Dutch Lipid Clinic Network Criteria for familial hypercholesterolaemia

Clinical background

Familial hypercholesterolaemia (FH) is an autosomal dominant genetic disorder that leads to raised LDL cholesterol from birth and a markedly increased risk of premature coronary artery disease. The prevalence in northern European populations is estimated at approximately 1 in 200, but fewer than 1% of these patients are diagnosed in most countries [1]. In untreated heterozygous FH, men develop coronary artery disease on average before the age of 55 and women before 60 [1].

The Dutch Lipid Clinic Network Criteria (DLCN) were developed to standardise the clinical diagnosis of heterozygous FH in settings where genetic testing is not immediately available. The criteria are based on a composite of family history, the patient's own history, clinical findings, untreated LDL cholesterol and the result of genetic analysis. The score stratifies the patient as definite, probable, possible or unlikely FH and is used to prioritise further investigation and treatment, and to identify relatives who should be offered cascade screening [1].

Calculating the Dutch Lipid Clinic Network Criteria

The DLCN score is the sum of the single highest-scoring finding within each of five categories. Within each category it is therefore not the sum of all positive findings that counts, but only the finding that gives the highest score.

$$\text{DLCN} = \max(H) + \max(P) + \max(C) + f(\text{LDL-C}) + D$$

where $\max(H)$ is the highest score within family history (0–2), $\max(P)$ is the highest score within the patient's own history (0–2), $\max(C)$ is the highest score within clinical examination (0–6), $f(\text{LDL-C})$ is the untreated LDL cholesterol level converted into points (0–8) and $D$ is a DNA mutation (0 or 8). The LDL-C score is determined by:

$$f(\text{LDL-C}) = \begin{cases} 8 & \text{if LDL-C} \geq 8{.}5 \text{ mmol/L} \ 5 & \text{if } 6{.}5 \leq \text{LDL-C} < 8{.}5 \ 3 & \text{if } 5{.}0 \leq \text{LDL-C} < 6{.}5 \ 1 & \text{if } 4{.}0 \leq \text{LDL-C} < 5{.}0 \ 0 & \text{if LDL-C} < 4{.}0 \end{cases}$$

A distinctive feature of the DLCN is that an identified functional mutation in LDLR, APOB or PCSK9 gives 8 points, which on its own places the patient in the probable FH category and, combined with any other positive finding, in the definite FH category. This reflects the fact that genetic confirmation is the strongest single diagnostic element.

The criteria were developed by a Dutch network of lipid clinics and published alongside an international expert panel's guidelines for heterozygous FH [2]. They were subsequently adopted by the European Atherosclerosis Society as the clinical diagnostic tool in the 2013 European consensus statement [1]. The DLCN was not derived from a statistical model on a defined cohort but rests on expert consensus. The major validation in an unselected general population came from Copenhagen, where 69,016 participants in the Copenhagen General Population Study were classified with the DLCN criteria and the prevalence of definite or probable FH (>5 points) was found to be approximately 1 in 200 [1].

Interpretation in practice

A patient who falls into the probable or definite FH category should be offered genetic testing where available, and if a pathogenic mutation is identified, first-degree relatives should be offered cascade screening with a lipid profile and, where possible, genetic testing [1].

Validation and performance

In the Japanese validation study by Tada et al., 857 patients referred to a university hospital between 2010 and 2022 were analysed [3]. The proportion of patients with a pathogenic variant in FH-related genes was 77.1% in the definite FH group, 28.7% in probable FH, 13.0% in possible FH and 1.2% in unlikely FH (P for trend <0.001). Cardiovascular disease was significantly associated with the diagnostic category, with an adjusted OR of 9.1 for definite FH, 4.2 for probable FH and 2.8 for possible FH compared with unlikely FH [3]. This confirms that DLCN stratification works as intended for risk separation.

In the Italian LIPIGEN study of 1,377 adults with genetically confirmed FH, only 37.9% were classified as definite FH and 28.5% as probable FH by the DLCN [4]. As many as 43.4% of patients lacked data for at least one of the eight criteria, and approximately 10% lacked data for four or more. The study also showed that the use of measured versus estimated untreated LDL-C values could substantially change the score [4].

An Austrian study of 469 patients who underwent genetic testing at two specialist lipid clinics found that the DLCN performed no better than an untreated LDL-C threshold of 190 mg/dL (4.9 mmol/L) alone in predicting genetically confirmed FH [5]. The sensitivity of the DLCN was 53.8% compared with 84.9% for LDL-C alone, while specificity was 84.1% and 39.0% respectively. Substantial discrepancies were observed between assessments made by the treating physician and retrospective scoring, and a median of 3 of 8 criteria were missing in clinical records [5].

Limitations

The DLCN was developed for adult patients with suspected heterozygous FH. It is not validated for children, in whom both the LDL-C thresholds and the clinical findings differ, and separate paediatric criteria have been proposed [1].

The commonest problem in practice is incomplete data. In both the Italian and the Austrian cohort a large proportion of criteria were missing, which systematically lowers the score and leads to genuine FH cases being missed [4, 5]. Family history data and the findings of clinical examination (tendon xanthomas, corneal arcus) in particular are often unavailable or not documented in the record [4].

The LDL-C score requires an untreated value. If the patient is already on lipid-lowering treatment, the value must be corrected, usually by multiplying by 1.3 for statin treatment, which rests on assumptions about treatment effect and introduces uncertainty [4]. Alternatively, the untreated LDL-C can be estimated from the treated value and the known treatment effect, but this method has been shown to affect the DLCN score substantially [4].

The DLCN does not identify all patients with genetically confirmed FH. In the LIPIGEN study, just over 60% of genetic FH cases were classified as possible or unlikely FH [4], and in the Austrian cohort sensitivity was only 53.8% [5]. This is particularly marked for patients with moderately raised LDL-C who carry mutations with a milder phenotype, such as certain PCSK9 variants [3].

Finally, the tool has not been shown to outperform a simple LDL-C threshold for screening in the general population [5]. Its strength lies rather in combining several sources of information into an integrated clinical assessment, and in providing a structured basis for prioritising genetic testing and cascade screening of relatives.

References

1. Nordestgaard BG et al. Familial hypercholesterolaemia is underdiagnosed and undertreated in the general population: guidance for clinicians to prevent coronary heart disease: consensus statement of the European Atherosclerosis Society. European Heart Journal 2013. PMID: 23956253
2. Civeira F, International Panel on Management of Familial Hypercholesterolemia. Guidelines for the diagnosis and management of heterozygous familial hypercholesterolemia. Atherosclerosis 2004. PMID: 15177124
3. Tada H et al. Validation of the 2022 Clinical Diagnostic Criteria of Familial Hypercholesterolemia in Japan. Journal of Atherosclerosis and Thrombosis 2024. PMID: 37967952
4. Casula M et al. Evaluation of the performance of Dutch Lipid Clinic Network score in an Italian FH population: The LIPIGEN study. Atherosclerosis 2018. PMID: 30270079
5. Ferch M et al. Performance of LDL-C only compared to the Dutch Lipid Clinic Network Score for screening of familial hypercholesterolaemia: the Austrian experience and literature review. European Journal of Preventive Cardiology 2025. PMID: 39535057