Diagnosis of Homozygous Familial Hypercholesterolaemia (HoFH) relies on untreated low-density lipoprotein-cholesterol (LDL-C) which is often unknown. We determine whether untreated LDL-C can be imputed from treated LDL-C in HoFH.

Two groups with HoFH were identified: Group 1 (n = 193) from Canada, Brazil and South Africa; Group 2 (n = 206) from the HoFH International Clinical Collaboration. Pre- and post-treatment LDL-C and lipid lowering therapy (LLT) intensity from Group 1 were used to develop a regression model and applied to treated LDL-C in Group 2 to impute pre-treatment LDL-C. The same process was performed in reverse. A final imputation model was created from combining both groups.

There was a curvilinear relationship between the expected and observed % lowering of LDL-C on LLT (r = 0.3923, p < 0.0001, Standard Error [SE] = 23 %). Using this relationship, LDL-C was imputed from treated values and showed significant correlation with pre-treatment LDL-C (r = 0.71, p < 0.001; mean values 13.4 ± 4.7 [Standard Deviation] and 13.6 ± 7.3 mmol/L, respectively, ns). Concordance between actual and imputed values ≥ 10 or <10 mmol/L was 80 %. Whereas 36 % of patients had treated LDL-C ≥ 10 mmol/L, 64 % had treated or imputed pre-treatment LDL-C ≥ 10 mmol/L.

In HoFH, the response to LLT can be quantified and used to impute untreated LDL-C from treated LDL-C. Imputation may augment awareness of possible HoFH in treated subjects lacking records of untreated LDL-C.