PRIMARY HYPOTHYROIDISM

Clinical epidemiology and causal factors

The prevalence of hypothyroidism during pregnancy in Western countries is estimated to be 0.3-0.5% for overt hypothyroidism (OH) and 2-3% for subclinical hypothyroidism (SCH) 208-211. On a worldwide basis, the most important cause of maternal thyroid deficiency remains iodine deficiency, which is known to affect over 1.2 billion individuals 139. Thyroid antibodies are found in 5-15% of normal women in the childbearing age and, when the iodine nutrition status is adequate, the main cause of hypothyroidism during pregnancy is chronic autoimmune thyroiditis.

Data on the prevalence of thyroid autoimmune features in pregnant women with a diagnosis of hypothyroidism have recently been reviewed by Glinoer 212. Ten studies were analyzed, of which 6 were retrospective (see Table 14-6). The overall prevalence of hypothyroidism was 2.2% to 3.4%, depending on study design. Some of these studies encompassed women already known to be hypothyroid, while in other studies the diagnosis was based on screening by serum TSH measurements, which was usually carried out before 20 weeks of gestation. It is also important to note that the cut-off level for an abnormal serum TSH differed among studies (from >3 mU/L to >6 mU/L). Interestingly, in a recent study by Casey et al. where the lowest serum TSH cut-off limit was employed, the highest prevalence of gestational hypothyroidism was observed 213. Another variable was the timing of serum TSH determinations, which varied from 5 to 20 weeks of gestation, a factor that may also have impacted on final prevalence figures. Altogether, the analysis of the data showed that thyroid antibodies were resent in 25% to 77% of hypothyroid pregnant women, with a mean prevalence of 46%. In those studies where epidemiologic information was available on groups of control women, the data showed that thyroid autoimmunity was 5.2-fold more frequent in women with a diagnosis of hypothyroidism, compared with euthyroid controls (mean of 48.5% versus 9.2%). It is important to emphasize that the prevalence of thyroid antibodies in pregnant hypothyroid women depends on the severity of thyroid dysfunction. For instance in the study by Allan et al., the prevalence of thyroid antibodies reached 55% among the women with a modest elevation in serum TSH (6-10 mU/L), while it exceeded 80% among the women with a markedly elevated serum TSH (10-200 mU/L) 211.

Thus, evidence from literature amply confirms that chronic autoimmune thyroiditis represents the main cause of gestational hypothyroidism. Other causes, though much less frequent, include the previous radical treatment of hyperthyroidism using radioiodine ablation or surgery, as well as surgery for thyroid tumors. A special mention should be made of recent studies showing that in pregnant women with diabetes mellitus type 1, thyroid dysfunction, mainly subclinical hypothyroidism, may even be more prevalent (27%-45%) 214, 215. A hypothalamic-hypophyseal origin of hypothyroidism is rare, and can include lymphocytic hypophysitis occurring during pregnancy or postpartum 216. Other rare causes to consider in the differential diagnosis of hypothyroidism are those associated with the presence of TSH receptor "blocking" antibodies. In such patients, hypothyroidism is presumably caused by interference in TSH to TSH-receptor interactions. Even though extremely uncommon, the clinical significance of this problem in the pregnant state is that blocking antibodies may be transferred to the fetus and thus cause intrauterine or transient neonatal hypothyroidism 156, 217-221.

Foot-note to Table 14-6:

The first column "Type of Study" indicates whether the study was prospective (P) or retrospective (R). The second "Country" indicates where the study was performed (and, in parentheses, the period during which the study was conducted). The third column "Hypothyroid pregnant women": lists the number of pregnant women with a diagnosis of hypothyroidism (prevalence is indicated in percent, when available) and also the cut-off limit used for serum TSH. The fourth column "Timing of Screening" shows at what time in gestation screening was carried out. The fifth column "Prevalence of Thyroid Antibodies" shows the prevalence data in percent, with an indication of the type of antibody measured. Finally when available, the prevalence of thyroid antibodies in the control pregnant women is also indicated (versus C). (from Glinoer, Ref 212)

Clinical and diagnostic features

Symptoms and signs may raise clinical suspicion of hypothyroidism during pregnancy (weight increase, sensitivity to cold, dry skin, etc.) but others may go unnoticed (asthenia, drowsiness, constipation, etc.). Because many women remain asymptomatic, particular attention is required from the obstetrical care providers for this condition to be diagnosed and to evaluate more systematically thyroid function when women attend the prenatal clinic for the first time. Only thyroid function tests confirm the diagnosis. A serum TSH elevation suggests primary hypothyroidism and measurement of serum free T4 levels further distinguish between subclinical hypothyroidism (SCH) and overt hypothyroidism (OH), depending on whether free T4 is normal or clearly below normal for gestational age. Determination of thyroid antibodies, thyroperoxidase (TPO-Ab) and thyroglobulin (TG-Ab) antibodies, confirms the autoimmune origin of the disorder 228-230.

Microchimerism

A fascinating new topic in the field of autoimmunity and pregnancy is that of fetal microchimerism, i.e. the migration of fetal cells into maternal blood and the prolonged engraftment of fetal progenitor cells into maternal tissues. Several studies have recently confirmed that microchimerism occurs within the thyroid gland in women with Hashimoto's and Graves' disease. Although the biological consequences of persisting fetal microchimerism are not yet clearly understood and only beginning to be explored, fetal cells engrafted into maternal tissues may possibly play a role in the etiology of autoimmune thyroid diseases, and perhaps also in the modulation of autoimmunity during pregnancy. The contribution of natural microchimerism to the origin or exacerbation of autoimmune diseases has been widely documented, although not yet universally accepted 231-236. For instance in two recent clinical studies, the authors failed to observe an association between previous pregnancy, parity and thyroid antibodies, hence arguing against a key pathogenic role of microchimerism as a trigger of chronic thyroid autoimmunity 237, 238.

Repercussions of hypothyroidism on the outcome of pregnancy

Despite the known association between decreased fertility and hypothyroidism, the latter condition does not preclude the possibility to conceive. This is probably the main reason why, until a few years ago, hypothyroidism had been considered – wrongly – to be relatively rare during pregnancy 239-243. In a recent study by Abalovich at al, 34% of hypothyroid women became pregnant without thyroxine treatment: 11% of them had OH and 89% SCH 225. When hypothyroid women become pregnant and maintain the pregnancy, they carry an increased risk for early and late obstetrical complications (see Table 14-7 and Table 14-8).

Foot-note to Table 7: the percentages listed were taken (or recalculated, when possible) from the studies shown in the references. ** SCH: subclinical hypothyroidism; ** OH: overt hypothyroidism.

Foot-note to Table 8: the percentages listed were taken (or recalculated, when possible) from the studies given in the references. * LBW: low birth weight; ** SCH: subclinical hypothyroidism; ** OH: overt hypothyroidism; *** O.R.: Odds Ratio; *** R.R.: Relative Risk.

Table 14-7 & Table 14-8 clearly show that both obstetrical and fetal complications occur with an increased frequency in pregnant women with hypothyroidism. As expected, these complications are both more frequent and more severe in women with OH than SCH. Most importantly, adequate treatment with thyroid hormone greatly reduces risks of a poorer obstetrical outcome 225, 244, 251-253. In the study of Abalovich, it was also shown that it was not so much the diagnosis of OH versus SCH that mattered with regard to pregnancy outcome, but primarily the adequacy of thyroxine treatment 225. When treatment was not adequate, pregnancy ended with abortion in 60% & 71% of OH & SCH women respectively, with an increased prevalence of preterm deliveries. Conversely, in hypothyroid pregnant women receiving adequate treatment, the frequency of abortions was minimal and pregnancies were carried to term without complications (see Figure 14-13). Concerning risks of premature delivery in untreated hypothyroidism women, a recent study confirmed that gravidas with high TSH levels had a greater than 3-fold increase in risk of very preterm delivery (<32 completed weeks) 227, 254.

Management and therapy of gestational hypothyroidism

Administration of L-thyroxine is the treatment of choice for maternal hypothyroidism, when the iodine nutrition status is adequate. Hypothyroid pregnant women require larger thyroxine replacement doses than do non-pregnant patients. Women who already take thyroxine before pregnancy usually need to increase their daily dosage by 30-50%, on average, above preconception dosage. Several reasons explain the incremented thyroid hormone requirements: the rapid rise in TBG levels resulting from the physiological rise in estrogen concentrations, the increased distribution volume of thyroid hormones (vascular, hepatic, and the fetal-placental unit), and finally the increased placental transport and metabolism of maternal T4 60, 228, 229, 255. Treatment should be initiated with a dose of 100-150 µg/day or titrated according to body weight. In non pregnant women, the full replacement thyroxine dose is 1.7-2.0 µg/kg bw/day. During pregnancy, because of the increased requirements, the full replacement thyroxine dose should be increased to 2.0-2.4 g/kg bw/day 60-62, 229, 255, 256. In a newly diagnosed hypothyroid patient, a full replacement thyroxine dose should be instituted immediately, assuming there are no abnormalities in cardiac function. In initially severe hypothyroidism, therapy may be initiated by giving for the first few days a thyroxine dose corresponding to two times the estimated final replacement daily dose, in order to normalize more rapidly the extra-thyroidal thyroxine pool. In women who already take thyroxine before conception, the need to adjust the preconception daily dosage may become manifest as early as by 4 to 6 weeks gestation, hence justifying the early adaptation of thyroid hormone replacement to ensure that maternal euthyroidism is maintained during the first months of pregnancy. An alternative (recommended by some thyroidologists) is to anticipate the expected increase in serum TSH by raising the thyroxine dose already before conception or as soon as the pregnancy is confirmed. Among the possible preventive strategies to avoid the risk of maternal hypothyroidism during early gestational stages, it can be recommended to adjust the preconception thyroxine dose with the aim to maintain serum TSH near the low-normal range 257, 258. It is important to note that 25% of hypothyroid women who are able to maintain a normal serum TSH level in the first trimester, and 35% of those who maintain a normal serum TSH level until the second trimester without increasing their daily dosage, will still require an increment in thyroxine replacement during late gestation to remain euthyroid 229, 248, 259. If a pregnancy is planned, patients should have thyroid function tests measured soon after the missed menstrual period. If serum TSH is not increased at that time, tests should be repeated at 8-12 weeks and then again at 20 weeks, as the increase in hormone requirements may not become apparent until later during gestation.

The magnitude of thyroxine increment during pregnancy depends on the etiology of hypothyroidism, namely the presence or absence of residual functional thyroid tissue. Women without residual functional thyroid tissue (after radioiodine ablation, total thyroidectomy, or due to congenital agenesis of the gland) require a greater increment in thyroxine dosage than women with Hashimoto's thyroiditis, who usually have some residual thyroid tissue. As a simple rule of thumb, it has been suggested that the increment in thyroxine can be based on the initial degree of TSH elevation. For women with a serum TSH between 5-10 mU/L, the average increment in thyroxine dosage is 25-50 µg/day; for those with a serum TSH between 10-20 mU/L, 50-75 µg/day; and for those with a serum TSH >20 mU/L, 75-100 µg/day 229. Serum free T4 and TSH levels should be measured within one month after the initiation of treatment. The overall aim is to achieve and maintain normal free T4 and TSH levels throughout pregnancy. Ideally, thyroxine doses should be titrated to reach a serum TSH value <2.5 mU/L (see Table 14-9). As discussed already, because it is sometimes difficult to correctly interpret the results of free T4 and TSH measurements in the context of pregnancy, it is useful to optimize monitoring of treatment by establishing laboratory-specific reference ranges for serum free T4 and trimester-specific reference ranges for serum TSH. Once parameters of thyroid function have been normalized by treatment, they should be monitored every 6-8 weeks. If they remain abnormal, thyroxine dosage should be adjusted and tests repeated after 30 days, and so on until normalization. After parturition, most patients need to decrease thyroxine dosage, over a period of ~4 weeks postpartum. It should also be remembered that women with thyroid autoimmunity features are at risk of developing postpartum thyroiditis, a syndrome that may justify differences in the pre- and post-pregnancy thyroxine requirements. It is therefore important to continue monitor thyroid function tests for at least during six months after the delivery 260.

A recent study has examined how closely the management of hypothyroidism in the general pregnancy population satisfies recently issued guidelines 261. This observational retrospective study concerned 389 women, at 5 recruitment centers in the USA. The results showed that 43% of serum TSH levels measured in 1st trimester were at or above the recommended upper limit of 2.5 mU/L. In 2nd trimester, 33% of serum TSH measurements were at or above the recommended upper limit of 3.0 mU/L. When using a less restrictive upper limit, defined as a serum TSH value above the 98th percentile of normal, the data showed that 20% of serum TSH values in 1st trimester and 23% in 2nd trimester were above the upper normal limit. The authors concluded that future strategies should focus on more effectively monitoring thyroxine treatment during pregnancy.

Fetal-neonatal consequences of maternal hypothyroidism

Role of thyroid hormone during fetal brain development

Alterations of thyroid function in newborns soon after birth have been recognized to be associated with neurodevelopment abnormalities for over a century. This led to two major concepts in endocrinology, namely those of endemic cretinism due to severe iodine deficiency and sporadic congenital hypothyroidism (CH), which, in turn, led to the development in the late 1970s of systematic screening programs for the early diagnosis and treatment of sporadic CH. During the largest part of the 20th century, it was considered that the developing brain did not need thyroid hormone until after birth and this concept was reinforced by the notion of an efficient uterine-placental barrier, preventing the transfer of physiologically relevant amounts of maternal thyroid hormone to the fetal compartment. At the end of the 1980s, however, it was shown that thyroid hormone (TH) of maternal origin was present in human fetal blood up to birth, and also that maternal TH had a protective effect on experimentally-induced CH in animal models 83-85. Recent availability of highly sensitive assays for the measurement of minute amounts of TH in fetal tissues have allowed to show that physiologic amounts of free T4 are present in coelomic and amniotic fluids surrounding the developing embryo already in first trimester. Also, specific nuclear receptors are present in fetal brain as early as ~8 weeks post-conception 262-267. The ontogenic pattern of TH concentrations and activity of iodothyronine deiodinases have been investigated in different cerebral areas of human fetuses as early as 11-18 weeks post-conception. Results of these studies have indicated the presence of increasing T4 and T3 concentrations in fetal brain, as well as a complex interplay between changing activities of the specific "D2" and "D3" iodothyronine deiodinases during gestation. This dual enzymatic system is currently interpreted to represent a regulatory pathway, allowing to fine tune the availability of T3 required for normal brain development and avoid the presence of excessive amounts of T3. In summary, a large body of both human and experimental evidence strongly suggests that thyroid hormone is an important factor contributing to normal fetal brain development 268-273. At early stages of pregnancy, the presence of TH in fetal structures can only be explained by the transfer of maternal TH to the fetal compartment, since fetal thyroid hormone production does not become efficient until mid-gestation.

Clinical studies on the role of maternal hypothyroidism for the psycho-neurological outcome in the progeny

Because of the heterogeneity of what is commonly referred to as gestational hypothyroidism, different clinical conditions must be considered. Thyroid insufficiency varies widely with regard to time of onset (first trimester versus later), degree of severity (SCH versus OH), progressive aggravation with gestation time (depending on the cause), and adequacy of treatment. To reconcile these variable clinical conditions into a global view of the repercussions of maternal hypothyroidism on the progeny is difficult. However, a common pattern clearly emerges. Several clinical studies have investigated the psychological and intellectual outcome in the offspring of pregnant women with thyroid insufficiency, both in conditions of hypothyroidism with an adequate iodine nutritional status and women with mild-moderate iodine deficiency. Overall, the results show that there is a significantly increased risk of impairment in neuro-psychological developmental indices, IQ scores, and school learning abilities in the offspring of untreated hypothyroid mothers.

Three decades ago, Evelyn Man and colleagues published a series of articles suggesting that children of mothers with inadequately treated hypothyroidism had significantly lower IQs than those born to adequately treated patients or normal controls. These pioneering data did not gain much clinical attention, probably because the prevailing dogma, at that time, was that maternal TH did not cross the placenta 274-276.

The first large-scale prospective study on outcome of children born to mothers with thyroid deficiency was reported by Haddow et al. in 1999 224. Sixty two women were identified retrospectively, who had hypothyroidism around mid-gestation with elevated serum TSH and low free T4. The children of these women were matched with 124 control children and all underwent extensive neuropsychological testing at ~8 years of age. Main results showed that the mean full-scale IQ of children born to mothers with untreated maternal thyroid disease was 7 IQ points lower than the mean IQ of children born to both controls and hypothyroid but thyroxine-treated mothers (see Table 14-10). Furthermore, three times as many children born to mothers with untreated hypothyroidism had IQs that were 2 standard deviations below the mean IQ of the controls. The conclusion was that undisclosed (and hence untreated) hypothyroidism occurring during the first half of pregnancy (and presumably prolonged thereafter) was associated with a risk of a poorer outcome in the progeny and a 3-fold increased predisposition for having learning disabilities later in life. It should however be noted that in this study, the severity of maternal hypothyroidism varied from OH to probable SCH.

Although still incompletely published, a large set of data was reported at the 2004 annual meeting of the American Thyroid Association by Rovet et al. 277. The interest of this remarkable work is double. First, the authors investigated children born to hypothyroid mothers who had been treated during pregnancy, but in whom thyroxine administration remained suboptimally adapted (mean TSH levels between 5-7 mU/L). Second, children were tested with extremely refined techniques and followed up to 5 years of age. Results showed that some components of intelligence were affected, while others not. At preschool age, the study-case children had a mild reduction in global intelligence that was inversely correlated with third trimester"s maternal TSH. On the other hand, there was no negative impact on language, visual-spatial ability, fine motor performance, or preschool ability. The conclusion was that the offspring of women with suboptimal treatment of maternal hypothyroidism may be at risk for subtle and selective clinically relevant cognitive deficits, which depend specifically on severity and timing of inadequate maternal thyroid hormone availability.

An interesting - although still intriguing - issue is that of potential repercussions for fetal development of isolated hypothyroxinemia in pregnant women. The first study to deal specifically with this issue was published by Pop et al. in 1999. The authors investigated development outcome in children born to mothers who had early (i.e. 1st trimester) low serum free T4 levels without TSH elevation 278. These women corresponded to the lowest 10th decile of free T4 values at 12 weeks gestation (<10.4 pMol/L) among 220 apparently normal pregnancies. Main results suggested that early maternal free T4 was associated with a lower developmental index in the children at 10 months of age. The study was criticized because the scoring method used was not discriminant enough and the study did not include a control group. The same group of authors published later a 2nd study based on similar selection criteria, but with a larger cohort and a control group, and also more refined motor and mental evaluations in infants at 1 and 2 years of age 279. Main results showed that children of mothers with prolonged hypothyroxinemia (until 24 weeks or later) had an 8-point to 10-point deficit on mental and motor developmental scales. Of interest, those infants of women who presented early hypothyroxinemia but recovered spontaneously normal free T4 levels during later gestation had a normal development, suggesting that prolonged hypo-T4 was needed to impair fetal neuro-development. Finally in 2006, the same group of investigators confirmed their earlier findings in another study that examined the neurobehavioral profile of neonates born to mothers with isolated early hypothyroxinemia 280. The results indicated that neuro-developmental difficulties could be identified as early as 3 weeks of age.

The significance of early isolated maternal hypothyroxinemia was also investigated by Casey et al. in a large-scale study of >17,000 women 213. The authors identified 233 women (1.3%) with isolated low free T4 in 1st half of pregnancy. Assessing adverse effects of this biochemical abnormality on parameters of pregnancy outcome, the authors were unable to identify any excessive adverse pregnancy outcome feature and concluded that their findings “cast further doubt on the biologic significance isolated maternal hypothyroxinemia”. Finally, a study is ongoing in Wales (under the leadership of John Lazarus) on this topic and, from preliminary data disclosed at recent meetings, it appears that this condition (isolated hypo-T4 with a normal serum TSH) seems to be relatively frequent but its consequences are still unclear. As already mentioned above, in the studies by Pop et al. referred to above, it is not clear why these women had a low free T4 and also why the majority recovered spontaneously a normal thyroid function after the 1st trimester of gestation. Our current opinion is that one must distinguish between isolated low free T4 and low free T4 in the contest of hypothyroidism, especially for the purpose of defining who should benefit from early treatment with thyroxine. This opinion is shared by Mitchell & Haddow who recently showed that in hypothyroid pregnant women, low free T4 concentrations were almost never observed without a concomitant elevation in serum TSH 281.

Clinical studies of the psycho-neurological outcome in the progeny associated with maternal hypothyroidism due to iodine deficiency

Consequences of maternal hypothyroidism on the progeny must be considered separately because iodine deficiency (ID) exposes both mother and fetus to thyroid underfunction 86. In 1994, Bleichrodt & Born published the results of a meta-analysis of 19 studies of infants" outcome in relation to ID 282. The main result was to demonstrate a shift in frequency distribution of IQ towards lower values, with a mean overall reduction of 13.5 IQ points in both neuro-motor and cognitive functions. Because this meta-analysis encompassed public health conditions with more or less severe iodine deficiency, the results cannot be fully extrapolated to mild-moderate ID. Therefore, the main results of a series of clinical studies (reported between 1989-1996) that have investigated the late outcome in children born to mothers who reside in areas with only mild-moderate ID were examined in a review article by Glinoer & Delange 86. Main results are summarized in Table 14-11, showing that many of these infants do present school learning difficulties.

A recent publication dealing with the outcome of children of mothers exposed to mild-moderate ID during pregnancy was performed in Sicily by Vermiglio et al. 283. The authors showed that these children had a greater than 10 point average deficit in global IQ, compared with children born to mothers from another area in Sicily without ID. Furthermore, they reported the presence of attention deficit & hyperactivity disorder (ADHD) in 69% of children born to the mothers who presented hypothyroxinemia during early gestation. Finally, Kasatkina et al. studied children born to mothers in whom gestational low free T4 values were corrected by thyroxine administration and showed that it improved the neuro-intellectual prognosis of the offspring evaluated at the ages of 6, 9, and 12 months 284.

Comments and perspectives

Clinical studies discussed in the sections above were carried out in conditions where thyroid underfunction was associated with either iodine deficiency or thyroid autoimmunity during pregnancy. Together, they clearly underline the notion that when thyroid failure occurs during pregnancy, this abnormality may be associated with impairment in the normal development of the fetal brain. Alterations in neuro-psycho-intellectual outcome in the progeny may ultimately result from a combination of detrimental effects related to maternal hypothyroidism per se, as well as from impaired fetal thyroid function per se, or the two in combination. With regard to maternal hypothyroidism, detrimental effects of hypothyroxinemia are predominant when present already during early gestational stages, although deleterious effects may also be due to maternal hypothyroidism occurring during late gestational stages and perhaps also to hypothyroidism taking place (or persisting) during the postpartum period, indirectly through a "lesser well-being" status, which is so characteristic of patients with unrecognized hypothyroidism. One important lesson to be learned from recent studies is that an altered neuro-psycho-intellectual development may occur in offspring even in the presence of only mild and perhaps transient maternal hypothyroxinemia, and obviously in the absence of detectable abnormalities of neonatal thyroid function after birth. In most clinical circumstances where a woman"s thyroid function is defective, hypothyroxinemia is not restricted to 1st trimester and hypothyroidism tends to worsen as gestation progresses, especially when left undiagnosed and untreated. The fetus may therefore be deprived of adequate amounts of thyroid hormones in the early stages of brain development, as well as during later neurological maturation and development, thereby reinforcing the negative consequences of early maternal hypothyroxinemia. Thus, any delay in diagnosing and treating maternal hypothyroidism may ultimately cost "IQ" points to the progeny, with the educational, socioeconomic, and public health consequences that may be foreseen.

Globally in maternal-fetal thyroid disease, three sets of clinical disorders ought to be considered, illustrated schematically in Figure 14-14. For infants with a defect in thyroid ontogeny leading to congenital hypothyroidism, the participation of maternal hormones to the fetal T4 circulating environment remains unaffected, and therefore the risk of brain damage results exclusively from insufficient fetal thyroid hormone production. In contrast, when only the maternal thyroid gland is deficient, such as in women with autoimmune hypothyroidism, it is both the severity and temporal occurrence of maternal underfunction that drive the resulting consequences for fetal neuronal development. Finally in iodine deficiency, both maternal and fetal thyroid functions are affected, and it is the degree and precocity of iodine deficiency during pregnancy that drives the potential repercussions for fetal neurological development 86.

Screening for hypothyroidism during pregnancy

There are no recommendations for universal screening for thyroid dysfunction in women before or during pregnancy. As the overall benefits of screening for thyroid dysfunction (primarily hypothyroidism) have not yet been universally justified by current evidence-based medicine, recent international guidelines have recommended "aggressive" case finding among the following groups of women who are at high risk, preferably already prior to pregnancy or in early gestation (see Table 14-12) 1.

Among possible screening algorithms, the following scheme has been proposed and is outlined in Figure 14-15.

The first step in the algorithm is to measure serum TSH and thyroid antibodies in early gestation. Because isolated hypothyroxinemia may occur in some women (without concomitant rise in serum TSH), it is reasonable to include systematically a free T4 determination. Ideally, both TG-Ab and TPO-Ab should be determined; however, if for economical reasons only one antibody can be measured, then it is preferable to measure TPO-Ab because it yields the best diagnostic score. When serum TSH is elevated or free T4 clearly below normal, and irrespective of the presence (or absence) of thyroid autoimmunity (TAI), women should be considered to have thyroid underfunction and treated with thyroxine. The next step concerns those women with TAI and normal thyroid function. We propose to base the medical attitude on serum TSH levels during early pregnancy. When serum TSH is <2.5 mU/L (most frequently associated with low antibodies titers and normal free T4 levels), thyroxine treatment is not systematically warranted, and serum TSH and free T4 should be monitored later during gestation. For women with TAI and a serum TSH that lies within the normal range in early gestation, but is already slightly shifted to higher "normal" values, i.e. between 2.5-4.0 mU/L (most frequently associated with higher antibody titers and low-normal free T4 levels), obstetric care providers should consider thyroxine treatment. It is important to keep in mind that serum TSH is down-regulated under the influence of peak hCG values in the 1st half of gestation, and also that the thyroid deficit tends to deteriorate as gestation progresses in TAI-positive women. Because the potential deleterious effects, for both mother and progeny, are not due to high serum TSH per se but to low free T4 concentrations, clinical judgment should be based on serum free T4. If low or low-normal for gestational age, thyroxine treatment is probably justified. In daily practice, when such a scheme is systematically applied, most - if not all - of the pregnancies followed are successful and uneventful 225, 292. Even though more prospective studies are needed to assess the final clinical relevance of the proposed scheme, the recent study of Negro et al. provided strong arguments in favor of early thyroxine administration in women with AITD and normal thyroid function during early gestation 202. Based on many indirect arguments, it is our personal belief that no harm can be done and that thyroxine treatment can only be beneficial, in such conditions, for both the patient and offspring. Finally, systematic screening for AITD during early pregnancy also allows to delineating women who are prone to developing thyroid dysfunction after parturition. Thus, even when no specific treatment is warranted during gestation, systematic screening is useful to clinicians for organizing the monitoring of potential postpartum thyroid dysfunction 293, 294.

Recommendations and "take home" messages

Because of the high frequency of autoimmune thyroiditis in young women, because subclinical hypothyroidism (SCH) often remains undiagnosed, because potential obstetric repercussions are associated with un(der)treated hypo-thyroidism, and finally because of the consequences of maternal hypothyroxinemia on fetal development, there is a justification to propose systematic screening for autoimmune thyroiditis and hypothyroidism in pregnancy.

SCH has been shown to be associated with an adverse outcome in both mother and offspring. Therefore, maternal hypothyroidism should be avoided.

Women with autoimmune thyroiditis who are euthyroid in early stages of pregnancy are still at risk of developing hypothyroidism later; therefore, their serum TSH levels should be monitored.

Subclinical hypothyroidism (serum TSH concentration above the upper limit of the reference range with a normal free T4) has been shown to be associated with an adverse outcome for both the mother and offspring. Thyroxine treatment has been shown to improve obstetrical outcome, but has not been proved to modify long term neurological development in the offspring. Given that potential benefits outweigh potential risks, thyroxine replacement is recommended in women with SCH.

If hypothyroidism has been diagnosed before pregnancy, it is recommended to adjust rapidly preconception thyroxine doses to reach a TSH level prior to pregnancy not higher than 2.5 mU/L. The thyroxine dose often needs to be incremented by 4-6 weeks gestation, and may require a 30-50% increment in dosage.

If overt hypothyroidism is diagnosed during pregnancy, thyroid function tests should be normalized as rapidly as possible. Thyroxine doses should be titrated to rapidly reach and thereafter maintain serum TSH concentrations of less than 2.5 mU/L in the first trimester (or 3 mU/L in 2nd & 3rd trimesters) or to trimester-specific ranges. Thyroid function tests should be re-measured within 30-40 days.

After delivery, most hypothyroid women need a decrease in the thyroxine dosage they received during pregnancy.

Finally, there is a consensus among obstetrical providers and endocrinologists against advising to interrupt a pregnancy, even when hypothyroidism has been diagnosed late during pregnancy.