Validation of the familial hypercholesterolemia diagnosis in the Danish National Patient Registry and the Danish familial hypercholesterolemia registry

Abstract Background Familial hypercholesterolemia (FH) is a common inherited disease characterized by elevated low-density lipoprotein (LDL)-cholesterol and a high risk of premature atherosclerotic cardiovascular disease. The Danish FH registry was established in 2020 to monitor diagnostics, detection and treatment quality of patients with FH and this nationwide registry together with the Danish National Patient Registry (DNPR) include more than 10,000 cases of FH. The usability of large registries relies on the validity of registered information. However, the validity of registered diagnoses in the DNPR and the Danish FH registry are currently unknown. Purpose We aimed to investigate the positive predictive value (PPV) of FH cases registered in the DNPR and Danish FH registry Method In this retrospective study, we retrieved a nationwide random sample of 800 registered FH diagnoses (ICD-10: DE780B and DE780B1+2) in the DNPR corresponding to 7.7% of all registered FH diagnoses in the country. The diagnosis was validated against information retrieved through medical records. All individuals were categorized according to the Dutch Clinical Lipid Network (DLCN) criteria. Individuals with a DLCN-score ≥ 6 were considered valid FH cases and as well as those carrying a pathogen genetic FH variant defined as pathogen genetic variants in the LDL receptor-, Apo lipoprotein B-, and Proprotein convertase subtilisin/kexin type 9 genes. Subsequently, we categorized individuals into those fulfilling a diagnosis of genetic FH or clinical FH. Positive predictive values (PPV) with 95% confidence intervals were used as measures of validity. The regional council approved the study (23/19548). Results We found a PPV of 63.8% (95%CI: 60.7; 67.0) for any FH diagnosis in the DNPR (510/800), whereas the PPV was 88.1% (95%CI: 82.2; 92.3%) for a genetic FH diagnosis (141/160). A primary diagnosis of FH established at a Department of Cardiology had a PPV of 71.0% (95%CI: 67.2; 74.6%) for any FH diagnosis (417/587) and the PPV was 88.3% (95%CI: 82.6; 92.8%) for a genetic FH diagnosis (113/128), respectively. Diagnoses of genetic FH was under-registered with only 216 subjects coded with a DE780B1 diagnosis corresponding to 55.6% (95%CI: 50.7; 60.5) of 388 subjects carrying a pathogenic FH variant according to the medical record. The Danish FH registry covered 59.1% (95%CI: 55.6; 62.5%) of all registered FH diagnoses in our study population retrieved from the DNPR (473/800). The PPV of individuals registered in the Danish FH registry with a DLCN 6 was 83.0% (95%CI: 78.9; 86.5) (318/383). Conclusion The PPV of an overall FH diagnosis in the DNPR was 63.8%, but restricting to a primary diagnosis of FH from cardiology units may be used to improve the validity of registered FH diagnoses. The PPV of FH cases in the Danish FH registry was 83.0%. Thus, diagnostic practices need to be optimized to monitor the marked focus on identifying far more patients with FH.