| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Special Articles |
Departamento de Endocrinología, Instituto de Investigaciones Biomédicas "Alberto Sols," Consejo Superior de Investigaciones Científicas and Facultad de Medicina, Universidad Autónoma de Madrid, 28029 Madrid, Spain
Address correspondence and requests for reprints to: G. Morreale de Escobar, Instituto de Investigaciones Biomédicas "Alberto Sols," Laboratorio B-16, Arturo Duperier, 4, 28029 Madrid, Spain. E-mail: gmorreale{at}iib.uam.es
 |
Abstract |
|---|
 Top Abstract IntroductionMaternal thyroid hormone...Information obtained in...A unifying hypothesis for...References |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaternal thyroid hormone...Information obtained in...A unifying hypothesis for...References |
|---|
As a consequence of the report by Haddow et al. (1) and other (3) recent studies, the possibility of avoiding potentially preventable alterations of neuropsychological development by screening pregnant women for maternal thyroid hormone deficiencies is being discussed. Even before this new evidence was presented relating maternal thyroid status with mental development of the child, screening for thyroid disfunctions during pregnancy had already been advocated because of their frequency and risk for the mother’s health and for the outcome of pregnancy (4).
Because such programs are not easily implemented, we believe the present comments may be pertinent to obtain the maximum benefits for the largest number of children.
|
Maternal thyroid hormone deficiencies and fetal neurodevelopment |
|---|
|
TopAbstractIntroductionMaternal thyroid hormone...Information obtained in...A unifying hypothesis for...References |
|---|
We will review, albeit in very condensed form, relevant information obtained from epidemiological, clinical, and basic science studies that strongly suggest that maternal thyroid status, especially in early pregnancy, is causally related to the survival and neuropsychological development of the offspring.
Reports from human populations with severe iodine deficiency (ID). After prolonged observations of cretins born in areas with endemic goiter, an important role of maternal thyroid function was suspected "it is of the greatest importance to inquire into the ante-natal history of all backward children and to examine the mother for thyroid defect" (5).
Not until 1965, however, were data presented that suggested that the
inability of the mothers to increase their low circulating
T4 during pregnancy was causally related to the
birth of cretins (6). Thereafter, numerous groups (see
Table 1
) presented convincing evidence
that maternal hypothyroxinemia early in pregnancy is not only the cause
of reproductive failure and the birth of neurological cretins, but also
of less severe mental deficits that affect a large proportion of the
apparently "normal" (noncretin) population of the same area. These
deficits, as well as the birth of cretins, are irreversible
consequences of the ID and can only be prevented within the first
months of gestation with an adequate supply of iodine. This finding is
in conceptual agreement with the very early (first trimester)
development of cerebral structures that are characteristically affected
in neurological cretinism. Two other findings are quite important in
the present context: namely, that the motor and cognitive impairment of
the progeny was correlated with the degree of maternal
hypothyroxinemia, and not with circulating T3 or
TSH levels; and that these hypothyroxinemic women are not clinically
hypothyroid, because of their relatively normal circulating
T3.
|
summarizes relevant information from
studies carried out in areas without severe ID, which have suggested an
important role of maternal T4 in the outcome of
the pregnancy and in the neuropsychological development of the progeny,
and which have provided increasing evidence that is the critical factor
is maternal hypothyroxinemia early in gestation.
|
Table 2
also includes reports regarding the increased proportion of
unsuccessful or complicated pregnancies in hypothyroxinemic women
(i.e. spontaneous abortions, premature births, major
complications at delivery, perinatal deaths, congenital malformations)
and their relation to first trimester low free
T4 (FT4), but not to
FT4 near term; early treatment to prevent such
complications has been stressed (22).
During the last decade, Pop et al. (23) have drawn attention to an average impairment of 10.5 IQ points in the offspring of mothers with high thyroid peroxidase antibody (TPO-Ab) titers during pregnancy. They have later reported (3) that in normal pregnancies a maternal FT4, at 12 weeks of gestation, which is equal to or less than the 10th percentile of first trimester values (10.4 pmol/L in their series), is associated with distinctly impaired psychomotor infant development at the age of 10 months, whether or not TSH and TPO-Abs were elevated. They have confirmed this is an ongoing prospective study (24), comparing developmental indices, determined at 3 weeks and 1 and 2 yr of age, with maternal FT4. Smit et al. (25) have found a similar relationship between first trimester FT4 and the early neurodevelopment of children born from treated hypothyroid women. These studies (3, 25) did not find significant correlations when the neurodevelopmental outcome was related to maternal TSH, or to FT4 later in gestation.
The finding that FT4 levels after midgestation are not correlated with the developmental outcome does not mean that after midgestation the protective effect of maternal T4 has been lost, or that the fetal brain no longer requires thyroid hormone. Prompt postnatal treatment of athyrotic babies prevents the severe mental retardation that was associated with this condition when onset of treatment was delayed (30). This very positive result of Neonatal Screening Programs has been quoted as argument that the human fetal brain does not require thyroid hormone for normal development before birth (31). But, an alternative explanation is that the fetal brain has been protected by the normal thyroid status of their mothers. Indeed, when maternal thyroid status is not normal throughout gestation and fetal thyroid function is also impaired, neurological damage and mental retardation of the child can be as severe as in neurological cretinism, despite immediate postnatal treatment of the baby with T4 (17, 32, 33, 34, 35).
Reports of poor developmental outcome in many babies faced with a premature interruption of the maternal supply of thyroid hormone—occurring when their thyroid is still quite immature (36)—also indicate that the fetal brain needs thyroid hormone throughout gestation and that a normal supply of maternal T4 has an important protective role after midgestation. The data from two such studies (37, 38) show a direct relationship between the degree of neonatal hypothyroxinemia and future neurodevelopment of the progeny. Both neurological dysfunction at 5 yr of age and school failure at 9 yr of age were significantly related to lower T4 levels during the neonatal period (37). Even after correction for other perinatal confounding factors, there was a statistically significant 30% increase in the odds of neurological dysfunction and school failure for each decrease of the T4 level by 1 SD. In the second study (38), the odds ratio for increased cerebral palsy in preterm infants with neonatal hypothyroxinemia increased more than 4-fold (after correction for many possible confounding factors), with a mean reduction of 7 points in the mental developmental score at 2 yr of age. The postnatal FT4 of premature babies, lower than that of fetuses of comparable age still in utero, is usually not accompanied by an elevation of circulating TSH (39), and a correlation is not found between neurodevelopmental outcome and neonatal TSH levels. Thus, in these babies, the critical factor for the prediction of future neurodevelopmental problems is the degree of their neonatal hypothyroxinemia, and not hypothyroidism, as detected by elevated circulating TSH. The degree of the hypothyroxinemia not only depends on gestational age and immaturity of the thyroid gland of the newborn, but is also increased by a low iodine intake (40). There is, however, no information regarding a possible relationship with the thyroxinemia of their mothers up to birth, but it seems plausible that maternal hypothyroxinemia, especially when related to ID, would only increase the difficulties encountered by the newborn in meeting postnatal hormone requirements, including those of the developing brain.
Information from basic research. The possible protective role
of the maternal transfer of thyroid hormone to the fetus was actually
not actively disputed until the mid-1960s and early 1970s
(41, 42, 43). However, during the following decades the
general consensus became that the placenta is virtually impermeable to
the iodothyronines and that the small amounts possibly transferred
would be of no physiological importance, either in health or disease,
ideas that were supported by experiments involving sheep. The
prevailing idea that there is no biologically significant transfer of
thyroid hormones throughout gestation (including the first trimester)
and that these hormones would not be required for normal early (or
late) fetal brain development is likely to have contributed to a
general lack of understanding and acceptance of the results from the
epidemiological and clinical studies summarized in Tables 1 and 2. A
convincing explanation could not be proposed for the relation between
early (mostly first trimester) maternal hypothyroxinemia and poor
neurodevelopmental outcome of the child. It was difficult to reconcile
the very severe, and mostly irreversible, neurological damage of
cretins (with normal thyroid function when iodine is supplied), with
the successful prevention of severe brain damage by early postnatal
treatment of congenital hypothyroidism (CH) babies, including
athyrotics. The lack of protective effects of maternal
T3, compared with T4,
observed in areas with severe ID, also posed many unanswered
questions.
New findings, mostly obtained during the last 15 yr in experimental
animals and in man, now clearly point to a role for the maternal
transfer of thyroid hormone in brain development throughout fetal life
and offer plausible explanations to the above questions. Despite these
new findings and despite increasing acceptance of an important role for
maternal T4 in the prevention of severe brain
damage in cases of CH, it is still stated that "Whether thyroid
hormone is needed during the 1st trimester is less certain. If it is,
it must be supplied by the mother, because none is secreted by the
fetus until the middle trimester" (44). Because this
point is essential for an understanding of the possible importance of
first trimester FT4 in fetal neurodevelopment, we
will briefly review in Table 3 much of
the information now available for experimental animals and for man.
|
|
Information obtained in experimental animals |
|---|
|
TopAbstractIntroductionMaternal thyroid hormone...Information obtained in...A unifying hypothesis for...References |
|---|
After onset of FTF. The maternal to fetal transfer of thyroid hormone is not interrupted and continues to contribute to the thyroid hormone available to fetal tissues at term. Of special interest in the present context is the finding that a normal level of maternal T4 is sufficient to protect a hypothyroid fetus preferentially from cerebral T3 deficiency until birth. It is important to realize that maternal T4 and T3 are not equivalent with regard to this preferential protection of the hypothyroid fetal brain from T3 deficiency. Without correction of the low maternal T4, normal levels of T3 in the maternal or fetal circulation have no protective effect because during fetal and postnatal development cerebral structures of the rat depend entirely on the local generation of T3 from T4 by type II 5'-iodothyronine deiodinase (D2), the activity of which is inversely related to the availability of T4. Changes in the activity of 5-iodothyronine deiodinase (D3), which inactivates both T3 and T4, also play a role. During these periods of development, the contribution of systemic T3 to the amount of T3 in fetal cerebral structures is negligible. Fetal brain T3 levels are also protected from excessive maternal circulating T4, whereas cerebral T3 homeostasis is not ensured when maternal circulating T3 is excessive. Such results suggest that over-treatment of the mother with T4 is potentially less damaging for the fetal rat brain than maternal hypothyroxinemia.
There is increasing evidence that the ontogenically programmed expression of D2 and D3, and the responses of these enzyme activities to thyroid hormone deficiency or excess, are the main mechanisms involved in the attainment of adequate concentrations of T3 in different cerebral structures at different stages of development (54, 55), despite the fact that at each time point the circulating concentration of T3 reaching all tissues is the same. This seems to be a general principle during development, even for other tissues and species (56, 57, 58). There is also an increasing consensus that an excess of thyroid hormones, especially T3, may have adverse effects on the developing fetus and that throughout gestation the maternal-fetal unit has numerous mechanisms to avoid this. This would explain the high expression and activities of D3 in many structures of the maternal-fetal unit (58, 59, 60, 61, 62), especially the uterus and placenta (63). Sulfation of the iodothyronines, their deiodination, and further desulfation by sulfatases also seem to be playing important roles in tailoring the amounts of T3 to changing temporal and spatial requirements during development (64, 65).
The abundant information obtained in pregnant sheep and newborn lambs has been reviewed by others (64, 66, 67, 68, 69). Most of the available information has been obtained after midgestation, corresponding to late fetal and early neonatal brain development in man, and it supports the existence of numerous mechanisms that would regulate the amounts of circulating thyroid hormone in the fetus and newborn (69, 70, 71). In this species, placentation is epitheliochorial, quite different from that of man and of the rat. Despite this, limited transfer does exist in this species, as shown by the presence, before onset of FTF, of T3 in the fetal brain, some of which is bound to the nuclear TR (72, 73). As in the rat, the hypothyroxinemia of iodine-deficient ewes before midgestation is accompanied by a reduced brain weight and changes in fetal brain morphology (68).
In summary, experimental findings, mostly obtained in the rat, provide an explanation for the lack of major irreversible brain damage at birth in the case of a human CH fetus from a normal mother. They also explain the early irreversible damage caused by ID, when both the mother and fetus are hypothyroxinemic throughout gestation (74, 75, 76), as well as the mechanisms involved in the preferential protective role, for the fetal brain, of T4 over T3.
Information obtained in man. Although man has an hemochorial
placenta similar to that of the rat, transfer of thyroid hormones from
mother to embryo and fetus may be quite different in both species.
However, there are many findings that suggest that conclusions derived
from experiments in the rat may be relevant to our understanding of
early human brain development. These similarities are in Table 3.
Mechanisms preventing excessive amounts of thyroid hormones from
reaching the fetal circulation are also operative in the human
maternal-fetal unit. They involve both deiodinating enzymes,
sulfotransferases, and sulfatases (64, 65, 115, 116, 117, 118, 119, 120).
Despite these mechanisms, maternal iodothyronines do reach the fetus.
Thyroid hormones, specifically T4, are already
available to embryonic and fetal tissues before the onset of FTF
(defined as the onset of secretion of the iodothyronines by the fetal
thyroid). In man this occurs at midgestation (
18–22 weeks),
coinciding with the development of the pituitary-portal vascular
system. T4 is found in first trimester coelomic
fluid, which bathes the yolk sac, from the earliest date studied,
namely at 6 weeks gestational age, in concentrations that correlate
significantly with maternal circulating levels (83),
whereas T3 is barely detectable. Although
coelomic fluid T4 concentrations are low compared
with those in adult blood, the FT4 concentrations
are comparable with those that are biologically effective in adults
(84). T3 has been quantified
in purified extracts from human fetal brain as early as 9–10 weeks
gestation (93, 94), and by midgestation the concentration
of T3 in the fetal brain reaches 34% of adult
values and is, therefore, much higher than the level previously
inferred from the very low circulating fetal T3
[<10% of adult values (121)]. This cerebral
T3 is likely to have been generated locally from
coelomic fluid T4 of maternal origin. Brain D2
and D3 may already have important roles in the human fetal cortex
before midgestation, with enzyme activities responding to maternal
hypothyroxinemia (98).
Nuclear TRs are present in the brain of 10-week-old fetuses, increasing rapidly by 16 weeks, a period of very active cortical neurogenesis (93, 94). TR occupancy by T3 was 25% throughout this period (94). As a result, the number of T3-occupied TR units in the whole fetal brain increases about 500-fold between 10 and 18 weeks, a finding that confirms that maternal thyroid hormone does reach the human brain early in gestation. It has later been found that TRs are already expressed by 8 weeks of gestation, with ontogenic changes specific for brain regions and receptor isoforms (95).
As in the rat, maternal to fetal transfer of thyroid hormones continues until birth. Early evidence of this, following administration of radiolabeled or stable iodothyronines, was confirmed by Vulsma et al. (110). They showed that cord blood T4 levels in newborns with a total organification defect were not negligible, with values that were 20–50% of those of normal newborns. This T4 had to be of maternal origin because children with this congenital defect are unable to synthesize thyroid hormones. If data obtained in the rat can be extrapolated to man, this amount of T4 in the fetal circulation ought to protect the brain of the CH fetus from T3 deficiency until birth (103), although signs of thyroid hormone deficiency might be apparent in other tissues that are more dependent on systemic T3 (i.e. delayed skeletal maturation, lung maturation, etc.).
Thus, all the conditions that in adults are known to result in biological effects of thyroid hormone are found in the human fetal brain, namely functioning D2, a maternal and/or fetal source of its substrate (T4), and specific nuclear receptors partially occupied by the T3 generated from T4.
There is still very little information that might identify an early biological effect resulting directly from this T3-TR interaction in vivo. A recent study performed in vitro with organ cultures from 13- to 23-week human brains reports that the addition of T3 stimulates cytoskeletal proteins, preferentially actin (101), but it remains to be shown that a deficiency of thyroid hormones in vivo would affect this process and do so during the first trimester. Until a cerebral end point of thyroid hormone action early in development is characterized, the long-lasting dispute concerning the mechanism(s) by which maternal hypothyroxinemia would negatively affect the fetal brain remains unresolved. Usually considered the most likely possibilities are: 1) poor placental function and maternal adaptation to pregnancy; 2) low levels of FT4 available to the first trimester embryonic tissues; or 3) both. In our opinion, the evidence that is accumulating points to possibility 3, with a major role for 2. A major role for possibility 1 would not explain the marked neurodevelopmental damage of the neurological cretins whose mothers usually have normal circulating T3 and are not clinically hypothyroid.
|
A unifying hypothesis for the findings summarized in Tables 1–3
|
|---|
|
TopAbstractIntroductionMaternal thyroid hormone...Information obtained in...A unifying hypothesis for...References |
|---|
shows
the proposed timing of insults to the developing brain in these
conditions, as well as in CH and prematurity.
|
attempts to quantify
the magnitude of the cognitive limitation of the children born from
mothers with inadequate T4,
FT4, TSH, or anti-TBO antibody titers during
pregnancy, as assessed from data provided by the authors of the
different studies, or from our calculations, using statistics reported
in the studies or provided by personal communication (Dr. V. Pop). We
have "normalized" the IQ scores given in different studies to those
of the test norm, taken as 100 ± 15, to calculate the frequency
of children with IQ scores less than or equal to 85 (mean-1
SD; feeble-minded or mentally defectives) or less than or
equal to 70 (mean-2 SD; mentally retarded). Most studies
provide information regarding the frequency of neurodevelopmental
scores that are less than or equal to 85, but only a few indicate
separately the frequency of those less than or equal to 70. The
metanalysis of results from studies in iodine-deficient areas do not
include cretins, but only the apparently "normal" population. All
of the studies have applied corrections for likely confounding
factors.
|
show an important difference with
respect to Studies I-IV, namely, the 5- to 7-fold smaller number of
children at risk who are, thus, identified. This is related to the
lower frequency with which an increased second trimester TSH is found
in pregnant women, compared with the frequency with which women are
selected using the other criteria: ID, low maternal
T4 or FT4, and/or
autoimmune thyroid disorders. Table 5 also highlights another important
issue; although the mean decreases in IQ points referred in the
different studies are less severe than previously reported for
inadequately treated CH babies (30), the number of
potentially affected children might be 150–200 times greater.
|
Thus, we conclude that present information permits an answer regarding points 1 and 2 of the aim of the present comments, and we, therefore, suggest that the very pertinent question asked by Utiger (44) regarding the frequency of hypothyroidism in pregnant women ought to be modified and referred to the frequency of first trimester maternal hypothyroxinemia.
An even more important reason for our present comments is that the major cause of maternal hypothyroxinemia worldwide is ID, as clearly stressed by Utiger (44) in his recent editorial accompanying the paper by Haddow et al. (1). Contrary to the usual findings in patients with primary hypothyroidism, low circulating T4 levels are not necessarily accompanied by elevated TSH in situations of ID, especially if it is mild or moderate. The thyroid gland is able to maintain euthyroidism by responding to ID through intrathyroidal autoregulatory mechanisms that do not require an increased circulating TSH, such as an increased thyroidal blood-flow and thyroid volume, increased thyroidal iodine clearance, and preferential thyroidal synthesis and secretion of T3 over T4, and increased intrathyroidal half-life of iodine-containing compounds. Circulating T4 decreases and serum thyroglobulin (Tg) increases, but TSH is usually normal both in goitrous and nongoitrous subjects (127), very probably because circulating T3 is normal or elevated. Enlargement of the thyroid and increased circulating Tg are more reliable parameters of ID than increased TSH (128). Even in areas of very severe ID, where neurological cretins were born, TSH levels (if increased), were not as high as in clinically hypothyroid patients from iodine-sufficient areas (129, 130) and were not related to the developmental outcome of the child (15).
The frequency of maternal hypothyroxinemia is likely to be much higher in areas of ID. In areas of grade III (severe) ID, 43% of the pregnant women had a low PBI (<6 µg/dL), with the protein-bound iodine values being positively correlated with neurodevelopmental outcome. Even in more developed countries, the frequency of maternal first trimester hypothyroxinemia may be higher than in Holland where Pop et al. (3) carried out their study, because of lower iodine intakes. Thus, in the very thorough studies performed by Glinoer at al (125) in Brussels, where ID is moderate (grade II; median urinary I in pregnant women being 56 µg/L), up to 30% of the women had low first trimester FT4 concentrations, which is almost 10 times the frequency of elevated TSH (2.3%) and 6 times the frequency of high titers of thyroid autoantibodies (5.2%). In Madrid (131), where ID is milder than in Brussels (the median urinary I being 90 µg/L throughout pregnancy), the number of women with first trimester FT4 levels that are less than the 10th percentile values for women receiving an adequate iodine supplement was increased 2-fold, again mostly without an increase in circulating TSH more than or equal to the 98th percentile, or an increase of women with positive thyroid autoantibodies. The median urinary I of these supplemented women was 190 µg/L, which would correspond to a daily 24-h excretion of 270 µg I, provided we assume a mean diuresis of l.4 l. Although their FT4 decreased during pregnancy, as repeatedly reported by others (125), it was significantly higher, throughout gestation, than that of the women who did not receive the I supplement. Indeed, in each trimester FT4 was significantly correlated with the urinary I, independently of gestational week or T4-binding globulin concentrations. None of the women on I supplements had goiter at delivery, in contrast to 24% of those not receiving them. This information would suggest that the iodine requirements of the pregnant woman are 200–300 µg per day, at least in areas where the previous borderline ID has not permitted the accumulation of iodine stores sufficient to face her increased needs and those of the fetus. These amounts are almost double those recommended for children and for nonpregnant or nonlactating adults.
The studies from Brussels and Madrid suggest that these women are likely to be hypothyroxinemic during pregnancy, unless given iodine supplements from the onset or before (44, 132, 133). As pointed out by Utiger (44), the United States might also be facing more problems related to ID than previously realized. His comment is based on the recent study regarding iodine nutrition in the United States (134), which has decreased markedly during the last 20 yr: the median urinary I has decreased from 320 µg/L to 145 µg/L. As many as 15% of women of childbearing age had concentrations below 50 µg/L. North Americans are, therefore, also included in the recommendation that iodine supplements be taken during pregnancy, from the very beginning, or even better, before its onset (44).
Moreover, as already discussed, ID does not necessarily cause hypothyroidism with increased TSH and may not be detected using an elevated TSH as criterion for selection of pregnant women with increased risk of progeny with neurodevelopmental deficits. Nonetheless, it may result in decreased maternal FT4 in a relatively large number of women and to the ensuing increased risk of psychomotor defects (3). That this relative ID is so easily preventable, and at such a small cost, makes it all the more frustrating that the general public and the medical community are not fully aware of the problems that ID may continue to pose for future generations.
Although our present comments have been focused mainly on the relationship between inadequate maternal thyroid function, especially hypothyroxinemia, and neurodevelopmental problems of her child, other negative effects cannot be overlooked, such as an increased rate of spontaneous abortions, placental disruptions, fetal distress, malformations, prematurity, decreased birth weight, poorer perinatal outcome, and pregnancy-induced hypertension (20, 22, 26, 27, 135, 136, 137, 138). Those caused by autoimmune thyroid disorders (AITDs) have been extensively studied and reviewed by Glinoer (4, 125, 139), both with respect to subclinical and overt hypothyroidism and to subclinical and overt hyperthyroidism and transient gestational hyperthyroidism. His conclusion is that systematic screening for autoimmune disorders and for TSH more than 4 or less than 0.1 mU/mL early in pregnancy is justified by present evidence, considering both their frequency and the increased risk of adverse pregnancy outcome and perinatal morbidity, and the benefits for the child and for the future health of the mother of adequate treatment. We should again like to remark that this conclusion was already reached by Glinoer (4, 125, 139) without taking into account that high titers of maternal anti-TPO antibodies during pregnancy are also a marker of impaired neuropsychological development of the child (23), or that the latter outcome is very poor in the progeny of untreated mothers with high titers of TSH receptor blocking antibodies (35, 140, 141).
To screen or not to screen, that is the question. Presently available epidemiological and experimental evidence strongly supports the need for an increasing widespread attention to maternal thyroid status during pregnancy. Glinoer (125, 139) has already proposed systematic screening algorithms (4) for AITDs and subclinical and overt hypo- and hyperthyroidism early in pregnancy, as well as the ensuing treatment protocols. The proposed screening program is based on the determination of positive anti-TPO antibodies (also positive anti-Tg antibodies if economically feasible) and TSH more than 4 or less than 0.1 mU/mL, preferably at 12 weeks of gestation. These algorithms do not include tests for TSH receptor blocking antibodies, despite the high risk of the very severe neurodevelopmental damage to the offspring, possibly because of their low incidence [1:180 000 pregnancies (142)] and because the bioassays developed so far are not easily performed. In view of the evidence we have reviewed here, however, which shows a relationship between early maternal hypothyroxinemia and poor neurodevelopmental outcome, we believe that inclusion of screening for low FT4 as well would considerably increase the benefits of the program proposed by Glinoer (4).
The cut-off points of the FT4 values needed to establish the degree of hypothyroxinemia at different gestational ages (and using different commercially available kits) remains to be defined, ensuring it is done in pregnant women with a confirmed iodine intake of 200–300 µg/day (145–220 µg I/L urine) or more. Inclusion in the reference group of women on an insufficient iodine intake would underestimate the lowest 10th percentile.
How to correct the hypothyroxinemia in time to avoid neurodevelopmental problems might also require further controlled studies or screening trials, as stressed in a recent editorial by Pop et al. (123). However, despite the different methodologies and study designs used in the studies performed by Man and Serunian (20, 21) and by Pop et al. (3) in women who were reportedly iodine sufficient (from Rhode Island and The Netherlands), there are some striking similarities in the IQ scores of their children, and the frequency of values less than or equal to 85. But, only the studies from Man and Serunian (20, 21) included a group of women whose low BEI was promptly corrected by treatment with thyroid extracts. It is interesting that none of the children born from these mothers had IQ scores less than or equal to 85, and the mean IQ was actually higher than that of the progeny of women without hypothyroxinemia, strongly suggesting that treatment would be effective.
A clear definition of the cut-off values of the proposed screening tests and the most appropriate treatment of positive cases were also lacking when neonatal thyroid screening programs were started: despite that they have become a clear example of success in the prevention of mental retardation. The proposed biochemical markers are all presently available as blood spot tests (143). Thus, linking thyroid screening of pregnant women to the logistic facilities developed for local neonatal thyroid screening programs might considerably improve the cost to benefit ratio, because the biochemical tests per se do not constitute the major expense of such programs.
Glinoer’s proposal (4) for screening at 12 weeks of gestation seems reasonable for hypothyroxinemia as well, because an earlier first prenatal visit is highly unlikely. Blood samples would then be taken for the determination of FT4, anti-TPO (and possibly anti-Tg) antibodies, and for TSH. Women who were not screened during this period should still be tested later in gestation because potential benefits might still be obtained, especially if the condition is caused by an autoimmune disorder that may also affect the fetal thyroid. Screening before pregnancy would probably benefit those women presenting with high antibody titers, abnormal values of TSH, or hypothyroxinemia, but may well be insufficient: pregnancy itself may be the factor that contributes to an inadequate availability of T4 for the fetus, unmasking an underlying thyroid failure, or an iodine intake insufficient to meet the increased maternal and fetal needs.
For the future assessment of the results of the inclusion of FT4 in the screening program, it would be important to evaluate the iodine intake of the pregnant woman by determining iodine and creatinine in casual urine samples taken on the same day as the blood samples. An inexpensive method for the determination of urinary iodine, easy to perform and not requiring costly hardware, is presently available, making these determinations easy to incorporate into any biochemical clinical laboratory or neonatal thyroid screening program (144). We are currently attempting to develop a method for the determination of iodine in urine spots on filter paper. These data would later disclose the proportion of women from different regions in whom hypothyroxinemia is related to ID and, therefore, easily prevented or corrected without further treatments.
At this first prenatal visit, all pregnant women would immediately start supplementing their usual diet, throughout pregnancy and lactation, with 200–300 µg I/day, either in the form of KI or KIO3 tablets or vitamin-mineral mixtures (44, 145). Even if the urinary iodine excretion value later indicates that the iodine intake was already adequate for pregnancy, the supplementation would not harm. It has, indeed, been shown that the administration of much larger amounts of iodine during any stage of pregnancy, such as those delivered by ingestion or injection of iodized oils, does not harm fetal neurodevelopment (146, 147). If the TSH data at screening suggest thyroid hyper function, supplementation could be stopped. Inclusion of iodine supplementation into the program is likely to be much more effective in ensuring permanent correction of maternal iodine deficiency than mere information programs to the general public and the medical profession.
If the results of the screening tests indicate abnormal values of either thyroid autoantibodies, TSH, or both, the algorithms outlined by Glinoer (4) could be implemented. If the only abnormal value is a FT4 that is low for gestational week, a new blood sample would be withdrawn within 2 weeks for the determination of serum FT4. Correction, or amelioration, of the FT4 levels would suggest that supplementation of the diet with iodine is sufficient. If, despite iodine supplementation, FT4 remains persistently lower than the 10th percentile value for normal pregnant women with a confirmed adequate iodine intake at comparable weeks of gestation, additional treatment with T4 might be considered pertinent to ensure the FT4 levels that have been found to be normal in iodine-sufficient women at the same stage of pregnancy.
Summary. Screening programs for the identification of pregnant women with hypothyroxinemia, hypothyroidism, AITDs, or hyperthyroidism are likely to pose more organizational problems than the screening of neonates for CH. However, the much greater frequency with which alterations of maternal thyroid function are detected, which potentially jeopardize the outcome of pregnancy and increase the risk of neurodevelopmental impairment of the offspring, ought to encourage their implementation. Fukushi et al. (143) have recently reported that a screening program of the thyroid function of pregnant women from the Sapporo region in Japan was initiated in 1991, and data are available for 70,632 women. The authors conclude it is a useful program, despite the fact that thyroid disease among pregnant women is likely to be lower in Japan than in other countries (hypothyroxinemia due to ID is highly unlikely). These programs are likely to greatly increase the awareness of the general medical profession of the importance of the prevention of ID during pregnancy and also lead to a much needed closer collaboration between gynecologists and endocrinologists for adequate treatments following well-defined protocols (2, 4).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received January 10, 2000.
Revised July 3, 2000.
Accepted August 2, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaternal thyroid hormone...Information obtained in...A unifying hypothesis for...References |
|---|
