Genome Sequencing Array Analysis for the Diagnosis of Familial Hypercholesterolemia

Background: Familial hypercholesterolemia (FH) is an underdiagnosed condition that leads to elevated LDL cholesterol levels and significantly increases the risk of early atherosclerosis-related events such as myocardial infarction and stroke. Current diagnosis relies on clinical scoring systems with limited accuracy and gene sequencing. 

Methods: This study tested the novel CARRENAL DNA array designed to detect mutations associated with FH. The array consists of 42,615 SNPs within lipid metabolism-related genes. Seventy patients suspected of having FH were tested using the CARRENAL array and two clinical scoring systems, the Simon Broome Criteria and the Dutch Lipid Clinic Network Score.

Results: The gold standard for diagnosing familial hypercholesterolemia (FH) is Next-Generation Sequencing (NGS). 13 out of 70 patients underwent NGS in addition to the  CARRENAL array. Mutations were identified in 10 of the 13 patients tested by NGS, with a diagnostic concordance between NGS and the array in 4 cases and a different LDL-R mutation indicated by the array compared to NGS was found in one case. The CARRENAL array missed FH mutations in 6 cases that were detected by NGS. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the microarray were 40%, 100%, 100%, and 33.3%, respectivelyUsing the Simon Broome criteria, 6 of the 70 patients (8.6%) fulfilled the criteria of “definite FH”, 51 patients (72.6%) had “probable FH”, and 13 patients (18.6%) were “unlikely” to have FH. In comparison to NGSc testing, the clinical diagnosis of “definite FH” according to Simon Broome achieved a sensitivity, specificity, PPV, and NPV of 14.29%, 94.5%, 40%, and 81.25%, respectively. Including participants with “probable FH” altered these results to a sensitivity of 100%, specificity of 23.64%, PPV of 25%, and NPV of 100%.According to the Dutch Lipid Clinic Network Score (DLCNS) criteria, 15 of the 70 patients (21.4%) were diagnosed with “definite FH”, 20 patients (28.6%) with “probable FH”, 26 patients (37.1%) with “possible FH”, and 9 patients (12.9%) as “unlikely” to have FH. In comparison to genetic testing, the clinical diagnosis of “definite FH” according to DLCNS achieved a sensitivity, specificity, PPV, and NPV of 100%, 98.18%, 93.3%, and 100%, respectively. When considering cases with probable FH, the results changed to a sensitivity of 100%, specificity of 61.82%, PPV of 40%, and NPV of 100%.

Conclusion: The sensitivity of CARRENAL array is not sufficient for routine use. Similarly, and in agreement with previous data, the specificity and the positive predictive value of the clinical scores is low. Currently, LDL-cholesterol remains the most important factor for the identification of FH with genetic sequencing as the gold standard. Further research towards sequential laboratory testing, e.g. array analysis followed by NGS in selected cases, is wanted to reduce the cost of providing the definitive diagnosis of FH.