Abstract
Background
The diagnosis of subclinical hypothyroidism is defined as the presence of an elevated thyroid stimulating hormone (TSH) with a normal free thyroxine (FT4) level. The commonly used direct analogue immunoassays for the measurement of FT4 have been shown to have poor performance at the upper and lower limits of the FT4 reference interval.
Purpose
The purpose of this pilot study was to investigate the percentage of individuals classified as having subclinical hypothyroidism with a standard immunoassay, that actually have low free thyroid hormone levels by mass spectrometry measurements.
Design
Outpatient samples with elevated TSH values and normal FT4 concentrations as per standard immunoassay methods were collected. FT4 and free triiodothyronine (FT3) analyses were performed on these samples using liquid chromatography–tandem mass spectrometry (LC-MS/MS).
Results
Sixty five percent (n=26) of patients (n=40) had (LC-MS/MS) FT4 or FT3 or both FT4 and FT3 values below mass spectrometry reference limits.
Conclusions
Our findings indicate that the direct analogue immunoassay method for FT4 measurement results in a significant proportion of patients being misclassified as having subclinical hypothyroidism.
Keywords: Mass spectrometry, Free thyroid hormone, subclinical hypothyroidism
1. Background
Subclinical hypothyroidism is defined as increased serum thyroid stimulating hormone (TSH) levels with normal serum free thyroxine (FT4) concentrations [1]. The prevalence of subclinical hypothyroidism in the population without known thyroid disease has been reported to be 4% to 10% [2,3]. Debate still exists with regards to the clinical importance of and therapy for elevation of serum TSH (in particular elevated levels <10 mIU/l) and the exact upper limit of normal for the serum TSH level that also varies with age [1]. Most clinical laboratories perform TSH and FT4 measurements on immunoassay (IA) platforms [4]. Whilst TSH analyses on immunoassay platforms are considered quite reliable, the validity of FT4 analysis by direct analogue immunoassay has been questioned for many years [5,6]. Significant limitations of the currently used FT4 immunoassays that have been described are the influence of changes in binding protein concentrations and a weak inverse linear log relationship to TSH in hypo- and hyperthyroid individuals [7–11]. These assays appear to perform best in euthyroid individuals but poorly in individuals with thyroid disease [12]. FT4 immunoassays' poor performance at low concentrations may lead to misclassification of patients as having subclinical hypothyroidism, when in fact they have FT4 levels lower than the reference interval when measured by tandem mass spectrometry. Liquid chromatography–tandem mass spectrometry (LC-MS/MS) following ultrafiltration of the sample has been shown to perform better in such circumstances, and in the case of FT4, agrees better with the gold standard equilibrium dialysis assay [10]. Numerous pharmacological agents affect thyroid hormone measurements and may further confuse interpretation of results beyond the problems encountered with immunoassay measurements. Drugs may affect thyroxine binding globulin levels (for example estrogens), thyroid hormone binding (for example carbamazepine), TSH levels (for example glucocorticoids), conversion of T4 to T3 (for example amiodarone) and may even induce thyroid disease (for example lithium) [13–16].
2. Materials and Methods
2.1 Sample collection
Outpatient serum samples received at the NIH Clinical Center (NIH-CC) for 4 months in 2013 with elevated TSH values and FT4 concentrations within the laboratory reference interval were selected for inclusion in the study. All samples were collected between 8–11am. Samples were stored at −70 degrees Celsius until MS analysis. Patients' samples were excluded if they had known thyroid disease, were receiving thyroid hormone replacement therapy or other medications that are known to affect FT4 values or cause elevation in TSH values. Samples from patients older than sixty years and those that were positive for anti thyroid peroxidase antibodies were also excluded. The study was approved by the Institutional Review Board of the NIH (Clinical Protocol number 93-CC-0094).
2.2 Immunoassay measurements
Samples were analyzed at NIH-CC Department of Laboratory Medicine. TSH and FT4 samples were processed according to usual laboratory procedures. FT4, reference interval 9.8–19.4 pmol/l (0.76– 1.50 ng/dl), was measured by a direct (analogue) immunoassay method on a Dimension Vista (Siemens Healthcare Diagnostics). TSH (reference interval, RI: 0.40–4.00 mIU/l) and free triiodothyronine (FT3, RI: 2.8– 6.5 pmol/l or 1.8– 4.20 pg/ml) were also measured on the Vista. Anti–thyroid peroxidase antibody (ATPOA) (RI: < 35 IU/ml) testing was performed on Immulite XPI 2000 (Siemens Healthcare Diagnostics). Two or three levels of commercially available internal quality control material were analysed at the start of each run
2.3 LC-MS/MS measurement
The FT4, reference interval used was 17.4 –30.9 pmol/l (1.35—2.40 ng/dl) and for FT3, reference interval 2.3–9.5 pmol/l (1.5–6.1 pg/ml). Samples were analysed in batches. Analyses were performed as per methods previously published [8,10,17]. Briefly four hundred microliters of human plasma/serum was filtered through a Centrifree YM-30 ultrafiltration device by centrifugation, and 450 μl of deuterium labeled internal standard, T4-d5 in methanol was then added to 150 μl of ultrafiltrate for deproteinization. After vortexing and centrifugation, 500 μl of supernatant was diluted with 400 μl of distilled de-ionized water and a 650 μl aliquot was injected onto a C-18 column. After washing, the switching valve was activated and the analytes were eluted from the column with a water/methanol gradient into the MS/MS system. Quantification by multiple reaction-monitoring (MRM) analysis was performed in the negative mode(ESI−). Three levels of internal quality control were analysed at the beginning and end of each run.
2.4 Statistical analysis
Non-normally distributed data (MS FT4 values) was normalised by log-transformation before analysis and back-transformed for data presentation. We used the Kolomogorov-Smirnov test to test for normality, and we used Pearson's correlation coefficient, Bland–Altman difference plots, and Passing Bablock regression analysis to evaluate the methods. Statistical analysis was performed on Medcalc Version 12 (MedCalc Software). A pf <0.05 was considered statistically significant.
3. Results
Fifty seven samples with increased TSH and immunoassay FT4 within normal reference limits were collected. Following exclusion of patients ≥ 60 y and those positive for ATPOA, a total of 40 samples from patients aged 6 – 59 y were included in the study. TSH values ranged between 4.3 and 8.2 mIU/l. Analysis of variance (ANOVA) revealed no statistically significant difference for values between males and females. See Table 1 for summary of results.
Table 1.
Summary of results
| Test | All patients (n=40) | Female (n=24) | Male (n=16) | ANOVA P value |
|---|---|---|---|---|
| Age (years) | Mean : 39.7 (24.5–44.9) SD: 16.2 |
Mean: 39.3 (32.3–46.4) SD: 16.6 |
Mean:40.3 (31.7–48.8) SD:16.1 |
P=0.32 |
| TSH (mlU/L) | Mean : 5.8 (95% CI, 5.3–6.3) SD: 1.5 |
Mean: 6.0 (95% CI, 5.3–6.7) SD:1.7 |
Mean: 5.5 (95% CI, 4.8 – 6.2) SD:1.2 |
P=0.36 |
| FT4 (IA) (ng/dL) | Mean: 1.0 (95% CI, 0.9–1.1) SD:0.2 |
Mean: 1.1 (95% CI,0.9–1.2) SD:0.2 |
Mean : 0.9 (95% CI,0.9–1.0) SD:0.1 |
P=0.07 |
| FT3 (IA) (pg/mL) | Mean: 3.0 (95% CI, 2.9–3.2) SD:0.4 |
Mean: 3.0 (95% CI, 2.8–3.2) SD: 0.4 |
Mean: 3.0 (95% CI, 2.8–3.3) SD: 0.4 |
P=0.73 |
| FT4 (MS) (ng/dL) | Mean: 1.4 (95% CI, 1.3–1.6) SD: 0.6 |
Mean: 1.5 (95% CI, 1.3–1.7) SD: 0.7 |
Mean: 1.3 (95% CI, 1.2–1.5) SD: 0.3 |
P=0.39 |
| FT3 (MS) (pg/mL) | Mean: 1.9 (95% CI, 1.7–2.1) SD:0.6 |
Mean: 2.0 (95% CI, 1.8–2.2) SD:0.6 |
Mean: 1.8 (95% CI, 1.4–2.1) SD:0.5 |
P=0.26 |
Abbreviations: SD (standard deviation); TSH (Thyroid stimulating hormone); FT4 (free thyroxine) ; FT3 (free tri-iodothyronine); IA (immunoassay) ; MS (mass spectrometry)
The CVs for the immunoassay methods used were as follows: CV of TSH 3–6% at a concentration of 5.8 mIU/l was 3–6%; FT4 at 12.4 pmol/l (0.96 ng/dl) was 4 –6%; FT3 at 19.2 pmol/l (1.49 pg/ml) was 6.9–7.2%. The CVs for the LC-MS/MS assays were: FT4- 4.1–6.6% at concentrations of 8.5 pmol/l(0.66 ng/dl) and 33.8 pmol/l (2.62 ng/dl); FT3- ≤ 9% at concentrations between 0.23 pmol/l (0.15 pg/ml) and 3.44 pmol/l (2.22 pg/ml).
Sixty–five percent (n=26/40) of patients had mass spectrometry FT4 or FT3 or both FT4 and FT3 values below mass spectrometry reference limits. (Fig. 1 summarises findings). Sixteen of the 26 patients (62%) had only low MS FT4 results. Three of the 26 patients (12%) had only low MS FT3 results and 7 patients (27%) had both low MS FT4 and MS FT3 results. Fifty eight percent (n=23/40) of patients that would be classified as subclinical hypothyroidism as per immunoassay FT4 measurements had LC-MS/MS FT4 values that were below the reference interval. Patients with LC-MS/MS FT4 results below the reference interval had on average values 16% below the lower limit of the LC- MS/MS specific reference interval. The majority had values that were greater than 10% below the lower limit. Thirteen patients (n=13/23) had MS FT4 results that were >10% below the low reference limit. Nine patient (9 of 23) had MS FT4 values >15% below the low reference limit.
Figure 1.
Pearson's correlation coefficient (n=34) between IA FT4 and MS FT4 was 0.55 with a 95% confidence interval (CI) of 0.28–0.73 and between IA and MS for FT3 it was 0.30 (95% CI 0.14–0.45). Regression analysis in the population studied (Fig. 2 and 3) for mass spectrometry versus immunoassay showed poor correlations between the two methods for both FT4: Slope 2.85 (95% CI 1.80 to 6.29), intercept −1.40 (95% CI −4.56 to −0.40) and FT3: Slope 1.29 (95% CI 0.66 to 2.09), intercept −2.16 (95% CI −4.72 to −0.12)
Figure 2.
Correlation between Immunoassay and LC-MS/MS for FT4
Abbreviations: IA: Immunoassay; MS: Mass spectroscopy; FT4: Free thyroxine
Figure 3.
Correlation of Immunoassay and LC-MS/MS FT3
Abbreviations: IA: Immunoassay; MS: Mass spectroscopy; FT3 Free triiodothyronine
There was significant bias between MS and immunoassay values, with FT4 immunoassay values being on average 48 % lower than MS values and FT3 immunoassay values on average 36 % higher than MS results. Bland Altman percentage difference plots show poor agreement between IA and MS data (Figs. 4 and 5); the 95% limits of agreement for IA FT4 and MS FT4 are between 13.4% and −82.3%. The 95% limits of agreement for IA FT3 and MS FT3 are between −7.3% to 102%.
Figure 4.
Bland-Altman Percentage Plot for IA FT4 versus MS FT4 (ng/dL)
Abbreviations: IA: Immunoassay; MS: Mass spectroscopy; FT4: Free thyroxine
Figure 5.
Bland-Altman Percentage Plot for IA FT3 versus MS FT3 (pg/mL)
Abbreviations: IA: Immunoassay; MS: Mass spectroscopy; FT3: Free triiodothyronine
4. Discussion
The findings of this preliminary study reveal that the direct analogue immunoassay method for FT4 measurement results in a significant proportion of patients being misclassified as having subclinical hypothyroidism. This is important as these individuals already have biochemically overt hypothyroidism by MS measurement. They should be followed up closely, in conjunction with clinical findings and symptoms, with a lower threshold for deciding on replacement therapy.
Mass spectrometric analysis has become the reference/ gold standard method for the measurement of various analytes. This method offers elevated analytical specificity and sensitivity [18]. Analysis of FT4 using LC-MS/MS with deuterated internal standard following sample ultrafiltration has been shown to correlate better with log TSH and with the gold standard equilibrium dialysis method in comparison to FT4 measured by direct immunoassay [8,10]. On this basis the authors believe that the “correct” FT4 (i.e., measurement by LC-MS/MS) is the one that correlates best with log TSH. Also FT4 measurement by MS is more accurate (particularly at extremes of range) than immunoassay measurement. Hence the significant difference in free thyroid hormone results and their interpretation observed between the two assays in this study is not merely a function of the different reference intervals but of the performance of the assays. In addition for this study we utilised, previously established method specific reference intervals for FT4 and FT3 tests on both immunoassay and LC-MS/MS to interpret patient values [19–21]. The LC-MS/MS reference intervals were derived from testing on over 1400 healthy children, aged 1–18 y [19] and was also confirmed in a healthy adult population (20–60 y; n=300) (unpublished data). The immunoassay FT4 and FT3 reference intervals provided by the manufacturer for the Siemens Dimension Vista (n=199) [20] were confirmed by the laboratory using healthy volunteers (n=20). The quoted reference intervals for FT4 (0.76–1.50 ng/dl) used in this study also concur with other reference interval studies by Lepoutre et al .10.3 –16.8 pmol/l (0.8–1.3 ng/dl) performed using the same immunoassay platform (n=129) [21].
We used outpatient samples and excluded patients on medications known to affect thyroid function tests. As TSH increases with age we also limited the age range from 6–59 years. It is of interest to note that the two patients excluded from final analysis due to positive ATPOA were defined as having subclinical hypothyroidism using immunoassay FT4 values but analysis with LC-MS/MS also revealed that both individuals had FT4 levels below the reference interval for the mass spectrometric method.
An important limitation of this study is the possible bias that may have been introduced by selecting those samples with FT4 values within the IA reference interval (and increased TSH), that were first analysed by immunoassay. We intend to perform a follow-up study where FT4 testing is performed first by LC-MS/MS and then selected samples run by immunoassay to confirm the findings of the current study. We intend to confirm the results of this study using a larger cohort whose only apparent abnormality is that they have been classified as having subclinical hypothyroidism.
Highlights
MS analysis of outpatient samples with elevated TSH and normal immunoassay FT4 values was performed.
This study revealed that 65% of these patients had low FT4 values by LC-MS/MS measurement.
MS analysis of free thyroid hormones is recommended where subclinical hypothyroidism is suspected.
Consider MS analysis of FT4 & FT3 when immunoassay results are discrepant to TSH/clinical findings.
Acknowledgements
This research was supported (in part) by the Intramural Research Project of the NIH Clinical Center
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Gharib H, Michael Tuttle R, Baskin HJ, Fish LH, Singer PA, McDermott MT. Subclinical Thyroid Dysfunction: A Joint Statement on Management from the American Association of Clinical Endocrinologists, the American Thyroid Association, and The Endocrine Society. J Clin Endocrinol Metab. 2005;90:581–585. doi: 10.1210/jc.2004-1231. [DOI] [PubMed] [Google Scholar]
- 2.Canaris GJ, Manowitz NR, Mayor G, Ridgway EC. The Colorado thyroid disease prevalence study. Arch Intern Med. 2000;160:526–534. doi: 10.1001/archinte.160.4.526. [DOI] [PubMed] [Google Scholar]
- 3.Hollowell JG, Staehling NW, Flanders WD, et al. Serum TSH, T4, and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III) J Clin Endocrinol Meta. 2002;87:489–499. doi: 10.1210/jcem.87.2.8182. [DOI] [PubMed] [Google Scholar]
- 4.Baloch Z, Carayon P, Conte-Devolx B, et al. Laboratory medicine practice guidelines: laboratory support for the diagnosis and monitoring of thyroid disease. Thyroid. 2003;13:3–126. doi: 10.1089/105072503321086962. [DOI] [PubMed] [Google Scholar]
- 5.Midgley JE, Christofides ND. Point: Legitimate and illegitimate tests of free-analyte assay function. Clin Chem. 2009;55:439–441. doi: 10.1373/clinchem.2008.116525. [DOI] [PubMed] [Google Scholar]
- 6.Wilcox RB, Nelson JC. Counterpoint: Legitimate and illegitimate tests of free-analyte assay function: we need to identify the factors that influence free-analyte assay results. Clin Chem. 2009;55:442–444. doi: 10.1373/clinchem.2008.120154. [DOI] [PubMed] [Google Scholar]
- 7.Toldy E, Locsei Z, Szabolcs I, Bezzegh A, Kovacs GL. Protein interference in thyroid assays: an in vitro study with in vivo consequences. Clinica Chimica Acta. 2005;352:93–104. doi: 10.1016/j.cccn.2004.08.014. [DOI] [PubMed] [Google Scholar]
- 8.Van Deventer HE, Mendu DR, Remaley AT, Soldin SJ. Inverse Log-Linear Relationship between Thyroid-Stimulating Hormone and Free Thyroxine Measured by Direct Analog Immunoassay and Tandem Mass Spectrometry. Clin Chem. 2011;57:122–127. doi: 10.1373/clinchem.2010.154088. [DOI] [PubMed] [Google Scholar]
- 9.Van Deventer HE, Soldin SJ. The expanding role of tandem mass spectrometry in optimizing diagnosis and treatment of thyroid disease. Advances in Clinical Chemistry. 2013;61:127–152. doi: 10.1016/b978-0-12-407680-8.00005-1. [DOI] [PubMed] [Google Scholar]
- 10.Soldin SJ, Soukhova N, Janicic N, Jonklaas J, Soldin OP. The measurement of free thyroxine by isotope dilution tandem mass spectrometry. Clin Chim Acta. 2005;358:113–118. doi: 10.1016/j.cccn.2005.02.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Soldin SJ, Cheng LL, Lam Y, Werner A, Le AD, Soldin OP. Comparison of FT4 with log TSH on the Abbott Architect ci8200: Pediatric reference intervals for free thyroxine and thyroid-stimulating hormone. Clinica Chimica Acta. 2010;411:250–252. doi: 10.1016/j.cca.2009.11.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tractenberg RE, Jonklaas J, Soldin SJ. Agreement of Immunoassay and Tandem Mass Spectrometry in the Analysis of Cortisol and Free T4: Interpretation and Implications for Clinicians. International Journal of Analytical Chemistry. 2010;(2010):7. doi: 10.1155/2010/234808. Article ID 234808. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Surks MI, Sievert R. Drugs and thyroid function. N Engl J Med. 1995;333:1688–1694. doi: 10.1056/NEJM199512213332507. [DOI] [PubMed] [Google Scholar]
- 14.Surks MI, DeFesi CR. Normal serum free thyroid hormone concentrations in patients treated with phenytoin or carbamazepine: a paradox resolved. JAMA. 1996;275:1495–1498. [PubMed] [Google Scholar]
- 15.Harjai KJ, Licata AA. Effects of amiodarone on thyroid function. Ann Intern Med. 1997;126:63–73. doi: 10.7326/0003-4819-126-1-199701010-00009. [DOI] [PubMed] [Google Scholar]
- 16.Lazarus JH. The effects of lithium therapy on thyroid and thyrotropin-releasing hormone. Thyroid. 1998;8:909–913. doi: 10.1089/thy.1998.8.909. [DOI] [PubMed] [Google Scholar]
- 17.Gu J, Soldin OP, Soldin SJ. Simultaneous quantification of free triiodothyronine and free thyroxine by isotope dilution tandem mass spectrometry. Clin Biochem. 2007;40:1386–1391. doi: 10.1016/j.clinbiochem.2007.08.007. [DOI] [PubMed] [Google Scholar]
- 18.Wu AHB, French D. Implementation of liquid chromatography/mass spectrometry into the clinical laboratory. Clinica Chimica Acta 420. 2013;420:4–10. doi: 10.1016/j.cca.2012.10.026. [DOI] [PubMed] [Google Scholar]
- 19.Soldin OP, Jang M, Guo T, Soldin SJ. Pediatric reference intervals for Free Thyroxine and Free Triiodothyronine. Thyroid. 2009;19:669–702. doi: 10.1089/thy.2009.0037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Siemens Dimension Vista Package insert FT4 September 2009. Siemens Healthcare Diagnostics; Tarrytown, NY: [Google Scholar]
- 21.Lepoutre T, Daumerie C, Gruson D. Measurement of TSH, FT4 and FT3 with immunoassays based on LOCI technology. Endocrine Abstracts. 2012;29:1675. [Google Scholar]