|Home | About | Journals | Submit | Contact Us | Français|
The clinical value of free thyroxine (FT4) and free triiodothyronine (FT3) analysis depends on the reference intervals with which they are compared. We determined age- and sex-specific reference intervals for neonates, infants, and children 0–18 years of age for FT4 and FT3 using tandem mass spectrometry.
Reference intervals were calculated for serum FT4 (n=1426) and FT3 (n=1107) obtained from healthy children between January 1, 2008, and June 30, 2008, from Children's National Medical Center and Georgetown University Medical Center Bioanalytical Core Laboratory, Washington, DC. Serum samples were analyzed using isotope dilution liquid chromatography tandem mass spectrometry (LC/MS/MS) with deuterium-labeled internal standards.
FT4 reference intervals were very similar for males and females of all ages and ranged between 1.3 and 2.4ng/dL for children 1 to 18 years old. FT4 reference intervals for 1- to 12-month-old infants were 1.3–2.8ng/dL. These 2.5 to 97.5 percentile intervals were much tighter than reference intervals obtained using immunoassay platforms 0.48–2.78ng/dL for males and 0.85–2.09ng/dL for females. Similarly, FT3 intervals were consistent and similar for males and females and for all ages, ranging between 1.5pg/mL and approximately 6.0pg/mL for children 1 month of age to 18 years old.
This is the first study to provide pediatric reference intervals of FT4 and FT3 for children from birth to 18 years of age using LC/MS/MS. Analysis using LC/MS/MS provides more specific quantification of thyroid hormones. A comparison of the ultrafiltration tandem mass spectrometric method with equilibrium dialysis showed very good correlation.
In 1990 the American Thyroid Association (ATA) recommended the use of the sensitive thyrotropin (TSH) and free thyroxine (FT4) tests to evaluate thyroid function (1). In the case of hypothyroidism, serum TSH levels are high and FT4 levels low, while in subclinical hypothyroidism serum TSH levels are high but FT4 levels are within the normal reference interval. FT4 testing is critical when pituitary or hypothalamic disease are known or suspected.
FT4, the unbound form of the thyroid hormone thyroxine (T4), is representative of thyroid status. Therefore, serum FT4 levels reflect the health of the thyroid gland and can assist in thyroid disease diagnosis (2). During fetal development, thyroid hormones are important in brain cell migration as well as differentiation of neurons, oligodendrocytes, astrocytes, and microglia (3). Therefore, adequate levels of maternal T4 are important for appropriate fetal neurodevelopment (4–9). Neonatal screening can help in the timely diagnosis of newborn thyroid diseases such as congenital hypothyroidism and thus prevent developmental or growth problems (10).
Approximately 20% of triiodothyronine (T3), the more active thyroid hormone, is synthesized by the thyroid gland, while 80% is formed in the peripheral tissues from the 5′-deiodination of T4. About 99.9% of T3 is normally found in the circulation bound to transport proteins. T3 has a lower affinity for binding to transport proteins and globulins than T4 (11).
Population-based reference intervals are a tool for interpretation of individual patient laboratory test results. The clinical value of FT4 results depends crucially on the reference intervals with which they are compared. Accuracy of FT4 and FT3 measurements by the direct analogue immunoassay methods has been questioned by many investigators because FT4 measurements are still vulnerable to method-dependent artifacts, particularly in some specific populations (12–17). For this reason we elected to assess the pediatric reference intervals using a specific and robust ultrafiltration tandem mass spectrometry (MS/MS) procedure developed in our laboratories (15,17).
Blood samples were obtained from Children's National Medical Center's (CNMC), Washington, DC, healthy outpatient population from January 2008 to June 2008. All personal identifiers except age and sex were removed and hormone profiles were performed on plasma or serum samples. Institutional Review Board approval was obtained. Serum TSH levels were within the normal reference intervals for all children (testing performed at CNMC; data not presented). All samples were kept at −80°C until analysis.
3,3′,5-Triido-l-thyronine (T3) and l-thyroxine (T4) were purchased from Sigma (St. Louis, MO). Deuterium-labeled T4 (T4-d5): T4-[tyrosine 2H5] was obtained from IsoSciences (King of Prussia, PA) and served as the internal standard. High performance liquid chromatography (HPLC) grade methanol and ammonium hydroxide were purchased from Fisher Scientific (Fair Lawn, NJ). HPLC grade glacial acetic acid was obtained from JT Baker (Phillipsburg, NJ). Acetonitrile was purchased from Burdick and Jackson (Muskegon, MI). Distilled de-ionized water was prepared from Millipore RiOs and Milli-Q Gradient ultrapure water system (Billerica, MA). Standards for the calibration curve in the range of 0.1–2.5ng/dL (1.5–38.5pmol/L) for FT3 and 0.5–5ng/dL (6.3–63.0pmol/L) for FT4 were prepared by adding both T3 and T4 spiking solutions to distilled de-ionized water. In-house quality control solutions at three concentrations (low, medium, high) were prepared in the same way to evaluate within-day and between-day precision as well as the method accuracy. A diluted solution of 2.5ng/dL (31.5pmol/L) T4-d5 in methanol was used as working internal standard.
Thyroid hormones were analyzed by methods developed in our laboratories and described in detail previously (15,17). Briefly, 400μL of serum or plasma were centrifuged using 30kDa cut-off ultrafiltration devices (Millipore, Bedford,MA) to yield ultrafiltrates (200μL) prepared at either 25°C or 37°C. One hundred fifty microliters of ultrafiltrate was treated with 450μL methanol containing deuterium-labeled internal standards. The solution was vortexed and centrifuged, 500μL of the supernatant was treated with 400μL of water, and 650μL was injected onto the C-18 column (Supelco LC-18-DB [3.3cm×3.0mm, 3.0μm ID]). The thyroid hormones were quantified in the negative mode employing multiple reaction mode monitoring on the API-5000 tandem mass spectrometer (Applied Biosystems/MDS Sciex, Foster City, CA) equipped with TurboIonSpray source was operated in the negative ion multiple reaction monitoring mode to perform the analysis. The HPLC system consisted of three Shimadzu LC-20AD pumps, a Shimadzu SIL-HTA autosampler, and a Shimadzu DGU-20A5 degasser (Shimadzu Scientific Instruments, Columbia, MD). The liquid chromatography tandem mass spectrometry (LC/MS/MS) procedure involves an online extraction/cleaning of the injected samples followed by an activation of a built-in Valco switching valve for subsequent sample introduction into the mass spectrometer.
Ultrafiltration was performed on all samples at 25°C. For comparison purposes the process was repeated on 28 samples at 37°C. This provided a factor allowing conversion of reference intervals derived at 25°C to those at 37°C. Results and comparisons are provided in Figs. 1 and and22.
Precision was evaluated by analyzing in-house quality control samples at three different concentration levels in duplicates for both FT3 and FT4 both within and between-day. Accuracy was evaluated by performing recovery studies. The between-day precision was measured on 20 different days for FT3 and 30 different days for FT4.
The 2.5 and 97.5 percentiles were determined and reference intervals were established for all analytes on this cohort of healthy outpatient pediatric population. Linear regression analysis was performed by using GraphPad Prism version 3.02 (GraphPad Software, San Diego, CA) (Figs. 1–3).
Reference intervals for FT4 and FT3 at both 25°C and 37°C are shown in Tables 1 and and2.2. Results obtained for 28 samples comparing ultrafiltration at 25°C vs. 37°C are shown in Figs. 1 and and2.2. Comparison of the ultrafiltration MS/MS procedure with equilibrium dialysis MS/MS procedure is shown in Fig. 3 (both performed at 37°C).
A narrow range of reference intervals was found for FT4 for both males and females. The values did not change according to sex. In the first year of life, the upper bound for FT4 is slightly higher than FT4 levels for children of both sexes older than 1 year of age (Table 1). A narrow range of reference intervals was found for FT4 for both males and females at both ultrafiltration temperatures 25°C and 37°C.
The lower bounds for FT3 levels were very consistent among males and females and did not fluctuate with age. The upper bounds of the reference intervals, however, show a slight increase of value in the female group 8 to <13 years and male group 8 to <12 years, as well as males 16 to <18 years (Table 2).
The upper FT4 value reached a maximum of 1.9ng/dL when ultrafiltration was performed at 25°C (2.8ng/dL if performed at 37°C) in the 1 month to <1 year group of both sexes, and decreased to 1.6ng/dL (2.4ng/dL at 37°C) in the 1- to 3-year group of both sexes, and then stabilized to a value of 1.6ng/dL (2.4ng/dL at 37°C) for both males and females. The lower values for all ages were slightly higher than the published values conducted by immunoassays (18–20). The initial decrease is consistent with the published literature. However, in the published literature, the upper ranges for free T4 continue to decrease with age, as opposed to our values, which do not continue to decrease with increasing age, but rather stay consistent after age 3 (18–20). Our results do not indicate a sex difference in the values of serum FT4, which is inconsistent with previously published values (18,19).
Free T3 lower reference ranges were very consistent with no patterns regarding sex or age. This consistency contrasts with the published literature values, which show an increase in the lower bound values with increasing age, and a decrease in the upper bound values after the first month of life (20).
The temperature at which ultrafiltration is performed merits some discussion. When our ultrafiltration MS/MS method was first developed it was compared to the then gold standard Nichols immunoassay kit in which equilibrium dialysis was performed at 37°C. We found a slope of 1 with an excellent correlation with the Nichols approach only if the ultrafiltration was performed at 25°C (17). In a small study, use of 37°C for ultrafiltration increased values of FT4 by approximately a factor of 1.5 (17). Once the Nichols kit was discontinued large reference laboratories were forced to use a different dialysis membrane and their results at 37°C were also found to be approximately 1.5 times higher than the classical Nichols equilibrium dialysis immunoassay method performed at 37°C. In a larger study reported in this paper we have now found that switching the ultrafiltration from 25°C to 37°C increased our results by a similar factor of 1.5 (Figs. 1 and and2)2) for both FT4 and FT3.
In this study we have also compared the ultrafiltration MS/MS method with an equilibrium dialysis MS/MS method. Results of this comparison at 37°C are shown in Fig. 3 and show good agreement.
Dr. O.P. Soldin is partially supported by grants 5U10HD047890-03 NIH/NICHD Obstetrics-Fetal Pharmacology Research Unit Network (OPRU) and by the Office of Research on Women's Health. Dr. S.J. Soldin is partially supported by NIH GCRC Grant M01-RR-020359. Ms. M. Jang was a Colaco scholarship summer student with Drs. Soldin.
No competing financial interests exist.