Oral biotin and falsely low TSH in specified immunoassays
conclusionConfirmed in some historical assays and platforms, but not in every TSH test. This current conclusion keeps dose, sampling time, method and reagent version together.
boundaryDo not expand this result into claims that biotin interferes with every TSH test, treats thyroid disease, establishes a patient misdiagnosis rate, or supplies a universal stopping interval.
Four measurement-assessment dimensions
| Measurement conclusion | Falsely low TSH readings were observed in some assays after oral biotin. Even within sandwich methods, findings differed by platform, reagent version, dose and sampling time. This is not evidence that biotin treats thyroid disease. |
|---|---|
| Evidence certainty | Analytical interference is directly observed under specified conditions; universal interference rates and stopping intervals remain unverified |
| Applicability | Detailed direct evidence is from healthy adults; no automatic extrapolation to patients, renal impairment or pregnancy |
| Safety | Safety is Warning because falsely low TSH can plausibly mislead interpretation. The direct observations concern selected-assay bias and reference-limit crossing, not rates of actual patient harm. Bias magnitude and duration in renal impairment, pregnancy, children or high-dose treatment populations are not established by these healthy-adult data. No individual supplement stopping interval, testing schedule or treatment change is prescribed. |
Because this is a measurement question, the A-F efficacy grade, numeric efficacy score, and seven efficacy axes do not apply. Document quality B is not a measurement-effect or harm-probability grade.
Useful facts when choosing a product
- The Li study gave 10 mg/day of biotin for 7 days to 6 healthy adults. Measurements were compared before exposure, on day 7 approximately 2 hours after the last dose, and on day 14 after 7 days off. The 18 serum time-point samples are not 18 participants.
- Mean TSH on Roche cobas e602 was 1.80→1.21→2.06 mIU/L. The formal contrast was −0.72 mIU/L, 95% CI −1.13 to −0.32, P=0.006. The simple before/after difference of −0.59 mIU/L from the displayed means is a separate calculation.
- Mean TSH on Ortho VITROS 5600 was 1.64→0.10→1.89 mIU/L; the formal contrast was −1.67 mIU/L, 95% CI −2.08 to −1.26, P<0.001. All 6 exposed participants had readings below 0.15 mIU/L, below this assay reference lower limit of 0.47 mIU/L. This observation is not a population misdiagnosis rate or incidence of actual hyperthyroidism.
- With the same Li specimens, the Siemens Dimension Vista 1500 contrast was −0.11 mIU/L, 95% CI −0.49 to 0.26, P=0.56, and Abbott ARCHITECT 2000 was −0.06 mIU/L, 95% CI −0.44 to 0.31, P=0.74. Vista TSH also used biotin, yet no significant change was observed at this exposure. Nonsignificance is not interpreted as zero bias or equivalence.
- The same 10 participants in Zhang used 5 mg/day and 10 mg/day for 7 days each. Roche cobas e602 TSH was lower at 2 hours with 5 mg, P=0.005, and at 2 and 8 hours with 10 mg, each P<0.001. No significant baseline difference was reported at 26 or 50 hours with either dose. Exact mean differences and CI remain unverified.
- TSH interference was not observed under the tested conditions on Zhang’s Beckman UniCel DxI 800 or Abbott ARCHITECT 2000. These TSH methods did not use biotin-streptavidin. Interference affecting other thyroid analytes on Beckman was not transferred to TSH.
- Roche Elecsys TSH document 2024-10 V2.0 (08429324501V2.0) lists testing up to biotin ≤1200 ng/mL on cobas e411/e601/e602. Acceptance is deviation ≤0.02 µIU/mL at TSH ≤0.2 µIU/mL, and ≤10% above that level. This is an analytical performance document for that reagent, not a human oral safe dose, guarantee of zero interference, or verification of every laboratory’s installed version.
- Serum-biotin units in Li were retained as pg/mL according to the published correction. The day 7 value of 3601 was assigned for analysis because all measurements exceeded 3600; it is not an exact measured concentration. TSH remains in mIU/L and is not combined with serum biotin.
- Exact supplement chemical form, batch, purity and total dietary exposure, historical reagent lots and tolerance thresholds, some registration histories/protocols/SAPs, and long-term or special-population data remain unverified. New studies, reagent documents, corrections, retractions, or numerical/classification errors trigger an update with accumulated history at the same assigned ID and URL.
Chamgap Semantic Classification Code
Permanent code issued
S.biotin.oral.serum-tsh-immunoassay-bias.false-low.pre-exposure-washout-or-biotin-independent-methodSubstances and nutrients > Biotin > Oral > Serum TSH immunoassay bias > Falsely low > Pre-exposure, washout, or biotin-independent method
Measurement-only claim. Platform, reagent version, dose, and time from last dose to blood draw remain separate; no universal TSH-test inference is made. The code identifies the intervention and claim; it never contains the evidence grade, score, or safety result.
Exact Claim Classification
These independent facets prevent evidence from different forms, routes, populations, effects, and comparators from being mixed.
| Intervention class | S · Supplement or nutraceutical |
|---|---|
| Canonical ingredient or intervention | Biotin |
| Source or part used | Oral supplemental biotin; manufacturing source and precise chemical form unverified |
| Formulation or processing | Study-specific biotin supplements; tablet/capsule, exact composition and batch unverified |
| Route | Oral |
| Dose | 5 mg/day or 10 mg/day; separate study, platform and sampling time strata |
| Duration | 7 or 8 days of exposure; separate treatment duration from last-dose-to-draw time |
| Population | Detailed direct evidence is from healthy adults; no automatic extrapolation to patients, renal impairment or pregnancy |
| Effect or condition | Falsely low serum TSH bias in specified immunoassays |
| Primary endpoint | TSH assay bias relative to pre-exposure/washout or a biotin-independent method; reference-limit crossing separately |
| Comparator | Same participants before exposure/after washout and comparator platforms on same-time serum; not a randomized placebo trial |
| Duplicate-detection key | biotin|oral|serum-tsh|false-low-assay-bias|adult-exposure|paired-pre-washout-or-biotin-independent-method|platform-reagent-dose-time-stratified |
Measurement values by platform, dose, and timing
The Li study gave 10 mg/day of biotin for 7 days to 6 healthy adults. Measurements were compared before exposure, on day 7 approximately 2 hours after the last dose, and on day 14 after 7 days off. The 18 serum time-point samples are not 18 participants.
The formal estimates below compare day 7 − mean(day 0, day 14) within the same participants. They are not simple baseline changes or effects between randomized groups. The 95% CI come from the reported model and are not relabeled SD or SE.
Mean TSH on Roche cobas e602 was 1.80→1.21→2.06 mIU/L. The formal contrast was −0.72 mIU/L, 95% CI −1.13 to −0.32, P=0.006. The simple before/after difference of −0.59 mIU/L from the displayed means is a separate calculation.
Mean TSH on Ortho VITROS 5600 was 1.64→0.10→1.89 mIU/L; the formal contrast was −1.67 mIU/L, 95% CI −2.08 to −1.26, P<0.001. All 6 exposed participants had readings below 0.15 mIU/L, below this assay reference lower limit of 0.47 mIU/L. This observation is not a population misdiagnosis rate or incidence of actual hyperthyroidism.
With the same Li specimens, the Siemens Dimension Vista 1500 contrast was −0.11 mIU/L, 95% CI −0.49 to 0.26, P=0.56, and Abbott ARCHITECT 2000 was −0.06 mIU/L, 95% CI −0.44 to 0.31, P=0.74. Vista TSH also used biotin, yet no significant change was observed at this exposure. Nonsignificance is not interpreted as zero bias or equivalence.
The same 10 participants in Zhang used 5 mg/day and 10 mg/day for 7 days each. Roche cobas e602 TSH was lower at 2 hours with 5 mg, P=0.005, and at 2 and 8 hours with 10 mg, each P<0.001. No significant baseline difference was reported at 26 or 50 hours with either dose. Exact mean differences and CI remain unverified.
TSH interference was not observed under the tested conditions on Zhang’s Beckman UniCel DxI 800 or Abbott ARCHITECT 2000. These TSH methods did not use biotin-streptavidin. Interference affecting other thyroid analytes on Beckman was not transferred to TSH.
Ylli’s study of 13 adults reported 10 mg/day for 8 days and measurements before dosing and at 2 and 5 hours on day 1 and day 8. The abstract supports Roche TSH changes and maximal interference at 2 hours. Numerical full-table bias was not verified, so percentages from search snippets were not imported.
Roche Elecsys TSH document 2024-10 V2.0 (08429324501V2.0) lists testing up to biotin ≤1200 ng/mL on cobas e411/e601/e602. Acceptance is deviation ≤0.02 µIU/mL at TSH ≤0.2 µIU/mL, and ≤10% above that level. This is an analytical performance document for that reagent, not a human oral safe dose, guarantee of zero interference, or verification of every laboratory’s installed version.
Serum-biotin units in Li were retained as pg/mL according to the published correction. The day 7 value of 3601 was assigned for analysis because all measurements exceeded 3600; it is not an exact measured concentration. TSH remains in mIU/L and is not combined with serum biotin.
Current measurement conclusion and platform boundaries
Falsely low TSH readings were observed in some assays after oral biotin. Even within sandwich methods, findings differed by platform, reagent version, dose and sampling time. This is not evidence that biotin treats thyroid disease.
Safety is Warning because falsely low TSH can plausibly mislead interpretation. The direct observations concern selected-assay bias and reference-limit crossing, not rates of actual patient harm. Bias magnitude and duration in renal impairment, pregnancy, children or high-dose treatment populations are not established by these healthy-adult data. No individual supplement stopping interval, testing schedule or treatment change is prescribed.
Platform-specific counterevidence and limits. With the same Li specimens, the Siemens Dimension Vista 1500 contrast was −0.11 mIU/L, 95% CI −0.49 to 0.26, P=0.56, and Abbott ARCHITECT 2000 was −0.06 mIU/L, 95% CI −0.44 to 0.31, P=0.74. Vista TSH also used biotin, yet no significant change was observed at this exposure. Nonsignificance is not interpreted as zero bias or equivalence. TSH interference was not observed under the tested conditions on Zhang’s Beckman UniCel DxI 800 or Abbott ARCHITECT 2000. These TSH methods did not use biotin-streptavidin. Interference affecting other thyroid analytes on Beckman was not transferred to TSH. Roche Elecsys TSH document 2024-10 V2.0 (08429324501V2.0) lists testing up to biotin ≤1200 ng/mL on cobas e411/e601/e602. Acceptance is deviation ≤0.02 µIU/mL at TSH ≤0.2 µIU/mL, and ≤10% above that level. This is an analytical performance document for that reagent, not a human oral safe dose, guarantee of zero interference, or verification of every laboratory’s installed version.
Review performed and remaining limitations
Exact supplement chemical form, batch, purity and total dietary exposure, historical reagent lots and tolerance thresholds, some registration histories/protocols/SAPs, and long-term or special-population data remain unverified. New studies, reagent documents, corrections, retractions, or numerical/classification errors trigger an update with accumulated history at the same assigned ID and URL. single-assistant self-review. This is not independent external clinical review or journal certification and does not guarantee an exhaustive literature search.
Search scope and limitations. Executed searches, access and selection records, and the platform-dose-time matrix are retained in the transfer package.
Evidence Table
| Study | Design | Sample | Funding | Endpoint | Result | Weight |
|---|---|---|---|---|---|---|
| Li 2017 oral-exposure repeated-measures study | Nonrandomized within-person pre-exposure, post-exposure, and washout comparison in 6 healthy adults, with cross-platform testing of the same specimens. | 6 participants; 18 time-point specimens are not 18 participants. | NIH, university, and participating-laboratory support plus author-specific commercial relationships are disclosed as reported. | Platform-specific formal repeated-measures TSH contrasts and reference-limit crossing. | Mean TSH on Roche cobas e602 was 1.80→1.21→2.06 mIU/L. The formal contrast was −0.72 mIU/L, 95% CI −1.13 to −0.32, P=0.006. The simple before/after difference of −0.59 mIU/L from the displayed means is a separate calculation. Mean TSH on Ortho VITROS 5600 was 1.64→0.10→1.89 mIU/L; the formal contrast was −1.67 mIU/L, 95% CI −2.08 to −1.26, P<0.001. All 6 exposed participants had readings below 0.15 mIU/L, below this assay reference lower limit of 0.47 mIU/L. This observation is not a population misdiagnosis rate or incidence of actual hyperthyroidism. With the same Li specimens, the Siemens Dimension Vista 1500 contrast was −0.11 mIU/L, 95% CI −0.49 to 0.26, P=0.56, and Abbott ARCHITECT 2000 was −0.06 mIU/L, 95% CI −0.44 to 0.31, P=0.74. Vista TSH also used biotin, yet no significant change was observed at this exposure. Nonsignificance is not interpreted as zero bias or equivalence. | Direct oral-exposure serum-TSH measurement evidence; not used for efficacy grading. |
| Zhang 2020 fixed-sequence oral-exposure study | Within-person fixed-sequence comparison in the same 10 healthy adults using 5 mg/day and 10 mg/day for 7 days each. | 10 participants; exact analyzable specimen counts by platform and time are unverified. | Authors reported no funding or conflicts; this declaration is not independent verification. | Platform-specific TSH change by dose and time from last dose to blood draw. | The same 10 participants in Zhang used 5 mg/day and 10 mg/day for 7 days each. Roche cobas e602 TSH was lower at 2 hours with 5 mg, P=0.005, and at 2 and 8 hours with 10 mg, each P<0.001. No significant baseline difference was reported at 26 or 50 hours with either dose. Exact mean differences and CI remain unverified. TSH interference was not observed under the tested conditions on Zhang’s Beckman UniCel DxI 800 or Abbott ARCHITECT 2000. These TSH methods did not use biotin-streptavidin. Interference affecting other thyroid analytes on Beckman was not transferred to TSH. | Direct oral-exposure measurement evidence interpreted only within its timing and platform strata. |
| Ylli 2021 oral-exposure study | Prospective within-person exposure in 13 adults; extraction was limited to abstract and figure captions. | 13 enrolled; detailed TSH analysis denominator unverified. | Detailed funding and product roles were not verified in the accessible record. | Roche and Abbott TSH before dosing and at 2 and 5 hours. | Ylli’s study of 13 adults reported 10 mg/day for 8 days and measurements before dosing and at 2 and 5 hours on day 1 and day 8. The abstract supports Roche TSH changes and maximal interference at 2 hours. Numerical full-table bias was not verified, so percentages from search snippets were not imported. | Directional corroboration; full numerical tables were not verified. |
| Roche 2024-10 V2.0 reagent analytical-performance document | Manufacturer analytical-interference validation document; not a human oral-exposure study. | Human participant count not applicable. | Manufacturer document. | Biotin test range and acceptance deviation for the specified reagent version. | Roche Elecsys TSH document 2024-10 V2.0 (08429324501V2.0) lists testing up to biotin ≤1200 ng/mL on cobas e411/e601/e602. Acceptance is deviation ≤0.02 µIU/mL at TSH ≤0.2 µIU/mL, and ≤10% above that level. This is an analytical performance document for that reagent, not a human oral safe dose, guarantee of zero interference, or verification of every laboratory’s installed version. | Defines the current reagent-version boundary; not counted as an independent human replication. |
Receipt — 12 References
Evidence access cutoff: 2026-09-15. Each citation states its actual access level (full-text transcription, abstract, index, or other) and remaining limitations. Bibliographic checking does not certify the study results or treatment efficacy.
Technical integration by: Codex · Evidence date: 2026-09-15 · Corrections: 1
Correction log — 1
Corrections applied to this verdict, in chronological order. Changes are logged, not erased.
- 2017-11-21 · Published unit correction applied — The serum-biotin unit in Li was retained as pg/mL under the published correction; TSH in mIU/L was unchanged.
Cite this measurement assessment
[Chamgap] Oral biotin and falsely low TSH in specified immunoassays — measurement-only current value · Efficacy grade N/A · Safety warning. 12 cited sources checked. Source: https://chamgap.com/en/verdicts/general/oral-biotin-tsh-falsely-low-immunoassay-bias/ · CC BY 4.0What this document does and does not do
Chamgap is an information source. It reports what research has and has not confirmed; it does not tell readers what to take or buy. That decision belongs to readers and, when needed, medical or legal professionals. This verdict reflects literature available up to the search date and may change as new research appears. Nothing here is medical advice.