How the genetic diagnosis of familial hypercholesterolemia can be guided by clinical features. Preliminary evidences from an Italian single-center experience

Abstract

Aim

Familial Hypercholesterolemia (FH) is highly prevalent in humans. To date, the need for genetic testing is indicated by widely used clinical scores (different for adults and children) that support the diagnosis of FH. The performance of genetic testing is a matter of debate and is likely dependent on different clinical, ethnic and environmental contexts, with sparse data from Italy. Our aims were to evaluate: (i) the performance of genetic testing in Italian patients with a clinical diagnosis of FH; (ii) the role of clinical features on the genetic testing performance.

Methods

Pathogenic (P) or likely pathogenic (LP) variants in ABCG5, ABCG8, APOB, APOE, LDLR, LDLRAP1, LIPA, PCSK9 were investigated by WES and confirmed by Sanger sequencing.

Results

The pick-up rate was 39.8%, progressively increasing (p < 0.001) with increasing clinical score. Discrimination of clinical score was excellent (AU-ROC = 0.897; 95% CI: 0.831, 0.963; p = 3.4 × 10^− 11). Positive predictive value was 95% in people with a clinical score ≥ 8 and only 45% with score 5–7. Negative predictive value was 95% in people with clinical score < 4 and only 55% with score 5–7. Adding age at diagnosis, BMI, and triglycerides (that differed significantly between positive and negative genetic testing patients) to the clinical score tended to improve AU-ROC.

Conclusion

Our preliminary evidences suggest that patients with intermediate levels on the currently available clinical scores are best suited for genetic testing for familial hypercholesterolemia (FH) and that the discriminatory power of these clinical scores might be improved by adding further clinical characteristics.