Abstract
Background: Biotin has been reported to interfere with several commonly used laboratory assays resulting in misleading values and possible erroneous diagnosis and treatment. This report describes a prospective study of possible biotin interference in thyroid-related laboratory assays, with a comparison of different commonly used assay platforms.
Materials and Methods: Thirteen adult subjects (mean age 45 ± 13 years old) were administered biotin 10 mg/day for eight days. Blood specimens were collected at three time points on day 1 and on day 8 (baseline, two, and five hours after biotin ingestion). Thyrotropin (TSH), free triiodothyronine (fT3), free thyroxine (fT4), total triiodothyronine (TT3), total thyroxine (TT4), thyroxine binding globulin (TBG), and thyroglobulin (Tg) levels were analyzed with four different platforms: Abbott Architect, Roche Cobas 6000, Siemens IMMULITE 2000, and liquid chromatography with tandem mass spectrometry (LC-MS/MS). TSH, fT3, fT4, TT3, and TT4 were measured with Abbott Architect and Roche Cobas 6000. fT3, fT4, TT3, and TT4 were also measured by LC-MS/MS. Tg was measured by Siemens IMMULITE 2000. TBG was assessed with Siemens IMMULITE 2000.
Results: Significant changes in TSH, fT4, and TT3 measurements were observed after biotin exposure when the Roche Cobas 6000 platform was used. Biotin intake resulted in a falsely lower Tg level when measurements were performed with Siemens IMMULITE 2000. At the time points examined, maximal biotin interference was observed two hours after biotin exposure both on day 1 and day 8.
Conclusions: A daily dose of 10 mg was shown to interfere with specific assays for TSH, fT4, TT3, and Tg. Physicians must be aware of the potential risk of erroneous test results in subjects taking biotin supplements. Altered test results for TSH and Tg can be particularly problematic in patients requiring careful titration of levothyroxine therapy such as those with thyroid cancer.
Keywords: biotin, fT4, hyperthyroidism, thyroglobulin, TSH, TT3
Introduction
Precise measurement of hormone ligands is essential for the routine diagnosis and management of a variety of endocrine disorders. Recently, biotin has been reported to interfere with several commonly used laboratory assays resulting in either misleadingly high or low laboratory results. A safety communication from the Food and Drug Administration (FDA) advised both patients and health care providers of the lack of sufficient information on the management of biotin intake relative to when blood is to be drawn for hormone assays to avoid incorrect test results (1).
Biotin is a water soluble B vitamin (B7) that is plentiful in both plant and animal foods such that biotin deficiency is extremely rare. Nevertheless, in addition to its presence in multivitamin preparations, it has become commonly used as a supplement advertised for skin, nail, and hair health. While the daily requirement is only 0.03 mg (2), over-the-counter preparations containing 2–20 mg are readily available and higher doses of 100–300 mg have been used for the treatment of multiple sclerosis and other neurologic pathologies (3,4).
Biotin interference with measurement of thyroxine (T4), triiodothyronine (T3), or thyrotropin (TSH) can lead to erroneous values that incorrectly indicate the presence of hyperthyroidism (5) prompting physicians to undertake inappropriate management decisions. This is particularly problematic in clinical circumstances that often demand precise titration of levothyroxine dosage such as in children, in the elderly, in pregnancy, and in patients with thyroid cancer. In addition, the monitoring of thyroid cancer patients for residual disease or recurrence is based, in large part, on serial measurements of thyroglobulin (Tg), and biotin-induced false elevations or depressions in serum Tg could lead to altered management with important clinical consequences (5). Other laboratory examinations that are potentially susceptible to biotin interference include troponin I and troponin T, crucial for acute myocardial infarction diagnosis (1), and assays for hepatitis B, hepatitis C, HIV serological markers, osteocalcin, gastrin, PTH, and testosterone (6–9).
The objective of the present report is to provide the results of a prospective investigation of possible biotin interference in thyroid-related laboratory assays, with a comparison of different commonly used assay platforms.
Materials and Methods
Thirteen adults were recruited into the study (nine women and four men), age 28 to 67 years old (mean age 45 ± 13 years old). Three subjects were post-thyroidectomy (2 for multinodular goiter and one for thyroid cancer) and 10 subjects had normal thyroid function. None was taking drugs that influence thyroid function other than three subjects, who required levothyroxine post-thyroidectomy and continued their treatment during the study. Before study initiation, all subjects had thyroid hormone levels within the normal reference range. Three subjects had Tg levels <1 ng/mL and four subjects had a Tg level between 1 and 10 ng/mL.
Pregnant and postpartum lactating women were excluded from enrollment. Other exclusion criteria were overt hyperthyroidism, renal and/or liver insufficiency, and positive anti-Tg antibodies. Renal or liver insufficiency was defined as a prediagnosed condition of renal and/or liver insufficiency, a glomerular filtration rate <60 mL/min/1.73 m2, or aspartate transaminase/alanine transaminase above normal range. Serum obtained from the collected blood specimens was stored at −80°C until further examined.
The effect of a single 10 mg dose of biotin was investigated on day 1 with serum samples obtained at the following time points: baseline before biotin ingestion (A), two hours after biotin ingestion (B), and five hours after biotin ingestion (C). Whether longer term exposure to biotin influenced analyte measurements was assessed by blood samples drawn after one week of biotin intake. On day 8, serum samples were obtained at baseline 24 hours after the last dose of biotin and before the next dose (D), 2 hours after biotin ingestion (E), and 5 hours after biotin ingestion (F). Blood was drawn in the fasting state for the studies on day 1 and day 8 to avoid additional biotin ingestion with food.
Biotin was extracted from 200 μL serum by protein precipitation using 300 μL methanol containing 25 ng/mL biotin-d4. Extracted biotin was measured by liquid chromatography with tandem mass spectrometry (LC-MS/MS) using an AB Sciex Triple Quad 6500 equipped with an ESI-Turbo V source (AB Sciex, Concord, ON, CA) and Shimadzu LC-20AD HPLC (Shimadzu Instruments, Columbia, MD). Chromatographic separation of biotin was performed using an Agilent Poroshell EC-C8 2.7 μm 2.1 × 30 mm analytical column at a flow rate of 0.5 mL/min in water, with mobile phase A being 0.1% formic acid in 2% methanol. Mobile phase B is 98% methanol in water with 0.1% formic acid. Elution gradient was as follows: 0–3 minutes, 5% mobile-B; 3–5 minutes, 75% mobile-B; 5–7 minutes, 95% mobile-B; 7–9 minutes, 95% mobile-B; 9–9.01 minutes, and 5% mobile-B. Total analysis time per injection was 12 minutes. The column temperature was set to 40°C, and the injection volume was 150 μL. Mass spectrometry analysis of biotin was operated in positive mode. Biotin was monitored by multiple reaction monitoring. Optimized mass transitions were 245/227 (Q1/Q3) for biotin and 249/231 (Q1/Q3) for biotin-d4. Dwell time was 150 msec for biotin and 50 msec for biotin-d4. Analytical measurement range for biotin was 1–40 ng/mL, with r > 0.99 in linear regression. Three levels of QCs were used at 2, 15, and 30 ng/mL. Coefficient of variation was <5% at all levels. Limit of detection was 0.3 ng/mL, and limit of quantitation was 0.6 ng/mL. Measurement of serum samples from 60 healthy donors (not taking biotin supplements) revealed a 2.5–97.5% reference interval using the percentile approach of <0.3–1.03 ng/mL.
Serum aliquots were analyzed for TSH, free T3 (fT3), free T4 (fT4), total T3 (TT3), total T4 (TT4), thyroxine binding globulin (TBG), and Tg levels with four different platforms, chosen based on the use of a streptavidin/biotin system. Roche Cobas 6000 and Siemens IMMULITE 2000 use streptavidin/biotin binding; LC-MS/MS and Abbott Architect do not. TSH, fT3, fT4, TT3, and TT4 were measured in both the Abbott Architect and Roche Cobas 6000 systems. For Abbott Architect, the normal reference range was TSH 0.4–4.2 μIU/mL, fT3 2.5–3.9 pg/mL, TT3 87–178 ng/dL, TT4 5–12.2 μg/dL, and fT4 0.8–1.5 ng/dL. For Roche Cobas 6000, the normal reference range was TSH 0.27–4.20 μIU/mL, fT3 2.0–4.4 pg/mL, TT3 80–180 ng/dL, fT4 0.93–1.7 ng/dL, and TT4 4.5–12.5 μg/dL. fT3, fT4, TT3, and TT4 also were measured by liquid chromatography with tandem mass spectrometry LC-MS/MS and Tg was measured by Siemens IMMULITE 2000. TBG was assessed with Siemens IMMULITE 2000. Normal range when measured with LC-MS/MS was fT3 1.8–7.6 pg/mL, TT3 62–202 ng/dL, fT4 1.18 2.88 ng/dL, and TT4 4.6 12.9 μg/dL (10). Tg normal range measured with Siemens was 1.6–60 ng/mL.
This study was approved by the Institutional Review Board of MedStar Health Research Institute, study ID 2018-30.
Sample size and statistical analysis
Our study was designed to explore the relationships between biotin and thyroid-related analytes based on a convenience sample of consecutive patients who visited our clinic during the study period 2018–2020. It was powered to detect a moderately strong linear relationship between Y (TSH, etc.) and biotin. A sample size of 13 provided 80% power at an alpha level of 0.05 using an F-test for partial correlation to detect an R2 of 0.45 (squared partial correlation) in a linear regression model of the outcome measure on biotin levels. Data were summarized as mean ± standard deviation for each measurement at every time point. Absolute and percent changes from baseline (A) were computed and tested using paired t-tests. Correlation coefficients between biotin levels and hormone levels were computed for each time point separately and by combining all time points. Associations were also examined and are presented using descriptive graphs. Linear mixed model analyses were conducted to assess the relationship between biotin level and each hormone measurement using repeated measures data. The dependent variables for these models were hormone measures regressed on biotin levels and fixed time effects with baseline (A) as the reference time point and random effects at patient level. Statistical analyses were conducted using Stata (11).
Results
Biotin
On day 1, the mean baseline (A) biotin concentration was 0.48 ng/mL. Mean biotin levels at two (B) and five hours (C) after 10 mg biotin ingestion increased to 59.3 and 26.8 ng/mL, respectively (Table 1). All subjects averred full adherence with dosage of 10 mg biotin/day from day 2 to day 7. On day 8, the mean baseline biotin level (D) 24 hours after the last biotin exposure was 13.5 ng/mL. Biotin levels at two (E) and five hours (F) after an additional 10 mg biotin dose were 68.2 and 43.7 ng/mL, respectively. Blood levels of biotin were higher on day 8 than on day 1 at the same relative time points as follows: baseline (D) was 28 times higher than baseline (A), two hours after 10 mg biotin ingestion (E) was 1.2 times higher compared with (B), and sample (F) five hours after biotin ingestions was 1.6 times higher than (C). Biotin half-life varied between 1.9 and 6.9 hours on day 1 and day 8, respectively.
Table 1.
Changes After Biotin Ingestion
| Time points | Biotin Mean (SD) | Tg Mean (SD) | TBG Mean (SD) | TSH Roche Mean (SD) | TSH Abbott Mean (SD) | |
|---|---|---|---|---|---|---|
| D1, 0h | 0.48 (0.36) | 9.17 (8.20) | 17.23 (4.44) | 1.62 (0.96) | 1.20 (0.60) | |
| D1, 2h | 59.34 (17.56) | 3.35 (3.55) | 17.55 (3.28) | 0.99 (0.52) | 1.06 (0.55) | |
| D1, 5h | 26.85 (13.02) | 3.83 (3.02) | 18.06 (3.50) | 1.22 (0.73) | 1.13 (0.67) | |
| D8, 0h | 3.49 (7.62) | 6.63 (5.56) | 17.26 (4.44) | 1.68 (0.91) | 1.29 (0.60) | |
| D8, 2h | 68.24 (13.91) | 2.88 (2.37) | 17.48 (3.48) | 0.99 (0.54) | 1.04 (0.53) | |
| D8, 5h | 43.75 (14.32) | 3.40 (3.03) | 18.87 (4.95) | 1.10 (0.70) | 1.22 (0.65) | |
| Time points | fT3 |
fT4 |
||||
|---|---|---|---|---|---|---|
| Roche Mean (SD) | Abbott Mean (SD) | LC-MS/MS Mean (SD) | Roche Mean (SD) | Abbott Mean (SD) | LC-MS/MS Mean (SD) | |
| D1, 0h |
2.86 (0.30) |
3.19 (0.33) |
3.03 (1.24) |
1.22 (0.26) |
0.94 (0.12) |
2.11 (0.93) |
| D1, 2h |
3.02 (0.29) |
3.21 (0.40) |
3.13 (1.63) |
1.37 (0.30) |
0.95 (0.12) |
2.42 (1.25) |
| D1, 5h |
2.89 (0.30) |
3.18 (0.37) |
2.96 (1.31) |
1.31 (0.30) |
0.95 (0.13) |
2.05 (0.65) |
| D8, 0h |
2.89 (0.26) |
3.11 (0.32) |
2.83 (1.02) |
1.23 (0.22) |
0.93 (0.12) |
2.01 (0.84) |
| D8, 2h |
2.99 (0.34) |
3.22 (0.30) |
2.75 (0.87) |
1.37 (0.28) |
0.93 (0.11) |
1.94 (0.73) |
| D8, 5h | 2.96 (0.30) | 3.18 (0.28) | 2.89 (0.97) | 1.35 (0.28) | 0.95 (0.13) | 2.15 (1.07) |
| Time points | T3 |
T4 |
||||
|---|---|---|---|---|---|---|
| Roche Mean (SD) | Abbott Mean (SD) | LC-MS/MS Mean (SD) | Roche Mean (SD) | Abbott Mean (SD) | LC-MS/MS Mean (SD) | |
| D1, 0h |
104.4 (17.4) |
95.8 (16.9) |
110.8 (17.3) |
7.17 (1.54) |
7.76 (1.54) |
7.71 (1.72) |
| D1, 2h |
137.2 (36.2) |
94.8 (22.7) |
114.4 (22.1) |
7.35 (1.48) |
8.11 (1.67) |
7.53 (1.57) |
| D1, 5h |
114.3 (15.7) |
92.5 (20.7) |
108.7 (18.6) |
7.20 (1.54) |
7.76 (1.59) |
7.64 (1.72) |
| D8, 0h |
106.5 (17.6) |
89.6 (23.9) |
109.7 (16.6) |
6.98 (1.46) |
7.55 (1.53) |
7.26 (1.54) |
| D8, 2h |
130.5 (27.2) |
93.5 (17.9) |
105.7 (11.7) |
7.08 (1.50) |
7.74 (1.56) |
7.36 (1.46) |
| D8, 5h | 121.0 (21.6) | 90.5 (18.4) | 100.7 (13.7) | 7.34 (1.43) | 7.88 (1.59) | 7.22 (1.50) |
Normal range of analytes measured with Roche Cobas 6000: TSH 0.27–4.20 μIU/mL, fT3 2.0–4.4 pg/mL, TT3 80–180 ng/dL, fT4 0.93–1.7 ng/dL, TT4 4.5–12.5 μg/dL; Abbott Architect: TSH 0.4–4.2 μIU/mL, fT3 2.5–3.9 pg/mL, TT3 87–178 ng/dL, TT4 5–12.2 μg/dL, fT4 0.8–1.5 ng/dL; LC-MS/MS: fT3 1.8 7.6 pg/mL, TT3 62–202 ng/dL, fT4 1.18 2.88 ng/dL, TT4 4.6 12.9 μg/dL; Siemens IMMULITE 2000: Tg 1.6–60 ng/mL.
D1, 0h—day 1, baseline; D1, 2h—day 1, two hours after biotin; D1, 5h—day 1, five hours after biotin; D8, 0h—day 8, baseline; D8, 2h—day 8, two hours after biotin; D8, 5h—day 8, five hours after biotin.
fT3, free T3; fT4, free T4; LC-MS/MS, liquid chromatography with tandem mass spectrometry; T3, triiodothyronine; T4, thyroxine; TBG-thyroxine binding globulin; Tg, thyroglobulin; TSH, thyrotropin; TT4, total T4.
Thyroglobulin
On day 1, ingestion of a single tablet of biotin was associated with a reduction of Tg levels by 56% after two hours (p = 0.002) and 51% after five hours compared with baseline (A) (p = 0.006). On day 8, baseline Tg levels (D) were 29% lower compared with the baseline on day 1 (A) (p = 0.02). Tg levels at two (E) and five hours (F) after the 10 mg biotin dose were 62% and 56% lower compared with baseline (A), p = 0.004 and p = 0.003, respectively (Table 1). Mean Tg levels at each time point are presented in Figure 1 and Table 2. When all the time points were analyzed together, a significant negative correlation between Tg and biotin levels was observed (r = −0.35; p = 0.002) (Fig. 2). The linear mixed model of Tg controlling for fixed time effects estimated that for each unit increase of 1 ng/mL of biotin, average Tg declined by 0.055 ng/mL (p = 0.057). One patient had a Tg level below 1 ng/mL after biotin ingestion and in eight patients Tg levels dropped below 5 ng/mL after biotin.
FIG. 1.
Changes in Tg, TSH, fT4, and T3 levels after biotin ingestion. Normal reference range: Roche Cobas 6000: TSH 0.27–4.20 μIU/mL, fT3 2.0–4.4 pg/mL, TT3 80–180 ng/dL, fT4 0.93–1.7 ng/dL, TT4 4.5–12.5 μg/dL; Abbott Architect: TSH 0.4–4.2 μIU/mL, fT3 2.5–3.9 pg/mL, TT3 87–178 ng/dL, TT4 5–12.2 μg/dL, fT4 0.8–1.5 ng/dL; LC-MS/MS: fT3 1.8 7.6 pg/mL, TT3 62–202 ng/dL, fT4 1.18 2.88 ng/dL, TT4 4.6 12.9 μg/dL; Siemens IMMULITE 2000: Tg 1.6–60 ng/mL. fT4, free T4; LC-MS/MS, liquid chromatography with tandem mass spectrometry; fT3, free T3; T3, triiodothyronine; T4, thyroxine; Tg, thyroglobulin; TT4, total T4.
Table 2.
Thyroglobulin Changes Measured with Siemens IMMULITE 2000
| Difference between time points | Mean (SD) | % | p-Value |
|---|---|---|---|
| D1, 2h–D1, 0h | −5.82 (5.45) | −56 | 0.002 |
| D1, 5h–D1, 0h | −5.34 (8.20) | −51 | 0.006 |
| D8, 0h–D1, 0h | −2.54 (3.41) | −29 | 0.02 |
| D8, 2h–D1, 0h | −6.28 (6.47) | −62 | 0.004 |
| D8, 5h–D1, 0h | −5.77 (5.68) | −56 | 0.003 |
FIG. 2.
Correlation of biotin levels with TSH, T3, fT4, Tg. Normal reference range: Siemens IMMULITE 2000: Tg 1.6–60 ng/mL; Roche Cobas 6000: TSH 0.27–4.20 μIU/mL, TT3 80–180 ng/dL, fT4 0.93–1.7 ng/dL, TT4 4.5–12.5 μg/dL.
Thyrotropin
TSH levels were measured with two instruments, Roche Cobas® 6000 and Abbott Architect. In measurements obtained by Roche Cobas, a single 10 mg biotin dose on day 1 resulted in reduced TSH levels by 34% after two hours (B) and 20% after five hours (C) (Table 3). No significant differences were observed between baseline TSH levels on day 1 (A) and on day 8 (D). On day 8, TSH levels at two (E) and five hours (F) after 10 mg biotin ingestion were 26% and 21% lower, respectively, compared with baseline (A) (Table 3 and Fig. 1). When all the time points were analyzed together, a significant negative correlation between TSH and biotin levels was observed (r = −0.29; p = 0.01) (Fig. 2). The mixed model controlling for time estimated that for each unit increase of 1 ng/mL of biotin, average TSH (by Roche) was reduced by 0.008 μIU/mL (p = 0.018). All the values measured were within the reference range.
Table 3.
Thyrotropin Changes After Biotin Ingestion
| |
TSH measured with Roche Cobas 6000 |
TSH measured with Abbott Architect |
||||
|---|---|---|---|---|---|---|
| Difference between time points | Mean (SD) | % | p-Value | Mean (SD) | % | p-Value |
| D1, 2h–D1, 0h | −0.63 (0.51) | −34 | <0.001 | −0.14 (0.20) | −10 | 0.022 |
| D1, 5h–D1, 0h | −0.41 (0.96) | −20 | 0.006 | −0.08 (0.37) | −4 | 0.475 |
| D8, 0h–D1, 0h | 0.06 (0.42) | 22 | 0.643 | 0.09 (0.27) | 26 | 0.271 |
| D8, 2h–D1, 0h | −0.63 (0.59) | −26 | 0.002 | −0.16 (0.34) | −3 | 0.114 |
| D8, 5h–D1, 0h | −0.52 (0.55) | −21 | <0.001 | 0.02 (0.50) | 15 | 0.883 |
TSH measured with Abbott Architect two hours after biotin ingestion was reduced by 10% on day 1 (p = 0.022), but neither the correlation between TSH levels and biotin nor the mixed model showed significant associations.
Total T4
TT4 levels were measured with three platforms (Roche Cobas 6000, Abbott Architect, and LC-MS/MS) (Table 1). No major significant differences were observed between the measurements at any time points except for a slight 0.3% increase in measured TT4, two hours after the first biotin ingestion when assessed with the Abbott instrument (p = 0.03). However, no significant relationships were observed between TT4 levels and biotin using correlations or the mixed model adjusted for time points.
Free T4
fT4 levels were measured with three instruments (Roche Cobas 6000, Abbott Architect, and LC-MS/MS) (Table 1). No significant differences in fT4 measurements were observed between the time points by either Abbott Architect or LC-MS/MS. However, measurements by Roche Cobas 6000 indicated increased fT4 levels at two and five hours postbiotin ingestion, both on day 1 and day 8 (day 1, 12% increase after two hours [B], 7% after five hours [C] compared with baseline [A], and day 8, 13% increase after two hours [E], 11% after five hours [F] compared with baseline [A]) (Table 4 and Fig. 1). A significant positive correlation was observed between fT4 and biotin levels (r = 0.30; p = 0.008) (Fig. 2). The mixed model also showed a significant effect of biotin levels on fT4 levels when measured by Roche Cobas 6000 (p = 0.004). All patients had normal fT4 values at baseline. In two of them, fT4 increased above the normal range after biotin ingestion.
Table 4.
Free Thyroxine Changes After Biotin Ingestion
| Difference between time points | fT4 measured with Roche Cobas 6000 |
fT4 measured with Abbott Architect |
fT4 measured with LC-MS/MS |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Mean (SD) | % | p-Value | Mean (SD) | % | p-Value | Mean (SD) | % | p-Value | |
| D1, 2h–D1, 0h | 0.15 (0.09) | 12 | <0.001 | 0.15 (0.09) | 1 | <0.001 | 0.31 (0.85) | 17 | 0.214 |
| D1, 5h–D1, 0h | 0.08 (0.06) | 7 | <0.001 | 0.02 (0.06) | 2 | 0.502 | −0.07 (0.65) | 4 | 0.719 |
| D8, 0h–D1, 0h | 0.01 (0.13) | 2 | 0.815 | −0.01 (0.08) | −1 | 0.721 | −0.10 (0.65) | 0.3 | 0.581 |
| D8, 2h–D1, 0h | 0.15 (0.10) | 13 | <0.001 | −0.01 (0.12) | −0.3 | 0.776 | −0.18 (0.93) | −3 | 0.229 |
| D8, 5h–D1, 0h | 0.13 (0.08) | 11 | <0.001 | 0.02 (0.08) | 2 | 0.502 | 0.04 (1.17) | 11 | 0.913 |
Total T3
TT3 levels were measured with all three instruments (Roche Cobas 6000, Abbott Architect, and LC-MS/MS) (Table 1). No significant changes were seen in TT3 levels between the different time points as measured by either LC-MS or Abbott Architect. However, after a 10 mg biotin exposure, TT3 measured with Roche 6000 was increased by 35% and 11% after two (B) and five hours (C), respectively, on day 1 and by 27% and 16% after two (E) and five hours (F), respectively, on day 8 (Table 5 and Fig. 1). A significant positive correlation was observed between TT3 levels and biotin when all measurements were analyzed together (r = 0.62; p < 0.001) (Fig. 2). A very strong statistically significant relationship was seen in the mixed model controlling for time, with TT3 (measured with Roche) increasing by 1.1 ng/dL for each 1 ng/mL increase in biotin (p < 0.001). All patients had normal TT3 values at baseline; three patients had increased TT3 above the normal range after biotin ingestion.
Table 5.
Triiodothyronine Changes After Biotin Ingestion
| |
Roche Cobas 6000 |
Abbott Architect |
LC-MS/MS |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Difference between time points | Mean (SD) | % | p-Value | Mean (SD) | % | p-Value | Mean (SD) | % | p-Value |
| D1, 2h–D1, 0h | 32.8 (40.5) | 35 | 0.013 | −1.0 (9.9) | −2 | 0.723 | 3.6 (16.6) | 4 | 0.451 |
| D1, 5h–D1, 0h | 9.9 (11.0) | 11 | 0.007 | −3.2 (10.7) | −4 | 0.299 | −2.1 (10.6) | −2 | 0.489 |
| D8, 0h–D1, 0h | 2.1 (10.1) | 3 | 0.469 | −6.2 (14.2) | −7 | 0.144 | −1.1 (20.6) | 1 | 0.854 |
| D8, 2h–D1, 0h | 26.1 (26.2) | 27 | 0.004 | −2.3 (16.9) | −2 | 0.434 | −5.9 (17.8) | −4 | 0.275 |
| D8, 5h–D1, 0h | 16.6 (12.1) | 16 | <0.001 | −5.3 (9.6) | −6 | 0.069 | −10.1 (17.9) | −8 | 0.064 |
Free T3
For each time point, fT3 levels were measured using Roche Cobas 6000 (Abbott Architect, LC-MS/MS) (Table 1). No significant differences in fT3 levels were observed at the different time points as measured by LC-MS/MS or Abbott Architect. Measurements by Roche Cobas 6000 indicated a slight increase of 6% in fT3 at two hours after biotin intake only on day 1. A positive correlation was observed between biotin and fT3 levels by Roche (r = 0.27; p = 0.02), but no significant association between biotin and fT3 (Roche) was observed in the mixed model.
Thyroxine binding globulin
TBG levels in all subjects were within the normal reference range as measured with Siemens IMMULITE 2000 and no changes were detected after biotin exposure (Table 1).
Discussion
As indicated by earlier reports (8,12–14) and confirmed by our results, the presence of increased levels of biotin in blood subjected to assay platforms using streptavidin-biotin binding can result in laboratory test results that can be either falsely higher (in a competitive assay) or falsely lower (in a sandwich assay).
In competitive binding assays, excess biotin binds to the streptavidin-coated microparticles competing with the biotinylated antibody (Fig. 3). As a consequence, after the wash step, less biotinylated antibody will be available to bind to the labeled analyte. A competitive binding assay will interpret the lower labeled analyte level as a higher analyte level giving a falsely higher result. In sandwich assays, there is a direct relationship between the tested analyte and the level of labeled antibody present after the wash. Thus, extraneous biotin would compete for streptavidin binding with the biotinylated antibody resulting in lower levels of labeled antibody and a falsely lower level of the analyte. In our study, a streptavidin/biotin system was used for the Roche and Siemens platforms, but not for the Abbott and LC-MS/MS platforms.
FIG. 3.
Mechanism of biotin interference in laboratory assays.
Various factors may contribute to the artifactually falsely high or low results, including the degree of blood biotin elevation based in turn on the amount of biotin ingested, the time interval from biotin ingestion to blood specimen collection, the biotin interference threshold, and the patient's own relative biotin metabolism.
As influenced by the presence of high biotin levels, thyroid function test results could mimic hyperthyroidism with low TSH, high TT3, high fT3, high fT4, and positive anti-TSH receptor antibodies (15–18). Our results confirmed this trend for TSH, TT3, and fT4 particularly when the Roche instrument was used. After biotin ingestion in some patients, TT3 and fT4 increased above the normal range. While TSH decreased, they remained within the normal reference range, notwithstanding that the TSH normal reference range remains subject to debate (19,20).
It is known that biotin serum levels peak approximately one to three hours after an oral dose, with steady-state levels achieved in approximately three days if biotin consumption is continued (21) The biotin half-life in our study was in accordance with the biotin half-life reported in the literature (ranging from 2 to 15 hours) (7,21,22).
We noted no further interference with the assay performance of TSH, TT3, and fT4 24 hours after the last biotin ingestion, a time when average levels of circulating biotin are lower than the biotin threshold.
Given the increasingly common use of biotin supplements in high dosage and the prevalence of hypothyroidism with dependence on periodic measurement of thyroid function tests for adjustment of T4 dosage, there is significant potential for clinical mismanagement of these patients based on misleading test results. The same concern applies to patients on levothyroxine post-thyroidectomy, or patients with thyroid hyperfunction treated with radioiodine or antithyroid drugs. Determination of biotin intake would be particularly important in situations requiring more exact titration of levothyroxine dosage, such as in pregnant women, in children, in the elderly, and in patients being monitored for residual or recurrent thyroid cancer (18).
In its role as a tumor marker, measurement of Tg, in precise, sensitive, and reproducible assays, is indispensable to the follow-up of patients with a history of thyroid cancer. Falsely high or low Tg levels could lead physicians to erroneously assume either worsening or improving disease status, respectively (23). Few data are available regarding biotin interference with Tg measurement, and the findings of our study are of special importance in this regard. The reports extant indicating a falsely lower Tg level due to biotin interference (16,17,24) have been largely case reports. There is, however, one study that observed lower serum Tg levels after 10 mg of biotin in a cohort of 10 subjects. While we utilized a Siemens 2000 IMMULITE platform, these authors used a Beckman UniCel DxI 800 and Roche Cobas e602 platform (24). Highly sensitive Tg assays have become available for monitoring patients with differentiated thyroid cancer (25–27). The newer methodology may improve the clinical utility of Tg assay for the detection of residual disease or disease recurrence (28). While the method uses a streptavidin/biotin system in which biotin interference could be expected, we have not found any information on the magnitude of Tg change due to biotin (29).
The biotin interference threshold depends on the test and the platform used (30). A lower threshold would predispose to interference of longer duration after discontinuation of the biotin supplement (21). This can be illustrated with a Tg assay characterized by a lower interference threshold of 5 ng/mL compared with a different assay with a threshold of ∼20 ng/mL (30). The impact on Tg results in our study was that biotin interference was still present after 24 hours when the average circulating biotin level was 15 ng/mL. Our linear mixed model estimates an average change of the analyte based on a 1 ng/mL change in biotin level. The model is not intended to be applied to specific biotin levels. Since only the average change in analyte is described, it is conceivable that the interference can become progressively more significant with higher biotin levels or less significant with biotin levels below the threshold.
LC-MS/MS is a highly sensitive, specific, and precise platform that does not use streptavidin/biotin linkage and offers higher accuracy than other immunoassays (12). To our knowledge, this is the first study to analyze biotin interference comparing the results obtained with Roche Cobas 6000 and Abbott Architect immunoassays, with LC-MS/MS results obtained for the same analytes and time points.
Although measurement of Tg and TSH by LC-MS/MS was not possible, levels of Tg and TSH after biotin administration were compared with baseline levels in the same subjects (14). Furthermore, a mixed model controlling for time was applied to compare the different time points for each analyte.
Proper identification and management of biotin interference are crucial for the correct interpretation of laboratory results. Despite an FDA warning (1), surveys indicate that only 6% of a cohort of people presenting at a hospital had been informed of the potential of biotin interference; and 30% of the patients were not aware of their daily dosage of biotin (31,32). In a recent survey, 19% of the physicians prescribing biotin were unaware of laboratory interference and almost half of the physicians did not ask patients to discontinue biotin before laboratory testing (33). Indeed, biotin levels measured in random samples taken at an emergency room revealed that 7% of the subjects had biotin levels that could potentially interfere with certain test results (31).
Of the different approaches that can be taken to avoid biotin interaction, it should be recommended that individuals taking 10 mg a day of biotin suspend intake for 24 hours before venipuncture if TSH, fT4, and TT3 are analyzed and, more than 24 hours if Tg is being analyzed. Optimally, biotin intake should be recorded in the patients' medical record.
When the allowance of an interval of several days between recommended discontinuation of biotin intake and venipuncture with assay performance that would be required for biotin washout is not feasible, alternative approaches to analyte measurement can be performed. These include testing in a platform that does not use streptavidin/biotin bonding (e.g., LC-MS/MS, Abbott Architect), using a protocol to deplete plasma biotin by means of streptavidin-coated microparticles (34,35) or using improved new-generation platforms that have a much higher threshold for biotin interference (36,37).
In summary, this study demonstrated the interference of serum biotin with measurement of thyroid-related hormones, TSH, fT4, TT3, and Tg, after a commonly ingested oral daily dose of 10 mg biotin. Clinicians who manage patients with thyroid disorders must be aware of the risks of erroneous test results when monitoring thyroid hormone replacement therapy in their patients. Future studies to further delineate the scope of this issue could use larger cohorts of subjects and explore the utility of a newer and different methodology or assay platforms for the various analytes affected by biotin.
Acknowledgments
This work was supported by The Catherine Heron and Al Schneider Fellowship in Thyroid Cancer.
Authors' Contributions
D.Y.: substantial contributions to the conception and design of the work, interpretation of data, and final approval of the version to be published.
S.J.S.: substantial contributions to the design of the work, data acquisition (laboratory analysis of the samples), and final approval of the version to be published.
B.S., B.W., G.N., H.N., and D.R.M.: substantial contributions in the laboratory analysis of the samples.
M.M.: substantial contributions in the statistical analysis and interpretation of the data.
D.W.: substantial contributions in the conception of the study and in patient recruitment.
C.J.G.-L.: substantial contributions in the conception of the work and in patient recruitment.
J.K.-G.: substantial contributions to the conception and design of the study, interpretation of data, and final approval of the version to be published.
K.D.B.: substantial contributions to the conception and design of the study, data interpretation, and final approval of the version to be published.
L.W.: substantial contributions to the conception and design of the study, data interpretation, and final approval of the version to be published.
Author Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.
References
- 1.FDA safety communication. Available at https://www.fda.gov/medical-devices/safety-communications/update-fda-warns-biotin-may-interfere-lab-tests-fda-safety-communication. (accessed May16, 2021)
- 2. Institute of Medicine 1998 Dietary Reference Intakes for Thiamin, Riboflavin, Niacin, Vitamin B6, Folate, Vitamin B12, Pantothenic Acid, Biotin, and Choline. Washington, DC: The National Academies Press [PubMed] [Google Scholar]
- 3.Couloume L, Barbin L, Leray E, et al. 2020High-dose biotin in progressive multiple sclerosis: a prospective study of 178 patients in routine clinical practice. Mult Scler 26:1898–1906 [DOI] [PubMed] [Google Scholar]
- 4.Battini R, Olivieri G, Milone R, et al. 2020Spinal cord demyelination in children: a diagnostic challenge in neuropaediatrics for a good outcome. Brain Dev 42:457–461 [DOI] [PubMed] [Google Scholar]
- 5.Lim SK, Pilon A, Guechot J. 2017Biotin interferes with free thyroid hormone and thyroglobulin, but not TSH measurements using Beckman-Access immunoassays. Ann Endocrinol (Paris) 78:186–187 [DOI] [PubMed] [Google Scholar]
- 6.Bayart JL, Favresse J, Stoefs A, et al. 2020Biotin interferences: have we neglected the impact on serological markers? Clin Chim Acta 503:107–112 [DOI] [PubMed] [Google Scholar]
- 7.Samarasinghe S, Meah F, Singh V, et al. 2017Biotin interference with routine clinical immunoassays: Understand the causes and mitigate the risks. Endocr Pract 23:989–998 [DOI] [PubMed] [Google Scholar]
- 8.Favresse J, Burlacu MC, Maiter D, Gruson D. 2018Interferences with thyroid function immunoassays: Clinical implications and detection algorithm. Endocr Rev 39:830–850 [DOI] [PubMed] [Google Scholar]
- 9.Kodani M, Poe A, Drobeniuc J, Mixson-Hayden T 2020 Determination of potential biotin interference on accuracy of results of serologic assays for various viral hepatitis markers. J Med Virol [Epub ahead of print]; DOI: 10.1002/jmv.25880 [DOI] [PubMed] [Google Scholar]
- 10.Sun Q, Avallone L, Stolze B, Araque KA, Özarda Y, Jonklaas J, Parikh T, Welsh K, Masika L, Soldin SJ. 2020Demonstration of reciprocal diurnal variation in human serum T3 and rT3 concentration demonstrated by mass spectrometric analysis and establishment of thyroid hormone reference intervals. Ther Adv Endocrinol Metab 11:2042018820922688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Available at www.stata.net. (accessed May16, 2021)
- 12.Araque KA, Klubo-Gwiezdzinska J, Nieman LK, Welsh K, Soldin SJ. 2019Assessment of thyroid function tests and harmonization: opinion on thyroid hormone harmonization. Ther Adv Endocrinol Metab 10:2042018819897049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Biscolla RPM, Chiamolera MI, Kanashiro I, Maciel RMB, Vieira JGH. 2017. A single 10 mg oral dose of biotin interferes with thyroid function tests. Thyroid 27:1099–1100 [DOI] [PubMed] [Google Scholar]
- 14.Li D, Radulescu A, Shrestha RT, et al. 2017Association of biotin ingestion with performance of hormone and nonhormone assays in healthy adults. JAMA 318:1150–1160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sharma A, Baumann NA, Shah P. 2017Biotin-induced biochemical graves disease: a teachable moment. JAMA Intern Med 177:571–572 [DOI] [PubMed] [Google Scholar]
- 16.Wijeratne NG, Doery JC, Lu ZX. 2012Positive and negative interference in immunoassays following biotin ingestion: a pharmacokinetic study. Pathology 44:674–675 [DOI] [PubMed] [Google Scholar]
- 17.Barbesino G2016Misdiagnosis of Graves' disease with apparent severe hyperthyroidism in a patient taking biotin megadoses. Thyroid 26:860–863 [DOI] [PubMed] [Google Scholar]
- 18.Kummer S, Hermsen D, Distelmaier F. 2016Biotin treatment mimicking Graves' disease. N Engl J Med 375:704–706 [DOI] [PubMed] [Google Scholar]
- 19.Soldin S, Sun Q, Stolze B. 2018TSH should not be used as a single marker of thyroid function. Ann Thyroid Res 4:151–154 [Google Scholar]
- 20.Ling C, Sun Q, Khang J, Felipa Lastarria M, Strong J, Stolze B, Yu X, Parikh TP, Waldman MA, Welsh K, Jonklaas J, Masika L, Soldin SJ. 2018Does TSH reliably detect hypothyroid patients? Ann Thyroid Res 4:122–125 [PMC free article] [PubMed] [Google Scholar]
- 21.Grimsey P, Frey N, Bendig G, et al. 2017Population pharmacokinetics of exogenous biotin and the relationship between biotin serum levels and in vitro immunoassay interference. Int J Pharmacokinet 2:247–256 [Google Scholar]
- 22.Chun KY2017Biotin interference in diagnostic tests. Clin Chem 63:619–620 [DOI] [PubMed] [Google Scholar]
- 23.Haugen BR, Alexander EK, Bible KC, et al. 20162015 American Thyroid Association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer: the American Thyroid Association guidelines task force on thyroid nodules and differentiated thyroid cancer. Thyroid 26:1–133 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zhang Y, Wang R, Dong Y, Huang G, Ji B, Wang Q. 2020Assessment of biotin interference in thyroid function tests. Medicine (Baltimore) 99:e19232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Trimboli P, Imperiali M, Piccardo A, et al. 2018Multicentre clinical evaluation of the new highly sensitive Elecsys® thyroglobulin II assay in patients with differentiated thyroid carcinoma. Clin Endocrinol (Oxf) 88:295–302 [DOI] [PubMed] [Google Scholar]
- 26.Broughton MN, Nome R, Sandven I, Paus E, Bjøro T. 2016Characterization of a new highly sensitive immunometric assay for thyroglobulin with reduced interference from autoantibodies. Tumour Biol 37:7729–7739 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Groen AH, Klein Hesselink MS, Plukker JT, et al. 2017Additional value of a high sensitive thyroglobulin assay in the follow-up of patients with differentiated thyroid carcinoma. Clin Endocrinol (Oxf) 86:419–424 [DOI] [PubMed] [Google Scholar]
- 28.Rotteveel-de Groot DM, Ross HA, Janssen MJR, et al. 2016Evaluation of the highly sensitive Roche thyroglobulin II assay and establishment of a reference limit for thyroglobulin-negative patient samples. Pract Lab Med 5:6–13 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Trambas C, Lu Z, Yen T, Sikaris K. 2018Characterization of the scope and magnitude of biotin interference in susceptible Roche Elecsys competitive and sandwich immunoassays. Ann Clin Biochem 55:205–215 [DOI] [PubMed] [Google Scholar]
- 30.Li DFA, Cervinski MA, Lynch KL, Kyle PB. 2020AACC guidance document on biotin interference in laboratory tests. J Appl Lab Med 5:575–587 [DOI] [PubMed] [Google Scholar]
- 31.Katzman BM, Lueke AJ, Donato LJ, Jaffe AS, Baumann NA. 2018Prevalence of biotin supplement usage in outpatients and plasma biotin concentrations in patients presenting to the emergency department. Clin Biochem 60:11–16 [DOI] [PubMed] [Google Scholar]
- 32.John JJ, Cooley V, Lipner SR. 2019Assessment of biotin supplementation among patients in an outpatient dermatology clinic. J Am Acad Dermatol 81:620–621 [DOI] [PubMed] [Google Scholar]
- 33.Waqas B, Wu A, Yim E, Lipner SR 2020 A survey-based study of physician practices regarding biotin supplementation. J Dermatolog Treat [Epub ahead of print]; DOI: 10.1080/09546634.2020.1770178 [DOI] [PubMed] [Google Scholar]
- 34.Mrosewski I, Urbank M, Stauch T, Switkowski R. 2020Interference from high-dose biotin intake in immunoassays for potentially time-critical analytes by Roche: evaluation of a countermeasure for worst-case scenarios. Arch Pathol Lab Med 144:1108–1117 [DOI] [PubMed] [Google Scholar]
- 35.Trambas C, Lu Z, Yen T, Sikaris K. 2018Depletion of biotin using streptavidin-coated microparticles: a validated solution to the problem of biotin interference in streptavidin-biotin immunoassays. Ann Clin Biochem 55:216–226 [DOI] [PubMed] [Google Scholar]
- 36.Ruth I, Mathieu E, Burniat A, Fage D, Cotton F, Wolff F. 2020Interferences in free thyroxine concentration using the Roche analytical platform: improvement of the third generation? Clin Chem Lab Med 58:e36–e39 [DOI] [PubMed] [Google Scholar]
- 37.Bowen R, Benavides R, Colon-Franco JM, et al. 2019Best practices in mitigating the risk of biotin interference with laboratory testing. Clin Biochem 74:1–11 [DOI] [PubMed] [Google Scholar]