Vitamin B12 assays compared by use of patients' sera with low vitamin B12 content.

We compared four radioisotope dilution (RD) methods and a microbiological assay for measuring concentrations of vitamin B12 in a selected panel of serum samples from patients known to be deficient in the vitamin. Low (less than 100 ng/L) and borderline (100-180 ng/L) results were similar between methods, but use of the manufacturers' recommended ranges for borderline results would have changed the diagnostic classifications for 22 of 38 samples. Results of all the RD methods inter-correlated well, but less so with the microbiological assay. Borderline, nondiagnostic results were common to all methods, and no apparent advantage was gained from using the microbiological assay.