Establishment of Reference Intervals for Thyroid- Stimulating Hormone and Free Thyroxine in Amniotic Fluid Using the Bayer ADVIA Centaur

Transcription

1 Clinical Chemistry / THYROID HORMONES IN AMNIOTIC FLUID Establishment of Reference Intervals for Thyroid- Stimulating Hormone and Free Thyroxine in Amniotic Fluid Using the Bayer ADVIA Centaur Nikola A. Baumann, PhD *, and Ann M. Gronowski, PhD Key Words: Reference intervals; Thyroid hormone; Amniotic fluid DOI: 1.139/69A5AV266W23AUA Abstract: Thyroid hormone concentrations in amniotic fluid (AF) can aid in the diagnosis of fetal hypothyroidism. However, the availability of reference intervals for thyroid hormones in AF using current methods is limited. The purpose of this study was to validate the Bayer ADVIA Centaur (Bayer Healthcare, Tarrytown, NY) third-generation thyroid-stimulating hormone (TSH), total thyroxine (T 4 ), and free thyroxine ( ) assays for use with AF and establish reference intervals for these thyroid hormones in AF. Third-trimester AF samples were assayed for TSH, T 4, and. Reference intervals were calculated using nonparametric analyses. The reference intervals were as follows: TSH (n = 131),.4 to.51 µiu/ml ( miu/l), with a median of.1 µiu/ml (.1 miu/l); (n = 133), less than.1 to.77 ng/dl ( pmol/l), with a median of.26 ng/dl (3.35 pmol/l). T 4 in AF was undetectable using the Centaur assay, reinforcing the importance of validating different manufacturers immunoassays for use with nonserum specimens. These reference intervals represent the first study performed to date using a third-generation TSH assay and a sensitive assay. When thyroid disease arises during pregnancy, the mother can be evaluated by using maternal serum. Evaluation of the fetus for suspected thyroid abnormalities is far more difficult. Proper fetal diagnosis is especially important because thyroid disorders can lead to growth retardation and possible intrauterine death. Approximately 3% of fetal thyroid disorders are associated with goiter, 1 which can cause neck dystocia and respiratory distress. Evaluating thyroid abnormalities in the fetus is clinically difficult; yet, because in utero treatment is available, proper diagnosis is important. Although a fetal goiter can be distinguished during ultrasound examination, ultrasound does not differentiate between hyperthyroidism and hypothyroidism. For knowledge of specific thyroid derangements, direct sampling of cord blood in utero can be performed, but percutaneous umbilical blood sampling is a technically demanding procedure that poses risk of fetal bradycardia, hemorrhage, and possible fetal death. 2 Although the risk of fetal death is low, reported as.5% to 1% or less, 3 amniocentesis is a far simpler and safer procedure, with less risk to the fetus. Studies indicate that from the 12th week of gestation onward, there are steady increases in the fetal serum concentrations of total thyroxine (T 4 ), free thyroxine ( ), thyroidstimulating hormone (TSH), and thyroxine-binding globulin. 2,4 The increasing concentrations of fetal cord blood thyroid hormones correlate with increasing concentrations of amniotic fluid TSH and. 5 Furthermore, these increases are independent of maternal concentrations of the same hormones, indicating that there is no significant transplacental passage of TSH and thyroxine. 2,5 A previous study from this laboratory established reference intervals for thyroid hormones in amniotic fluid using the 158 Am J Clin Pathol 27;128: Downloaded 158 from DOI: 1.139/69A5AV266W23AUA

2 Clinical Chemistry / ORIGINAL ARTICLE Abbott AxSym method for TSH, total T 4, and. 6 In that study, clinical case reports from the literature were also examined, and in all cases of fetal hypothyroidism, the amniotic fluid TSH and concentrations were outside the newly established reference intervals. The data supported the use of thyroid hormone concentrations in amniotic fluid as an aid in the diagnosis of fetal hypothyroidism. However, that study was limited because it used a secondgeneration TSH assay. Today, more sensitive third-generation TSH assays and more sensitive assays are routinely used. Furthermore, it is important to establish and/or validate reference intervals when changing assay methods or when assessing different patient populations. Recent studies comparing automated TSH methods have shown that TSH methods do not provide comparable results across the analytic measurement range. 7 The present study expands on previous studies, establishing amniotic fluid reference intervals for TSH, T 4, and using more sensitive assays on the ADVIA Centaur automated immunoassay analyzer (Bayer Healthcare, Tarrytown, NY). Materials and Methods Samples Leftover amniotic fluid samples sent to the Barnes- Jewish Hospital chemistry laboratory (St Louis, MO) for fetal lung maturity testing from July 22 through November 25 and stored at 2 C were used. All samples were visually inspected for the presence of blood. Amniotic fluid specimens with visible blood contamination were excluded from the study. The study was approved by the Washington University Human Studies Committee (St Louis). Validation of Assays for Use With Amniotic Fluid The ADVIA Centaur automated immunoassay analyzer was used for all thyroid hormone measurements. The ADVIA TSH assay is a sandwich immunoassay with an analytic sensitivity of.2 µiu/ml (.2 miu/l). The ADVIA total T 4 assay is a competitive immunoassay with an analytic sensitivity of.3 µg/dl (3.87 nmol/l). The ADVIA assay is a competitive immunoassay with an analytic sensitivity of.1 ng/dl (1.29 pmol/l). This assay cannot be diluted, and our laboratory has established linearity in plasma to 4.5 ng/dl (58.1 pmol/l). Validation of the ADVIA Centaur thyroid hormone assays using amniotic fluid as a matrix was performed. Serum (2% vol/vol) with a known concentration of TSH, total T 4, or was added to amniotic fluid and then serially diluted using amniotic fluid with a low concentration of TSH, total T 4, or as the diluent. Samples were analyzed in duplicate, and recovery was calculated. As expected, because of binding protein equilibrium, cannot be accurately recovered from diluted specimens. Precision Studies Intra-assay precision for the TSH and assays was determined by measuring amniotic fluid samples 1 consecutive times. Precision was established at 2 concentrations of TSH and. Analyte Stability Studies The effect of the freeze-thaw process on the stability of TSH and was determined by measuring the hormones in 9 specimens before (within 24 hours of collection) and after freezing (24-48 hours) at 2 C. Results obtained before and after freezing were analyzed by using the t test. A P value of less than.5 was considered statistically significant. Calculation of Reference Intervals Thorough chart review for all cases was performed, and samples were excluded if there was a patient history of fetal anomalies, multiple gestations, or any evidence of thyroid disease. Gestational ages ranged from 31.3 to 39.4 weeks (median, 36 weeks). The level of TSH was measured in 131 specimens and of in 133 specimens. Central 95% reference intervals for TSH and were calculated using nonparametric analysis. 8 This process ranks data according to numeric value and calculates percentiles as a function of these ranked numbers. The concentrations at the 2.5 and 97.5 percentiles provide an estimate of the central 95% reference interval for the population based on the given set of data. Results The ADVIA Centaur TSH assay was validated for use with amniotic fluid. Recovery of serum TSH diluted in amniotic fluid was measured at 12 concentrations in the range between.8 µiu/ml (.8 miu/l) and 37 µiu/ml (37 miu/l) Figure 1A. The TSH concentrations (µiu/ml; to convert to miu/l, multiply by 1.) and percentages of recovery were as follows, respectively:.8 and 13%;.1 and 111%;.13 and 127%;.21 and 11%;.35 and 19%;.64 and 18%; 1.22 and 88%; 2.37 and 86%; 4.68 and 78%; 9.3 and 84%; 18.5 and 94%; and 37. and 89%. The mean ± SD recovery of serum TSH was 99% ± 15%. Recovery of serum T 4 when diluted in amniotic fluid is shown in Figure 1B. The T 4 concentrations (µg/dl; to convert to nmol/l, multiply by 12.9) and percentages of recovery were as follows, respectively:.9 and 182%; 1.8 and 164%, Downloaded from Am J Clin Pathol 27;128: DOI: 1.139/69A5AV266W23AUA 159

3 Baumann and Gronowski / THYROID HORMONES IN AMNIOTIC FLUID A B Observed TSH (µiu/ml) Observed T 4 (µg/dl) Expected TSH (µiu/ml) Expected T 4 (µg/dl) Figure 1 Recovery of serum thyroid-stimulating hormone (TSH; A) and total thyroxine (T 4 ; B) in amniotic fluid. Serum (2% vol/vol) with a known concentration of TSH or T 4 was added to amniotic fluid and then serially diluted using amniotic fluid as the diluent. The dashed line represents a slope of 1. A, Slope =.99; y-intercept =.12. Inset graph shows recovery in the low TSH concentration range. Samples were analyzed in duplicate. B, Slope = 1.518; y-intercept =.22. Values are given in conventional units; to convert to Système International units, for TSH (miu/l), multiply by 1.; for T 4 (nmol/l), multiply by and 158%; 3.5 and 159%; and 4.3 and 158%. The mean ± SD recovery of serum T 4 was 164% ± 1%. The recovery data demonstrated overrecovery of T 4, suggesting that the amniotic fluid matrix may affect the measurement of serum T 4 with this assay. in serum (2% vol/vol) was diluted into amniotic fluid, and recovery was measured using the ADVIA Centaur assay. recovery was 195% in this sample. This indicates that serum can be recovered and measured in amniotic fluid using the ADVIA Centaur assay. However, as expected, detailed recovery studies could not be performed because cannot be diluted and accurately measured owing to the effect of dilution on binding protein equilibrium. 9 Intra-assay precision studies were performed on patient amniotic fluid samples. The coefficients of variation (CVs) for the TSH and assays were less than 1%. The intra-assay CVs for the TSH assay in 1 samples were 9.% and 4.1% at TSH concentrations of.6 µiu/ml (.6 miu/l) and.11 µiu/ml (.11 miu/l), respectively. The assay had intraassay CVs of 9.6% and 9.8% at concentrations of.34 ng/dl (4.39 pmol/l) and.52 ng/dl (6.71 pmol/l), respectively (n = 1). It is interesting that we found that T 4 could not be detected in any patient amniotic fluid samples using the ADVIA Centaur method. This did not seem to be simply a matrix effect because serum T 4 could be diluted into amniotic fluid and measured (Figure 1B). This inability to measure amniotic fluid T 4 was further confirmed with 1 amniotic fluid samples that were measured using the ADVIA Centaur and the Abbott AxSym (Abbott Laboratories, Abbott Park, IL) instruments. Amniotic fluid samples analyzed using the Abbott AxSym total T 4 assay gave results between 1.8 and 2.9 µg/dl ( nmol/l), whereas all samples had undetectable T 4 (<1. µg/dl [12.9 nmol/l]) by the ADVIA Centaur assay. Amniotic fluid specimens were stored frozen ( 2 C) before analysis. Thus, it was necessary to determine the stability of TSH and after specimens were subjected to a freeze-thaw cycle. We analyzed 9 amniotic fluid specimens before (within 24 hours of collection) and after freezing (24-48 hours) at 2C. A single freeze-thaw cycle did not have a significant effect on thyroid hormone concentrations in amniotic fluid. Postfreeze TSH and results were within a mean ± SD of 21% ± 27% (P =.22) and 11% ± 21% (P =.8) of the original result, respectively Figure 2. Samples from 133 patients met selection criteria (see the Materials and Methods section) and were subsequently thawed and assayed immediately for TSH and (not all samples could have both assays performed owing to insufficient sample volume). The frequencies of amniotic fluid TSH and results are shown in Figure 3. Central 95% reference intervals, median values, and ranges for TSH and were calculated and are given in Table Am J Clin Pathol 27;128: Downloaded 16 from DOI: 1.139/69A5AV266W23AUA

4 Clinical Chemistry / ORIGINAL ARTICLE A.5 B TSH (µiu/ml).3.2 (ng/dl) TSH Prefreeze TSH Postfreeze. Prefreeze Postfreeze Figure 2 Stability of thyroid-stimulating hormone (TSH; A) and free thyroxine ( ; B) in amniotic fluid. Amniotic fluid specimens were assayed for TSH (A) and (B) before (within 24 hours of collection) and after freezing (24-48 hours) at 2 C. Values are given in conventional units; to convert to Système International units, for TSH (miu/l), multiply by 1.; for free T 4 (pmol/l), multiply by A 4 B Frequency 2 1 Frequency TSH (µiu/ml) (ng/dl) Figure 3 Frequencies of amniotic fluid thyroid hormone results. A, Thyroid-stimulating hormone (TSH). B, Free thyroxine ( ). Values are given in conventional units; to convert to Système International units, for TSH (miu/l), multiply by 1.; for free T 4 (pmol/l), multiply by Table 1 Reference Intervals for TSH and Free T 4 Calculated Using Nonparametric Analysis * No. of Samples Median Range Central 95% Reference Interval TSH (µiu/ml) Free T 4 (ng/dl) < < TSH, thyroid-stimulating hormone; T 4, thyroxine. * Values are given in conventional units; to convert to Système International units, for TSH (miu/l), multiply by 1.; for free T 4 (pmol/l), multiply by Discussion We established normal amniotic fluid reference intervals for TSH and. To our knowledge, this is the first study that has established these reference intervals using the ADVIA Centaur automated immunoassay analyzer. Our reference intervals for TSH and compare favorably with those in previous third-trimester studies. 5,6,1-12 An interesting and unexpected result from this study was the observation that amniotic fluid T 4 cannot be measured using the ADVIA Centaur. This does not seem to be simply a Downloaded from Am J Clin Pathol 27;128: DOI: 1.139/69A5AV266W23AUA 161

5 Baumann and Gronowski / THYROID HORMONES IN AMNIOTIC FLUID matrix effect of the fluid because serum with a known concentration of total T 4 can be recovered when diluted into amniotic fluid (recovery approximately 16%). Furthermore, 1 samples with undetectable total T 4 concentrations using the ADVIA Centaur had detectable T 4 concentrations using the Abbott AxSym analyzer. This observation could be the result of failure of the protein binding inhibitor used in the Centaur assay to release T 4 bound to proteins in amniotic fluid. This observation highlights the potential obstacles in adapting assays for use with a novel matrix such as amniotic fluid. It also emphasizes the importance of validating reference intervals before transferring them between platforms and methods. In addition, because of the inability to dilute specimens and perform in-depth validation experiments and the lack of information available on T 4 binding capacity in amniotic fluid or available information on amniotic fluid binding proteins, we believe that TSH is the preferred analyte for assessment of thyroid status in amniotic fluid when using the ADVIA Centaur immunoassay platform. Total T 4 has been shown to be a useful analyte using other manufacturer s assays; however, all assays must be validated before specimens are analyzed for diagnostic purposes. It is also possible that the ADVIA Centaur T 4 assay does not recognize sulfated T 4, which is a metabolite of T 4 in neonates and amniotic fluid. 13,14 This finding reinforces the importance of validating different manufacturers immunoassays for use with nonserum specimens. The samples used in this study were obtained from a population of women undergoing amniocentesis for fetal lung maturity analysis and do not necessarily represent a normal healthy population. Although this is important, we believe that the reference intervals are still useful as supported by a previous study. 6 Six specimens from women with known thyroid anomalies were excluded. All 6 of these women had amniotic fluid TSH and concentrations within our established reference intervals and, to the best of our knowledge, delivered normal infants. This supports the idea that amniotic fluid thyroid hormone concentrations reflect the thyroid status of the fetus and are independent of the thyroid status of the mother. A potential limitation of the present study is that amniotic fluid specimens were stored frozen for up to 3 years before analysis. However, we demonstrated that TSH and were not significantly affected by a freeze-thaw cycle (Figure 2). Previous studies 15 have also reported that TSH and thyroxine are stable in serum for up to 5 years at 2 C, lending support to the idea that these analytes are stable in specimens stored frozen for extended periods. The usefulness of thyroid hormone measurement in amniotic fluid lies in its use not only for diagnosis but also for management. Many authors have treated cases of fetal hypothyroid goiter with intrauterine thyroxine, using ultrasound examination to show regression of fetal goiter and amniotic fluid to show decreases in amniotic fluid TSH and/or increases in amniotic fluid T 4 levels. Application of our amniotic fluid reference intervals may provide a more relevant end point for therapy than ultrasound monitoring because the regression of goiter does not necessarily confer a euthyroid state. In addition, amniotic fluid is a simple sample to obtain in this setting because it is readily accessible when intrauterine thyroxine injections are performed. Fetal hyperthyroidism is currently diagnosed through findings of fetal goiter, tachycardia, advanced bone maturation, and abnormal cord serum testing. The usefulness of the amniotic fluid thyroid hormone reference ranges for diagnosing hyperthyroidism is unclear. We know of no published case reports of fetal hyperthyroidism in which amniotic fluid thyroid tests were performed. It is not possible to know how useful amniotic fluid TSH and measurements are in diagnosing and monitoring fetal hyperthyroidism until several cases of fetal hyperthyroidism are reported. We established reference intervals for TSH and in third-trimester amniotic fluid by using the ADVIA Centaur. This work expands on previous studies demonstrating that reference intervals for thyroid hormones (especially TSH) in amniotic fluid can be of significant use in the diagnosis and management of fetal hypothyroidism, reducing the need for umbilical blood sampling. From the Department of Pathology and Immunology, Division of Laboratory Medicine, Washington University School of Medicine St Louis, MO, Supported in part by Bayer Healthcare by supplying reagents. Presented in part at the American Association for Clinical Chemistry Annual Meeting; July 25; Orlando, FL. Address reprint requests to Dr Baumann: Dept of Pathology, University of Illinois Medical Center at Chicago, 84 S Wood St, Room 21-G, Chicago, IL Acknowledgment: We thank Mark H. Wener, MD, University of Washington, Seattle, for testing samples using the Abbott AxSym instrument. * Dr Baumann is currently with the Department of Pathology, University of Illinois Medical Center at Chicago. References 1. Volumenie JL, Polak M, Guibourdenche J, et al. Management of fetal thyroid goiters: a report of 11 cases in a single perinatal unit. Prenat Diagn. 2;2: Thorpe-Beeston JG, Nicolaides KH, McGregor AM. Fetal thyroid function. Thyroid. 1992;2: Daffos F. Fetal blood sampling. Annu Rev Med. 1989;4: Ballabio M, Nicolini U, Jowett T, et al. Maturation of thyroid function in normal human foetuses. Clin Endocrinol (Oxf). 1989;31: Yoshida K, Sakurada T, Takahashi T, et al. Measurement of TSH in human amniotic fluid: diagnosis of fetal thyroid abnormality in utero. Clin Endocrinol (Oxf). 1986;25: Am J Clin Pathol 27;128: Downloaded 162 from DOI: 1.139/69A5AV266W23AUA

6 Clinical Chemistry / ORIGINAL ARTICLE 6. Singh PK, Parvin CA, Gronowski AM. Establishment of reference intervals for markers of fetal thyroid status in amniotic fluid. J Clin Endocrinol Metab. 23;88: Rawlins ML, Roberts WL. Performance characteristics of six third-generation assays for thyroid-stimulating hormone. Clin Chem. 24;5: Solberg HE. Establishment and use of reference values. In: Burtis CA, Ashwood ER, eds. Tietz Textbook of Clinical Chemistry. Philadelphia, PA: Saunders; ADVIA Centaur Assay Manual. Tarrytown, NY: Bayer Healthcare; Kourides IA, Heath CV, Ginsberg-Fellner F. Measurement of thyroid-stimulating hormone in human amniotic fluid. J Clin Endocrinol Metab. 1982;54: Kourides IA, Berkowitz RL, Pang S, et al. Antepartum diagnosis of goitrous hypothyroidism by fetal ultrasonography and amniotic fluid thyrotropin concentration. J Clin Endocrinol Metab. 1984;59: Hollingsworth DR, Alexander NM. Amniotic fluid concentrations of iodothyronines and thyrotropin do not reliably predict fetal thyroid status in pregnancies complicated by maternal thyroid disorders or anencephaly. J Clin Endocrinol Metab. 1983;57: Chopra IJ, Santini F, Hurd RE, et al. A radioimmunoassay for measurement of thyroxine sulfate. J Clin Endocrinol Metab. 1993;76: Polk DH. Thyroid hormone metabolism during development. Reprod Fertil Dev. 1995;7: Vasiliades J, Hilgers T, Gentrup B. Long term stability of hormones in serum [abstract]. Clin Chem. 1995;41(suppl):S Perelman AH, Johnson RL, Clemons RD, et al. Intrauterine diagnosis and treatment of fetal goitrous hypothyroidism. J Clin Endocrinol Metab. 199;71: Bruner JP, Dellinger EH. Antenatal diagnosis and treatment of fetal hypothyroidism: a report of two cases. Fetal Diagn Ther. 1997;12: Perrotin F, Sembely-Taveau C, Haddad G, et al. Prenatal diagnosis and early in utero management of fetal dyshormonogenic goiter. Eur J Obstet Gynecol Reprod Biol. 21;94: Abuhamad AZ, Fisher DA, Warsof SF, et al. Antenatal diagnosis and treatment of fetal goitrous hypothyroidism: case report and review of the literature. Ultrasound Obstet Gynecol. 1995;6: Downloaded from Am J Clin Pathol 27;128: DOI: 1.139/69A5AV266W23AUA 163