Objective
The aims were to establish a gestational-age specific curve for serum total thyroxine (T4) levels and to compare pregnancy outcomes of euthyroid women with those identified to have subclinical hypothyroidism (SCH) defined by an elevated thyroid-stimulating hormone (TSH) level in conjunction with either total T4 or free T4 determinations.
Study Design
Over a 2.5 year period, serum thyroid analytes were measured in all women presenting for prenatal care. After exclusion of women with overt thyroid disorders, the normal distribution of serum total T4 levels were determined by quantile curves for those screened in the first 20 weeks and who were delivered of a singleton infant weighing at least 500 g. Pregnancy outcomes for women with an elevated TSH and normal total T4 concentrations were analyzed and compared with those of women identified to have SCH defined by normal free T4 levels.
Results
Of 17,298 women tested, serum total T4 increased into the second trimester and plateaued around 16 weeks. The upper threshold for total T4 ranged from 12.6 to 16.4 μg/dL, and the lower threshold ranged from 5.3 to 8.0 μg/dL. Women identified to have SCH defined by serum free T4, total T4, or both were at risk for preterm delivery ( P = .007) and placental abruption ( P = .013) when compared with euthyroid women.
Conclusion
When combined with elevated TSH levels, free or total T4 determinations are equally sensitive to identify women with SCH who are at increased risk for preterm birth and placental abruption when compared with euthyroid women.
Current guidelines from the American College of Obstetricians and Gynecologists state that in women suspected to have hypo- or hyperthyroidism, an assessment of thyroid gland function during pregnancy by measurement of serum levels of thyroid stimulating hormone (TSH) and free thyroxine (T4) should be conducted. These paired results are used to classify pregnant women as being euthyroid or to have either overt or subclinical hyper- or hypothyroidism.
Pregnancy-induced alterations in thyroid-binding protein concentrations can interfere with the immunoassay for serum free T4 determinations, and this has led some to question its accuracy in pregnancy. For example, in 1 study of 2 commercially available immunoassays, free T4 levels were falsely low in almost 70% of women diagnosed with hypothyroidism. Because of these findings, a Task Force of the American Thyroid Association and the American Association of Clinical Endocrinologists recently recommended that during pregnancy, thyroid function assessment should be determined with serum TSH and total T4 instead of free T4 values.
There are several factors to consider before such recommendations are adopted. First, there are limited data regarding normative serum total T4 values across pregnancy. Several small studies indicate that total T4 levels are trimester specific and that these values increase during the first trimester and remain elevated until delivery. Although normal values for serum free T4 during pregnancy have been determined in several large investigations, there have been no similar studies to establish norms for total T4 levels.
A second consideration is that obstetrical health care providers are not accustomed to utilizing total T4 assay results to evaluate pregnant women with possible thyroid disease. Third, the American College of Obstetricians and Gynecologists has promulgated guidelines that recommend thyroid analyte screening with serum TSH followed by serum free T4 if this is abnormally high or low. Finally, most recent studies of pregnant women have utilized TSH and free T4 analytes to identify and classify thyroid disorders.
For the reasons cited in the previous text, the present study was designed to provide gestational age–specific normative values for serum total T4 in a cohort of more than 17,000 women presenting for prenatal care up to 20 weeks’ gestation. Another aim was to compare adverse pregnancy outcomes in women with subclinical hypothyroidism defined by either serum free T4 or total T4 levels.
Materials and Methods
This is a secondary analysis of a prospectively collected database initially designed to estimate normative values for thyroid hormone analytes in a large population of pregnant women attending prenatal clinics at Parkland Hospital (Dallas, TX). The protocol was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center and Parkland Hospital.
The details of this process and thyroid-analyte assays have been previously described. Briefly, excess serum obtained for prenatal serological screening from Nov. 1, 2000, to April 14, 2003, was delivered to our immunochemistry research laboratory for thyroid analyte studies. Using a chemiluminescent assay (Immulite 2000 Analyzer), we determined serum concentrations for thyrotropin (TSH), and free T4 and total T4.
In the initial phase of these studies, serum free T4 was reflexively determined in women found to have abnormally high or low serum TSH levels. Those who also had an abnormally high (TSH >3.0 mU/L) and a low serum free T4 (<0.9 ng/dL) were contacted and referred to an obstetrical complications clinic for further evaluation for possible overt thyroid dysfunction. These women were excluded from this study, whereas all other women were enrolled.
For the purposes of the current study, normal reference ranges for serum TSH and free T4 included values between the 2.5th and 97.5th percentiles. For TSH, these values were 0.03-4.13 mU/L, and for free T4 these values were 0.9-2.0 ng/dL and were established from prior studies. Pregnancy outcomes were analyzed for all women screened prior to 20 weeks’ gestation who were delivered at Parkland Hospital of a singleton neonate weighing 500 g or more. To establish normal values for serum total T4, quantile curves were calculated to establish the 2.5th and the 97.5th percentiles for each week up to 20 gestational weeks.
Pregnancy outcomes were retrieved from a computerized perinatal database that has been in use since 1983 as previously described. Briefly, abruption was defined either clinically or by pathologic diagnosis. Preterm birth was defined as gestational age of less than 37 weeks at delivery. Respiratory distress syndrome was defined when an infant required a ventilator at more than 24 hours of life. Necrotizing entercolitis included only those infants that required surgery. And intaventricular hemorrhage included only those that were deemed grade 3 or 4.
These data are routinely entered for all women at the time of delivery at Parkland Hospital, and the data are verified by research nurses. For this study, women were identified to have subclinical hypothyroidism if they had an abnormally elevated serum TSH level in combination with a normal serum free T4 or normal total T4 level. Maternal demographics and selected perinatal outcomes were compared between women who were considered euthyroid and those who were classified to have subclinical hypothyroidism either by normal serum free or total T4 levels.
The 2.5th, 25th, 50th, 75th, and 97.5th percentiles of total T4 levels by gestational age at laboratory draw were estimated using quantile regression. All patients, regardless of outcomes, are included in this estimation process, so the percentiles represent those of the population not identified as being hypothyroid. Extremes provide a statistical threshold for evaluation as possibly at-risk patients. Consequently, these estimates provided thresholds for classifying patients into euthyroid or subclinical hypothyroid using the total T4 levels.
Two-group comparisons between subclinical hypothyroid classification (by either free T4 or total T4) and those euthyroid by both free T4 and total T4 are made by using the Pearson χ 2 for categorical variables and a Student t test for continuous measures. Statistical computations were performed using SAS version 9.2 (SAS Institute, Cary, NC). Two-sided values of P < .05 were judged statistically significant.
Results
During the study period, there were a total of 17,298 women who met the criteria for inclusion. The 2.5th and 97.5th threshold serum levels for total T4 values are presented in Table 1 according to gestational age in weeks. The upper threshold for total T4 ranged from 12.6 to 16.4 μg/dL and the lower threshold ranged from 5.3 to 8.0 μg/dL between 6 and 20 weeks’ gestation. The median total T4 value peaked at 16 weeks’ gestation. A quantile regression curve was created according to gestational age, and the 2.5th and 97.5th percentiles are represented in Figure 1 . This figure shows that total T4 values slowly rise into the second trimester during which these values peaked and remained elevated at midpregnancy.
| Gestational age, wks | 2.5th percentile | 50th percentile | 97.5th percentile |
|---|---|---|---|
| 6 | 5.27 | 8.79 | 12.64 |
| 7 | 5.61 | 9.32 | 13.50 |
| 8 | 5.95 | 9.85 | 14.36 |
| 9 | 6.29 | 10.37 | 15.20 |
| 10 | 6.58 | 10.83 | 15.90 |
| 11 | 6.77 | 11.19 | 16.33 |
| 12 | 6.83 | 11.40 | 16.40 |
| 13 | 6.84 | 11.50 | 16.23 |
| 14 | 6.87 | 11.55 | 15.99 |
| 15 | 6.99 | 11.58 | 15.81 |
| 16 | 7.18 | 11.59 | 15.69 |
| 17 | 7.39 | 11.57 | 15.60 |
| 18 | 7.60 | 11.50 | 15.50 |
| 19 | 7.80 | 11.40 | 15.40 |
| 20 | 8.00 | 11.30 | 15.30 |
In Table 2 , selected maternal demographic characteristics of pregnant women identified to have subclinical hypothyroidism (SCH) by either free T4 or total T4 assay values are shown, and each is compared with euthyroid women. Women with SCH diagnosed by either thyroxine analyte were more likely to be older and have a higher body mass index compared with the euthyroid cohort. Pregnancy outcomes within the 3 groups are presented in Table 3 . Women with SCH defined by either T4 analyte were at higher risk for suffering placental abruption or delivery at a gestation of 34 weeks or less. There were no significant differences in neonatal outcomes between these groups as shown in Table 4 .
| Demographic | SCH by free T4 (n = 404), % | P value a | Euthryoid (n = 15,689), % | P value b | SCH by total T4 (n = 367), % |
|---|---|---|---|---|---|
| Age, y | 26.9 ± 5.9 | < .001 | 25.5 ± 5.6 | < .001 | 26.8 ± 6.0 |
| ≥35 | 44 (11) | .008 | 1161 (1) | .007 | 41 (11) |
| Ethnicity | < .001 | < .001 | |||
| Hispanic | 341 (84) | 13,471 (86) | 308 (84) | ||
| African American | 27 (7) | 1588 (10) | 26 (7) | ||
| White | 16 (4) | 321 (2) | 13 (4) | ||
| Other | 20 (5) | 309 (2) | 20 (5) | ||
| Nulliparity | 145 (36) | .916 | 5671 (36) | .801 | 135 (37) |
| Body mass index, kg/m 2 | 32.1 ± 6.3 | .163 | 31.7 ± 5.5 | .280 | 32.1 ± 6.4 |
a P value when comparing SCH by free T4 in comparison with euthyroid women
b P value when comparing SCH by total T4 in comparison with euthyroid women.
| Demographic | SCH by free T4 (n = 404), % | P value a | Euthryoid (n = 15,689), % | P value b | SCH by total T4 (n = 367), % |
|---|---|---|---|---|---|
| Hypertension | |||||
| Gestational | 41 (10) | .397 | 1401 (9) | .345 | 38 (10) |
| Severe | 23 (6) | .774 | 842 (5) | .765 | 21 (6) |
| Placental abruption | 4 (1) | .026 | 52 (0) | .015 | 4 (1) |
| Weeks gestation at delivery | |||||
| <37 | 27 (7) | .390 | 891 (6) | .251 | 17 (5) |
| ≤34 | 18 (4) | .011 | 385 (2) | .008 | 17 (5) |
| ≤32 | 10 (2) | .068 | 218 (1) | .088 | 9 (2) |
| Cesarean delivery | 108 (27) | .317 | 3853 (25) | .707 | 87 (24) |
| Repeat | 59 (15) | .156 | 1923 (12) | .528 | 49 (13) |
| Primary | |||||
| Dystocia | 16 (4) | .566 | 716 (5) | .853 | 16 (4) |
| Fetal distress | 17 (4) | .933 | 647 (4) | .629 | 17 (5) |
| Other | 16 (4) | .713 | 567 (4) | .632 | 15 (4) |
a P value when comparing SCH by free T4 in comparison with euthyroid women
b P value when comparing SCH by total T4 in comparison with euthyroid women.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
