Molecular genetic analysis

Genomic DNA was extracted from peripheral blood using the Maxwell RSC kit (Promega GmbH, Madison, Wisconsin, United States), and its quality and quantity were then measured using a Quantus fluorometer (Promega GmbH). Next‑generation sequencing (NGS) analysis of the proband genomic DNA was conducted using the Ampliseq for Illumina on‑demand kit (Illumina, San Diego, California, United States). After preparing the libraries, the samples were amplified and sequenced using the MiSeq sequencer (Illumina). Bioinformatic analyses were conducted using the BaseSpace and Variant Interpreter applications (Illumina) in relation to the human genome hg19. The nomenclature of the identified changes followed the 2016 Human Genome Variation Society. The interpretation of the significance of the detected genetic variants and their connection with the observed disease was classified using current databases, including Varsome, ClinVar, Franklin, GeneBe, expert recommendations, and scientific publications. The presence of pathogenic variants in the probands was confirmed by Sanger sequencing using a 3500 Genetic Analyzer (Thermo Fischer Scientific, Waltham, Massachusetts, United States). The multiplex ligation‑dependent probe amplification (MLPA) test for detecting deletions or duplications in the LDLR gene was carried out using SALSA MLPA Probemix for the LDLR gene (MRC‑Holland, Amsterdam, the Netherlands).

Statistical analysis

The patients with a confirmed pathogenic genetic mutation for FH were compared with those without such confirmation. Categorical variables are shown as numbers and / or percentages. Continuous variables are presented as mean (SD) or median (interquartile range [IQR]). Differences between the groups were analyzed using the t test for normally distributed continuous variables, while the Wilcoxon rank‑sum test was employed for non‑normally distributed continuous variables. The Pearson χ² test was used to evaluate categorical variables. In the case of ordinal variables with more than 2 values, the distribution of FH pathogenic mutations was compared using the Cochran–Armitage trend test.

Univariable and multivariable logistic regressions identified predictors of the outcome of a genetic test confirming pathogenic FH mutation. The multiple regression model included factors identified by the stepwise regression model with a P value threshold (0.25 to enter, 0.1 to leave). A 2‑sided P value below 0.05 indicated significance. JMP package, version 17.1.0 (JMP Statistical Discovery, Cary, North Carolina, United States) was utilized for all statistical analyses.

Model performance

Within the primary cardiovascular prevention Kordian program, a high FH risk group was identified and referred to a tertiary center for comprehensive diagnostics, including genetic testing. The patients with confirmed pathogenic mutations responsible for FH accounted for 4.6% of the entire population included in the Kordian program, 15.1% of the population referred to the reference coordinating center, and 45.6% of the population selected by specialists for genetic testing. This shows reasonable specificity of the proposed algorithm. Given the estimated prevalence of FH in the Polish population of 0.4%, it can be concluded that the Kordian program initially identified a group of patients with nearly a 12‑fold higher risk of FH than the general population. Furthermore, PCP narrowed down the population to individuals with the risk increased by more than 38 times. The specialists referring the participants for genetic tests helped identify a population with a 114 times higher FH risk than the general Polish population. Projects similar to Kordian, involving PCP and centers for lipid disorder diagnosis and treatment, were conducted in other European countries, with detection rates of pathogenic FH mutations similar to ours. For example, a program with the acronym ARIAN was implemented in Spain. After excluding secondary causes of high LDL‑C levels (above 6.5 mmol/l), patients with DLCN score of at least 6 points were referred for genetic testing. Under the ARIAN program, 153 patients were referred for genetic testing; 67 were confirmed to have a pathogenic FH mutation.⁷ The patients with confirmed FH accounted for nearly 44% of those tested, allowing the program to identify high‑risk FH individuals as effectively as Kordian. Additionally, researchers from the United Kingdom found pathogenic FH mutations in 36.5% of 635 patients who met the Simon Broome FH criteria (definite and probable FH).⁸ Among patients in the Korean population eligible for genetic testing under the same classification, 32% were found to have pathogenic FH mutations in a cohort of 97 individuals.⁹ In a London hospital, among 204 patients qualifying for genetic testing under the Simon Broome criteria, nearly 55% had pathogenic FH mutations.¹⁰ In our group, this value of 45.6% corresponds with that reported by researchers in London. This Figure ranks among the highest recorded in the literature, supporting the assumptions behind the methodology used to select patients for genetic testing in the Kordian program.

Data from the Polish population suggest that over half of the patients with a probable diagnosis of FH have been diagnosed and treated solely as part of secondary prevention.² Therefore, developing algorithms for primary prevention, such as those implemented in the Kordian program, is essential for FH mutation carriers. We demonstrated that an integrated primary care–specialist pathway using clinical DLCN scoring effectively identifies FH patients before cardiovascular events occur. Notably, the PCPs conducted DLCN score assessments in a large cohort of patients. The predicted cost of performing a DLCN assessment in the PCP outpatient setting is not high, as it relies on medical history, physical examination, and lipid level measurements; genetic testing is not essential for diagnosing FH. PCPs should have access to a reference center where FH genetic testing can be performed under the National Health Fund contract, and the number of such centers is increasing.

Genetic findings

The first gene linked to FH was the LDLR gene, discovered in the 1970s.¹¹ In our group, 40 patients (70%) had mutations in this gene, aligning with data from other populations. Genetic screening in the Dutch population found that about 67% of FH patients carry an LDLR mutation.¹² In the French population, this percentage was 73.9%,¹³ in the Taiwanese population 89.89%,¹⁴ and it exceeded 94% in an Italian cohort.¹⁵

The second gene linked to FH development is APOB.¹⁶ In our group, 28% of the patients had APOB mutations, a rate higher than in other cohorts. In studies on the Dutch population, this percentage was approximately 8%.¹⁷ A slightly higher mutation rate in this gene was reported in Asian populations, occurring in up to 20% of individuals with a confirmed genetic FH background.¹⁸ This discrepancy may be due to our study analyzing the entire APOB gene, rather than the exon 26 fragment containing the p.Arg3527Gln mutation.¹⁹ Our observations align with data on FH mutations in Poland, where LDLR mutations are the most common.^19-21

Clinical features

The patients with a confirmed pathogenic FH mutation had slightly different clinical characteristics than those without such a confirmation. The patients with a confirmed genetic diagnosis of FH were younger and had lower BMI than those whose diagnosis of FH was not confirmed. The latter showed coexistence of other metabolic phenotypes, as indicated by higher values of triglycerides, alanine transaminase, glucose, and BP. Similar observations were published in 2024 in Ireland, where patients with lipid disorders confirmed to have the genetic basis of FH were less likely to be diagnosed with hypertension, but more likely to have a positive family history of premature CVD.²² Data from the New Zealand population also indicated that patients with high LDL‑C level and genetic FH diagnosis were, similarly to our population, younger and more likely to have relatives with high LDL‑C levels.²³ PCPs started lipid‑lowering therapy with a similar frequency in both our study subgroups. Conversely, ACEIs/sartans were prescribed twice as often in the patients without FH mutations, while β-blockers were used nearly 4 times more frequently as compared with the genetically confirmed FH group. This may indirectly imply that the patients without confirmed FH had more comorbidities, such as arterial hypertension and overweight or obesity. This suggests that their high LDL‑C level is rather of a polygenic than monogenic origin and is part of broader metabolic syndrome.

Analysis of the features that significantly influenced the likelihood of confirming the genetic diagnosis of FH showed that, in this study group, a very important factor was having a first‑degree relative with an LDL‑C level exceeding 4.9 mmol/l. This observation aligns with data from some other available publications.^24,25 A very high LDL‑C level in a relative is a factor often recognized in the literature as significantly increasing the risk of FH. Interestingly, this factor serves as an alternative element to the DLCN risk assessment scales for FH. One such scale was established by the American Heart Association (AHA) and is used to assess the likelihood of FH in specific patients.²⁵ Furthermore, according to this AHA definition, a patient in whom this factor has been identified and who is concurrently diagnosed with an elevated LDL‑C level, also above 4.9 mmol/l, can be diagnosed with FH. Increased LDL‑C level in a relative is also 1 of the 5 criteria of the Simon Broome Familial Hypercholesterolemia Register diagnostic criteria for FH.²⁶ This classification does not account for LDL‑C, but rather for total cholesterol; the limit for FH diagnosis was set at 7.5 mmol/l. Importantly, this classification also includes a criterion for the occurrence of CVD in slightly younger patients than in the case of the DLCN scale; the criterion is met by second‑line relatives who were diagnosed with heart attack under the age of 50 years. Specialists qualifying the patients for genetic testing during the Kordian program adopted a similar reduction in age as a factor significantly increasing the risk of FH in a patient, however, in this case, the age was reduced to 40 years, resulting in the frequency of genetic confirmation of the diagnosis of FH approaching 50% of the population referred for genetic testing, even if this group included patients with 3 points on the DLCN scale. This result emphasizes the crucial importance of family history, as well as the clinical expertise of the specialists working within the Kordian program.

FH is an autosomal‑dominant trait; therefore, an equal prevalence of mutations in both sexes is expected. Interestingly, there were more women than men in the final Kordian cohort; we believe this finding reflects selection bias and likely greater willingness among women to participate in genetics studies.²⁶

Limitations

This study has several shortcomings that warrant discussion. First, the FH screening was conducted using the Kordian program’s primary‑prevention assumptions; therefore, patients with a prior cardiovascular event were excluded. Additionally, the percentage of individuals from the initial Kordian population with identified FH mutations was unexpectedly high (4.6%). Selection bias is the most plausible explanation for this result. It seems that the PCPs were more inclined to encourage patients with particularly high LDL‑C levels and a positive history of CVD to participate. Simultaneously, these patients were probably more willing to accept such an invitation. It may also be worth emphasizing that this is not an epidemiologic study, and that some degree of bias is likely unavoidable. The study lacks information on patient rates of decline to participate, which could provide additional insight into the magnitude of potential selection bias. Furthermore, an inclusion criterion of having 3–5 points on the DLCN scale, along with a strongly positive family history of early CVD or high LDL‑C level in children, was an arbitrary decision based on the clinical judgement of the study team and was not evidence‑based. Additionally, our study did not confirm the pathogenic nature of the identified mutations through functional experiments or by examining their segregation with the FH phenotype within families. Nonetheless, the fact that all of them were classified as pathogenic in the genetic databases and previously reported as FH‑causing variants strongly supports their causal role. Furthermore, the program did not gather detailed data on all initial study participants, including DLCN classification points.

Conclusions

Collaboration between PCPs and specialists with access to genetic testing, guided by a well‑structured algorithm using the DLCN scale, may enable effective identification of patients with FH mutations before the onset of CVD. The results of this study could be valuable for future planning of a nationwide program to screen for FH in adults and support the implementation of primary cardiovascular prevention in this very high‑risk group of mutation carriers. We suggest that PCPs across Poland widely adopt the DLCN scale in their practice, including individuals before a CVD diagnosis, as it is supported by the presented data and aligns with current international diagnostic guidelines.²⁷ This, together with the recent introduction of compulsory screening for high LDL‑C level in 6‑year‑old children in Poland,²⁸ should accelerate the identification of FH mutation carriers in our society.