Biotin

Systematic PubMed review of case reports and RCTs of biotin for hair and nail growth (Patel, Swink, Castelo-Soccio 2017). Yield: 18 reported cases of biotin use for hair and nail changes across the entire literature.

All 18 cases improved clinically on biotin. In all 18, the patient had an underlying pathology explaining poor hair or nail growth: inherited biotin metabolism defects, acquired deficiency, brittle nail syndrome, or uncombable hair syndrome. The authors concluded that these cases are uncommon and that evidence for supplementation in healthy individuals is lacking.

This is the paper the supplement industry cites, and it says the opposite of what the label implies. An uncontrolled case series in patients whose deficiency was the cause of the hair problem tells you that correcting a deficiency works. It carries no information about a biotin-replete adult. Zero of 18 cases were healthy supplement users, and the review found no RCT to include.

• Patel DP, Swink SM, Castelo-Soccio L (2017). Skin Appendage Disord PMID 28879195

Updated systematic review of oral biotin for hair growth or hair quality, case reports and case series excluded (Yelich 2024). Only three studies in the world literature met inclusion.

The highest-quality of the three was double-blind and placebo-controlled and found no difference between the biotin and placebo groups on hair growth. The other two studied narrow populations (patients on isotretinoin, and women after sleeve gastrectomy), were open to multiple biases, and neither produced results favouring biotin. The authors describe a large discrepancy between public perception and the scientific literature.

Seven years after Patel 2017, a fresh search still turns up three studies total, and the one with a control group is null. This is the single most important fact about biotin for hair and it is almost never quoted in marketing. It is a review-level finding, so the underlying trial's n and duration are not restated here; we cite it as the review reports it rather than inflating it into a primary result.

• Yelich A, Jenkins H, Holt S, Miller R (2024). J Clin Aesthet Dermatol PMID 39148962

Prospective cohort, n=156 women who underwent laparoscopic sleeve gastrectomy for obesity with 12-month follow-up (Sen and Turkcapar 2021). Hair loss was self-reported. 22 patients with post-operative biotin deficiency were prescribed biotin 1,000 mcg/day for 3 months; a separate 29 patients with normal biotin levels took 1,000 mcg/day on their own initiative for an average of 2.5 months.

Hair loss occurred in 72% of patients (112/156) after surgery, typically starting at months 3 to 4 and lasting 5.5 +/- 2.6 months, with no permanent alopecia. Among the deficient group taking biotin, 5/22 (23%) reported a marked decline in hair loss. Among the biotin-replete group taking the same dose, 11/29 (38%) reported the same. The difference between deficient and replete users was not significant (P = 0.2). The authors concluded biotin provides low efficacy.

This is the closest thing in the literature to the experiment that matters: same dose, same setting, deficient versus not. If the deficiency-correction story generalised, the deficient arm should have done clearly better. It did not, and numerically did worse. The outcome is subjective self-report with no placebo arm and no hair counts, and post-surgical telogen effluvium resolves on its own, so the 23% and 38% figures are best read as spontaneous recovery in both groups rather than as a drug effect in either. This is an open-label cohort, not evidence that biotin helps anyone.

• Sen O, Turkcapar AG (2021). J Laparoendosc Adv Surg Tech A PMID 32762597

Prospective non-randomised crossover trial, n=6 healthy adults (2 women, 4 men, mean age 38, range 31 to 45), given biotin 10 mg/day for 7 days (Li 2017, JAMA; NCT03034707). 11 analytes measured on 37 immunoassays across 4 diagnostic platforms: 23 assays using biotin-streptavidin chemistry and 14 assays without it as negative controls. Analytes included TSH, total T4, total T3, free T4, free T3, PTH, prolactin, NT-proBNP, 25-hydroxyvitamin D, PSA and ferritin.

Interference appeared in 9 of 23 (39%) biotinylated assays and in 0 of 14 non-biotinylated controls (P = 0.007). The direction depends on assay architecture: 5 of 8 (63%) competitive immunoassays read falsely high, and 4 of 15 (27%) sandwich immunoassays read falsely low. Effects were present at day 7 and gone by day 14, one week after stopping.

This is the strongest harm evidence on the page and the reason for the grade. The dose is not a research curiosity, it is what a 10,000 mcg hair gummy delivers. The design includes a within-assay negative control and a washout, which rules out chance drift. Because competitive assays skew high and sandwich assays skew low, the classic pattern is a suppressed TSH with a raised free T4, which reads as Graves disease on paper. n=6 is small and the authors call it preliminary, but the mechanism is analytical chemistry rather than biology, so a small n is far less limiting than it would be for an efficacy claim.

• Li D, Radulescu A, Shrestha RT, et al. (2017). JAMA PMID 28973622

Analytical validation study on 23 plasma samples from multiple sclerosis patients and healthy volunteers on high-dose biotin, plus 39 unsupplemented control patients (Piketty 2017). Plasma biotin ranged 31.7 to 1,160 mcg/L. Free T3, free T4, TSH, PTH, 25-hydroxyvitamin D, FSH, LH, prolactin, C-peptide and cortisol were measured before and after a streptavidin-coated microparticle depletion step, and again on assays that do not use streptavidin-biotin capture. Supported by a case plus systematic review of factitious Graves disease from biotin (Elston 2016).

Most hormone results in the high-biotin plasmas were abnormal and normalised after biotin was removed. Results measured on non-streptavidin platforms were normal from the start, apart from two slight PTH elevations. In the 39 unsupplemented patients, the depletion step changed nothing except an 8.4% drop in PTH. In Elston 2016, a patient had markedly abnormal thyroid function tests that did not fit the clinical picture and that normalised after stopping biotin far faster than the half-life of T4 would permit, which is only explicable as assay interference.

Three independent lines converge: the abnormality tracks the biotin, disappears when the biotin is physically removed, and never appears on platforms that do not use the biotin-streptavidin bond. This is as close to proof of causation as a lab-interference claim gets. The clinical consequence is documented, not theoretical: patients have been worked up and treated for Graves disease that they did not have. The 31.7 to 1,160 mcg/L range here comes largely from multiple sclerosis dosing, well above hair-supplement levels, so this study establishes the mechanism and direction rather than the threshold. Li 2017 supplies the threshold at consumer doses.

• Piketty ML, Prie D, Sedel F, et al. (2017). Clin Chem Lab Med PMID 28222020
• Elston MS, Sehgal S, Du Toit S, Yarndley T, Conaglen JV (2016). J Clin Endocrinol Metab PMID 27362288

Two studies. Frame 2019: plasma from healthy volunteers taking biotin 10 mg/day, mixed with plasma of known elevated troponin, plus biotin-spiked specimens, tested on three assays (Roche Gen5 cTnT, contemporary Gen4 cTnT, and a high-sensitivity cTnI). Mumma 2020: biotin measured in 797 zero-hour and 646 three-hour samples from 850 US emergency department patients with suspected acute coronary syndrome, plus 2,023 random US laboratory network samples, with modelled AMI misclassification risk for the Gen5 cTnT assay.

Frame 2019: daily 10 mg biotin produced significant interference in the Gen5 cTnT assay. The interference thresholds were 31 ng/mL for Gen5 cTnT, 315 ng/mL for the contemporary cTnT, and above 2,000 ng/mL for hs-cTnI, so the falsely low troponin is achievable with over-the-counter supplements on one platform and essentially unreachable on another. Mumma 2020: biotin above 20 ng/mL was present in 1/797 (0.13%) zero-hour ED samples and 15/2,023 (0.74%) random laboratory samples; modelled false-negative AMI risk at the 19 ng/L cutoff was 0.026% at zero hours, falling to 0.0064% at one hour and 0.00048% at three hours.

We report both because they answer different questions and the honest answer needs both. Frame shows the mechanism reaches consumer doses on a widely used platform, which means a specific patient on a 10,000 mcg hair gummy can have a falsely reassuring troponin. Mumma shows that across an unselected chest-pain population the aggregate miss rate is very low, largely because serial sampling rescues most cases. Neither cancels the other. The individual risk is what matters at the point of care, and troponin is the test where a false negative is measured in myocardium. Both are analytical and modelling studies, not outcome trials, and no study has counted actual missed infarctions attributable to biotin.

• Frame IJ, Joshi PH, Mwangi C, et al. (2019). Am J Clin Pathol PMID 30715102
• Mumma B, Diercks D, Twerenbold R, et al. (2020). Clin Chem Lab Med PMID 32804676

Three sources. Wolf 1991: worldwide neonatal screening survey, 8,532,617 newborns screened across 14 countries, 142 infants identified. Mock 2002: randomised placebo-controlled trial, n=26 pregnant women with raised urinary 3-hydroxyisovaleric acid (10 in early pregnancy at 6 to 17 weeks, 16 in late pregnancy at 21 to 37 weeks), 300 mcg/day biotin or placebo for 14 days. Zempleni 2008: review of biotin homeostasis and the causes of acquired deficiency.

Wolf 1991: profound biotinidase deficiency occurs in about 1:112,271 newborns, partial deficiency about 1:129,282, combined about 1:60,089, with roughly 1 in 123 people a heterozygous carrier. Mock 2002: 3-hydroxyisovaleric acid excretion fell on biotin and rose on placebo in both early pregnancy (5 vs 5 women, P < 0.006) and late pregnancy (8 vs 8 women, P < 0.002), establishing that the raised excretion of normal pregnancy reflects genuinely reduced biotin status. Zempleni 2008: acquired deficiency arises from inadequate intake, drug-vitamin interactions including long-term anticonvulsants, increased catabolism in pregnancy and smoking, and defects in the sodium-dependent multivitamin transporter, holocarboxylase synthetase or biotinidase.

Deficiency is diagnosed by cause and by metabolite, not by a hair complaint. Biotinidase deficiency is caught on newborn screening in the US and much of Europe and is treated with pharmacological biotin for life; it presents in infancy with seizures, hypotonia and dermatitis alongside alopecia, not as adult thinning. Marginal deficiency in pregnancy is real but was detected by urinary 3-hydroxyisovaleric acid, a functional marker, and the corrective dose in Mock 2002 was 300 mcg, one thirtieth of a typical hair gummy. Serum biotin alone is a poor test. Anticonvulsant and alcohol-related deficiency are mechanistically supported, with the alcohol transporter work done in cell and animal models rather than in humans, so we flag that as weaker. None of these routes describes the adult with androgenetic alopecia who is buying biotin off a shelf.

• Wolf B (1991). J Inherit Metab Dis PMID 1779651
• Mock DM, Quirk JG, Mock NI (2002). Am J Clin Nutr PMID 11815321
• Zempleni J, Hassan YI, Wijeratne SS (2008). Expert Rev Endocrinol Metab PMID 19727438

Consecutive case series of women presenting with a complaint of hair loss, serum biotin measured, trichograms performed (Trueb 2016). Complemented by a retail survey of 176 skin, hair and nail supplements from 7 stores containing 255 distinct ingredients (Perez-Sanchez 2020).

Trueb 2016: biotin deficiency by serum level was found in 38% of women complaining of hair loss. Of the 24% showing diffuse telogen effluvium on trichogram, 35% had associated seborrhoeic-like dermatitis. Only 11% of the patients with low serum biotin had any personal history of a recognised risk factor for deficiency. The author explicitly rejects the practice of treating women complaining of hair loss with indiscriminate oral biotin. Perez-Sanchez 2020: the retail survey flagged nutrient overdosing, absent warning labels and no independent third-party testing as recurring problems in this product category.

This is the honest complication on the page and it cuts against the simple story in both directions. A 38% low-serum rate is far higher than the rarity of clinical deficiency, which tells you serum biotin is a weak standalone marker rather than that a third of women are deficient: only 11% had a plausible cause, there was no matched control group of women without hair loss, and no functional marker such as 3-hydroxyisovaleric acid was measured. We cite it because omitting it would be cherry-picking, and because its own author draws the same conclusion we do. The practical point is that a low serum biotin in isolation does not establish that biotin is causing the hair loss, and does not justify a 10,000 mcg dose.

• Trueb RM (2016). Int J Trichology PMID 27601860
• Perez-Sanchez AC, Burns EK, Perez VM, Tantry EK, Prabhu S, Katta R (2020). Cureus PMID 32874806