Concordance of Lipoprotein(a) measurements in mg/dL and nmol/L: Insights from the Lp(a)HORIZON trial

BACKGROUND AND AIM: Challenges in measurement standardization have complicated both research and clinical care. We assessed the concordance of Lp(a) concentration measured in mg/dL and nmol/L using the Roche Tina-quant Lipoprotein(a) Gen 2 mass and molarity Assay™ with values calculated using two published conversion methods in the Lp(a)HORIZON screening program. METHODS: Lp(a) concentrations were analyzed from 14,247 screened patients for the Lp(a) HORIZON trial. Of those, 13,666 had Lp(a) measurement in both mg/dL and nmol/L. Agreement between directly measured Lp(a) nmol/L values and nmol/L values converted from mg/dL using two published conversion methods - the Copenhagen General Population Study (CGPS) formula and the until recently widely used conversion factor of 2.5 - was assessed. The accuracy of conversion formula was evaluated by comparing converted nmol/L values with measured nmol/L values, using Bland-Altman plots, and by examining concordance with the Lp(a)HORIZON trial selection thresholds (70 mg/dL and 90 mg/dL). RESULTS: Overall, the CGPS formula showed moderately greater agreement with observed nmol/L measurements (Mean Absolute Error (MAE) = 16.95, Normalized Root Mean Squared Error (NRMSE) = 0.21) compared with the conversion factor of 2.5 (MAE = 42.30, NRMSE = 0.44). For the CGPS and conversion factor of 2.5, the accuracy between directly measured and converted values for selection thresholds used in the Lp(a)HORIZON trial was 97% and 92% at the 70 mg/dL, and 94% and 85% at 90 mg/dL thresholds, respectively. CONCLUSIONS: The CGPS conversion formula showed a better agreement with measured Lp(a) values and more accurately reproduced the Lp(a)HORIZON selection thresholds than the conversion factor of 2.5.