Thyroid Disorders and Infertility
The prevalence of hypothyroidism in women in the reproductive age (2040 years) varies between 2% and 4%.[29,30] In this age group, autoimmune thyroid disease (AITD) is the most common cause of hypothyroidism.[31,32] Hypothyroidism is associated with a broad spectrum of reproductive disorders ranging from abnormal sexual development through menstrual irregularities to infertility. The impact of hypothyroidism on the menstrual cycle has been identified since the 1950s and leads to changes in cycle length and blood flow.[33,34] Joshi et al. found 68% of menstrual abnormalities in 22 women with hypothyroidism compared to only 12% in 49 controls. In the study by Krassas et al., the prevalence of menstrual irregularities (mainly oligomenorrhoea) reached 23% among 171 hypothyroid patients, while being only 8% in 214 controls (P < 0·05). There were also 12% of women presenting with amenorrhoea in the hypothyroid group, and none among the control subjects. The authors also showed an association between the severity of menstrual abnormalities and higher serum TSH levels. The discrepancy in the frequencies of menstrual abnormalities between both studies was due to an earlier diagnosis of thyroid diseases in the study by Krassas et al. Severe hypothyroidism is commonly associated with ovulatory dysfunction due to numerous interactions of thyroid hormones with the female reproductive system. Both hyperprolactinaemia, due to increased TRH production, and altered GnRH pulsatile secretion, leading to a delay in LH response and inadequate corpus luteum, have been reported.[37,38,39] Thyroid responsitivity by the ovaries could be explained by the presence of thyroid hormone receptors in human oocytes. Thyroid hormones also synergize with the FSH-mediated LH/hCG receptor to exert direct stimulatory effects on granulosa cell function (progesterone production), and in in vitro studies effects on differentiation of the trophoblast have been shown. Recently, Cramer et al. showed that serum TSH levels were a significant predictor of failure of IVF, as TSH levels were significantly higher among women who produced oocytes that failed to be fertilized. Another pathway through which hypothyroidism may impact on fertility is by altering the peripheral metabolism of oestrogen and by decreasing SHBG production. Both pathways may result in an abnormal feedback at the pituitary level. Independently of hormonal changes, hypothyroidism can also lead to menorrhagia by altered production of coagulation factors (decreased levels of factors VII, VIII, IX and XI).
Studies on the incidence of infertility in hypothyroid patients are scarce. Ideally, this should be evaluated prospectively by comparing the incidence with a matched control group. Unfortunately, such data are not available and most studies deal with the prevalence of infertility in cross-sectional studies of hypothyroid patients or in selected populations presenting at fertility clinics. In a study by Joshi et al., 6% of hypothyroid women had infertility compared to 5% in a euthyroid group with a goitre and 2·4% in healthy control women. It should be noted, however, that in this study, the number of patients was small, the thyroid antibody status unknown and the control population not clearly defined. In another study by Lincoln et al., serum TSH levels were determined in 704 infertile women without known thyroid disorder: 2·3% had an increased serum TSH level, representing both overt and subclinical hypothyroidism. The study comprised no control subjects, but the prevalence of elevated serum TSH was comparable to that found in the female population in the reproductive age. In a study by Arojoki et al., the prevalence of an increased serum TSH was 4% and that of overt hypothyroidism 3·3% in 299 infertile women. The highest percentage of women with an increased serum TSH was observed in the group with ovulatory dysfunction (6·3%), compared to 4·8% in the idiopathic group, 2·6% in the tubal infertility group and none in the endometriosis group. There was no statistical difference when comparing the frequency of hypothyroidism between the different groups of infertility.
In clinical practice, patients with overt thyroid failure are probably detected before they are referred to infertility treatment clinics, thereby introducing a bias in the true prevalence of hypothyroidism in infertility disorders. In the case of hypothyroidism, L-thyroxine should be administered because euthyroidism normalizes PRL and LH levels, reverses menstrual abnormalities and increases spontaneous fertility. Given the potential implications of hypothyroidism on ovulatory function, screening for thyroid insufficiency is certainly indicated in women with ovulatory dysfunction.
Subclinical hypothyroidism (SCH) has recently been challenged as data have indicated that physiological free T4 (FT4) variations are narrower in one individual than those observed within the reference range of a population. These data might reflect an abnormally low FT4 value for patients who present a mildly increased serum TSH.[47,48] Some authors have proposed restricting the upper normality limit of serum TSH to 2·5 mU/l. Today, however, there is no agreement among endocrinologists about the most appropriate (i.e. physiologically relevant) upper limit of normality for serum TSH.
To date, studies investigating the association between SCH and infertility are still based on the previous upper serum TSH levels and some older studies are even based on the presence of an abnormal serum TSH after a TRH stimulation test. Table 1 summarizes the most relevant studies on the prevalence of SCH in women with infertility. In the study by Bohnet et al., SCH was considered as an infertility factor by itself. Therefore, 11 of these 20 women were treated with 50 µg/day of L-thyroxine. This normalized their mid-progesterone secretion and 2/11 became pregnant. Bals-Pratsch et al. did not observe corpus luteum insufficiency in infertile women with SCH. Gerhard et al. reported a positive correlation between basal TSH, LH and testosterone concentrations in the early follicular phase. Women with an elevated serum TSH had a lower pregnancy rate than those women with a normally stimulated serum TSH. Eighty out of 185 infertile women had an abnormal TRH test, but among the latter only one woman had an increased basal serum TSH (0·5%). In the study by Shalev et al., the prevalence of SCH was 0·67% in 444 infertile women, all with ovulatory dysfunction. Grassi et al. investigated 129 women of infertile couples with ovulatory dysfunction, male and idiopathic infertility. Six patients (4·6%) had a basal serum TSH level > 4·5 mU/l and five of these six women had AITD. The authors noted that the mean duration of infertility was significantly longer in the patients with thyroid disorders (abnormal TSH and/or AITD) compared to those without (3·8 vs. 2·6 years; P = 0·005). Arojoki et al. found elevated serum TSH levels (> 5·5 mU/l) in 4% women presenting with infertility for the first time. The prevalence of having an increased serum TSH was highest in the group with ovulatory dysfunction (6·3%). Prior to infertility examinations, 10 of 299 women were already receiving L-thyroxine for primary hypothyroidism. The incidental finding of an elevated serum TSH value in patients with infertility was therefore reduced to four of 299 women (1·3%), and this was in the range of the prevalence of SCH in the general population in Finland (23%). In a casecontrolled study, Poppe et al. investigated the prevalence of SCH (TSH > 4·2 mU/l) in women of infertile couples (n = 438) who presented for the first time at the centre of reproductive medicine. The control population consisted of 100 fertile women matched for age. The prevalence of an increased serum TSH was comparable in both the study group and controls (< 1%; P = ns). Overall, the studies investigating the association between SCH and infertility were poorly controlled. Considering the largest cohorts published, the prevalence of SCH in infertile women ranged from 1% to 4% and most cases with SCH were associated with ovulatory dysfunction.
Recently, Raber et al. investigated prospectively a group of 283 women referred for infertility. All patients underwent a TRH stimulation test (SCH was defined as a serum TSH > 15 mU/l). Women with a diagnosis of SCH were treated with L-thyroxine and followed prospectively over a 5-year period. Of the referred women, 34% had SCH, an unusually high prevalence reflecting the specific referral pattern. Among the women who became pregnant during the follow-up period, over 25% still had SCH at conception. Furthermore, the women who never achieved a basal serum TSH < 2·5 mU/l or a TRH-stimulated TSH < 20 mU/l became pregnant less frequently than those who did. More frequent abortions were also observed in the women with a higher basal serum TSH (independent of the presence of autoimmunity). In the studies summarized in Table 1 , three included a control (fertile population) and the prevalence of SCH was comparable between the study group and controls. The prevalence of SCH was considerably higher in the studies based on a TRH-stimulation test to detect SCH compared with the studies that were based only on the upper limit of basal serum TSH. This difference might indicate that in older studies, using less sensitive measurements of serum TSH, the actual TSH reference levels are perhaps slightly too high in the setting of infertility.
In the general population, the prevalence of subclinical hyperthyroidism is approximately 1·5%.[29,30] Joshi et al. found menstrual irregularities in 65% of hyperthyroid women, compared to 17% in healthy controls. Krassas et al. observed irregular cycles in 46 of 214 hyperthyroid women (22%); 24 women had hypomenorrhoea, 15 polymenorrhoea, five oligomenorrhoea, two menorrhagia and none had amenorrhoea. The prevalence of menstrual abnormalities was 2·5 times higher than in the control population (8%). SHBG production increases in hyperthyroid women, the metabolism of oestrogen is altered and the conversion of androgens to oestrogens is increased. Hyperthyroxinemia increases the gonadotrophin response to GnRH and baseline gonadotrophin concentrations are also frequently elevated. The decrease in menstrual flow may also relate to effects on haemostatic factors, including the synthesis of factor VIII.[3,58] Despite these metabolic changes, hyperthyroid women usually maintain ovulation, according to endometrial biopsies. Treatment of hyperthyroidism frequently corrects these cycle changes. The precise impact of hyperthyroidism on fertility remains ill-defined. As was the case for hypothyroidism, most studies on the prevalence of hyperthyroidism in infertility are derived from uncontrolled and retrospective cohort studies. In only one small study in 53 hyperthyroid patients was it shown that 5·8% of them were infertile. Women with increased thyroid hormones, suggestive of hyperthyroidism and infertility, should undergo a complete laboratory work-up and receive appropriate treatment (avoiding radioiodine). Thyroid function should be corrected, especially before an ART procedure is planned.[59,60]
The prevalence of AITD is 510 times higher in women than in men, which might be explained by genetic factors, the effects of oestrogens and perhaps chromosome X abnormalities.[61,62,63,64] The importance of AITD is twofold. First, it is the most common autoimmune disorder in women, affecting 510% in the childbearing period; second, it is the most frequent cause of thyroid dysfunction, although AITD can also be present without thyroid dysfunction and therefore remain undiagnosed. Numerous studies have investigated the prevalence of AITD in women with infertility. The main studies are summarized in Table 2 .[55,65,66,67,68,69,70,71] The interpretation of these data is difficult because of selection bias (women with different causes of infertility or a selected cause such as ovulatory dysfunction), the retrospective setting and the different types of control population. There were differences in sample size as well as in the assays used for detecting thyroid antibodies. However, pooling the results of these studies favours a significantly increased incidence of AITD in women with infertility compared to controls, with an overall estimated relative risk of 2·1 (P < 0·0001). It should be mentioned that the calculated P-value is based on raw figures rather than on a systematic meta-analysis. In two of these studies, a strong association between AITD and a particular cause of infertility was identified. In the casecontrolled study by Poppe et al., comparing 438 women of infertile couples with 100 age-matched fertile women, the relative risk of AITD reached 3·57 (P = 0·016) when women had endometriosis as cause of infertility. Gerhard et al. showed that 44% of women with AITD had endometriosis, compared with only 9% in women without AITD. There is abundant evidence that endometriosis is associated with a variety of immunological changes. Auto-antibodies to endometrial antigens and deposits of complement compounds have been reported in a number of studies. Data also indicate that endometriosis is associated with immune cell depression (NK cells), decreased activity and cytotoxicity against autologous endometrium.[73,74] In a study by Janssen et al., a strong association was shown between AITD and women with polycystic ovarian syndrome (PCOS) as the cause of infertility. The prevalence of AITD was 26·9% compared to 8·3% in women without PCOS (P < 0·001). According to the authors, this association correlated in part with an increased oestrogen/progesterone ratio characteristic for this syndrome. Although AITD represents an organ-specific autoimmune disorder, associations between AITD and non-organ-specific autoimmune disease states have also been described, suggesting a shared immunogenetic background. The underlying pathogenic mechanisms associating AITD and infertility remain largely speculative, as neither animal models nor in vitro data on this issue are available.
Steroid hormone synthesis by oocytes is dependent on adequate levels of thyroid hormones for normal reproductive function. T3 modulates FSH and LH action on steroid biosynthesis and T3 binding sites have been identified in mouse and human oocytes.[40,41] Thyroid hormones enhance the action of oestrogen and potentiates oestrogen responses, as, for example, PRL production in the pituitary.[75,76,77] Whether T3-dependent impact on female reproduction or impaired T3 levels in case of AITD have any physiological relevance in infertility requires further investigation.
Successful implantation and development of an embryo is determined by progesterone and oestrogen on the one hand and adhesion molecules, growth factors and cytokines on the other. As already alluded to above, thyroid hormones play a role in the secretion/action of ovarian steroids. As comparable pregnancy rates have been reported between women with, or without, AITD with a normal thyroid function, it is considered that AITD per se does not alter the implantation of the embryo.[54,78] Different studies support the concept of an association between reproductive failure and the presence of non-organ-specific autoantibodies as well as organ-specific autoimmunity such as thyroid antibodies.[66,69,]
The particular association between AITD and altered pregnancy outcome was first described by Stagnaro-Green et al. Several studies were published later, increasing the statistical power and leading to the conclusion that AITD (without overt thyroid dysfunction) was significantly associated with a three- to fivefold increase in overall miscarriage rate.[80,81] The presence of AITD seems to be an independent risk factor for miscarriage in spontaneous pregnant women, even in the absence of anti-nuclear or anti-cardiolipin antibodies.[82,83]
Geva et al. were the first to describe the association between organ-specific autoantibodies (both thyroid and ovarian) and reproductive failure after IVF. The following studies gave somewhat controversial results.[68,84,] For example, in the study by Kutteh et al., this association not reproduced. It was, however, a retrospective study and thyroid antibodies were measured only during pregnancy, when the antibody titres decrease and patients might therefore have been misclassified. In the study by Muller et al., it was not clearly stated whether the patients underwent one or more IVF procedures, and this would be important to assess the success rate of an ART cycle. Recently, Poppe et al. performed a prospective study to delineate the impact of AITD on the outcome of ART. Four hundred and thirty-eight women were included, presenting for a first attempt of ART. The results showed that the pregnancy success rates were comparable between women with vs. without AITD, but that the miscarriage rate was significantly higher in women with AITD (53% vs. 23%; P = 0·016). Together with previously published studies, this study favours a negative impact of AITD on the outcome of pregnancy after ART. Figure 2 shows all studies published to date on the association between pregnancy outcome after ART and the presence of AITD.
Prevalence of miscarriage in four studies carried out in women who achieved pregnancy after ART: grey bars, women with AITD; black bars, control women without AITD. The number of patients included is given below the authors' names. Reproduced and modified from Stagnaro-Green and Glinoer, with permission of the authors. 
The underlying pathophysiological mechanism for the association between AITD and miscarriage remains hypothetical, as finding an association does not imply a causal relationship. Three working hypotheses have been proposed.[80,81,]  The first is that miscarriage is not directly related to the presence of AITD, but that the latter represents a marker of a more generalized autoimmune imbalance responsible for a greater rejection rate of the foetal graft. This notion is supported by the observation that women with recurrent abortions have an increased number of CD5/20-positive B cells compared to women with no or only one abortion. An abnormal T-cell function has also been reported in women with AITD, including higher numbers of endometrial T cells.[89,90] Furthermore, some animal models favour a role of thyroid microchimerism (i.e. the presence of foetal cells of paternal origin in the maternal thyroid) to facilitate foetal loss. Finally, aberrant immune recognition of thyroglobulin (Tg) and placental antigens by Tg antibodies has been shown in mice immunized with human Tg who experienced a decrease in foetal and placental weights. The second hypothesis postulates that, despite euthyroidism during pregnancy, the presence of AITD could be associated with a subtle deficiency in thyroid hormones or an inadequate response of the thyroid to adapt to the changes associated with increased oestrogen levels, such as ovarian hyperstimulation or pregnancy. Arguments in favour are that, in women with threatened abortion, significantly lower thyroid hormone levels were observed in those who subsequently did have a miscarriage compared with those who delivered successfully. Glinoer et al. showed a tendency towards higher serum TSH levels in women with AITD, compared to women without AITD during the first trimester, and demonstrated that serum FT4 was significantly lowered in such women during late gestational stages. In a recent meta-analysis it was shown that serum TSH was already slightly higher before pregnancy, albeit still within the normal range, in women with AITD compared to women without AITD. The pattern of changes in thyroid hormones after ART-induced pregnancies clearly shows a dissociation for both serum TSH (higher) and FT4 (lower) in women with AITD, compared to those without AITD. In the study by Bagis et al., only women with AITD who had a miscarriage showed differences in their median serum TSH value (higher) and FT4 (lower), compared to women without AITD. In studies investigating pregnancy loss in relation to the presence of antiphospholipid antibodies, it was shown that the time period of miscarriage corresponded to ± 22 weeks of gestation. By contrast, in women with AITD, miscarriages occur mainly in the first trimester of gestation. The third hypothesis is based on the significant association of AITD with infertility. It could be that women with AITD have a subtle underlying fertility problem, leading to conception at a later age (34 years older, on average). Increased age is an independent factor rendering women prone to pregnancy loss. The three hypotheses do not contradict one another, and it remains plausible that the increased risk of pregnancy loss associated with AITD is multifactorial, resulting eventually from a combination of several independent deleterious factors.[81,87]
Medical interventions to decrease the risk of a miscarriage in women with AITD have been based on the above-mentioned hypotheses. Successful modulation of the immune system in patients with AITD has been reported with the use of intravenous immunoglobulins.[97,98,99] Although these treatments were beneficial in terms of pregnancy outcome, these studies included only a small number of patients, women with other autoantibodies (beside thyroid), and often lacked appropriate controls. Only one of the three published studies was carried out in an ART setting. In the study by Vaquero et al., a comparison was made between the beneficial effects of immunoglobulins and thyroid hormone extracts on the outcome of pregnancy in women with AITD and a history of recurrent miscarriage. Thyroid hormones in 16 women yielded a significantly improved live-birth rate (81%), compared with 11 women treated with immunoglobulins during gestation (51%). Abalovich et al. showed that it was not the diagnosis of overt or subclinical hypothyroidism that mattered in relation to the pregnancy outcome but the adequacy of L-thyroxine treatment. Fifty-one pregnancies were conceived under hypothyroidism: 16 overt and 35 subclinical hypothyroidism. Ninety-nine pregnancies were conceived under euthyroidism while undergoing thyroid therapy. When treatment with L-thyroxine was inadequate, abortion occurred in 60% of overtly hypothyroid patients and in 71% of patients with SCH. When treatment was adequate, there were no abortions in any of the groups. In a recent study by Negro et al., spontaneously pregnant women (n = 984) comprising 115 with AITD (12%) were followed for thyroid function and pregnancy outcome. Women with AITD were randomly assigned to group A (n = 57) receiving L-thyroxine or group B (n = 58) without treatment; women without AITD served as controls. The median dosage of L-thyroxine was 50 µg/day in group A and remained unchanged throughout pregnancy. At the time of treatment onset (within 1 week after the first gynaecological visit), mean gestational ages were similar among the different groups. In group B, serum TSH levels increased progressively during gestation, with 19% of untreated women with AITD having a supranormal serum TSH at delivery. The spontaneous serum TSH increment was significantly less marked in controls than in group B. The controls and women in group A (treated) maintained normal serum FT4 levels, while FT4 levels decreased by 30% in group B women during gestation. l-Thyroxine treatment in women with AITD led to a significant decrease in the miscarriage rate from 13·8% to ± 3·5% and premature delivery from 22·4% to ± 7%. Some comments have been made on the limitations of this study. The study was not placebo controlled or double blinded, the mean serum TSH was already significantly higher before intervention in women with AITD compared with controls, and the rationale for the L-thyroxine dose was arbitrary. Despite these limitations, these results show, for the first time, in a prospective randomized study, the beneficial effects of L-thyroxine administration on the outcome of pregnancy in women with AITD. The only intervention study with L-thyroxine treatment in women with AITD pregnant after ART has also been published by Negro et al. In that study, the authors showed that the miscarriage rate was reduced to 33%, compared with 52% in untreated controls. The study, however, failed to reach statistical significance, probably because of the small number of patients and not because of the inclusion of women with different types of infertility, as the outcome of ART women is independent of infertility cause. The studies by Negro et al. should be confirmed by a double-blinded placebo-controlled trial, with an appropriate number of patients, before L-thyroxine can be systematically proposed for euthyroid women with AITD planning to undergo an ART procedure.
The medical preparation for an ART procedure is called controlled ovarian hyperstimulation (COH). This combines a treatment to down-regulate the pituitary gonadal axis using GnRH agonists or antagonists, and stimulation of the ovaries with (recombinant) FSH. When three or more large follicles are seen on echography, treatment is discontinued and 10 000 units of hCG are given to induce ovulation. Depending on the protocol used, 23 embryos are transferred when fulfilling strict morphological criteria. The COH procedure leads to elevated oestrogen and TBG levels, thus potentially influencing free thyroid hormone levels. Maternal FT4 levels play an important role in the placental physiology and in the early foetal development, especially in the first trimester when the foetal thyroid is not yet active.[42,104,105,106,107] Data on changes of thyroid hormone levels very early in the first trimester remain scarce. The optimal setting to investigate thyroid function in these early stages is in women who undergo an ART procedure, in which the biochemical stage of pregnancy can be determined precisely. Muller et al. have investigated thyroid function after COH. They measured thyroid function in the period immediately after COH, without mentioning timing in relation to either induction of ovulation or pregnancy outcome. After COH, a significant increase in serum TSH and a decrease in FT4 levels were noted compared to pre-COH levels. The authors explained these changes in thyroid function by a rapid increase in oestrogen levels and, in turn, a higher TBG production rate and hypersialylation. Any increase in serum TBG concentrations results in an increase in total T4, which tends to lower serum FT4, thereby stimulating serum TSH through the classical pituitarythyroid feedback mechanism.[24,25] In another article, Poppe et al. investigated the impact of COH on thyroid hormone during the first weeks of pregnancy and compared women with and without AITD. In this study, all patients underwent the same COH procedure and had normal baseline thyroid function (i.e. before COH). A significant impact of COH on thyroid function was evidenced, with a significant increase in serum TSH and FT4 levels compared with baseline values. After this initial peak, serum TSH and FT4 levels decreased to pretreatment levels. Thereafter, the levels of FT4 remained normal throughout the first trimester, although mean FT4 was clustered near the lower limit of normality in women with AITD. Follow-up of thyroid function over the first 10 weeks of pregnancy showed significantly higher TSH and lower FT4 values in women with AITD, compared to women without AITD. The conclusions of the study were that COH in preparation for pregnancy places an important strain on the maternal thyroid and also that during the first trimester (i.e. the period associated with the higher incidence of abortions), serum FT4 levels are always lower in the women with AITD. In contrast to the study by Muller et al., Poppe et al. could not show a drop in serum FT4 2 weeks after implantation. A possible explanation for this discordance may be related to the timing of thyroid hormone measurements in both studies, differences in oestradiol levels after oocyte pick-up and/or direct effects of hCG (used for the induction of ovulation) on thyroid function.[27,109]
It is currently thought that the FT4 surge occurring during the first trimester of pregnancy is biologically relevant for the development of the cerebral cortex in the foetus.[104,105,106] Thyroid hormones are found in embryonic cavities soon after conception and are freely available for entry into foetal tissues. Low maternal FT4 levels (even in the presence of normal serum TSH) could disrupt the local availability of T3 and may therefore interfere with normal neuro- and placental development. If thyroid function is further impaired, steroid and cytokine actions may fail to sustain a normal early pregnancy. However, defining what is a normal cut-off value for maternal serum FT4 levels remains hotly debated among experts. One hypothesis is that low T4 levels (mostly due to underlying AITD) may cause miscarriage. In the studies by Muller and Poppe, it was not possible to determine whether the AITD status or low FT4 level was the main determinant of the outcome after ART. In another recent study by Poppe et al., thyroid function was investigated after COH and compared between a group of women with ongoing pregnancies and a group with clinical miscarriages. Mean age and number of transferred embryos were similar in both groups. To avoid the confounding impact of concomitant thyroidal disorders, women who were taking thyroid hormones, antithyroid drugs or who had AITD were excluded. Forty-five women had ongoing pregnancies and 32 had a miscarriage after 511 weeks. Compared with baseline values, serum TSH and FT4 increased significantly 2 weeks after embryo transfer in both the ongoing pregnancy and the miscarriage group. This study indicates that in women without AITD, changes occurring in thyroid hormones after COH were not different with respect to the outcome of ART.
The negative health outcomes linked to thyroid abnormalities during pregnancy have resulted in an increased focus on the question of screening for thyroid dysfunction in the peripartum period. Because of the association between thyroid hormonal abnormalities and miscarriage, the presence of AITD and pregnancy loss, subtle degrees of thyroid dysfunction and the risk of decreased IQ in the offspring, the apparent merit of screening seems established. However, a screening programme can only reasonably be proposed if responses to questions such as the following are convincing. Is the prevalence of thyroid disease during pregnancy sufficient to merit screening? Are the repercussions of thyroid disease during pregnancy significant? Is there a screening test available that is reliable, inexpensive, easily accessible, and accurate? Have intervention strategies been shown to be safe and effective in decreasing the negative consequences of thyroid disorders before and during pregnancy? A certain number of positive answers to these questions have been delineated in the present review, but others still await the burden of proof. The frequency of thyroid disease during pregnancy is sufficient to warrant screening by measurement of TSH and probably also thyroid antibodies. Similarly, an adverse impact of thyroid disorders on both the mother and the foetus has clearly been documented. To date, in the only randomized prospective intervention trial that has been published recently, the authors have documented a significant decrease in the rate of spontaneous miscarriage and preterm delivery in euthyroid women with AITD treated with L-thyroxine since early gestation. However, although compelling, universal screening cannot be recommended on the basis of a single trial. A large-scale randomized trial, the 'Controlled Antenatal Thyroid Study', is under way, which could be of great importance for the screening guidelines. Until further information sheds light on the missing answers to the questions evoked above, the decision to organize a screening procedure must be left to individual judgement. In 2005, an international committee was set up under the auspices of the American Endocrine Society, with the aim to write consensus guidelines for 'thyroid disorders and pregnancy'. One of the authors of the present review (D.G.) was the representative of the European Thyroid Association and the proceedings of the committee's work should soon be endorsed. Some believed that the evidence was sufficient to justify screening all women before or during pregnancy, and this is already implemented as common practice in many centres throughout Europe. For other members, the main problem was that the lack of evidence documenting treatment efficacy still leaves the clinician in a quandary. Universal screening may be premature, but the association of thyroid abnormalities and untoward outcomes during pregnancy (and postpartum) is impossible to ignore. Therefore, the committee came to the consensual agreement that aggressive case-finding in high-risk populations may provide the appropriate balance between inaction and screening the entire population. Targeted screening should be recommended for women who are at an increased risk or have active thyroid disease. This should also be considered in women who have had gynaecological and/or obstetric complications associated with hypothyroidism but in whom treatment has yet to be shown to be beneficial. Thus, screening selected patient groups is probably justified, based on the significant risks to offspring, the probable benefit of treatment, and the probable low incidence of adverse outcomes from intervention. It should be noted that the committee did recommend screening as part of the work-up of infertility. Screening and treatment in women of infertile couples, as proposed by the authors of the present review, are summarized in Fig. 3 in algorithmic format. The major reasons to perform a screening in the setting of infertility are: the increased prevalence of AITD in infertile women compared with fertile women, the beneficial effects of L-thyroxine treatment (in case hypothyroidism) on surrogate endpoints (menstrual cycle, LH pulsatility and hyperprolactinaemia), potentially avoiding an ART procedure and the prevention of an evolution to overt thyroid dysfunction after COH in women with AITD.
Clin Endocrinol. 2007;66(3):309-321. © 2007 Blackwell Publishing
Cite this: Thyroid Disease and Female Reproduction - Medscape - Mar 01, 2007.