einstein. 2007; 5(1):51-55
51Normal values of thyroid-stimulating hormone and free thyroxin in pregnant women
Objective: To prepare a chart for normal values of thyroid-stimulating
hormone (TSH) and free thyroxin (free T4) and to verify the correlation
between them, in pregnant women. Methods: A prospective cross-
sectional study was performed on 850 pregnant women of the
Prenatal Outpatient Clinic of the Hospital do Servidor Público Estadual
– Francisco Morato de Oliveira (HSPE – FMO), from August 2003 to
August 2005. Serum TSH and free T4 levels were determined using
microparticle enzyme immunoassay (MEIA). Maternal age ranged from
18 to 47 years, mean of 32.5 ± 6.9 years. Mean gestational age was
19.1 ± 8.5 weeks. With linear regression analysis, the 50th and 97.5th
percentiles for TSH and the 2.5th, 50th and 97.5th percentiles for free T4,
at weeks 6 to 42 were calculated. The possible correlation between
these two indicators was analyzed. results: Mean TSH and free T4
values were 2.14 ± 1.51 µIU/ml and 0.98 ± 0.30 ng/dl, respectively.
The correlation between free T4 and TSH levels was positive and
statistically significant according to the Pearson’s correlation coefficient
(r = 0.069, p = 0.044). Conclusion: Identifying the normal TSH and
free T4 limits during pregnancy is the first step to safely determine the
real need for thyroid dysfunction treatment.
Keywords: Thyroid-stimulating hormone; Thyroxine; Pregnancy;
Objetivo: Construir um gráfico de normalidade para valores de
hormônio tireoestimulante (TSH) e de tiroxina livre (T4 livre) em
gestantes e verificar a possível correlação entre os valores de ambas.
Métodos: Estudo prospectivo transversal envolvendo 850 gestantes do
ambulatório de pré-natal do HSPE no período de agosto de 2003 a agosto
de 2005. A determinação dos valores séricos de TSH e T4 livre foram
realizadas pelo ensaio imunoenzimático de micropartículas (MEIA). A
idade materna variou de 18 a 47 anos, com média de 32,5 ± 6,9 anos.
A idade gestacional apresentou média de 19,1 ± 8,5 semanas. Foram
calculados, por meio de regressão linear, os percentis 50 e 97,5 para
os valores de TSH e os percentis 2,5, 50 e 97,5 para os de T4 livre,
entre 6 e 42 semanas. Analisou-se a possível correlação entre esses
dois indicadores. Foi considerado risco α (alfa) menor ou igual a 5% e
intervalo de confiança de 95%. resultados: Os valores médios de TSH
e T4 livre encontrados foram, respectivamente, 2,14 ± 1,51 µUI/ml e
0,98 ± 0,30 ng/dl. Por meio de regressão linear, ajustou-se uma reta
para os valores de TSH e T4 livre em função da idade gestacional. O
modelo foi significativo para os valores de TSH e T4 livre (p < 0,001 e
p = 0,003, respectivamente). Foram comparados os valores de T4 livre
e TSH pela medida de correlação de Pearson indicando uma correlação
positiva entre T4 livre e TSH (r = 0,069; p = 0,044). Conclusão: A
identificação dos limites de normalidade para o TSH e o T4 livre ao
longo da gestação é o primeiro passo para aquilatar a presença real
de afecções tireoidianas.
Descritores: Hormônio tireoestimulante; Tiroxina; Gravidez; Valores
Methodological advances have allowed the widespread
routine measurement of free T4 (the biologically active
fraction) and TSH(1-2).
The diagnosis and follow-up of thyroid disorders
during pregnancy are complex due to the physiological
changes occurring in pregnancy and which interfere in
thyroid physiology, as well as in the tests used to assess
it(1,3-5). Some studies show the relation of neurological
changes in children of women displaying subclinical
hypothyroidism during pregnancy(1,6-8). The definition
of subclinical hypothyroidism is not consensual in the
Normal values of thyroid-stimulating hormone and free
thyroxin in pregnant women
Valores de normalidade do hormônio tireoestimulante e da tiroxina livre em gestantes
Marcelo Antonini1, Silvio Martinelli2, Ana Paula Marques Fernandes Yoshizumi3, Sidney Antonio Lagrosa
Garcia4, Umberto Gazi Lippi5
Study carried out at Serviço de Ginecologia e Obstetrícia do Hospital do Servidor Público Estadual “Francisco Morato de Oliveira” – HSPE, São Paulo (SP), Brazil.
1 MD. Specialist, Hospital do Servidor Público Estadual “Francisco Morato de Oliveira” – HSPE, São Paulo (SP), Brazil.
2 Assistant Physician in charge of the Out patient’s Prenatal Clinic for Late Pregnancy, Hospital do Servidor Público Estadual “Francisco Morato de Oliveira” – HSPE, São Paulo (SP), Brazil.
3 Assistant Physician, Department of Gynecology and Obstetrics, Hospital do Servidor Público Estadual “Francisco Morato de Oliveira” – HSPE, São Paulo (SP), Brazil.
4 Head of the Department of Obstetrics, Hospital do Servidor Público Estadual “Francisco Morato de Oliveira” – HSPE, São Paulo (SP), Brazil.
5 Full Professor of Obstetrics of the Medical Course, Faculdade de Ciências Médicas da Universidade Metropolitana de Santos – UNIMES; Director of the Gynecology and Obstetrics Service, Hospital do
Servidor Público Estadual “Francisco Morato de Oliveira” – HSPE, São Paulo (SP), Brazil.
Corresponding author: Marcelo Antonni – Rua João Ramalho, 1074 – ap. 71 – Perdizes – CEP 05008-002 – São Paulo (SP), Brasil – Tel.: 11 3672-1877 – e-mail: email@example.com
Received on Jul 14, 2006 – Accepted on Oct 18, 2006
einstein. 2007; 5(1):51-55
52Antonini M, Martinelli S, Yoshizumi APMF, Garcia SAL, Lippi UG
literature, with TSH values ranging from 4.0 to 10.0.
Therefore, studies establishing the TSH and free T4
levels associated to higher perinatal risk and what are the
mid- and long-term repercussion are still necessary. In
order to answer these questions it is mandatory to know
the profile of the population, since several factors may
interfere in the free T4 and TSH levels.
The objective of this study was to prepare a chart of
normal values of the thyroid stimulating hormone (TSH)
and of free thyroxin (free T4) in normal pregnant women
of the Hospital do Servidor Público Estadual Francisco
Morato de Oliveira – HSPE, São Paulo and to verify the
correlation between these two measurements.
This is a prospective cross-sectional study involving 850
pregnant women attending the Prenatal Outpatient
Clinic of HSPE from August 2003 to August 2005.
The inclusion criteria were: medically and obstetrically
normal pregnant women; gestational age, based on the
date of the last menstrual period (LMP) and confirmed
by ultrasonography, up to 20 weeks; and maternal age
greater than 18 years.
The exclusion criteria were: patients with any clinically
suggestive thyroid dysfunction; patients who have used
drugs known to affect the thyroid function (levothyro-
xine, anti-thyroid drugs, iodine-containing drugs, lithium,
hydantoinates, estrogens and progestogens, iodinated
contrast media) up to six weeks prior to getting pregnant;
and patients who underwent surgery or radiotherapy
affecting thyroid function.
The pregnant women were informed about this study
and those agreeing to participate in it signed informed
consent forms, according to the Ethics and Research
Committee of HSPE, which approved the study.
Maternal age ranged from 18 to 47 years, mean ±
SD of 32.5 ± 6.9 years. The gestational age at the time
the laboratory tests were done ranged from 6.0 to 41.1
weeks, mean ± SD of 19.1 ± 8.5 weeks, respectively.
The majority of tests were performed between 12 and
14 weeks of gestation.
At the time of patient inclusion, a 6 ml blood sample
was obtained by peripheral venipuncture, into a dry tube,
for later serum TSH and free T4 assaying.
The serum levels of both hormones were measured
by microparticle enzyme immunoassay (MEIA)(9), in
ADVIA CENTAUR (BAYER), with the AxSYM System
hTSH II Ultrasensitive (Abbott) and AxSYM System
Free T4 (Abbott).
This study analyzed the TSH and free T4 levels in
different weeks of pregnancy and the correlation between
these two indicators. Data were recorded in a standar-
dized form and later transferred to the Microsoft Excel
For the statistical analysis, the means, standard
deviations and medians were calculated. The Pearson
correlation coefficient was used to determine the
correlation coefficient between TSH and free T4 levels.
The α (alpha) risk equal to or lower than 5% and the
95% confidence interval were utilized.
The data were depicted in box plot charts, for better
viewing of their distribution. The values were separated
into four groups (quartiles) of the same size and plotted
in the chart. The second quartile is equal to the median,
that is, it is the figure that halves the data in two equal
size groups. The box begins in the first quartile and ends
in the third one and contains 50% of the central data.
The box middle line is the median. Points marked with
circles or with asterisks are the outliers.
TSH levels in the population
Mean TSH of the studied population was 2.14 ± 1.51
µIU/ml, ranging from 0.01 to 13.0 µIU/ml. Table 1 depicts
the 50th and the 97.5th percentiles for TSH according to the
gestational age and figure 1, their graphic representation.
The 2.5th percentile was not utilized because it contained
negative values. By means of linear regression analysis,
a fitted straight line was obtained relating TSH levels
and gestational age. The model was significant (p <
0.001) and the line equation was: TSH = 1.7154 +
Figure 1. Percentiles (50th and 97.5th) for TSH according to gestational age
einstein. 2007; 5(1):51-55
53Normal values of thyroid-stimulating hormone and free thyroxin in pregnant women
Free T4 levels in the population
Mean free T4 of the studied population was 0.98 ± 0.30
ng/dl, ranging from 0.08 to 3.6 ng/dl. Table 2 depicts the
2.5th, 50th and 97.5th percentiles for free T4 relative to
gestational age and figure 2, its graphic representation.
The same model described for TSH was applied for
free T4, which was also significant (p = 0.003), and
the regression line equation was: free T4 = 1.0559-
In this sample, 93 (10.8%) patients had TSH levels
above the upper reference limit of the kit (4.0 µIU/ml).
Considering the 97.5th percentile as the upper normal
limit, 39 (4.58%) of the pregnant women in this group
had levels higher than that, and 38 had free T4 levels
between the 2.5th and the 97.5th percentiles, one of them
with levels lower than the 2.5th percentile.
Table 1. Percentiles (50th and 97.5th) for TSH (µUI/ml) according to gestational age
gestational age T S H
Table 2. Percentiles (2.5th, 50th and 97.5th) for free T4 (ng/dl) according to
Figure 2. Percentiles (2.5th, 50th and 97.5th) for free T4 according to gestational age
einstein. 2007; 5(1):51-55
54Antonini M, Martinelli S, Yoshizumi APMF, Garcia SAL, Lippi UG
Correlation between TSH and free T4
When the Pearson’s correlation coefficient was used
for comparing free T4 and TSH levels, a statistically
significant positive correlation was found (r = 0.069;
p = 0.044). Figure 3 displays the dispersion measurements
between TSH and free T4.
Figure 3. Dispersion measurements between TSH (µUI/ml) and free T4 (ng/dl)
The thyroid gland plays an important role in pregnancy
outcome as well as in fetal development. The changes
caused by pregnancy in the mother establish a new
homeostatic equilibrium starting when the zygote is
implanted and only ending at the puerperium(10-12).
According to the American College of Obstetrics and
Gynecology (ACOG), thyroid disorders are the second
most frequent endocrine disorder in child-bearing age
women(12). Several authors associate the presence of thyroid
disorders with higher rate of pre-eclampsia, spontaneous
abortion, premature delivery, intra-uterine fetal death and
disturbed fetal psychomotor development(13).
Therefore, understanding the nature and the magnitude
of the changes that occur in pregnancy is important to
differentiate pregnant women with thyroid diseases from
those with adaptative changes due to pregnancy. Hence the
importance of determining the normal thyroid function,
especially of free thyroxine (free T4) and of the thyroid
stimulating hormone (TSH) during pregnancy, as these
two are markers of the maternal thyroid function.
Technological advances made available very sensitive
and automated methods for the assessment of thyroid
function, such as the microparticle enzyme immunoassay
(MEIA) used in the present study. As this is an ultra
sensitive assay, minimal concentrations of TSH and
therefore, minor thyroid dysfunction may be detected(14).
The measurement of the free thyroxine fraction, in place of
the total thyroxine, is justified for its biological activity and
the bypassing of potential interference due to physiological
or pathological fluctuations in thyroglobulin levels, the
main protein responsible for T4 transport(1,13).
The mean TSH levels found in this study, 2.14 ± 1.51
µIU/ml, are similar to those found by Kabyemela et al.(15)
(1.99 ± 1.16 µIU/ml), in Tanzania, and by Sieiro Netto et
al. (2004)(16) (2.17 ± 1.29 µIU/ml), in Brazil. In another
study, also carried out in Brazil by Vieira et al.(1), the
mean TSH values (1.54 ± 0.80 µIU/ml) were lower than
those found in the present study.
The mean free T4 of 0.98 ± 0.30 ng/dl found in this
study is different from that found by Kabyemela et al.(15)
(0.80 ± 0.17 ng/dl) and by Vieira et al.1 (2004) (0.78 ±
0.16 ng/dl) however, were similar to those reported by
Sieiro Netto et al. (2004)(16) (1.03 ± 0.28 ng/dl). Both,
the TSH and free T4 levels may reflect differences in the
populations studied, the amount of ingested iodine, the
TSH assay technique, among others.
A significant positive correlation was found between
the free T4 and TSH levels (p < 0.01) due to the existing
feedback between the thyroid stimulating hormone and
thyroxine production. This correlation was not significant
in non-pregnant women, according to Panesar et al.(17).
In order to establish the normal TSH and free T4
limits, the 2.5th and the 97.5th percentiles for each of these
parameters were determined for different gestational ages.
Because the 2.5th percentile of TSH contained negative
values, it was not used in the analysis. The use of TSH
and T4 limits could be helpful in clinical practice. The
lower TSH limit, especially in the first trimester, may be
difficult to interpret as TSH is suppressed by the increase in
chorionic gonadotropin. Panesar et al.(17), like in this study,
use the 97.5th percentile of TSH as its upper limit. Haddow
et al.(7) found that 64% of the pregnant women with TSH
levels above the 98th percentile in the second trimester,
displayed hypothyroidism after 10 years. Of these pregnant
women with high TSH, 77% were unaware of any thyroid
dysfunction prior to that. Thus they recommend the TSH
98th percentile be used as threshold for treatment or
rigorous laboratory follow up regarding hypothyroidism.
The p2.5 and p97.5 values were within the normal
range established by the free T4 manufacturer. In the
study by Panesar et al.(17), all the values were within the
reference range during pregnancy.
Ninety three (10.8%) patients had TSH levels above the
upper normal limit established by the manufacturer (4.0
µIU/ml). However, considering the 97.5th percentile as the
upper normal limit, 39 (4.6%) pregnant women had high
TSH levels. Of these, 38 (4.5%) had free T4 levels between
the 2.5th and the 97.5th percentile, that is, subclinical
hypothyroidism. One patient had free T4 levels below the
2.5th percentile, consistent with overt hypothyroidism, even
if there was no clinical evidence. Hence, only 39 of these 93
patients would have the diagnosis of hypothyroidism made.
By using this criterion, no treatment would be prescribed
for 54 of these pregnant women.
einstein. 2007; 5(1):51-55
55Normal values of thyroid-stimulating hormone and free thyroxin in pregnant women
The determination of the normal TSH and free T4 limits
throughout pregnancy is the first step for the assessment
of actual thyroid dysfunction, allowing the diagnosis and
treatment of patients who, otherwise would not have
1. Vieira JGH, Kanashiro I, Tachibana TT, Griringhello MT, Hauache OM, Maciel
RMB. Definição de valores normais de tiroxina livre durante a gravidez. Arq
Bras Endocrinol Metab. 2004;48(2):305-9.
2. Ekins R. Measurement of free hormones in blood. Endocr Rev. 1990;11(1):4-46.
3. Burrow GN, Fisher DA, Larsen PR. Maternal and fetal thyroid function. N Engl
J Med. 1994;331(9):1072-78.
4. Glioner D, Nayer PH, Bourdoux P, Lamarca M, Robin C, Steirteglem AV, et al.
Regulation of maternal thyroid during pregnancy. J Clin Endocrinol Metab.
5. Glioner D. The regulation of thyroid function in pregnancy: pathways of endocrine
adaptation from physiology to pathology. Endocr Rev. 1997;18(3):404-33.
6. Pop VJ, de Rooy HA, Vader HL, van der Heide D, von Son MM, Komproe IH.
Low maternal free thyroxin concentrations during pregnancy are associated
with impaired psychomotor development in infancy. Clin Endorinol.
7. Haddow JE, Palomaki GE, Allan WC, Williams JR, Knight GJ, Gagnon J, et al.
Maternal thyroid deficiency during pregnancy and subsequent neuropsychological
development of the child. N Engl J Med.1999;341(8):549-55.
8. Zoeller RT. Transplacental thyroxin and fetal brain development. J Clin Invest.
9. Manual Técnico Operações AXSYM System ABBOTT T4 livre e hTSH II Ultra-
sensível. São Paulo: Abbott; 2003.
10. Peixoto S, Sankovisck M, Mendes ETR, Fernandes GL. Pré-natal. 3a ed. São
Paulo: Roca; 2004. p. 31-50.
11. Vermiglio F. Maternal hypothyroxinaemia during the first half of gestation in
a iodine deficient area with endemic cretinism and related disorders. Clin
12. American College of Obstetricians and Gynecologists – ACGO. Thyroid disease
in pregnancy. Int J Gynaecol Obstet. 2002;79(2):171-80.
13. Brent GA. Maternal thyroid function: interpretation of thyroid function tests in
pregnancy. Clin Obstet Gynecol. 1997;40(1):3-15.
14. Romaldini JH, Sgarbi JA, Farah CS. Disfunções mínimas da tireóide:
hipotireoidismo subclínico e hipertireoidismo subclínico. Arq Bras
Endocrinol Metabol. 2004;48:147-58.
15. Kabyemela EAR, Swinkels MJW, Chuwa LMM, Ross HA, Dolmans WMV,
Benraad TJ. Thyroid function studies in normal pregnant Tanzanian women.
Am J Trop Med Hyg. 1996;24(1):58-61.
16. Sieiro Netto L, Coeli CM, Micmacher E, Mamede SC, Nazar LK, Correa EK, et
al. Estudo longitudinal do eixo hipófise-tireóide durante a gravidez. Arq Bras
Endocrinol Metab. 2004;48(4):493-98.
17. Panesar NS, Li CY, Rogers MS. Reference intervals for thyroid hormones in
pregnant Chinese women. Ann Clin Biochem. 2001;38:329-32.