The Diagnostic Accuracy of Methylmalonic Acid to Detect Inadequate Vitamin B-12 Status Relative to 3cB-12 Was Higher Compared with That of Total Cobalamin and Total Homocysteine in a Cross-Sectional Survey of Apparently Healthy US Adults.

Background: Diagnosing low vitamin B-12 status is important to reduce megaloblastic anemia and neurological disorders. The lack of a highly specific and sensitive clinical marker and questions about appropriate cutoffs hamper early detection. Objectives: We evaluated the predictive power of 3 conventional markers for inadequate vitamin B-12 status [total cobalamin (B-12), methylmalonic acid (MMA), and total homocysteine (tHcy)] using the combined vitamin B-12 indicator, 3cB-12 (≤ Methods: Using NHANES 1999-2004 data for vitamin B-12, MMA, and tHcy for persons ≥20 y with normal renal function, we devised a training set (1999-2002, Results: In the training set, the area under the curve for MMA was 0.963, followed by B-12 (0.944) and tHcy (0.937); it further increased with 2- and 3-marker models (0.985-0.998). In the validation set, the diagnostic accuracy (Youden's index) for the new optimum cutoffs was 0.80 for MMA >229 nmol/L and 0.69 for vitamin B-12 <220 pmol/L and tHcy >10.4 μmol/L. Among the traditional cutoffs, Youden's index varied (0.30-0.79); vitamin B-12 <126 or <148 pmol/L and MMA >376 nmol/L had the highest specificity (>99%). The latter also had reasonable sensitivity (54.9%) and generated an apparent (2.35%) and a sensitivity/specificity-adjusted prevalence (2.48%) close to the 3cB-12 reference estimate (2.74%). Conclusions: When using a single marker, MMA >376 nmol/L may be a useful cutoff for healthy populations with low (<5%) prevalence of inadequate vitamin B-12 status to minimize false positives, whereas MMA >229 nmol/L may be preferred for screening higher-risk groups to detect more cases.