| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Hospital Saint-Pierre, Department of Internal Medicine, Thyroid Investigation Clinic, Université Libre de Bruxelles, Belgium
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. The Regulation of...III. Pathological Alterations of...IV. Conclusions and PerspectivesReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. The Regulation of...III. Pathological Alterations of...IV. Conclusions and PerspectivesReferences |
|---|
|
II. The Regulation of Thyroid Function in Normal Pregnancy |
|---|
|
TopAbstractI. IntroductionII. The Regulation of...III. Pathological Alterations of...IV. Conclusions and PerspectivesReferences |
|---|
A. The thyroid hormone transport proteins
Thyroid hormones (TH) are transported in serum noncovalently bound
to three proteins: T4-binding globulin (TBG), albumin, and
transthyretin (previously called prealbumin or TBPA) (4). The relative
distribution of TH among the binding proteins is directly related to
both their affinities and concentrations. In steady state conditions
the bound hormone fraction is in equilibrium with a free unbound
fraction, which represents a minute amount of the total circulating TH:
0.04% for T4 and 0.5% for T3 (5). Despite the
fact that TBG in serum is by far the least abundant of the three
transport proteins, about two thirds of the T4 in serum of
normal subjects is carried by TBG, owing to its extremely high affinity
for the hormone (6). In conditions with TBG excess, such as pregnancy,
the proportion of circulating T4 carried by TBG is even
greater, in excess of 75%, which indicates that TBG represents the
major thyroid hormone transport protein in pregnancy (7). Furthermore,
during pregnancy the respective affinities of the three binding
proteins for their hormonal ligands are not significantly modified, and
the circulating levels of both serum albumin and transthyretin remain
stable, with only a slight tendency to decrease near the end of
gestation, mainly as a result of passive hemodilution due to the
increased vascular pool (8, 9, 10). Thus, the major change for thyroid
hormone-binding proteins involves the marked and rapid increase in
serum TBG levels as a result of estrogen stimulation. Compared with
preconception concentrations (average 15–16 mg/liter), serum TBG
begins to increase in pregnancy after a few weeks and reaches a plateau
around midgestation, 2.5-fold higher than the initial value (between
30–40 mg/liter). Thereafter, the TBG concentration remains practically
unchanged until term (Fig. 1
) (11, 12, 13).
|
10–15%). Therefore, the net effect of
kinetic changes due to highly sialylated estrogen-specific fractions on
the global TBG clearance is relatively minor. Also, the work was
carried out in a heterologous experimental system, where human TBG was
injected into rats to perform the metabolic studies; this probably
explains the authors’ results of short TBG half-lives (1 day),
compared with the known physiological half-times of TBG determined in
the Rhesus monkey (>4 days) and the human (>5 day) (15, 20, 21). Thus, we infer from these data that a prolonged biological half-time per se cannot entirely account for the observation that, at least in primates, serum TBG starts to increase 24 h after the exposure to high estrogen (15). Ain et al. (22) also attempted to demonstrate directly an effect of estrogen on the synthesis and secretion of TBG, using a human hepatoma cell line Hep G2, which produces TBG. The authors failed to show an increase in the cytoplasmic TBG mRNA content after estrogen exposure. It should be kept in mind, however, that Hep G2 cells are not an ideal model for the study of estrogen effects on TBG synthesis because these tumor cells do not react to estrogen stimulation as do normal cells (4). TBG could already be produced at its maximal rate or the cells could simply be unresponsive to estrogen. Hence, until the debate can be solved by more definitive and direct arguments, we consider it safe to propose that the increase in serum TBG found in pregnancy might result from a combination of factors: increased TBG production by liver, prolonged half-life due to increased sialylation, and stabilization of the TBG molecule because more T4 is proportionally bound to it.
Irrespective of the precise molecular mechanisms that may explain the
TBG rise in pregnancy, it is important to note that the serum TBG
increase observed during the first part of gestation does not follow a
smooth curve. Determinations of TBG levels on a weekly basis in a large
number of pregnancies indicates that the overall profile of the TBG
rise in blood exhibits wide individual variation until the plateau is
attained, and also that the plateau value is variable individually
(Fig. 1
). Such variation can be partially explained by the fact that
preconception TBG levels are variable between 10–22 mg/liter
(reference range in a normal female population) (23), but probably also
because the effects induced by estrogen require that a certain
threshold, estimated to correspond to E2 concentrations in
the order of 500–1,000 ng/liter, be reached. Figure 1 indicates that
serum E2 exhibits wide individual variation in the early
stages of gestation, with the threshold range reached after as little
as 6 weeks or as long as 12 weeks in healthy, normally progressing
pregnancies. A final practical point to remember is that in women with
inherited partial TBG deficiency, estrogen stimulation associated with
pregnancy leads to variable modifications of TBG levels: no increase is
observed in some women, while in others TBG is increased compared with
preconception values, albeit to a much lesser extent than in women
without congenital TBG aberration (24, 25).
B. The thyroid hormones
1. Total thyroid hormones.
In pregnancy, the alterations in
total TH levels are the direct consequence of the marked increase in
serum TBG: total T4 and T3 levels increase
significantly during the first half of gestation. Levels of serum
T4 rise sharply between 6 and 12 weeks, progress more
slowly thereafter, and stabilize around midgestation; for serum
T3, the rise is more progressive (26). Both total
T4 and T3 reach their plateau values by 20
weeks and are maintained until term. Because of the 20-fold greater
affinity of TBG for T4 compared with T3,
changes in T4 levels follow the changes in TBG more
closely. It can be expected therefore that the
T3/T4 molar ratio should remain essentially
unaltered during pregnancy (27, 28, 29). Later in this review we will
discuss the importance of an increased T3/T4
ratio, as an indicator of thyroidal alterations due to iodine
deficiency during pregnancy.
These modifications represent the necessary adjustment from the
"old" (preconception) steady state equilibrium to the "new"
(gestational) equilibrium of the thyroidal economy. The changes are
initiated by the progressive expansion of the TBG extracellular pool,
which increases from 2,700 to 7,400 nmol over a trimester,
accompanied by a major increase in hormone-binding capacity of the
serum (8, 30). In the nonpregnant woman, approximately one third of
circulating TBG carries a T4 molecule; i.e.
the molar T4/TBG ratio is 0.35–0.40. To ensure homeostasis
of the free hormone concentrations during pregnancy, the extrathyroidal
T4 pool must increase in parallel (31, 32, 33). The thyroidal
adjustment therefore implies that, in the early stages of pregnancy, a
transient period takes place, during which T4 and TBG
concentrations are constantly changing.
This concept is fundamental to understanding the thyroidal alterations that are observed in pathological conditions such as iodine deficiency or hypothyroidism, characterized by the inability to achieve an adequate adjustment by the glandular machinery. Indeed, the adjustment of the thyroidal economy can be achieved only through a steady increase of T4 output by the gland during this period. To reach the new steady state, the hormonal output must steadily increase over a period of one trimester, with a constant daily enhancement over baseline T4 production values of 1–3% (34). When the new steady state has been reached, the overall production rate of TH should become similar to that prevailing before pregnancy (35).
How is the required thyroidal adjustment that takes place in the first
trimester of gestation regulated? Because the rapid rise in the serum
hormone-binding capacity due to increased serum TBG levels tends to
induce a trend toward slightly decreased free hormone concentrations,
the thyroidal adjustment is regulated primarily through the normal
pituitary-thyroid feedback mechanisms, i.e. by TSH
stimulation of the thyroid gland (Fig. 2). In healthy
pregnant women, the "extra load" on the thyroidal machinery is
relatively minor, and these physiological changes are unnoticeable: an
increase in serum TSH is not commonly observed. On the contrary, as
will be discussed later, in women with iodine deficiency or autoimmune
thyroiditis and subclinical hypothyroidism, the TSH surge is amplified,
and increases in serum TSH can be demonstrated, revealing the
underlying mechanisms of adaptation (36).
|
illustrates an interesting
study by Roti et al. (46) in which the authors compared
serum-free thyroid hormones in pregnant women at term and their
newborns, using ten different commercially available kits. The data
show the variability in free T4 and T3
concentrations obtained with different methods, but they also show that
free hormone levels are always significantly lower than in nonpregnant
women. Longitudinal studies based on reliable methodology
(i.e. methods that are not influenced by changes in serum
TBG and albumin levels) in large numbers of pregnant women without
iodine deficiency have confirmed that serum free T4 levels
are lower by an average of 10–15% at delivery, in comparison with
nonpregnant female subjects. Changes in free T3 levels
follow a parallel pattern. In most pregnant women, however, free
hormone levels are maintained within the nonpregnant reference range
(47, 48).
|
The calculation of free T4 indices (which is still very much in use in many countries) deserves a comment: these indices are established on the basis of the known physico-chemical properties of the thyroid hormone transport proteins, using the T4/TBG ratio or the T3 resin uptake test. One should remember that these estimations of free T4 concentrations from indirect calculations do not always provide reliable results in pregnancy. The free T4 index based on the T3 resin uptake test shows only small fluctuations in pregnancy, while the index based on the T4/TBG ratio yields values significantly lower than those found in nonpregnant women (23, 57).
3. Peripheral metabolism of thyroid hormones.
Three enzymes
catalyze the deiodination of thyroid hormones in human tissues (58).
Type I deiodinase, by outer ring deiodination of T4, is
responsible for the production of most of the circulating
T3. As already discussed, total T3 levels
follow, albeit less tightly, the increase in total T4
associated with the rise in TBG during the first half of pregnancy.
Furthermore, when the thyroid gland is more stimulated (such as in
iodine deficiency) during pregnancy, there is also preferential
secretion of T3 by the gland, presumably under the direct
influence of TSH. Concerning reverse T3 (formed by inner
ring deiodination from T4 or T4 sulfate as
substrates), maternal serum levels increase during pregnancy in
proportion to the increase observed for total T4 (59, 60).
Consequently, even though this has not been proven by direct evidence,
there is no argument to propose that the activity of type I deiodinase
should be altered in pregnant women. Type II deiodinase acts only on
the outer ring and prefers T4 and reverse T3 as
substrates. The enzyme is expressed in certain tissues (i.e.
pituitary gland, brain, brown adipose tissue) and also in the placenta.
Since the activity of type II deiodinase increases when the
availability of T4 decreases, it has been proposed that
type II deiodinase activity may represent a homeostatic mechanism for
maintaining T3 production in the placenta when maternal
T4 concentrations are reduced (i.e. during
hypothyroidism or iodine deficiency) (61). The placenta also contains
large amounts of type III deiodinase (62). This enzyme converts
T4 to reverse T3 and T3 to
T2. Placental type III deiodinase, by its extremely high
activity during fetal life, may explain the low T3 and high
reverse T3 concentrations, characteristic of fetal thyroid
hormone metabolism. The ontogeny of the three deiodinases in the
developing fetus involves complex metabolic pathways that are beyond
the scope of the present article. For detailed information, readers are
referred to two excellent and extensive recent reviews on this
important topic (50, 63).
Finally, elevated deiodination activity in the placenta probably plays an important role for the metabolism of maternal thyroid hormones. As was discussed earlier, the metabolic changes associated with the progression of gestation, in its first half, constitute a transient phase from a preconception steady state to a pregnancy steady state. The changes are accomplished through the gland’s increased hormonal output to reach and remain at the new equilibrium. Once the latter has been reached, one would expect the hormonal needs to revert to their initial levels. For instance, the increased daily dose of L-T4 necessary to maintain euthyroidism in hypothyroid-treated pregnant patients should only be transient; however, clinical experience clearly indicates that it is not the case. If the increased L-T4 dosage is not maintained in those patients during later stages of gestation, they rapidly become hypothyroid. This indicates that once the new steady state is reached, increased hormonal demands are maintained: this could be partially explained by factors such as transplacental passage of maternal hormones and increased turnover of T4 of maternal origin, due to the high activity of placental type III deiodinase. To date, the quantitative importance of changes in the peripheral metabolism of maternal thyroid hormones and the exact role of the placenta in this mechanism have not been fully elucidated.
C. The serum levels of thyroglobulin (TG)
Thyroglobulin is the protein matrix on which thyroid hormones are
synthesized in the thyroid gland. Even though the TG molecule has no
peripheral hormonal action, the serum levels of TG represent a
sensitive, albeit nonspecific, indicator of the activity or stimulation
state of the thyroid gland. Several studies have indicated that TG is
frequently elevated during pregnancy: the increase in TG can be
observed as early as the first trimester, but by later stages of
gestation and particularly near term is significantly more pronounced
(64, 65, 66, 67, 68). The alterations in serum TG associated with pregnancy were
first considered to result from transient thyroidal stimulation due to
the thyrotropic action of human (h) CG at the end of the first
trimester (69). This hypothesis is probably not correct because, as
indicated above, TG changes occur mainly in the late stages of
gestation (when hCG levels have decreased) and also because statistical
correlation between the increments in TG and peak hCG levels is lacking
(70).
Recently, TG changes in pregnancy have been investigated in greater
detail. These studies have revealed that the increase in TG is
correlated with other indices of thyroidal stimulation, such as slight
elevations in serum TSH (usually remaining within the normal range) and
an increase in the T3/T4 molar ratio,
suggesting preferential T3 secretion (34). Most
importantly, changes in TG are also associated with an increase in
thyroid volume (TV), and we have proposed that TG alterations may
constitute a sensitive biochemical marker to monitor the goitrogenic
stimulus frequently occurring during pregnancy in relation with iodine
deficiency (71). In the Brussels area, where the iodine intake is only
marginally low, between 50–100 µg/day, serum TG was found abnormally
elevated in more than 50% of women at delivery with values ranging
between 30 and 180 µg/liter, comparable to the TG values observed in
conditions of severe glandular stimulation such as GD (Fig. 4).
|
In pregnancy, the renal clearance of iodide increases significantly because of an increased glomerular filtration rate. Renal hyperfiltration and increased clearance, observed for iodide and several other molecules (both smaller and larger) begins in the early weeks of gestation and persists until term, thereby constituting an obligatory renal iodine "leakage" (72, 73, 74). The iodide loss tends to lower the circulating levels of inorganic iodide and induces, in turn, a compensatory increase in thyroidal iodide clearance, which reaches 60 ml/min and is accompanied by an absolute elevation of iodide entry into the gland (75, 76). These mechanisms indicate that the thyroidal activity is increased during pregnancy, as has been suggested by early studies using radiolabeled iodine administered to pregnant women, as well as histological studies of thyroid follicular cells obtained during pregnancy and showing marked functional activity (77, 78, 79).
A second mechanism of iodine deprivation in the mother occurs later in gestation, from the passage of a part of the available iodine from the maternal circulation to the fetal-placental unit. At midgestation, the fetal thyroid gland has already started to produce thyroid hormones that are indispensable for adequate development of the fetus (80, 81, 82). Hence, when iodine deprivation exists during the first half of gestation, it tends to become more severe in the final stages. The extent of the iodine passage from mother to fetus is not precisely established. Another interesting and unresolved question is the role of the placenta in transferring iodide: does it simply represent passive transfer or is there an "active pump" (83)? In the human the median urinary iodine excretion decreases by 10–15 µg/day in the second half of gestation compared with the first half, perhaps representing the fraction of iodide transferred (34). Since this difference has not been confirmed by other studies, it remains an open question for future work (84).
In the nonpregnant condition an adequate iodine intake is estimated to be 100–150 µg/day. Based on several studies, the consensus recommendation of the World Health Organization is that the iodine supply should be increased in pregnant and lactating women to at least 200 µg/day (85, 86). For pregnant women who reside in countries with an iodine-sufficient environment with an intake often more than 150 µg/day, the iodine losses in the urine and from transfer to the fetus are probably of little importance. Iodine deficiency disorders (IDD) do not present problems in the United States, Japan, or a limited number of countries in Europe (the Scandinavian countries, Switzerland, Austria), where national programs of dietary iodine supplementation have been in place for many years. In other areas of the world, however, IDD constitutes a serious public health issue (87). Available data indicate that 1 to 1.5 billion people are at risk of IDD. Among them, there are more than 500 million people who live in areas with overt iodine deficiency and a high prevalence of goiter.
Countries like Belgium, on the other hand, are representative of most European countries where systematic programs of dietary iodine supplementation have not been implemented and where the "natural" iodine supply is at, or below, the lower limit of adequacy. The average iodine intake in Belgium is below 100 µg/day (88). Inasmuch as there is no endemic goiter in the population, this restricted level of iodine intake is presumably sufficient to cover the usual needs of thyroid hormone production in normal adult subjects, at least as long as nothing intervenes to disrupt the fragile equilibrium. Pregnancy therefore acts as an indicator of the underlying iodine restriction by its increased hormonal demands and obligatory iodine losses, and gestation results in a relative iodine-deficient state. In countries with a more severe iodine deficiency, the repercussions of iodine deprivation during pregnancy are obviously further enhanced (89).
E. The hypothalamic-pituitary control of thyroid function and the
role of hCG
1. Hypothalamic-pituitary-thyroid axis (HPTA).
We have already
mentioned some arguments suggesting that elevated estrogen levels in
pregnancy may influence the HPTA, perhaps by acting directly at
different (and not yet clearly defined) levels in the thyroid gland
feedback-regulatory mechanisms. In his 1993 review in Endocrine
Reviews, Burrow (48) analyzed in detail the few available studies
in which the HPTA has been assessed, either by the administration of
T4 or T3 to pregnant women for short periods
with the aim of evaluating the TSH responses to induced hormonal
changes (75, 76, 90, 91) or after TRH administration (52, 92, 93, 94).
Unfortunately these studies, performed before 1980, employed the then
available assays which were unable to detect subtle serum TSH changes.
Overall, the conclusion drawn from this early work was that the
responsiveness of the HPTA can be considered to function normally in
pregnancy. If the above comments are interpreted with caution, we would
certainly agree with Burrow’s general conclusion (48).
2. Regulation of serum TSH.
A correct interpretation of the
modifications in serum TSH concentrations is crucial to correctly
assess the alterations in pregnancy-associated thyroid function
parameters. In earlier work, conflicting data have been reported: some
authors found no change in serum TSH in pregnancy (95, 96), while
others observed significant increases in TSH throughout gestation (93, 97). With the introduction 10 yr ago of sensitive immunoradiometric
techniques allowing for extremely precise determinations of TSH levels
within the normal range, new and important insights have been gained to
better define the patterns of serum TSH changes during pregnancy.
In the present review, by examining different periods during gestation, we shall address two main questions related to serum TSH alterations. During the first trimester when hCG levels reach their peak, is there a transient fall in basal and TRH-stimulated TSH?; if so, to what is the TSH decrease related and what is its clinical relevance? During the second half of gestation, do TSH levels remain stable (i.e. comparable to before pregnancy and also to before the hCG peak) or are there subtle but significant modifications in serum TSH? If the latter is true, what is the meaning of TSH changes in late gestation?
a. Transient fall in serum TSH in the first trimester.
The
first observation of a transient fall in serum TSH during the second
and third months of pregnancy in normal women was reported in 1976, and
the authors at that time postulated that TSH suppression might be
related to an intrinsic "TSH-like" activity of hCG. Unfortunately,
with the "crude" techniques available, the authors could not show a
reciprocal relationship between TSH and hCG levels in individual serum
samples, and they concluded that "it was unlikely that hCG alone was
responsible for the TSH suppression" (98). At that time, it was
commonly believed that the placenta produced large amounts of various
chorionic products, distinct from hCG, with thyroid-stimulating
activity. Among those, human chorionic TSH (hCT) was a favorite and it
was felt that hCT, alone or in concert with hCG, was responsible for
the biological thyrotropic activity observed (99). A few years later,
however, convincing evidence indicated that hCT was not a significant
factor as a thyroid-stimulating agent and that peak hCG levels in
normal pregnant women coincided with an important increase in the
bioassayable thyroid-stimulating activity (100).
The basis for these early studies on thyroid stimulators of placental origin stemmed from the clinical observations in the 1970s of an association of hyperthyroidism with molar pregnancy (101, 102, 103). It has since been amply confirmed that in various pathological conditions, such as molar pregnancy (104, 105), other trophoblastic disease (choriocarcinoma) (106, 107, 108, 109), and cancers of various origins (110, 111, 112, 113), elevated hCG levels could induce hyperthyroidism, characterized by the rapid appearance of thyrotoxic symptoms and their even more rapid disappearance after the surgical removal of the mole or cure of the tumor. Taken together, these observations have led to the concept that a substance secreted during pregnancy, and at particularly high levels in moles and choriocarcinomas, could be responsible for hyperthyroidism. Based on physico-chemical analyses of molar or tumor extracts, it was then shown that the thyroidal stimulator most probably was hCG (114, 115, 116). It was also suggested that the thyroid-stimulating effects found in these pathological circumstances could be due not only to the extremely high circulating hCG levels, but perhaps also to the presence of molecular variants of hCG with particularly potent thyrotropic activity (117, 118, 119).
To date, there is a bulk of compelling evidence to indicate that there
is indeed a transient fall in serum TSH near the end of the first
trimester in normal pregnancy, and that this partial TSH suppression is
associated with the elevation in circulating hCG. In 1985, Guillaume
et al. (120) reported a significant blunting of the TSH
response to TRH in six women who had higher hCG levels (64,000
IU/liter) at the end of first trimester, compared with 19 other
pregnant women with a similar gestational age, in whom the TSH response
to TRH was unaltered and hCG levels were comparatively lower (45,000
IU/liter). In 1988, Pekonen et al. (121) showed a negative
correlation between hCG and TSH levels in a small group of pregnant
women investigated immediately before and after abortion. These authors
were the first to demonstrate clearly, at the level of the individual,
a decrease in serum TSH associated with high hCG values. In our
prospective studies on maternal thyroid function in pregnancy, the
regulatory role of hCG was first investigated in a cohort of several
hundred women in whom TSH and hCG levels were systematically determined
between 8–14 weeks gestation (34). The results showed that a lowering
in serum TSH was coincident with the peak hCG values (Fig. 5). The profiles of changes in serum TSH and hCG were
clear mirror images, and there was a significant reciprocal correlation
between TSH and hCG in individual samples. The results also indicated a
linear relationship between hCG and free T4 concentrations
during early gestation. Thus, the lowering of TSH corresponds to a
transient and partial blunting of the pituitary-thyroid axis associated
with an increased hormonal output by the thyroid gland. From these
preliminary observations, we concluded that hCG is a thyroid regulator
in normal pregnancy (3, 34). Similar conclusions were reached by
Ballabio et al. (122), who proposed that hCG be considered
"a putative physiological regulator" of maternal thyroid function
in normal pregnancy.
|
) showed that in the first trimester, 18% of the women
had a transient subnormal serum TSH. Moreover in almost half of these
cases, serum TSH was transiently undetectable (<0.05 mU/liter). The
lowering in serum TSH was still observed in 5% of the women during the
second trimester. Women with a blunted TSH level (<0.20 mU/liter) had
a circulating hCG concentration significantly higher than in women with
normal TSH levels. Confirming our previous estimates, women with a
blunted TSH displayed, on average, hCG levels above 50,000 IU/liter.
Among the 62 women with a transient suppression of TSH in the first
trimester, seven (11%) had free T4 levels that were
transiently increased above the upper limit of normality; among the
latter, three women presented symptoms compatible with hyperthyroidism
associated with significant vomiting. In this cohort, it should also be
noted that nine women with a twin pregnancy were seen in the first
trimester, with five of them having transiently undetectable serum TSH
levels and two of five women exhibiting overt gestational
thyrotoxicosis. Thus, some women may have sufficient thyroid
stimulation to cause transient hyperthyroidism, as will be discussed in
detail later (see Sections III.C.2 and III.C.3)
|
we have collected the serum TSH values reported in 18 studies between
1969 and 1994, in which the authors have examined longitudinal changes
in serum TSH according to the trimesters of pregnancy. A note of
caution is worthwhile because of the inherent difficulties in the
analysis and comparison of such data. First, the techniques employed
for the determination of TSH in serum have obviously evolved
tremendously in quality and precision over the span of 25 yr. Second,
the number of pregnant women investigated in each study was extremely
variable, from small numbers to large groups. Third, the actual TSH
values were, in some cases, presented in a graph rather than being
available directly in the articles; it was therefore necessary in some
cases to recalculate the actual TSH values, as precisely as possible,
from the original reports. Fourth, it was not always clearly stated
whether the women investigated and presented as "normals," were
actually normal. For instance, the presence of a goiter and/or thyroid
antibodies was not systematically evaluated, and most studies did not
specify whether such women had been excluded from the reports. Finally
and most importantly, in many of the studies no comments were made
concerning iodine intake, and we can only infer from the authors’
geographical locations whether the study was performed in an
iodine-replete or -restricted area. Regardless of these inescapable
limitations, the table still yields interesting information. First,
nine of 18 studies clearly report an increase in serum TSH, within the
normal range, near term. Second, the majority of studies that show
serum TSH increases during late gestation originate from Europe: they
can therefore be considered to have been performed in areas with a
restricted iodine intake. In a few studies this question was addressed
directly. More specifically, both the Belgian (34) and Danish (131)
studies, which showed a systematic increase in serum TSH near term,
were carried out in well defined areas with a marginal iodine
deficiency. As will be discussed later, the same investigators also
showed that iodine supplementation during pregnancy significantly
prevented these alterations in serum TSH. Conversely, a Dutch study
(132), which failed to show a difference in serum TSH between the third
trimester and nonpregnant controls, was performed in an iodine-replete
area.
|
3. Thyrotropic action of hCG.
The thyrotropic action of hCG is
explained by the structural homology between the hCG and TSH molecules,
and between LH/CG and TSH receptors. Thus, hCG is able to bind to the
TSH receptor of thyroid follicular cells and exerts its stimulatory
effects by activating intracellular messengers, such as cAMP (133).
a. In vivo effects of hCG.
The role of hCG in regulating
maternal thyroid function in the first trimester of pregnancy has
already been discussed. The thyroid gland of normal pregnant women may
be stimulated by elevated circulating hCG levels to transiently secrete
slightly more T4 and induce in turn a partial suppression
of serum TSH. In up to one fifth of normal pregnancies, serum TSH may
be transiently suppressed in the first trimester to values below the
lower limit of normal.
An interesting question is whether it may be possible to distinguish,
among normally progressing pregnancies, those women who are prone to
blunt their serum TSH in the first trimester in response to the
increase in circulating hCG. We approached this question in two
clinical studies. In the first, the serum concentrations of intact hCG
and its free 
- and ß-subunits were measured in two groups of
normal pregnant women from the same cohort, subdivided on the basis of
whether or not they had a partially suppressed serum TSH (below 0.20
mU/liter) in the first trimester (Fig. 7). The results
showed that a low serum TSH was associated with significantly higher
levels of both intact hCG and free ß-hCG subunit, whereas there was
no significant difference in free -hCG subunit concentrations.
Furthermore, in women with a low serum TSH and high hCG production,
there was also a 20% increase in mean free T4 levels
during the first trimester (70). The hCG-induced stimulatory effects on
the maternal thyroid gland were transient inasmuch as the parameters of
thyroid function were similar in both the intially low TSH and the
normal TSH groups during the last trimester and at term. In the second
study, our aim was to define more precisely the quantitative
relationships between circulating hCG and thyroidal stimulation in the
first trimester. The levels of hCG, TSH, and free T4 in
early gestation were investigated in two groups of euthyroid women with
single or twin pregnancies in whom the gestational age was precisely
known because conception was obtained by in vitro
fertilization techniques (Fig. 8). Results showed that
peak hCG values in twin pregnancies were not only significantly higher
than in single pregnancies (in fact, almost double), but also of much
longer duration. Serum hCG values above 75,000 IU/liter lasted for less
than 1 week in single pregnancy, while up to 6 weeks in twin pregnancy.
Concerning the thyroidal repercussions, twin pregnancy was associated
with a more profound and frequent lowering in serum TSH (blunted TSH
values below 0.20 mU/liter were observed 3-fold more frequently). Also,
free T4 values remained unchanged in single pregnancy while
transiently rising in twin pregnancy (134).
|
|
b. In vitro effects of hCG.
Highly purified hCG increases
iodide uptake and cAMP production and induces growth in rat FRTL-5
thyroid cells (135, 136, 137). Recently, it has also been confirmed that
purified hCG interacts in vitro with the human TSH receptor,
thereby stimulating the human thyroid gland (138, 139, 140). Similarly,
serum of pregnant women has been shown to exert a thyroid-stimulating
activity in vitro (141). Thus, there is presently good
evidence that the effects of hCG reported in vivo correspond
to a TSH receptor-mediated thyroid-stimulating action in
vitro (142).
TSH is a glycoprotein hormone composed of two subunits linked together
to form the intact heterodimeric active molecule (143). The TSH
receptor located on the surface of thyroid epithelial cells belongs to
the family of receptors coupled to G proteins. The structure of the TSH
receptor has been identified and consists of three domains, a long
extracellular domain representing the N-terminal part of the molecule,
a transmembrane-spanning domain of seven peptides joined by intra- and
extracellular loops, and finally an intracellular C-terminal domain
coupled to the G proteins complex (144, 145). To explain the
thyrotropic effects of hCG, it is necessary to compare the structures
of the hCG molecule with our present knowledge of the TSH molecule and
its receptor. As in the case of TSH, hCG is also composed of two
noncovalently linked subunits. The -subunit is common to all members
of this family of hormones, whereas it is the ß-subunit that confers
its specificity to hCG (146, 147). There is a high structural homology
between the ß-subunits of hCG and TSH. As is the case for the
hormones, there is also a considerable homology between the LH/hCG and
TSH receptors (Fig. 9). The homology reaches 70% for
the transmembrane-spanning domains and 45% for the extracellular
domains of the receptors where the hormones bind (144, 148, 149). These
molecular homologies are now part of a novel endocrine concept,
referred to as "spill-over" syndromes (142).
|
). Human CG activates the same domain of the TSH
receptor as does TSH by a spill-over mechanism related to molecular
mimicry. It remains possible, e.g. in pathological (tumoral)
conditions, that variant hCG molecules or metabolites of circulating
hCG, exhibiting a more potent thyrotropic action, may play a role;
until now, however, there has been no clear evidence that such variants
or metabolites are involved in the hCG effects observed in normal
pregnancy.
|
). The same events underlie the pathological
repercussions, in both mother and offspring, when pregnancy occurs in
iodine-deficient areas (36, 49, 150, 151). These specific circumstances
will be discussed in detail in the second part of this review.
|
|
III. Pathological Alterations of Thyroidal Regulation Associated with Pregnancy |
|---|
|
TopAbstractI. IntroductionII. The Regulation of...III. Pathological Alterations of...IV. Conclusions and PerspectivesReferences |
|---|
In Europe where, in the majority of countries, there is usually only a moderate iodine restriction, pregnancy in otherwise healthy women is often associated with goitrogenesis but rarely with hypothyroidism. In other regions of the globe, with a more severe iodine deficiency, however, both maternal and neonatal hypothyroidism is frequently encountered, endemic cretinism representing the most dramatic expression of these alterations.
1. Consequences of iodine deficiency during pregnancy.
In most
European countries, populations do not benefit from a systematic
addition of iodine to the diet, and there is good and recent evidence
that nutritional allowances for an adequate daily iodine intake,
unanimously recommended by international agencies such as the United
Nations International Childrens Emergency Fund, International Council
for the Control of Iodine Deficiency Disorders, and World Health
Organization are far from being fulfilled: IDD persists and still
constitutes a serious public health hazard (85, 86, 87). In regions with a
marginally low iodine supply, it is particularly difficult to reach
firm conclusions concerning the adequacy of iodine intake, mainly
because important fluctuations occur in daily intake, both among
individuals and also from one day to another. Measuring urinary iodine
excretion levels reflects only the iodine intake of the most recent
previous days. What really matters, however, is the long-term iodine
balance, which determines the extent of intrathyroidal iodine stores.
In populations with a chronically reduced iodine supply, it is the
decreased availability of iodine that allows a better understanding of
thyroidal alterations associated with pregnancy, because borderline
iodine nutrition levels lead to increased thyroidal stimulation.
As a representative example of European countries, the average iodine
intake in Belgium is limited to between 50–100 µg/day. Figure 12 illustrates urinary iodine excretion levels
determined in pregnant women without iodine supplementation from the
Brussel’s area, showing that 85% of them have iodine intakes clearly
below the recommended amount. As a consequence, pregnancy acts to
reveal the underlying iodine restriction, and gestation results in a
state of increased relative iodine deficiency (153, 154).
|
2. Assessment of increased thyroidal stimulation.
Since the
early 1990s, the concept was developed that increased thyroidal
stimulation resulting from iodine restriction may lead to goiter
formation during pregnancy. Hence, pregnancy should be regarded as an
additional factor during a woman’s life (an event that may obviously
be repeated at short intervals) that may induce thyroidal pathology
when iodine intake is marginally low. It is therefore important that
clinicians correctly assess increased thyroidal stimulation (34, 160).
In practice, four simple biochemical parameters have been identified
and proven to be useful markers.
The first parameter is relative hypothyroxinemia. As already discussed, free T4 levels tend to decrease slightly, even in pregnant women who have an adequate iodine supply. In women with iodine restriction, however, the early rise in total T4 (associated with the rise in TBG) was shown to be inappropriately low, with free T4 and T3 levels progressively decreasing during the first part of gestation to stabilize at a low level (with an average decrement of 30%) in the second part of gestation (34, 131, 161). Under the environmental conditions that we investigated in Brussels before iodine supplementation was systematically introduced during pregnancy, it was observed that one third of pregnant women had free T4 values near or below the lower limit of normal (34). It was also shown that there was a tendency for individuals to exhibit variable patterns of glandular adaptation. For instance, a woman whose serum free T4 was already in the lower tertile of the population’s range during early gestation had a greater than 80% risk of remaining in the lower part of the range during late gestation. Conversely, a woman with a serum free T4 in the upper part of the population’s range during the last months of pregnancy had a greater than 90% chance of having a serum free T4 in the upper part of the range in early gestation, indicating that in this case thyroidal adaptation had taken place during the first trimester (32). That relative hypothyroxinemia was truly related to iodine restriction was confirmed by its partial correction when iodine supplementation was administered early enough during gestation (89, 123, 131).
The second parameter is preferential T3 secretion, reflected by an elevated molar ratio of total T3/T4 in serum. It was mentioned previously that, owing to differences in the respective binding affinities of TBG for T4 and T3, the T3/T4 ratio tends to remain unchanged during pregnancy. Under conditions of a normal iodine intake, the serum T3/T4 ratio ranges between 10–22 (x10-3) in euthyroid pregnant women (28, 59, 124, 162). In clinical and experimental conditions in which there is an increased stimulation of the thyroid gland, e.g. in GD (163) or after acute TSH stimulation (164), the T3/T4 ratio increases as the result of preferential T3 production by the gland. The T3/T4 ratio also depends upon the extent of iodine depletion (i.e. a small intrathyroidal iodine pool) and has been shown to be useful for evaluating the degree of thyroidal stimulation in endemic iodine deficiency (165).
In the pregnant women that we investigated in Brussels, the T3/T4 ratio was significantly increased and remained elevated throughout gestation in women without iodine supplements, whereas the administration of iodine was accompanied by a lowering of the ratio. In our experience, however, iodine supplements given alone (from the 15th week of gestation onward) were not sufficient to normalize the T3/T4 ratio, an indication that the intrathyroidal iodine pools remained relatively deprived, probably because the iodine supplements were used immediately for thyroid hormone production, rather than stored (166). It is also of interest to note that after parturition in untreated pregnancies, recovery of normal thyroid function may take months: at 6 months postpartum the ratio of T3/T4 was still elevated (167). These results suggest that the thyroidal alterations associated with pregnancy in iodine-restricted conditions not only persist after term, but may also have long lasting stimulatory effects on the thyroid gland, a consideration that may help explain why features of excessive glandular stimulation are frequently observed again in the same individuals in subsequent pregnancies, especially when the interval between pregnancies is brief.
The third parameter is related to changes in serum TSH. It was already mentioned that iodine restriction is associated with a significant increase in serum TSH after the first trimester. A progressive increase in serum TSH, until term, is observed in more than 80% of pregnancies under iodine-restricted conditions. Serum TSH changes usually remain within the normal range in women who are otherwise healthy. Albeit of relatively small amplitude, these modifications are statistically highly significant, with median TSH concentrations increasing from 0.75 mU/liter in the first trimester to 1.09 in the second, 1.28 in the third, and 2.08 mU/liter at term in Brussels (2, 34). Hence, serum TSH more than doubles during pregnancy when the iodine supply is limited, a clear indication of a sustained thyrotropic stimulation of the thyroid gland. At 6 months postpartum, it was observed that serum TSH levels had generally reverted to pregestational values (167). In comparison, in women who received iodine supplementation during pregnancy, the increment in serum TSH was markedly diminished by 50% or more at term (131, 166).
In areas with severe iodine deficiency such as in Ubangui (Republic of Zaïre), TSH modifications during pregnancy are not restricted to the normal range and are of a much greater amplitude. In such areas, maternal TSH values were found to exceed 100 mU/liter in some women at the time of delivery, confirming the intensity of chronic thyroidal stimulation (168). In comparison, pregnant women from the same villages, who received 1 ml of iodized oil in the second trimester of gestation, had significantly lower mean serum TSH values at delivery, never exceeding 20 mU/liter.
The fourth parameter is related to the changes in serum TG levels. It
was already mentioned that serum TG is frequently elevated in
pregnancy, particularly during the late stages of gestation near term
(34, 68, 160, 169, 170). An illustration that thyroidal stimulation is
associated with increased TG concentrations is given in Fig. 13. When we investigated pregnant women (selected
because they displayed increased thyroidal stimulation), who were given
or not given iodine supplements, a linear relationship was demonstrated
between the increments in serum TG and TSH: without iodine
supplementation, the relative increment in TSH reached 100% at term
(compared with values in the first trimester) and was associated with a
60% relative TG increment. Conversely, with iodine supplementation, TG
concentrations remained unchanged or even decreased. Moreover, in a
group of pregnant women who received a combination treatment (iodine +
L-T4) during pregnancy, initially elevated TG
levels not only decreased but normalized rapidly, in concomitance with
a reduction in TSH concentrations (166).
|
In summary, relatively simple laboratory tools and standardized criteria can be used to assess excessive thyroidal stimulation, based on the routine determination of serum total T4 and T3, free T4, TSH, and TG levels. Better understanding of the complex mechanisms that intervene to regulate thyroid function during pregnancy and the deviations from physiological adaptation observed in iodine-deficient conditions may be very valuable in assessing the alterations of thyroidal economy associated with pregnancy and also in monitoring their therapy and prevention.
3. Gestational goitrogenesis and its prevention by iodine
supplementation.
Several investigations have been carried out in
Europe in recent years to evaluate the modifications in TV associated
with gestation. Together these studies have amply confirmed the
original observations by Crooks et al. (171), who reported
as early as 1967 (in those early days employing palpation) a striking
difference in the incidence of goiter in pregnant women between
Aberdeen, Scotland (area of lower iodine intake) and Reykjavik, Iceland
(area of higher iodine intake) (171). The authors observed that the
incidence of gestational goiter was 3-fold greater in Scotland compared
with Iceland, and that it doubled during pregnancy in the lower, while
remaining virtually unchanged in the higher, iodine area.
Table 2 summarizes seven recent European studies in
which TV modifications associated with pregnancy have been evaluated
precisely, employing thyroid ultrasonography. In Finland (172) and
Ireland (84), where the iodine intake is considered adequate, the
increment in TV was small, on average 10–15%: such changes are
probably consistent with vascular swelling of the thyroid gland
("intumescence") during pregnancy. In Belgium (34) and Denmark
(130), areas with a restricted iodine intake, the increment in TV was
greater, reaching 25% on average. From our work, it became evident
that the size of the thyroid gland increases significantly when
pregnant women are not supplemented with iodine: an increase in TV was
observed in more than 80% of the women investigated and took place
gradually with increasing gestation time. Even though the increment in
TV, given as an average, may not seem spectacular, it is important to
consider the wide individual variation in TV modifications, with many
women exhibiting a doubling in TV at term (34, 49, 160). Moreover in
our experience, almost 10% of the women developed a goiter during
pregnancy (i.e. TV > 22 ml by ultrasonography), and
volumetric changes in the gland were associated with clear biochemical
evidence of increased thyroidal stimulation, hence strongly suggesting
that pregnancy truly induces goitrogenesis.
|
) (n = 13) pregnancies. Each point corresponds to the
mean ± SD of individual serum samples obtained at
each gestational age. Statistical differences were calculated using the
nonparametric Mann-Whitney rank test (*, P < 0.05;
**, P < 0.01; ***, P <
0.001). [Modified with permission from J. P. Grün et
al.: Clin Endocrinol (Oxf), in press (134). ©
Blackwell Science Ltd.]
