Association of Biotin Ingestion With Performance of Hormone and Nonhormone Assays in Healthy Adults

Key Points

Question  Does oral biotin supplementation interfere with hormone and nonhormone assays that use biotinylation in their design?

Findings  In this nonrandomized crossover study of 6 healthy adults (2 women, 4 men), 10 mg/d of biotin ingested for 1 week was associated with potentially clinically important assay interferences in some but not all biotinylated hormone and nonhormone assays studied.

Meaning  Oral biotin use may be associated with false hormone and nonhormone assay results.

Importance  Biotinylated antibodies and analogues, with their strong binding to streptavidin, are used in many clinical laboratory tests. Excess biotin in blood due to supplemental biotin ingestion may affect biotin-streptavidin binding, leading to potential clinical misinterpretation. However, the degree of interference remains undefined in healthy adults.

Objective  To assess performance of specific biotinylated immunoassays after 7 days of ingesting 10 mg/d of biotin, a dose common in over-the-counter supplements for healthy adults.

Design, Setting, and Participants  Nonrandomized crossover trial involving 6 healthy adults who were treated at an academic medical center research laboratory

Exposure  Administration of 10 mg/d of biotin supplementation for 7 days.

Main Outcomes and Measures  Analyte concentrations were compared with baseline (day 0) measures on the seventh day of biotin treatment and 7 days after treatment had stopped (day 14). The 11 analytes included 9 hormones (ie, thyroid-stimulating hormone, total thyroxine, total triiodothyronine, free thyroxine, free triiodothyronine, parathyroid hormone, prolactin, N-terminal pro-brain natriuretic peptide, 25-hydroxyvitamin D) and 2 nonhormones (prostate-specific antigen and ferritin). A total of 37 immunoassays for the 11 analytes were evaluated on 4 diagnostic systems, including 23 assays that incorporated biotin and streptavidin components and 14 assays that did not include biotin and streptavidin components and served as negative controls.

Results  Among the 2 women and 4 men (mean age, 38 years [range, 31-45 years]) who took 10 mg/d of biotin for 7 days, biotin ingestion–associated interference was found in 9 of the 23 (39%) biotinylated assays compared with none of the 14 nonbiotinylated assays (P = .007). Results from 5 of 8 biotinylated (63%) competitive immunoassays tested falsely high and results from 4 out of 15 (27%) biotinylated sandwich immunoassays tested falsely low.

Conclusions and Relevance  In this preliminary study of 6 healthy adult participants and 11 hormone and nonhormone analytes measured by 37 immunoassays, ingesting 10 mg/d of biotin for 1 week was associated with potentially clinically important assay interference in some but not all biotinylated assays studied. These findings should be considered for patients taking biotin supplements before ordering blood tests or when interpreting results.

Trial Registration  clinicaltrials.gov Identifier: NCT03034707

Inaccuracy of laboratory medicine diagnostic tests may be associated with ingestion of over-the-counter vitamin and herbal supplements.^1,2 One such example is interference associated with biotin. Biotin (vitamin B7), a water-soluble vitamin found in meat, fish, eggs, and dairy,³ serves as a catalyst for carboxylase enzymes important in macronutrient metabolism.⁴ Biotin supplements, especially very large doses (eg, 10-15 mg/d, or 333-fold greater than the dietary recommendation of 30 µg/d)⁵ have become popular for presumptive health benefits such as stimulating hair growth and treating certain medical conditions including biotinidase deficiency, diabetes, lipid disorders, and diabetic peripheral neuropathy.^6-9 Ingestion of up to 300 mg/d may be beneficial for multiple sclerosis.^10,11 Taking high-doses of biotin may result in inaccurate laboratory results because biotin is commonly used, in the form of biotinylated antibodies or analogues, as a critical component in immunoassays. These assays exploit the strong, stable, and specific binding between biotin and streptavidin to amplify the assay sensitivity for detecting low analyte levels.^12,13 Excessive biotin in a blood sample can compete with biotinylated components in the assay, potentially falsely decreasing results in sandwich immunoassays or falsely increasing results in competitive immunoassays (eFigures 1 and 2 in the Supplement).¹⁴

Due to a lack of systematic studies, little is known about whether or how performance of specific biotinylated immunoassays may be associated with biotin ingestion at doses common in over-the-counter supplements (10 mg/d) in healthy adults. Therefore, this study was designed to assess the association of short-term biotin ingestion for 7 days with performance of assays that measure 11 hormone and nonhormone analytes: thyroid-stimulating hormone (TSH), total thyroxine (T4), total triiodothyronine (T3), free T4, free T3, intact parathyroid hormone (PTH), prolactin, N-terminal pro-brain natriuretic peptide (NT-proBNP), 25-hydroxyvitamin D (25-OHD) in 6 healthy adults, and ferritin and prostate-specific antigen (PSA) in the 4 men, using 4 assay systems.

This study was approved by the institutional review board at the University of Minnesota. Six healthy adults responded to study announcement fliers, gave informed consent before participation, and were not compensated. Exclusion criteria including but not limited to conditions that potentially affect biotin or hormones were being pregnant or lactating, having known thyroid disease, undergoing thyroid hormone treatment, ingesting over-the-counter dietary or nutritional supplements (excluding standard multivitamin preparations containing no more than 100% of the daily value for biotin and calcium), working the night shift, smoking,¹⁵ being treated with anticonvulsant medicine,¹⁶ or lacking the capacity to consent. Participants were asked to stop taking multivitamins 2 weeks before study participation.

Participants were instructed to take 10 mg/d of biotin (Nature Made) at the same time each morning for 7 days. Blood specimens were collected by venipuncture at baseline prior to starting biotin, after 1 week of biotin supplementation (day 7), and 1 week after participants stopped taking biotin (day 14). On the last day of treatment (day 7), the blood specimen was drawn approximately 2 hours after taking their last dose. Blood specimens were labeled with participant study identification number and date, promptly centrifuged and processed to produce serum sample aliquots that were stored up to a year at −70°C until further analysis.

Eighteen serum samples, collected from the 6 participants at the 3 time points, were each split into 4 aliquots and sent for testing at 4 clinical laboratories using different diagnostic assay systems: Johns Hopkins Medical Institutions used Roche cobas e602 for measurement of all 9 hormones and ferritin; Children’s Mercy Hospital used the OCD Vitros 5600 for TSH, PTH, total T4, total T3, free T4, NT-proBNP, and ferritin and Siemens Immulite 2000 for prolactin. The University of Minnesota Medical Center used Siemens Vista Dimension 1500 for TSH, PTH, total T4, free T4, free T3, NT-proBNP, ferritin, and PSA and Siemens Advia Centaur XP for total T3 and PTH; Boston Medical Center used Abbott Architect 2000 for TSH, PTH, total T4, total T3, free T4, prolactin, 25-OHD, ferritin, and PSA. eTable 1 in the Supplement summarizes information on the 37 assays for the 11 analytes evaluated on 4 systems: 23 incorporated biotin and streptavidin components and 14 did not include those components, serving as negative controls. Analytical imprecision data for these 37 assays are included in eTable 2 in the Supplement. Two immunoassay principals were used (eFigures 1 and 2 in the Supplement): the sandwich immunoassay for TSH, PTH, prolactin, NT-proBNP, PSA, and ferritin and the competitive immunoassay for total T4, total T3, free T4, free T3, and 25-OHD.

Four analytes (NT-proBNP, ferritin, 25-OHD, and PSA) were performed as a second tier after the initial analysis of 7 analytes (TSH, intact PTH, total T4, total T3, free T4, free T3, prolactin). As a result, the sample size for some tests was less than 6 due to an insufficient volume in the blood sample or sex differences in the reference range. Prostate-specific antigen and ferritin were measured for only the 4 men.

Assays were performed for each analyte as a single batch, on automated systems, by clinical laboratory technologists blinded to the nature of the study. Serum biotin was measured by the Cambridge Biomedical Research Group (Boston, Massachusetts) using a microbial growth assay.¹⁷ Serum calcium was measured at Johns Hopkins Medical Institutions with the Roche cobas c701 chemistry analyzer.

The study was powered to detect changes larger than the expected assay coefficient of variations (imprecisions) at normal reference levels for common hormone tests (eTable 2 in the Supplement). Because significant biotin ingestion–associated changes were observed in some assays studied, study recruitment was terminated after data analysis from the first 6 participants.

Analyte levels below or above their reportable ranges were assigned the following values: for Ortho Vitros NT-proBNP, 11 pg/mL was used for levels reported to be lower than 11.1 pg/mL; for both Siemens Vista Dimension and Roche cobas NT-proBNP, 4 pg/mL was used for levels reported that were less than 5 pg/mL. For biotin more than 3600 pg/mL, 3601 pg/mL was used. These value assignments would underestimate any biotin interference that was present. (To convert biotin from pg/mL to nmol/L, multiply by 0.00409; prolactin from ng/mL to pmol/L, multiply by 43.478; free T3 from pg/mL to pmol/L, multiply by 1.54; total T3 from ng/mL to nmol/L, multiply by 1.54; free T4 from ng/dL to pmol/L, multiply by 12.871; total T4 from µg/dL to nmol/L, multiply by 12.871.)

Each combination of an analyte and a system was analyzed separately with repeated measures of an analysis of variance (ANOVA) (mixed linear model), for which the random effect was a participant and the within-participant fixed effect was time (day of study). The primary analysis compared study day 7, the last day participants took biotin, with the mean of baseline and study day 14, before participants took biotin and after they stopped taking biotin, using a contrast in the ANOVA. To test this contrast for each analyte and system, P < .05 was the criterion for statistical significance. For each combination of an analyte and a system, we also present comparisons of pairs of times using the Tukey honest significant difference post hoc test. Data are reported as mean and 95% CIs. The Fisher exact test was used to compare biotin interference outcomes between the biotinylated assays and nonbiotinylated assays. All tests were 2-sided. All analyses used JMP Pro v13 (SAS Institute Inc).

Six healthy adults enrolled in the study, 2 women and 4 men, with a mean age of 38 years (range, 31-45 years). Baseline analyte concentrations for the 6 participants were within the manufacturer’s reference ranges for 7 analytes measured by 29 assays, except for 4 analytes measured by 8 assays: (1) PTH by OCD Vitros 5600 and Siemens Advia Centaur XP; (2) total T4 by Abbott Architect; (3) prolactin by Roche cobas e602, Siemens Immulite 2000, and Abbott Architect; and (4) ferritin by the Roche cobas e602 and Abbott Architect. In each case, this involved no more than 1 or 2 participants. None of the participants had abnormal baseline results across all systems for a given analyte.

The mean baseline serum biotin concentration was 774 pg/mL (95% CI, 554-1004 pg/mL) compared with a mean study day-7 concentration of more than 3600 pg/mL (95% CI, 3601-3601, mg/mL; P value <.001) for all 6 participants (Figure 1A). On study day 14, the mean biotin concentration decreased to 1090 pg/mL (95% CI, 687-1493 pg/mL). Baseline and study day 14 biotin concentrations did not differ statistically (eTable 4 in the Supplement).

Biotin ingestion was associated with falsely decreased Roche cobas e602 TSH levels by a mean of 0.72 mIU/L (95% CI, −1.13 to −0.32 mIU/L; P = .006), a 37% reduction from baseline, although all results remained within the euthyroid reference range (Figure 1B). The interference was much greater when measured by the Vitros 5600 TSH assay (Figure 1C) for which TSH significantly decreased by a mean of 1.67 mIU/L (95% CI, −2.08 to −1.26 mIU/L; P < .001), a 94% reduction from baseline, with all results falsely decreased to below the reference range (ie, <0.15 mIU/L; reference range, 0.47-4.68 mIU/L).

Biotin ingestion was associated with significantly decreased OCD Vitros PTH results by a mean of 25.8 pg/mL (95% CI, −34.8 to −16.8 pg/mL; P < .001), a 61% reduction from baseline. In 2 participants with normal baseline PTH concentrations, biotin ingestion was associated with falsely decreased PTH results, slightly below the lower limit of the reference range at 7.0 pg/mL and 7.2 pg/mL (reference range, 7.5-53.5 pg/mL). Serum calcium concentrations in all participants remained stable within the reference range (Table 1).

Biotin ingestion was associated with statistically significant false increases in 4 assays: 3 Roche cobas e602 assays measuring total T3, free T3, and free T4; and the Siemens Vista Dimension 1500 measuring free T3 (Figure 2 and Figure 3). Roche cobas e602 total T3 concentrations falsely increased by a mean of 0.85 ng/mL (95% CI, 0.49-1.22 ng/mL; P = .001), whereas Siemens Vista free T3 concentration falsely increased by a mean of 0.78 pg/mL (95% CI, 0.50-1.06 pg/mL; P < .001; Table 2). In 3 participants, the Roche cobas e602 total T3 results and in 1 participant Siemens Vista free T3 result were higher than their respective reference ranges. Table 3 shows the results for prolactin, which did not have biotin-associated changes.

eFigure 3 in the Supplement shows that biotin ingestion was associated with falsely reduced OCD Vitros 5600 NT-proBNP results by an average of more than 13.9 pg/mL (95% CI, −24.7 to −3.12 pg/mL; P = .03) to less than 11.1 pg/mL in all participants. The actual reduction was underestimated because results while participants were taking biotin were below the assay’s reportable range. Biotin ingestion was associated with falsely increased Roche cobas 25-OHD results by a mean of 9.25 ng/mL (95% CI, 5.72-12.8 ng/mL; P < .001) higher than the baseline (eTable 3 in the Supplement). None of the 11 analytes measured by 37 assays differed between baseline and day 14, except ferritin measured by the Siemens Vista Dimension (eTable 4 in the Supplement). Ferritin at day 7 of biotin treatment did not differ significantly from the mean of baseline and day 14 or from baseline alone in all 4 systems, supporting that biotin ingestion was not associated with the difference between baseline and day 14.

Biotin interference was not observed in any of the 14 nonbiotinylated assays (Table 4). Biotin interference was not observed in 14 of the 23 (61%) biotinylated assays; however, it was observed in 9 of the 23 (39%) biotinylated assays: falsely decreasing results in 4 sandwich immunoassays (4 of 15 [27%]); falsely increasing results in 5 competitive immunoassays (5 of 8 [63%]). Biotin interference outcomes were significantly different between biotinylated assays (9 of 23 [39%]) and nonbiotinylated assays (none) (Fisher exact test, P = .007).

This study involving 6 healthy adults demonstrated that oral biotin was associated with potentially clinically important assay interference in some but not all biotinylated assays. Among the 23 biotinylated assays studied, biotin interference was of greatest clinical significance in the OCD Vitros TSH assay, where falsely decreased TSH concentrations (to <0.15 mU/L) could have resulted in misdiagnosis of thyrotoxicosis in otherwise euthyroid individuals. Likewise, falsely decreased OCD Vitros NT-proBNP, to lower than assay detection limits, could possibly result in failure to identify congestive heart failure.¹⁸ Because healthy study participants had normal baseline NT-proBNP, further study of patients with high baseline NT-proBNP concentration would be required to establish the effect of biotin interference on the diagnosis of heart failure. The smaller changes observed in other assays, namely OCD Vitros PTH; Roche cobas e602 TSH, total and free T3, free T4, and 25-OHD; and Siemens Vista free T3, although primarily producing false results within the reference range among participants while taking biotin, could lead to falsely normal or abnormal interpretation of the results for individuals starting from baseline levels closer to the reference range limits.¹⁹

Many of the studied biotinylated assays were not affected by 7 days of biotin, despite what would have been predicted (Table 4) based on the assay biotin and streptavidin components. For some assays (eg, Roche cobas e602 PTH, total T4, free T4, and prolactin and Siemens Vista 1500 TSH, PTH, free T4, and prolactin), there are reports of biotin interference.^20-24 For example, a falsely low Roche Elecsys PTH level of 48 ng/L was reported in a patient with hyperparathyroidism in chronic kidney disease (true PTH level 576 ng/L).²¹

Potential reasons for discrepancies between predicted and actual biotin interference in an individual blood sample include inherent differential biotin interference tolerance among biotinylated assays owing to assay design, endogenous levels of free biotin and biotin metabolites present in the sample,^21,25,26 the type of biotin supplement, the dose and duration of biotin ingestion,²⁰ time of blood draws after the last dose, and the analyte concentration. Supraphysiologic doses of biotin may increase blood concentrations by 1.5 to 163 times higher than normal, depending on dose and measurement time after administration.^20,21,27-29 Differential biotin interference tolerance among assays is likely due to the amounts of biotinylated antibodies or analogues used, the availability of streptavidin binding sites and areas in the assay reagents (eg, streptavidin-coated magnetic particles) (eFigures 1 and 2 in the Supplement), and possible effects from biotin metabolites.²⁰

The time required for patients to stop taking supplements with biotin to avoid assay interference appears to be variable and may depend on the patient population, time of blood collection relative to the last biotin dose, biotin dose, chronicity of biotin exposure, and half-life of free biotin and biotin metabolites in plasma. Maximal assay interference was demonstrated 2 hours after a single 30-mg biotin dose.³⁰ Peak biotin blood concentration occurred 1.25 hours and 1.5 hours after a 100-mg and a 300-mg single dose, respectively, with half-life up to 18.8 hours following a single 300-mg dose.²⁸ In contrast, a 1.8-hour half-life was reported from a 600-µg biotin dose.³¹ Biotin metabolite concentrations (ie, bisnorbiotin) were significantly higher following months of taking 100 mg of biotin 3 times a day than they were following a single 300-mg dose,²⁰ suggesting that biotin metabolites are affected by chronicity of biotin exposure. Spiking studies^24,32,33 showed interference at higher in vitro biotin concentrations than have been measured in vivo,²¹ indicating the importance of biotin metabolites.^20,21,26 Biotin assay interference following discontinuation of biotin has been demonstrated at 24 hours and at 16 hours following a single 30-mg dose and 3 daily 100-mg doses, respecively.^22,30 Falsely abnormal hormone levels returned to normal 3 days after ceasing to take 300 mg of biotin in 3 daily doses and 2 days after ceasing to take up to 300 mg of biotin.^34,35 In pediatric populations, biotin interference was found 2 days after the last dose of biotin (10 mg/d for 4 days), and disappeared at a week in infants and young children receiving between 2 and 15 mg/kg/d.^32,36 In the current study, mean biotin concentration returned to baseline and the biotin ingestion–associated interferences resolved 1 week after a 10-mg/d 7-day course.

To our knowledge, the current study is the first to systematically assess the association of biotin ingestion (10 mg/d for 7 days) in healthy adults with performance of 37 assays that measure 11 analytes over 4 major diagnostic systems (Table 4). The study has several limitations. First, only healthy adults with mostly normal analyte concentrations were evaluated. Second, the study was neither randomized nor blinded to the investigators or participants, although it was blinded to the clinical laboratories. The study did not have a placebo group, but the crossover design and repeated measures analysis allowed each participant’s baseline values to serve as his/her own controls, making it more efficient (smaller sample size is needed) than a randomized design, because between-person variability in overall level of an analyte is eliminated. Third, because no formal dose-response pharmacokinetic study of biotin at various doses was performed, the minimal dose and duration required to alter assay results remains undetermined. Fourth, the sample size was small; power may have been insufficient to detect smaller effects of biotin ingestion. Definitive studies of the effect of biotin on specific assays and analytes will require further investigation.

Despite these limitations, this study reinforces cautionary advice regarding potential limitations of assays that use biotin streptavidin binding for clinical evaluation of individuals who ingest large doses of biotin. A few assay manufacturers package inserts acknowledge biotin interference, recommending delayed sample collection after biotin intake. Based on these findings, manufacturers may need to consider modifying biotinylated assays to minimize the effects of biotin ingestion.^20,21,37 Laboratories could identify assays that contain biotinylated components.³⁸ Clinicians may want to ask about biotin ingestion even if assay results are not suspect because biotin interferences can cause either falsely normal or abnormal results.²⁴ It may be advisable for patients to stop taking biotin, preferably for a week as studied herein, before undergoing laboratory testing. Alternatively, in the presence of biotin ingestion, nonbiotinylated assays would be preferred. Future studies, including patients with normal and abnormal analyte concentrations, are recommended to further clarify the extent and pharmacokinetics of ingested biotin interference on various assay platforms.

In this preliminary study of 6 healthy adult participants and 11 hormone and nonhormone analytes measured by 37 immunoassays, ingesting 10 mg/d of oral biotin for 1 week was associated with potentially clinically important assay interference in some but not all biotinylated assays studied. These findings should be considered for patients taking biotin supplements before ordering blood tests or when interpreting results.

Corresponding Author: Danni Li, PhD, 420 Delaware St SE, MMC 609, Minneapolis, MN 55455 (dannili@umn.edu).

Correction: This article was corrected online November 21, 2017, for an incorrect unit of measure in the Statistical Analysis section of the text and in Table 1.

Accepted for Publication: August 27, 2017.

Author Contributions: Drs Li and Burmeister had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Li and Radulescu were both first coauthors.

Study concept and design: Li, Radulescu, Shrestha, Burmeister.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Li, Radulescu, Shrestha, Burmeister.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Hodges, Burmeister.

Obtained funding: Li, Burmeister.

Administrative, technical, or material support: Radulescu, Shrestha, Killeen, Fan, Garg, Sokoll, Burmeister.

Study supervision: Li, Radulescu, Shrestha, Burmeister.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Karger reports receiving grants and nonfinancial support from Siemens Healthcare Diagnostics. Dr Sokoll reports receiving grant support from Abbott Laboratories. Dr Killeen reports receiving personal fees from Roche Diagnostics and Abbott Diagnostics. No other disclosures or possible conflicts of interest were reported.

Funding/Support: Research reported in this publication was supported by grant UL1TR000114 from the National Center for Advancing Translational Sciences of the National Institutes of Health. Funds to support blood collections and biotin measurements were provided by the University of Minnesota Undergraduate Research Opportunities Program and the University of Minnesota Medical Center, respectively. Funds to support hormone and nonhormone assay measurements were provided by participating laboratories.

Role of the Funder/Sponsor: The funding agencies had no roles in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: The content herein is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Additional Contributions: We thank Bryani Lee, BS, University of Minnesota, for processing blood specimens; she received compensation through the University of Minnesota Undergraduate Research Opportunities Program. We also thank Kathy Larson, BS, Jina Forys, BS, and Karri Cargill, BS, University of Minnesota Medical Center Fairview; Deborah Boblitz, AA, MLT, and Phaedre Mohr, BS, Johns Hopkins Medical Institutions; and Amy Wiebold, BS, at the Children’s Mercy Hospitals and Clinics for testing the serum samples, none of whom received compensation for their roles in the study. We thank John Bantle, MD, University of Minnesota, for useful discussions; he received no compensation for his role in the study.