A direct streptavidin-binding assay does not accurately quantitate biotin in human urine

In human urine, the biotin concentration assayed directly using an avidin-binding assay (ABA) apparently overestimates "true" biotin concentration as measured by HPLC separation of biotin from biotin metabolites followed by ABA. Because biotin metabolites account for about half of biotin plus biotin metabolites in human urine, we speculate that the error might arise from biotin metabolites. We sought to test the following hypothesis: biotin measured by direct ABA routinely exceeds true biotin in urine due to biotin metabolites; however, if urinary biotin is quantitated using a streptavidin-binding assay (SABA) that does not detect biotin metabolites, results will agree with true biotin. An assay for biotin that uses europium coupled to streptavidin and time-resolved fluorescence was developed and validated. Urine samples were obtained from biotin-deficient, normal and biotin-supplemented adults. In 133 urine samples from 26 subjects, biotin by direct ABA correlated positively and significantly with biotin measured after HPLC separation (P < 0.001; r = 0.78). However, biotin by direct ABA routinely exceeded true biotin. The magnitude of the overestimate correlated strongly with biotin metabolites; r = 0.80 and P < 0.0001. In 92 samples from nine subjects, biotin by direct SABA correlated positively and significantly with true biotin (P = 0.001; r = 0.73) but exceeded true biotin by more than analytical error in 62 of the 92 samples. The error did not correlate significantly with total biotin metabolites. In 62 samples analyzed by both assays, biotin by direct SABA correlated weakly (r = 0.69) but significantly (P < 0.0001) with biotin by direct ABA. These studies provide evidence that direct SABA does not accurately quantitate biotin. Although the errors from direct ABA arise primarily from metabolites, the errors from direct SABA cannot be attributed primarily to biotin metabolites. Whether these interfering substances are biotin metabolites or other unknown substances, the substances are likely separated from the biotin fraction by HPLC.