RESUMO
BACKGROUND AND AIMS: There is controversy on the accuracy of different diagnostic criteria for familial hypercholesterolemia (FH). The aim of this study is to assess the performance of different clinical criteria used to identify individuals for FH genetic cascade screening in Brazil. METHODS: All index cases (IC) registered in the Hipercol Brasil program between 2011 and 2016 were analyzed. Inclusion criteria were age ≥18 years and elevated LDL-cholesterol (LDL-C) levels, with a conclusive result in the genetic test, whether positive or negative. Initially, we tested the multivariable association between clinical and laboratory markers and the presence of an FH causing mutation. Then, we analyzed sensitivity, specificity, positive and negative predictive values for the LDL-C quartile distribution, LDL-C as a continuous variable, as well as the performance measures for the Dutch Lipid Clinic Network (DLCN) score to identify a mutation. RESULTS: Overall, 753 ICs were included and an FH causing mutation was found in 34% (n = 257) of the subjects. After multivariable analysis, LDL-C as a continuous variable, tendon xanthomas and corneal arcus were independently associated with the presence of FH mutations. LDL-C values ≥ 230 mg/dL (5.9 mmol/L) had the best tradeoff between sensitivity and specificity to diagnose a mutation. The DLCN score presented a better performance than LDL-C to identify a mutation, area under the ROC curve were 0.744 (95% CI: 0.704-0.784) and 0.730 (95% CI: 0.687-0.774), respectively, p=0.014. CONCLUSIONS: In our population, LDL ≥230 mg/dL is a feasible criterion to indicate ICs to genetic testing.