High-dose biotin therapy leading to false biochemical endocrine profiles: validation of a simple method to overcome biotin interference.

BACKGROUND High-dose biotin therapy is beneficial in progressive multiple sclerosis (MS) and is expected to be adopted by a large number of patients. Biotin therapy leads to analytical interference in many immunoassays that utilize streptavidin-biotin capture techniques, yielding skewed results that can mimic various endocrine disorders. We aimed at exploring this interference, to be able to remove biotin and avoid misleading results. METHODS We measured free triiodothyronine (fT3), free thyroxine (fT4), thyroid-stimulating hormone (TSH), parathyroid homrone (PTH), 25-hydroxyvitamin D (25OHD), follicle-stimulating hormone (FSH), luteinizing hormone (LH), prolactin, C-peptide, cortisol (Roche Diagnostics assays), biotin and its main metabolites (liquid chromatography tandem mass spectrometry) in 23 plasmas from MS patients and healthy volunteers receiving high-dose biotin, and in 39 biotin-unsupplemented patients, before and after a simple procedure (designated N5) designed to remove biotin by means of streptavidin-coated microparticles. We also assayed fT4, TSH and PTH in the 23 high-biotin plasmas using assays not employing streptavidin-biotin binding. RESULTS The biotin concentration ranged from 31.7 to 1160 µg/L in the 23 high-biotin plasmas samples. After the N5 protocol, the biotin concentration was below the detection limit in all but two samples (8.3 and 27.6 μg/L). Most hormones results were abnormal, but normalized after N5. All results with the alternative methods were normal except two slight PTH elevations. In the 39 biotin-unsupplemented patients, the N5 protocol did not affect the results for any of the hormones, apart from an 8.4% decrease in PTH. CONCLUSIONS We confirm that most streptavidin-biotin hormone immunoassays are affected by high biotin concentrations, leading to a risk of misdiagnosis. Our simple neutralization method efficiently suppresses biotin interference.