We evaluated the data of 2011 patients with FH diagnosed clinically using the Japan Atherosclerosis Society (JAS) 2017 criteria at Kanazawa University Hospital between 1990 and 2020. All the patients in this study fulfilled at least two of the three essential clinical criteria stipulated by the JAS for FH diagnosis. The criteria are as follows: 1) LDL cholesterol level ≥180 mg/dl, 2) tendon xanthoma on the backs of the hands, elbows, knees, or other areas; Achilles tendon hypertrophy or Achilles tendon thickness on X-ray ≥ 9 mm; or xanthoma tuberosum, and 3) family history of FH or premature coronary artery disease diagnosed in a first- or second-degree relative. Nine hundred and sixty-one patients were excluded due to missing data (such as data on blood lipids and genetic analysis or data on homozygous and compound heterozygous FH. Finally, 1,050 patients were included in this study (Supplementaary Figure S1).

We defined hypertension as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or use of antihypertensive agents. Further, we used the definition of diabetes given by the Japan Diabetes Society (Araki et al., 2020). Smoking status was defined as a current smoking status. Cardiovascular disease (CVD) was defined as angina pectoris, myocardial infarction, or severe stenotic region(s) in the coronary artery (≥75% stenosis), identified on angiography or computed tomography. Serum levels of total cholesterol, triglycerides, and high-density lipoprotein cholesterol were determined enzymatically using automated instrumentation. LDL cholesterol level was calculated using the Friedewald formula if triglyceride level was <400 mg/dl; otherwise, it was determined enzymatically.

We assessed genotypes using a next-generation sequencer. In brief, the coding regions of LDLR, APOB, PCSK9, and LDLRAP1 were sequenced as described in a previous study (Tada et al., 2018). Further, copy number variations at the LDLR were assessed using the eXome Hidden Markov Model software as described in an earlier study (Yamamoto et al., 2016). We evaluated the pathogenicity of the genetic variants according to the standard American College of Medical Genetics and Genomics criteria (Richards et al., 2015). We classified pathogenic variants as missense variants or PTVs, which include frameshift variants, large deletion or duplication variants, nonsense variants, and splice site variants.

This study was approved by the Ethics Committee of Kanazawa University. All procedures were conducted in accordance with the ethical standards of the Human Research Committee (institutional and national) and the Helsinki Declaration (1975, revised in 2008). Informed consent for genetic analysis was obtained from all the study participants.

Categorical variables were reported as numbers and percentages, and they were compared using Fisher’s exact test or chi-square test. Normally distributed continuous variables were reported as means ± standard deviations. Not normally distributed continuous variables were reported as medians and interquartile ranges. Mean values of continuous variables were compared using Student’s t-test for independent variables, and median values were compared using non-parametric Wilcoxon–Mann–Whitney rank sum test. Chi-square or Fisher’s post-hoc test was used for categorical variables as indicated. Cox proportional hazard model was used to assess relationships between all the variables. Cumulative Kaplan–Meier survival curves starting at baseline were constructed to compare times to the first MACE. All statistical analyses were conducted using R statistics (https://www.r-project.org). p-values < 0.05 were considered statistically significant.

The clinical characteristics of the study participants are shown in Table 1. The mean age of patients was 49 years, and almost half of the patients were men. The median LDL cholesterol level at baseline was 244 mg/dl, and it decreased to 110 mg/dl at follow-up. A total of 776 patients (73.9%) had a family history of FH and/or premature CVD. Furthermore, 290 patients (27.6%) had a history of CVD. Upon division of patients into two groups based on the occurrence of MACEs, we observed significant differences in all variables (except family history of FH and/or premature CVD) between the two groups. In addition, we observed differences in the trend of characteristics; for example, we observed, following division of patients with pathogenic mutations into three groups based on genotype, that the proportion of patients with diabetes decreased (Supplementary Table S1). The medical treatments administered during follow-up are summarized in Table 2. Statin therapy, frequently followed by ezetimibe and colestimide therapy, was administered to most of the patients.

We identified 89 pathogenic variants in 665 patients. Of the pathogenic variants, 277 were classified as missense variants, and 388 were classified as PTVs (46 were frameshift variants, 69 were large deletion or duplication variants, 204 were nonsense variants, and 69 were splice site variants; Figure 1). The details are shown in Supplementary Table S2.

Based on genotype, we divided the patients into the following three groups: patients without pathogenic variants, patients with missense variants, and patients with PTVs. We found that LDL cholesterol level was highest in patients with PTVs, and patients with missense variants were found to have higher LDL cholesterol levels than patients without pathogenic variants (Figure 2A). The median LDL cholesterol levels of patients without pathogenic variants, patients with missense variants, and patients with PTVs were 216 mg/dl, 236 mg/dl, and 256 mg/dl, respectively. The LDL cholesterol levels of the three groups showed deformed trimodal distributions (Figure 2B).

Over the median follow-up duration of 12.6 years, 175 patients had MACEs, which include CVD-associated death, myocardial infarction, unstable angina, and staged revascularization (Table 3). Using Cox proportional hazard model, we assessed factors associated with MACEs and found that age (hazard ratio [HR] = 1.09, 95% confidence interval [CI] = 1.04–1.14, p = 2.2 × 10⁻¹⁴), male gender (HR = 1.58, 95% CI = 1.08–2.08, p = 0.0079), hypertension (HR = 3.11, 95% CI = 2.10–4.25, p = 6.9 × 10⁻⁶), diabetes (HR = 2.44, 95% CI = 1.46–3.52, p = 0.0021), smoking (HR = 2.56, 95% CI = 1.56–3.18, p = 0.00041), LDL cholesterol level (per 10 mg/dl; HR = 1.01, 95% CI = 1.00–1.02, p = 0.022), and prior CVD (HR = 3.45, 95% CI = 2.02–4.80, p < 2.2 × 10⁻¹⁶) are significantly associated with MACEs (Table 4). We also found that missense variants and PTVs are associated with MACEs (HR = 1.58, 95% CI = 1.08–2.08, p = 0.0033 and HR = 3.24, 95% CI = 2.12–4.40, p = 3.9 × 10⁻⁶, respectively).

We assessed the survival curve and found that patients with PTVs have the worst outcome of the three groups and that patients with missense variants have worse outcomes than patients without pathogenic variants (Figure 3).