Antithyroid Drug-induced Fetal Goitrous HypothyroidismKey Points
- Treating pregnant women with antithyroid
drugs (ATDs) puts the fetus at risk of overtreatment and thus
subsequent development of fetal hypothyroidism and goiter formation
- Fetal goitrous hypothyroidism can cause
severe pregnancy-related complications and potentially harm fetal growth
and neurological development
- Treatment of fetal goitrous
hypothyroidism with intra-amniotic levothyroxine achieves better results
than simply discontinuing maternal ATD treatment
- Awareness of the pregnancy-related changes to maternal thyroid status is essential when treating maternal hyperthyroidism
- Close monitoring of the maternal thyroid status, especially estimates of free T4 levels, is the best way to avoid overtreatment
- Centralized care of pregnant women with
Graves disease in specialized multidisciplinary units is urgently needed
to maintain optimal fetal development
Abstract and Introduction
Abstract Maternal overtreatment with antithyroid drugs
can induce fetal goitrous hypothyroidism. This condition can have a
critical effect on pregnancy outcome, as well as on fetal growth and
neurological development. The purpose of this Review is to clarify if
and how fetal goitrous hypothyroidism can be prevented, and how to react
when prevention has failed. Understanding the importance of
pregnancy-related changes in maternal thyroid status when treating a
pregnant woman is crucial to preventing fetal goitrous hypothyroidism.
Maternal levels of free T
4 are the most consistent indication
of maternal and fetal thyroid status. In patients with fetal goitrous
hypothyroidism, intra-amniotic levothyroxine injections improve fetal
outcome. The best way to avoid maternal overtreatment with antithyroid
drugs is to monitor closely the maternal thyroid status, especially
estimates of free T
4 levels.
Introduction In pregnant women, overtreatment with
antithyroid drugs (ATDs) puts the fetus at great risk. Iatrogenic fetal
hypothyroidism can impair the neurological development and growth of the
child.
[1-4] Furthermore, a fetal goiter can cause tracheal compression, which
increases the risk of developing polyhydramnios (owing to reduced
swallowing ability), premature labor (attributable to rupture of the
fetal membranes caused by the polyhydramnios), dystocia (because
ofhyperextension of the fetal neck) and airway obstruction at birth.
[5] Premature labor is the main cause of newborn morbidity and mortality in
all pregnancies. The risk connected with premature labor, which must be
prevented, increases if the fetus has other fetal diseases in addition
to hypothyroidism, such as iatrogenic hypothyroidism. Whether
discontinuation of maternal ATD treatment is sufficient when a fetal
goiter develops or if the fetus needs direct treatment with
intra-amniotic levothyroxine injections is a subject of debate. However,
the development of iatrogenic fetal goiters can be prevented if the
endocrinologist is aware of the changes in maternal thyroid status and
metabolism that occur during pregnancy.
This article reviews the reported cases of
fetal goiter formation attributable to overtreatment of maternal
hyperthyroidism with ATDs. The aim of this Review is to clarify if and
how such cases could have been prevented, and how to react when
prevention has failed.
Thyroid Status in Pregnancy During the first trimester, maternal serum concentrations of total T
4 rise because of a combination of the thyrotropic effect of human
chorionic gonadotropin (hCG) and the stimulatory effect of estrogen on
the concentration of thyroid binding globulin (TBG). Furthermore, the
hCG-induced stimulation of the TSH receptor leads to an increase in the
production of thyroid hormones, which then leads to a decrease in TSH
levels by negative feedback to the pituitary gland.
[6] Reference values based upon thyroid function variables of pregnant
women indicate that during the first trimester normal levels of maternal
total T
4 are increased by ~50% compared to the usual
nonpregnant values (which vary between laboratories because of the
different assays used).
[7-9] In the second and third trimesters, the hCG-induced stimulation of the
thyroid gland decreases, while the maternal level of total T
4 continues to be above, or at least in the high end, of the nonpregnant
reference values, and levels of TSH continue to be in the low end of the
range.
[10,11] Other changes also influence maternal thyroid status, such as increased
renal iodine clearance, increased blood volume, and placental-fetal
exchange and metabolism of thyroid hormones.
[12,13] The placental-fetal exchange and metabolism of thyroid hormones is particularly important to the fetus.
The fetus begins to metabolize thyroid
hormones early in the first trimester, but the production and secretion
of fetal thyroid hormones does not reach notable levels until
midgestation.
[14] Until then the fetus is dependent on the maternal supply of thyroid
hormones; even at term up to 30% of the fetal thyroid hormones are of
maternal origin.
[15-17] Although the supply of maternal T
4 is extremely important for the fetus, quantitatively, the levels of
fetal thyroid hormones are much lower than the maternal thyroid hormone
levels. This balance is secured by a preferential placental deiodination
(by type 3 deiodinases) of T
4 to the presumably inactive reverse T
3, which prevents fetal hyperthyroidism.
[13,18] However, once the fetal thyroid gland becomes functional, TSH-receptor
autoantibodies (TRAbs) that are able to cross the placenta freely will
affect the fetal thyroid gland (Figure 1). Fetuses of mothers with
TSH-receptor stimulating immunoglobulins (TSIs), as found in Graves
disease, are therefore at risk of developing hyperthyroidism—even when
the mother is euthyroid.
[19-21] ATDs (such as thionamides) given to the mother also cross the placenta
and block the activity of fetal thyroid peroxidase (and peripheral
deiodination when the mother is treated with propylthiouracil), which
increases the risk of developing fetal hypothyroidism and thus a fetal
goiter. Before the onset of fetal thyroid function there is no need to
assume that ATDs will have a direct effect on the fetus;
[22] however, iodide uptake and colloid formation begin as early as the eleventh week after conception.
[12] Treating maternal Graves disease with ATDs, therefore, requires a
careful balance between securing sufficient fetal production of thyroid
hormones (and supply of maternal T
4), while preventing fetal hyperthyroidism attributable to thyroid-stimulating antibodies.
(Enlarge Image)
| Figure 1.
Pathways of Fetal Goiter Development in Connection to Maternal Antithyroid Drug Treatment for Graves disease. Both maternal autoantibodies and ATDs can cross the placenta and exert their effect on the fetal thyroid gland, at worst, leading to fetal goiter formation and hypothyroidism, which puts the fetal outcome at risk. Abbreviations: ATD, antithyroid drug; TPO, thyroid peroxidase.
|
The changes in thyroid status during
pregnancy, especially in women with autoimmune diseases, complicate the
interpretation of maternal thyroid status and thus the need for
treatment with ATDs. The first case of a goiter in a newborn baby
attributable to maternal propylthiouracil treatment was published by
Eaton in 1945.
[23] Yet, the task of treating pregnant women with Graves disease still remains a puzzle to many physicians.
Fetal Goitrous Hypothyroidism In total, we found 48 cases of fetal goitrous hypothyroidism attributable to maternal ATD treatment reported in 20 case reports
[24-43] and seven larger investigations
[44-50] between 1980 and 2009. The cases were divided into two groups according
to intervention: group A with regulation of maternal ATD treatment
supplemented by invasive treatment with intra-amniotic levothyroxine
injections (23 cases), and group B with noninvasive regulation of the
maternal ATD dose only (25 cases). Details of the review criteria and
the individual patients are provided in Supplementary information
online.
Maternal Thyroid Status Across the two groups, 12 women had been
diagnosed with Graves disease before their current pregnancy and 11
women were diagnosed with this condition during their current pregnancy.
At the time of fetal goiter diagnosis, the average dose of
propylthiouracil was 289.0 mg/day (information from 19 of 23 patients)
in group A and 222.8 mg/day (information from 17 of 25 patients) in
group B. No correlation was found between the duration of maternal
thyroid disease and the dose of propylthiouracil. Seven women received
imidazole derivatives (such as methimazole and carbimazole) during the
first trimester, three of whom
[27,44] were switched to propylthiouracil during the second trimester.
In group A, when a goiter was discovered, ATD
treatment was discontinued in three patients, reduced in 12 patients,
reduced then discontinued in three cases, discontinued then restarted in
one case, and not changed in two patients (information from 21 of 23
patients). In group B, ATD treatment was discontinued in three patients,
reduced in 12, reduced then discontinued in two, and in three cases the
mother received supplemental levothyroxine treatment (information from
20 of 25 patients).
TRAb levels were only reported in 14 patients
(61%) from group A and in 12 patients (48%) from group B. Nine women in
group A were positive for TRAb, while five had negative or normal
levels. Five women in group B had positive TRAb levels, while seven had
negative or normal levels. Across the groups, only in eight patients did
the authors differentiate between TSIs and TSH-receptor binding
inhibitory immunoglobulins (TBIIs). Information on levels of TBIIs was
given in five patients,
[25,26,33,44] three of whom had both positive TBII and TSI levels. Three women in
group A (information from 11 of 23 cases) had a TSH level >4.5 mU/l
at the time of goiter discovery (four had a TSH level >2.5 mU/l),
while TSH levels were only reported in seven patients in group B, one
[49] of whom had a TSH concentration >2.5 mU/l (TSH <2.5 mU/l during the first trimester is currently recommended
[51]). Thus in most cases, the observed TSH levels did not indicate maternal hypothyroidism.
In the publications where both reference
ranges and analyses were provided (15 patients across both groups), data
showed that all women but two
[26] had levels of free T
4 below or in the lower part of the reference range at the time of goiter discovery. In some patients, free T
4 levels were low even when TSH levels were below or in the very low part of the reference range.
[28,29,38,42] Thus, maternal hypothyroxinemia seems to be the most reliable maternal indicator of fetal hypothyroidism.
Fetal Thyroid Status On average, the fetal goiters were detected
by ultrasound examination at gestational week 29 in both groups. The
earliest detection of a fetal goiter was in gestational week 19.
[38] Upon discovery of the fetal goiters, fetal
blood or amniotic fluid sampling was performed in all patients in group A
to confirm the diagnosis of fetal goitrous hypothyroidism. Fetal blood
sampling showed levels of TSH between 9.7 mU/l and 1,640.0 mU/l (median
38.0 mU/l; information from 19 of 23 patients). TSH levels from the
amniotic fluid sampling were only available in six patients, with an
average level of 4.6 mU/l. In group B, information was generally lacking
about fetal goiter size and particularly fetal thyroid status. Though
not surprising given the noninvasive approaches used in patients from
group B, this hindered a comparison of the severity of fetal
hypothyroidism between the two groups. However, in three studies
[28,43,46] in group B, fetal blood levels of TSH (range 40.2-56.0 mU/l) were given
at the time of goiter discovery, and did not differ from those of group
A. In addition, fetal goiter size did not seem to differ between the
two groups.
Intra-amniotic levothyroxine injections
(group A) were given between one and six times with an average dose of
279 |ig per injection (dose provided for 49 of 58 injections). A
decrease in goiter size was seen within 0.5-2.5 weeks after the first
injection where information was given on the gestational week of the
subsequent examinations (10 patients).
Fetal blood or amniotic fluid sampling was
only performed in eight patients in group A after the primary diagnosis
of fetal hypothyroidism. Although these tests were not always performed
at the first ultrasound examination following the first injection of
levothyroxine, those performed 1-7 weeks after the first injection all
showed normalization or improvement of the fetal thyroid status. In
group B, a decrease in goiter size was reported within 1- 9 weeks of the
maternal ATD treatment regimen being altered (information from 13
patients). However, in four patients in group B, the goiter size was
unchanged after 2- 5 weeks.
[28,39,49] Obstetric Outcome Some studies indicate that male fetuses are
more prone to develop goitrous hypothyroidism attributable to maternal
ATD treatment than female fetuses.
[52] Our review of the literature does not support this hypothesis, as we
found a total of15 female and 12 male fetuses with iatrogenic goiter (in
21 cases there was no information on the fetal sex).
In group A, the fetal goiters resolved in 10
patients, while in seven patients a goiter could still be seen or
palpated at birth (
Table 1). However, the goiters had decreased in size after treatment in 16 of
the 19 patients for whom this information was provided. This finding
should be viewed in light of the fact that as pregnancy progresses, the
fetal thyroid gland will physiologically increase in size. The three
cases that did not show a decrease of the absolute goiter size could,
therefore, reflect a relative decrease compared to gestational age. Six
neonates in group B had a goiter at birth, and seven did not, while no
information was provided for the remaining 12 neonates.
In group A, 16 neonates were euthyroid (73%),
while six were hypothyroid (information from 22 neonates). In group B,
11 neonates were euthyroid (50%), eight were hypothyroid, one was
subclinically hypothyroid and two were hyperthyroid (information from 22
neonates). Across the two groups, postnatal thyrotoxicosis developed in
six cases (
Table 1). This finding could be attributed to the slower clearance of the
maternal TRAbs than that of propylthiouracil. Thus one must be aware
that a neonate, though hypothyroid during intrauterine life because of
overtreatment with propylthiouracil, can develop autoimmune
thyrotoxicosis after birth.
The median fetal birth weight was 2,891 g in
group A and 3,115 g in group B (information from 17 and eight neonates,
respectively) (
Table 1). In three cases, the weight of the newborn compared to gestational age was below the Scandinavian growth curve minimum;
[53] however, only one case
[28] fell below an American gender-specific growth curve.
[54] Gestational age at birth was an average of 36.5 weeks in group A and
35.5 weeks in group B (information from 18 and six births, respectively)
(
Table 1). These results were in accordance with previous conclusions that there
is a connection between maternal hypothyroidism or thyroid
autoimmunity, and preterm delivery and small birth weights.
[55] Three neonates
[35,37,41] suffered from respiratory distress at birth (
Table 2); in two
[35,41] no goiter was seen as the possible cause of this distress. In one neonate,
[37] who was born at gestational week 30, the respiratory distress must be
attributed not only to a prominent goiter but also to prematurity of the
lungs at delivery.
More than half of the fetuses were delivered by cesarean section (information from 21 patients). In five patients
[29,33,38,40,44] a cesarean section was required because of breech presentation, which
has previously been correlated with maternal hypothyroxinemia and
hypothyroidism.
[56,57] In two women, cesarean sections were performed because of fetal distress,
[31,37] in one of whom this distress was probably attributable to the
intra-amniotic levothyroxine injection that was given within 24 h of the
mother going into preterm labor.
Severe Complications Reports were given on several of the
complications associated with fetal goitrous hypothyroidism including
polyhydramnios, hyperextension of the fetal neck, intrauterine growth
restriction (IUGR)
[28,29,43] and fetal hydrops
[32] (
Table 2). Although no statistical comparison was applied between group A and
group B because of the small sample size, the gravest outcomes were
undoubtedly seen in group B, with two stillbirths recorded (
Table 2).
[39,44] In neither of these two cases did the goiters occur earlier in the
pregnancy nor were they larger at the time of discovery than in other
cases from group A or group B. Furthermore, cases of retarded bone
development,
[36,44,46] advanced bone ossification
[30] and congenital malformations
[32] were reported across both groups (
Table 2).
A general lack of follow-up information
characterized the publications included in both groups. In group A,
follow-up examinations were performed between 2 weeks and 3 years after
birth in 12 children and found no neurological and/or motorical sequelae
in any of the children. The weight of one child was within the tenth
percentile at 3 years of age.
[28] Seven children in group B had neurological and/or motorical evaluations
performed between 12 days and 20 months after birth. All the children
were developing normally. No reports were given on neurological
retardation.
How to Avoid Maternal Overtreatment The cases presented in this Review included a
range of severe complications that have been associated with thyroid
dysfunction and autoimmunity during pregnancy:
[55,58-60] stillbirth, premature labor, increased risk of breech presentation,
IUGR, delayed bone development, polyhydramnios and fetal hydrops. This
observation stresses the importance of avoiding iatrogenic fetal
goitrous hypothyroidism.
Treatment with ATDs During Pregnancy Up to 0.4% of all pregnant women have been
reported as hyperthyroid (not including the more common transient
gestational thyrotoxicosis), with autoimmunity being the primary
etiology in 90% of cases.
[61,62] Untreated hyperthyroidism during pregnancy is associated with an
increased risk of preeclampsia, congestive heart failure, fetal
mortality, infants born small for gestational age and thyroid storm—the
risk of these complications increases with increasing maternal
autoantibody levels.
[60,63,64] Therefore, treatment of maternal hyperthyroidism is necessary.
Guidelines have so far recommended treatment
with ATDs rather than with radioiodine or surgery unless severe adverse
effects associated with the drugs occur.
[51] Supplementing ATD treatment with levothyroxine (the 'block-replace'
regimen administered to three patients in group B) is not recommended.
This treatment strategy can result in the use of higher doses of ATD to
keep the woman euthyroid compared to the use of ATDs alone, thus
increasing fetal ATD exposure.
[65] The ATDs that are most frequently used (in
pregnant women and the general population) are propylthiouracil and
methimazole. In contrast to findings from the 1970s,
[66] more recent research has indicated that placental transfer of methimazole is not greater than that of propylthiouracil.
[67-69] However, propylthiouracil is still recommended as the first-line ATD
during pregnancy, because of the possible association of methimazole
with congenital abnormalities. Several studies have reported congenital
abnormalities (especially aplasia cutis congenita and choanal atresia)
in fetuses exposed to methimazole during the first trimester of
pregnancy.
[70-73] Barbero
et al.
[74] found an odds ratio of 18 for choanal atresia among infants who were exposed to methimazole
in utero compared
to infants who were not exposed to this drug. These cases have led to a
general acknowledgment of a so-called 'methimazole embryopathy'. By
contrast, cases of congenital malformations connected to
propylthiouracil treatment are few. In one case reported by Yanai
et al.,
[12] a child was born with severe malformations after propylthiouracil
exposure. To our knowledge, one case of choanal atresia after
propylthiouracil treatment during pregnancy has been reported.
[75] In a study published in 2010, several instances of birth defects after
maternal propylthiouracil treatment were found, as well as a significant
association (
P <0.01) between methimazole treatment and choanal atresia.
[76] Thus, although methimazole-induced embryopathy is well established, a
teratogenic effect of propylthiouracil still cannot be eliminated.
Maternal adverse effects associated with
propylthiouracil treatment have lately been given much attention. In
spring 2010, the FDA issued a warning about the risk of liver failure in
connection to treatment with propylthiouracil.
[77] Leading up to this warning was an increasing awareness of the
hepatotoxicity of the drug, with several reported instances of liver
transplantation (with poor survival rates) following
propylthiouracil-induced acute liver failure in both children and
adults.
[78-81] Neonatal hepatitis after maternal propylthiouracil treatment during pregnancy has also been reported.
[82] However, isolated case reports of methimazole-associated liver failure can also be found.
[83-85] Although still controversial, a pragmatic
suggestion has been to treat pregnant women with propylthiouracil during
the first trimester of pregnancy (to avoid the teratogenicity of
methimazole) and then switch to methimazole treatment during the second
and third trimesters (to minimize the risk of hepatotoxicity associated
with propylthiouracil).
[81] However, more research is needed to establish whether or not such a
transition would do more harm than good to both mother and fetus.
Importance of Careful Monitoring—The Mother Development of a fetal goiter is a clear
indication of fetal thyroid gland dysfunction. However, slight maternal
overtreatment with ATDs without fetal goiter formation might still put
the fetus at risk of growth restriction and compromised neurological
development. Maternal ATD treatment spans a continuum from a
well-balanced treatment through slight overtreatment to the extreme
cases of gross overtreatment in which fetal goiter formation occurs.
Where on this continuum the risks of severe pregnancy-related
complications and long-term consequences begin is not evident. Pregnant
women should be kept on optimal ATD regimes by careful monitoring;
however, the risk of overtreatment with ATDs during pregnancy exists
even in skilled hands.
The cases presented in this Review
demonstrated no connection between the fetal thyroid status and the
maternal dose of ATD. Doses of propylthiouracil as small as 50 mg daily
caused fetal overtreatment in some patients. Maternal levels of TRAb
also do not seem to be connected to the development of the fetal
goiters. Cases of both negative and positive antibody levels were found;
in the latter, TSIs, TBIIs, or both were detected (this finding has no
implications for maternal TRAb-monitoring with regards to predicting
intrauterine or neonatal thyrotoxicosis according to the current
guidelines
[51]).
One must consider that the natural suppression of the immune system
during pregnancy will invariably lead to decreased antibody levels and
improvement, if not remission, of autoimmune disease—thus decreasing the
requirement for ATDs.
[45,86-88] Interestingly, even in the reviewed cases with highly suppressed
maternal TSH concentrations, the fetuses had severe hypothyroidism, and
in fact only a minority of the women had a TSH level >2.5 mU/l. Even
the women with very low TSH concentrations had free T
4 concentrations in the low part of the reference range. Maternal free T
4 levels were thus the only consistent indicator of maternal and fetal thyroid status (similar results were found by Momotani
et al.
[89]).
The apparent discrepancy between the levels of free T
4 and TSH can be ascribed to the unstable or latent reaction of pituitary TSH to the changes in T
4 levels that occur with the changes in thyroid hormone status during pregnancy, and to adjustments of ATD dose.
[90,91] Thus, in pregnant women with hyperthyroidism, overtreatment might be
caused by difficulties in interpreting the levels of free thyroid
hormones. Pregnancy-related hyperestrogenism induces a rise in TBG and
thus in total T levels. As the active component of the circulating
thyroid hormones is the free part, an estimate of the free T
4 concentration is essential. Free T
4 estimates can generally be derived in three ways: by one of the commercially available free T
4 methods used in many clinical biochemical laboratories, or by measurement of total T
4 and applying some correction for the binding proteins, such as a T
3 or T
4 uptake (resulting in a free T
4 index), or by direct measurement of TBG (giving a T
4:TBG ratio). However, the measurement of free T
4 is an estimate and not a quantitatively correct value, and none of the
methods correct sufficiently for the extreme increase in levels of TBG
that occur during pregnancy. However, a correct interpretation is more
likely if a free T
4 index (or T
4:TBG ratio) is used, rather than the so-called direct methods. Interpreting the latter form of free T
4 estimate is extremely difficult during pregnancy and further complicated by methodological differences between laboratories.
[90-93] The use of reference ranges that include
women who are not pregnant is probably a further explanation of maternal
overtreatment. Many authors have stressed this point and advocated the
introduction of local trimester-specific reference ranges according to
the methods used in the local laboratory.
[8,9] Vaidya
et al.
[94] even showed that using reference ranges that are not specific to the
gestational age would misdiagnose as euthyroid up to 30% of pregnant
women who were in fact hypothyroid. Complicating the use of reference
ranges further, Boas
et al.
[10] showed that the intraindividual variation in thyroid hormone levels
throughout pregnancy was considerably smaller than the interindividual
variation.
[10] The authors suggest that a woman's levels of thyroid hormones should be
evaluated in comparison to her own earlier levels instead of a
population-based reference range. Boas
et al.
[10] thus proposed a predictive model to calculate the woman's individual
euthyroid status throughout pregnancy. This model would be possible in
women treated for hyperthyroidism or hypothyroidism because of the
regular monitoring of their thyroid status and would ease the
interpretation of the individual thyroid hormone measurements.
In their interpretation of maternal thyroid
function, specialists who monitor pregnant women during ATD treatment
should pay the utmost attention to four factors. First, the latency of
reaction of TSH levels to alterations in free T
4 levels. Second, the correct interpretation of free T
4 and free T
3 estimates during pregnancy according to the applied method for
measurement. Third, trimester-specific reference intervals for all the
measured variables. Finally, the concept of intraindividual versus
interindividual variations of thyroid-related hormones.
The maternal free T
4 estimate was
the only reasonably consistent indication of maternal and fetal thyroid
status. To avoid maternal overtreatment with ATDs, close monitoring of
free T
4 levels, to keep them within the laboratory's trimester-specific reference range, is crucial (
Box 1).