shown in Fig. 2 uses the pretreatment LDL cholesterol and presence of tendon xanthomata in potential index cases to predict the likelihood of a mutation and value of genetic testing. Use of pretreatment LDL cholesterol as the primary criterion accords with earlier MEDPED recommendations (8) and is useful for identifying patients with intermediate likelihood of FH in whom genetic confirmation is required. Inclusion of tendon xanthomata as a second criterion further improves the specificity of the diagnosis of FH. Figure 2. View largeDownload slide Flowchart showing use of LDL cholesterol concentration and presence of tendon xanthomata in predicting the likelihood of an FH-causing mutation and assisting decisions on whether to undertake genetic testing. Genetic testing is arbitrarily not routinely recommended where probability of a mutation is <15%, is recommended when the probability of a mutation is 15% to 60%, and is optional when the probability of a mutation >60%. Figure 2. View largeDownload slide Flowchart showing use of LDL cholesterol concentration and presence of tendon xanthomata in predicting the likelihood of an FH-causing mutation and assisting decisions on whether to undertake genetic testing. Genetic testing is arbitrarily not routinely recommended where probability of a mutation is <15%, is recommended when the probability of a mutation is 15% to 60%, and is optional when the probability of a mutation >60%. In adult patients referred to a specialist lipid clinic, prediction of an FH-causing mutation, or the decision to request a genetic test, could simply be based on the pretreatment LDL cholesterol and presence of a personal (or family) history of tendon xanthomata. Whereas florid tendon xanthomata are currently less frequent among heterozygous FH (32), their earlier detection may be enhanced with ultrasonography (33). This may provide a more standardized approach to the diagnosis of FH and overcome the notable discordance between DLCN and SB clinical diagnostic criteria. Country-specific LDL cholesterol levels should replace the original MEDPED criteria (8, 27). Services with requisite laboratory and clinical resources may elect to genetically test all patients referred with a putative diagnosis of FH, but the cost effectiveness of this policy remains to be evaluated. Abbreviations: Abbreviations:   AHA American Heart Association   ASCVD atherosclerotic cardiovascular disease   AUC area under the curve   CAD coronary artery disease   CI confidence interval   DLCN Dutch Lipid Clinic Network   FH familial hypercholesterolemia   HDL high-density lipoprotein   LDL low-density lipoprotein   MEDPED Make Early Diagnosis to Prevent Early Deaths   NGS next-generation sequencing   NPV negative predictive value   OR odds ratio   PPV positive predictive value   ROC receiver operator characteristic   SB Simon Broome Acknowledgments We thank Professor Frank van Bockxmeer (Royal Perth Hospital, School of Biomedical Sciences, University of Western Australia) for providing expert clinical advice. Disclosure Summary: The authors have nothing to disclose. References 1. Austin MA , Hutter CM , Zimmern RL , Humphries SE . Familial hypercholesterolemia and coronary heart disease: a HuGE association review . Am J Epidemiol  . 2004 ; 160 ( 5 ): 421 – 429 . Google Scholar Crossref Search ADS PubMed   2. Soutar AK , Naoumova RP . Mechanisms of disease: genetic causes of familial hypercholesterolemia . Nat Clin Pract Cardiovasc Med  . 2007 ; 4 ( 4 ): 214 – 225 . Google Scholar Crossref Search ADS PubMed   3. Watts GF , Shaw JE , Pang J , Magliano DJ , Jennings GL , Carrington MJ . Prevalence and treatment of familial hypercholesterolaemia in Australian communities . Int J Cardiol  . 2015 ; 185 : 69 – 71 . Google Scholar Crossref Search ADS PubMed   4. Benn M , Watts GF , Tybjærg-Hansen A , Nordestgaard BG . Mutations causative of familial hypercholesterolaemia: screening of 98 098 individuals from the Copenhagen General Population Study estimated a prevalence of 1 in 217 . Eur Heart J  . 2016 ; 37 ( 17 ): 1384 – 1394 . Google Scholar Crossref Search ADS PubMed   5. Nordestgaard BG , Chapman MJ , Humphries SE , Ginsberg HN , Masana L , Descamps OS , Wiklund O , Hegele RA , Raal FJ , Defesche JC , Wiegman A , Santos RD , Watts GF , Parhofer KG , Hovingh GK , Kovanen PT , Boileau C , Averna M , Borén J , Bruckert E , Catapano AL , Kuivenhoven JA , Pajukanta P , Ray K , Stalenhoef AF , Stroes E , Taskinen MR , Tybjærg-Hansen A ; European Atherosclerosis Society Consensus Panel . Familial hypercholesterolemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease . Eur Heart J  . 2013 ; 34 : 3478 – 3490 . Google Scholar Crossref Search ADS PubMed   6. Hooper AJ , Nguyen LT , Burnett JR , Bates TR , Bell DA , Redgrave TG , Watts GF , van Bockxmeer FM . Genetic analysis of familial hypercholesterolaemia in Western Australia . Atherosclerosis  . 2012 ; 224 ( 2 ): 430 – 434 . Google Scholar Crossref Search ADS PubMed   7. Mortality in treated heterozygous familial hypercholesterolaemia: implications for clinical management. Scientific Steering Committee on behalf of the Simon Broome Register Group . Atherosclerosis  . 1999 ; 142 ( 1 ): 105 – 112 . 8. Williams RR , Hunt SC , Schumacher MC , Hegele RA , Leppert MF , Ludwig EH , Hopkins PN . Diagnosing heterozygous familial hypercholesterolemia using new practical criteria validated by molecular genetics . Am J Cardiol  . 1993 ; 72 ( 2 ): 171 – 176 . Google Scholar Crossref Search ADS PubMed   9. World Health Organization. Familial hypercholesterolemia (FH): report of a second WHO consultation. Available at: http://apps.who.int/iris/bitstream/10665/66346/1/WHO_HGN_FH_CONS_99.2.pdf. Accessed 10 January 2018. 10. Ahmad ZS , Andersen RL , Andersen LH , O’Brien EC , Kindt I , Shrader P , Vasandani C , Newman CB , deGoma EM , Baum SJ , Hemphill LC , Hudgins LC , Ahmed CD , Kullo IJ , Gidding SS , Duffy D , Neal W , Wilemon K , Roe MT , Rader DJ , Ballantyne CM , Linton MF , Duell PB , Shapiro MD , Moriarty PM , Knowles JW . US physician practices for diagnosing familial hypercholesterolemia: data from the CASCADE-FH registry . J Clin Lipidol  . 2016 ; 10 ( 5 ): 1223 – 1229 . Google Scholar Crossref Search ADS PubMed   11. Gidding SS , Champagne MA , de Ferranti SD , Defesche J , Ito MK , Knowles JW , McCrindle B , Raal F , Rader D , Santos RD , Lopes-Virella M , Watts GF , Wierzbicki AS ; American Heart Association Atherosclerosis, Hypertension, and Obesity in Young Committee of 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 . The agenda for familial hypercholesterolemia: a scientific statement from the American Heart Association . Circulation  . 2015 ; 132 ( 22 ): 2167 – 2192 . Google Scholar Crossref Search ADS PubMed   12. Chan DC , Pang J , Hooper AJ , Burnett JR , Bell DA , Bates TR , van Bockxmeer FM , Watts GF . Elevated lipoprotein(a), hypertension and renal insufficiency as predictors of coronary artery disease in patients with genetically confirmed heterozygous familial hypercholesterolemia . Int J Cardiol  . 2015 ; 201 : 633 – 638 . Google Scholar Crossref Search ADS PubMed   13. Haralambos K , Whatley SD , Edwards R , Gingell R , Townsend D , Ashfield-Watt P , Lansberg P , Datta DBN , McDowell IFW . Clinical experience of scoring criteria for familial hypercholesterolaemia (FH) genetic testing in Wales . Atherosclerosis  . 2015 ; 240 ( 1 ): 190 – 196 . Google Scholar Crossref Search ADS PubMed   14. Kinpara K , Okada H , Yoneyama A , Okubo M , Murase T . Lipoprotein(a)-cholesterol: a significant component of serum cholesterol . Clin Chim Acta  . 2011 ; 412 ( 19-20 ): 1783 – 1787 . Google Scholar Crossref Search ADS PubMed   15. Johansen CT , Dubé JB , Loyzer MN , MacDonald A , Carter DE , McIntyre AD , Cao H , Wang J , Robinson JF , Hegele RA . LipidSeq: a next-generation clinical resequencing panel for monogenic dyslipidemias . J Lipid Res  . 2014 ; 55 ( 4 ): 765 – 772 . Google Scholar Crossref Search ADS PubMed   16. Friedewald WT , Levy RI , Fredrickson DS . Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge . Clin Chem  . 1972 ; 18 ( 6 ): 499 – 502 . Google Scholar PubMed   17. Damgaard D , Larsen ML , Nissen PH , Jensen JM , Jensen HK , Soerensen VR , Jensen LG , Faergeman O . The relationship of molecular genetic to clinical diagnosis of familial hypercholesterolemia in a Danish population . Atherosclerosis  . 2005 ; 180 ( 1 ): 155 – 160 . Google Scholar Crossref Search ADS PubMed   18. Silva PRS , Jannes CE , Oliveira TGM , Miname MH , Rocha VZ , Chacra AP , Gurgel MHC , Montenegro RM , Rodrigues Sobrinho CRM , Bello Moreira AS , Assad MHV , Pinto MRC , Tada MT , Santos RD , Pereira AC , Krieger JE . Evaluation of clinical and laboratory parameters used in the identification of index cases for genetic screening of familial hypercholesterolemia in Brazil . Atherosclerosis  . 2017 ; 263 : 257 – 262 . Google Scholar Crossref Search ADS PubMed   19. Civeira F , Ros E , Jarauta E , Plana N , Zambon D , Puzo J , Martinez de Esteban JP , Ferrando J , Zabala S , Almagro F , Gimeno JA , Masana L , Pocovi M . Comparison of genetic versus clinical diagnosis in familial hypercholesterolemia . Am J Cardiol  . 2008 ; 102 ( 9 ): 1187 – 1193, 1193.e1 . Google Scholar Crossref Search ADS PubMed   20. Khera AV , Won HH , Peloso GM , Lawson KS , Bartz TM , Deng X , van Leeuwen EM , Natarajan P , Emdin CA , Bick AG , Morrison AC , Brody JA , Gupta N , Nomura A , Kessler T , Duga S , Bis JC , van Duijn CM , Cupples LA , Psaty B , Rader DJ , Danesh J , Schunkert H , McPherson R , Farrall M , Watkins H , Lander E , Wilson JG , Correa A , Boerwinkle E , Merlini PA , Ardissino D , Saleheen D , Gabriel S , Kathiresan S . Diagnostic yield and clinical utility of sequencing familial hypercholesterolemia genes in patients with severe hypercholesterolemia . J Am Coll Cardiol  . 2016 ; 67 ( 22 ): 2578 – 2589 . Google Scholar Crossref Search ADS PubMed   21. Clarke RE , Padayachee ST , Preston R , McMahon Z , Gordon M , Graham C , Crook MA , Wierzbicki AS . Effectiveness of alternative strategies to define index case phenotypes to aid genetic diagnosis of familial hypercholesterolaemia . Heart  . 2013 ; 99 ( 3 ): 175 – 180 . Google Scholar Crossref Search ADS PubMed   22. Ellis KL , Pang J , Chan DC , Hooper AJ , Bell DA , Burnett JR , Watts GF . Familial combined hyperlipidemia and hyperlipoprotein(a) as phenotypic mimics of familial hypercholesterolemia: frequencies, associations and predictions . J Clin Lipidol  . 2016 ; 10 ( 6 ): 1329 – 1337.e3 . Google Scholar Crossref Search ADS PubMed   23. Fouchier SW , Dallinga-Thie GM , Meijers JC , Zelcer N , Kastelein JJ , Defesche JC , Hovingh GK . Mutations in STAP1 are associated with autosomal dominant hypercholesterolemia . Circ Res  . 2014 ; 115 ( 6 ): 552 – 555 . Google Scholar Crossref Search ADS PubMed   24. Talmud PJ , Shah S , Whittall R , Futema M , Howard P , Cooper JA , Harrison SC , Li K , Drenos F , Karpe F , Neil HA , Descamps OS , Langenberg C , Lench N , Kivimaki M , Whittaker J , Hingorani AD , Kumari M , Humphries SE . Use of low-density lipoprotein cholesterol gene score to distinguish patients with polygenic and monogenic familial hypercholesterolaemia: a case-control study . Lancet  . 2013 ; 381 ( 9874 ): 1293 – 1301 . Google Scholar Crossref Search ADS PubMed   25. Iacocca MA , Hegele RA . Recent advances in genetic testing for familial hypercholesterolemia . Expert Rev Mol Diagn  . 2017 ; 17 ( 7 ): 641 – 651 . Google Scholar Crossref Search ADS PubMed   26. Alonso R , Mata P , Zambón D , Mata N , Fuentes-Jiménez F . Early diagnosis and treatment of familial hypercholesterolemia: improving patient outcomes . Expert Rev Cardiovasc Ther  . 2013 ; 11 ( 3 ): 327 – 342 . Google Scholar Crossref Search ADS PubMed   27. Watts GF , Gidding S , Wierzbicki AS , Toth PP , Alonso R , Brown WV , Bruckert E , Defesche J , Lin KK , Livingston M , Mata P , Parhofer KG , Raal FJ , Santos RD , Sijbrands EJ , Simpson WG , Sullivan DR , Susekov AV , Tomlinson B , Wiegman A , Yamashita S , Kastelein JJ . Integrated guidance on the care of familial hypercholesterolaemia from the International FH Foundation . Int J Cardiol  . 2014 ; 171 ( 3 ): 309 – 325 . Google Scholar Crossref Search ADS PubMed   28. Wiegman A , Gidding SS , Watts GF , Chapman MJ , Ginsberg HN , Cuchel M , Ose L , Averna M , Boileau C , Borén J , Bruckert E , Catapano AL , Defesche JC , Descamps OS , Hegele RA , Hovingh GK , Humphries SE , Kovanen PT , Kuivenhoven JA , Masana L , Nordestgaard BG , Pajukanta P , Parhofer KG , Raal FJ , Ray KK , Santos RD , Stalenhoef AF , Steinhagen-Thiessen E , Stroes ES , Taskinen MR , Tybjærg-Hansen A , Wiklund O ; European Atherosclerosis Society Consensus Panel . Familial hypercholesterolaemia in children and adolescents: gaining decades of life by optimizing detection and treatment . Eur Heart J  . 2015 ; 36 ( 36 ): 2425 – 2437 . Google Scholar Crossref Search ADS PubMed   29. Troeung L , Arnold-Reed D , Chan She Ping-Delfos W , Watts GF , Pang J , Lugonja M , Bulsara M , Mortley D , James M , Brett T . A new electronic screening tool for identifying risk of familial hypercholesterolaemia in general practice . Heart  . 2016 ; 102 ( 11 ): 855 – 861 . Google Scholar Crossref Search ADS PubMed   30. Lázaro P , Pérez de Isla L , Watts GF , Alonso R , Norman R , Muñiz O , Fuentes F , Mata N , López-Miranda J , González-Juanatey JR , Díaz-Díaz JL , Blasco AJ , Mata P . Cost-effectiveness of a cascade screening program for the early detection of familial hypercholesterolemia . J Clin Lipidol  . 2017 ; 11 ( 1 ): 260 – 271 . Google Scholar Crossref Search ADS PubMed   31. Knowles JW , Rader DJ , Khoury MJ . Cascade screening for familial hypercholesterolemia and the use of genetic testing . JAMA  . 2017 ; 318 ( 4 ): 381 – 382 . Google Scholar Crossref Search ADS PubMed   32. Kindt I , Mata P , Knowles JW . The role of registries and genetic databases in familial hypercholesterolemia . Curr Opin Lipidol  . 2017 ; 28 ( 2 ): 152 – 160 . Google Scholar Crossref Search ADS PubMed   33. Junyent M , Gilabert R , Zambón D , Núñez I , Vela M , Civeira F , Pocoví M , Ros E . The use of Achilles tendon sonography to distinguish familial hypercholesterolemia from other genetic dyslipidemias . Arterioscler Thromb Vasc Biol  . 2005 ; 25 ( 10 ): 2203 – 2208 . Google Scholar Crossref Search ADS PubMed   Copyright © 2018 Endocrine Society


This is a preview of a remote PDF: https://academic.oup.com/jcem/article-pdf/103/4/1704/24565765/jc.2017-02622.pdf

Chan, Dick C, Pang, Jing, Hooper, Amanda J, Bell, Damon A, Bates, Timothy R, Burnett, John R, Watts, Gerald F. A Comparative Analysis of Phenotypic Predictors of Mutations in Familial Hypercholesterolemia, The Journal of Clinical Endocrinology & Metabolism, 2018, 1704-1714, DOI: 10.1210/jc.2017-02622