AUTOIMMUNE THYROID DISEASE AND PREGNANCY

Effects of pregnancy on immune function

Many autoimmune diseases are affected by pregnancy. In a normal pregnancy, the maternal immune system undergoes a major adjustment to allow the maintenance of what may be immunologically considered a foreign body - with 50% paternal genes -, the developing fetus. Alterations in maternal immune system which permit the successful implantation of the fetal allograft have not yet been definitively identified, but the factors leading to this immune tolerance seem likely to be partially responsible for the generalized improvement in autoimmune thyroid diseases, characteristic of the pregnant state.

In normal pregnancy, along with the overall dampening of the immune system, maternal immune responses have been shown to shift, moving immune responses away from Th1 cell-mediated immunity and reducing antibody production, hence leading to a pattern were both arms of immune responses are reduced 145. Table 14-3 summarizes the main effects of pregnancy on lymphocyte subsets in patients with and without thyroid autoantibodies 146-149. Precise mechanisms by which thyroid antibodies, as well as those directed against other tissues, are suppressed during pregnancy, and often exacerbate after delivery, remain relatively obscure. Presumably, the rapid reduction in immune suppressor functions following delivery leads to the reestablishment and exacerbation of these conditions. The postpartum exacerbation of autoimmune thyroid disease is one of the most striking examples of this phenomenon. This pattern is especially well illustrated in patients with Hashimoto's disease, in euthyroid patients with positive thyroid antibodies who develop postpartum thyroid dysfunction, and in Graves' disease patients who frequently present exacerbations and recurrences of thyrotoxicosis after parturition 150-160.

Thyroid autoimmunity and disorders of female reproduction

Infertility

Infertility is defined as the absolute inability to conceive after one year of regular intercourse without contraception. The overall prevalence of infertility is estimated to range from 10% to 15% and has remained stable over the past few decades. The work up of infertile women usually identifies different causal factors, including male-factor infertility in 30%, female causes of infertility in 35%, a combination of both male and female infertility in 20%, and idiopathic infertility in 15%. Female causes of infertility comprise endometriosis, tubal occlusion and ovulation dysfunction. Among the factors that may negatively influence normal fertility, immunologic factors are known to play an important role in the reproduction processes of fertilization, implantation and early development of the embryo. Different investigations support the association between reproductive failure and abnormal immunological test results, including anti-phospholipid, anti-nuclear antibodies and organ specific autoimmunity, among which the presence of antithyroid antibodies 161, 162.

With regard to thyroid dysfunction, clinical hypothyroidism is clearly associated with female infertility and, in women in the reproductive age, autoimmune thyroid disease (AITD) is the most common cause of hypothyroidism. The association between subclinical hypothyroidism (SCH) and infertility has been evaluated in different studies, but most of the latter are retrospective and uncontrolled 163. The impact of AITD on infertility in women without thyroid dysfunction is even much less clear and the clinical relevance of such possible association remains controversial. In a series of recent studies by the group of Poppe in Brussels, new light was shed on these difficult issues. They performed a controlled prospective study of 438 consecutive couples consulting for infertility and showed that female infertility was significantly associated with AITD without thyroid dysfunction (the strongest association was found in women with endometriosis). In a follow-up study of infertile couples who benefited from Assisted Reproductive Techniques (ART), these authors showed that medically-assisted conception and onset of gestation were not hampered by AITD, but a successful outcome of the ongoing pregnancies was significantly reduced in those women with AITD due to greater early pregnancy loss (see Figure 14-11) 164-167.

The main practical question is whether one should give the benefit of thyroxine administration to infertile women who have positive thyroid antibodies with variable degrees of thyroid insufficiency. Obviously, overt thyroid dysfunction should be treated before conception or planned ART. Since SCH has a negative impact on the outcome of pregnancy after ART, thyroxine treatment should also be advised. Evidence on the treatment of isolated autoimmune features, but without thyroid dysfunction, was insufficiently documented until recently to advise prompt action (see later section on medical interventions).

Another interesting development concerns reproductive function in males with thyroid dysfunction. In males, hyperthyroidism causes alterations in spermatogenesis and fertility, and most studies show that hyperthyroid male patients have abnormalities in seminal parameters, mainly sperm motility. These abnormalities tend to improve and normalize when euthyroidism is restored by treatment. Concerning hypothyroidism in males, severe and prolonged thyroid insufficiency may impair reproductive function, particularly when its onset occurs in childhood. Severe juvenile hypothyroidism may also be associated with precocious puberty. Finally, patho-zoospermia and astheno-zoospermia seem more prevalent in infertile males who present features of AITD 168-170. An interesting study was recently published by Krassas et al. 171. Among 71 men with thyroid dysfunction (1/3rd with hyperthyroidism and 2/3rd with hypothyroidism), the authors found an elevated frequency of erectile dysfunction (56/71; 79%). Moreover, the restoration of a euthyroid status by thyroid treatment also restored a normal (or significantly improved) erectile function.

Miscarriage

Thirty-one percent of all pregnancies end in miscarriage. Generally, women who experience a single pregnancy loss do not routinely undergo an evaluation for the cause of miscarriage. Women who experience recurrent miscarriages (i.e. 0.3%-5% of women), which is defined as three or more spontaneous miscarriages without an intervening live birth, should thoroughly be evaluated for an underlying etiology (such as infections, auto-immune disorders, exposure to drugs, etc.) 172-174.

An association between AITD and miscarriage was first reported in 1990-91, as a serendipitous discovery 175, 176. Since then, an impressive number of studies have confirmed that women with AITD, without overt thyroid dysfunction, have a significantly increased risk of miscarriage. Furthermore, this risk was shown to be independent of the presence of antinuclear and anticariolipin antibodies and, in most studies, increasing age appeared as an independent risk factor for a miscarriage.

In Table 14-4, the information available from thirteen studies investigating the risk of a miscarriage in relation with the presence (versus the absence) of AITD has been compiled 177. The authors concluded that the overall risk of having a miscarriage was 3-fold to 5-fold greater in women with AITD (and apparent euthyroidism). In another review by Stagnaro-Green & Glinoer in 2004, a classification was attempted by examining separately an association between AITD & miscarriage (in 5 studies), between AITD & recurrent miscarriage (in 7 studies), and finally between AITD & early pregnancy loss after ART (in 5 studies) 178. Overall and with only few exceptions, all studies documented a statistically significant relationship between thyroid autoimmunity and an increased risk of pregnancy loss. Finally in 2004, Prummel & Wiersinga published a meta-analysis of both the case-controlled and longitudinal studies published since 1990, after the association between miscarriage and AITD was first described 179. The results of this meta-analysis amply confirmed that an association exists, with an overall increased relative risk of a miscarriage of 2.73 in women with AITD.

Foot-note to Table 14-4: summary of information provided by the analysis of 13 studies carried out over the last decade in three continents. Over 5,500 women were investigated, both as study cases and controls. Prevalence of AITD varied widely, from 6% in Brussels to 33% in Salt Lake City. Together, the main results (except in 2 studies) concurred to establish that AITD is significantly associated with an increased miscarriage rate.

To find an association between AITD and miscarriages does not imply a causal relationship, as underlying causal mechanisms might also be attributable to a combination of factors that would potentially lead to miscarriage by themselves. Three main hypotheses have been proposed. The first is that miscarriage is linked to a generalized immune imbalance. For instance, women who have had multiple miscarriages have an increased number of CD5/20+ B cells compared with women who have had one or none. Moreover, abnormal T-lymphocyte function has been reported in women with AITD, including a higher number of endometrial T cells. Finally, aberrant immune recognition of thyroglobulin (Tg) and placental antigens by antibodies to Tg has been demonstrated in mice immunized with human Tg, and resulted in decreased fetal and placental weights 180-182. Thus, based on this 1st hypothesis, AITD would merely represent a marker of an underlying, more generalized immune imbalance that, in turn, would explain a greater rejection rate of a fetal graft. In the second hypothesis, the presence of AITD is thought to be associated with inappropriate low levels of thyroid hormones for the given gestational period, despite apparent biological euthyroidism. Although this hypothesis has sometimes been undermined by the difficulty in defining strictly normal thyroid function tests during pregnancy, data to support it have been obtained from women at high risk of miscarriage, among whom thyroid hormone levels were significantly reduced only in those who subsequently actually miscarried. For instance in the study by Bagis et al., only women with AITD and who experienced a miscarriage showed a difference in median serum levels of TSH and T4 compared to women without AITD 183. Thus, based on this 2nd hypothesis, AITD would be associated with a subtle deficiency in thyroid function, i.e. a lesser ability to adapt adequately to the changes associated with the pregnant state, because of a reduced functional reserve characteristic of chronic thyroiditis. The third hypothesis is based on maternal age. Women with AITD are generally older than healthy controls and increased age is an independent risk factor for miscarriage. Thus, based on this 3rd hypothesis, AITD could act by delaying the occurrence of conception because of its known association with infertility. Thyroid antibody-positive women would tend to become pregnant only at an older age (3-4 years older, on the average) and be more prone to pregnancy loss. Overall, these hypotheses do not contradict one another, and it remains plausible that the increased risk of pregnancy loss associated with AITD results from a combination of several independently harmful factors 184-189. For more detailed insight into this complex topic, the readers are referred to the recent review by Poppe et al 190.

Women undergoing ART

Several studies have examined whether miscarriages were more frequent in infertile women who underwent ART, according to the presence versus absence of thyroid autoimmunity. While some studies showed a 2-fold to 3-fold difference in the miscarriage rate in thyroid antibody-positive versus antibody-negative patients, other studies did not confirm such findings 165, 184, 191-194. The largest of these series (retrospective) failed to demonstrate an adverse effect on the miscarriage rate in antibody-positive versus antibody-negative women undergoing ART 194. The prevalence of thyroid autoimmunity, in women undergoing ART, was examined in four studies and found to range between 14%-22%, a prevalence that was not statistically different from the prevalence of thyroid antibody in women who did not undergo ART 165, 192, 193, 195. Pregnancy rates have also been examined in women with or without thyroid autoimmunity undergoing ART and the results were also conflicting. In three of these studies, there was no difference in the overall pregnancy rate (see also Figure 14-11) 165, 184, 196. In other studies, however, pregnancy rates were found to be lower by 1.5-fold to 2-fold in thyroid antibody-positive women, compared with those without antibodies 191, 197. In summary, the literature on pregnancy loss related to women with thyroid autoimmunity undergoing ART is mixed. The methods of ART were not consistent between the series nor were the causes of infertility controlled for among the various studies. Given that the majority of the studies did find a relationship, there is at least a suggestion that such a relationship may exist, but without sufficiently clear evidence to draw a definitive conclusion.

Medical intervention in women with AITD

Medical intervention to reduce the miscarriage risk in women with AITD consists of immunomodulation or thyroxine administration. Successful modulation of the immune system was reported in patients with AITD who received immunoglobulins with (or without) additional heparin or aspirin 198-200. Although these treatments were beneficial in terms of pregnancy outcome, the studies included only a small number of patients, as well as women with auto-antibodies other than thyroid antibodies, and often lacked appropriate controls.

In the study by Vaquero et al, a comparison was made between the beneficial effects of immunoglobulins and thyroid hormone extracts (started before conception and continued until mid-gestation) on the outcome of pregnancy in women with AITD with a history of recurrent miscarriage 201. In the 16 women treated with thyroid hormones, significant improvement was observed in the live birth rate compared with that among the 11 women who received immunoglobulins (81% versus 51%). Criticisms of the study related to the small number of patients and the fact that thyroid hormone extract therapy was started before conception 188. Negro et al. performed two studies, one in women who became pregnant after use of ART and the second in spontaneously pregnant women 192, 202. In the first study, the miscarriage rate was reduced to 33% among thyroxine-treated women, compared with 52% among untreated controls, although this difference was not significant, probably because of the small number of patients. In the second study, the same group of authors followed up a large group of pregnant women, 12% of whom were positive for AITD. Half of the AITD-positive women were treated with thyroxine during gestation, while the other half was left untreated. The end points of the study were outcomes of pregnancy and changes in thyroid function, assessed by comparing the women with AITD with and without treatment and also the healthy pregnant controls. Striking reductions in the rates of miscarriage (by 75%) and premature delivery (by 69%) were reported among women with AITD who had received thyroxine since early gestation and throughout pregnancy. Furthermore, thyroxine-treated women with AIT maintained a euthyroid status, while free T4 decreased by 30% and TSH levels increased progressively during gestation in the untreated group; 19% of the latter women became subclinically hypothyroid at the time of parturition. The study was criticized because it was not placebo-controlled or double-blinded. It did, however, provide prospective data from the first randomized trial that confirmed the efficacy of thyroxine administration in pregnant, euthyroid women with autoimmune features 203.

These findings have implications for screening and medical intervention. For instance, if delayed conception plays a significant role to explain decreased fertility in women with AITD, it would certainly constitute an argument for screening systematically infertile women for the presence of mild thyroid underfunction that is so frequently associated with thyroid antibodies, particularly when women seek medical advice before IVF procedures. Such an approach was used in Finland in 2000 204. A study by Arojoki et al. showed a high prevalence of women with elevated serum TSH levels, an association between oligo-amenorrhea and abnormally elevated serum TSH values and an overall improvement in the success rate of induced pregnancies after thyroxine administration. Finally, women with AITD could be advised to plan for a pregnancy at a younger age, although this type of medical advice is more easily said than applicable in practice.

To conclude on this section, although a clear association exists between thyroid autoimmunity and pregnancy loss, systematic screening can not be universally recommended at present time, at least until adequately designed therapeutic trials will demonstrate beyond doubt a clear reduction in the rate of miscarriage with thyroxine treatment. This being said, more and more data point to the growing interest of screening as well as early thyroxine administration before/during pregnancy, and many centers, in Europe and elsewhere, already routinely screen women with infertility and/or miscarriage for the presence of thyroid autoimmunity and dysfunction.

Effects of pregnancy on thyroid function in women with thyroid auto-antibodies

Table 14-5 lists the various types of autoimmune thyroid disorders that can be found in the pregnant and postpartum population. These aspects are also discussed in greater detail below and postpartum thyroiditis is reviewed in Chapters 8 and 13.

The prevalence of AITD in the pregnant population is comparable to that found in the general female population with a similar age range, i.e. between 5-15% 205. In first trimester patients with gestational diabetes mellitus, the prevalence of thyroid antibodies is even higher (20-25%) 206, 207. Taken together, the high frequency of thyroid antibodies, increased risks of miscarriage, risks of developing hypothyroidism with the progression of gestation, and finally the observation that postpartum thyroiditis occurs in one half of women with AITD led us to recommend that all pregnant patients be screened for the presence of TPO antibodies during the 1st trimester of pregnancy 208.

More than a decade ago, we undertook a prospective study in women with AITD and a normal thyroid function in early pregnancy. The aim was to evaluate sequentially the changes in thyroid function occurring with progression of gestation to term, without medical intervention 209. The study showed that despite the expected decrease in antibody titers during gestation, thyroid function gradually deteriorated towards hypothyroidism in a significant fraction of such women (see Figure 14-12). In the 1st trimester, serum TSH (albeit within the normal range) was already significantly shifted to higher values in women with AITD, compared with normal pregnant controls. Serum TSH remained higher throughout gestation and at parturition 40% of AITD-positive women had a serum TSH >3 mU/L, with almost one-half of them above 4 mU/L. Thus, while women with AITD were able to maintain a normal thyroid function in early gestation (due to sustained thyrotropic stimulation), their mean serum free T4 levels were significantly reduced to (or below) the lower limit of the normal reference range at delivery. Average reduction in serum free T4 reached 30% and almost one half of these women had free T4 values in the hypothyroid range by the time of delivery, confirming that these women have a reduced functional thyroid reserve. The risk of progression to hypothyroidism could be predicted from serum TSH levels and TPO-Ab titers measured in early pregnancy. When serum TSH was already above 2.5 mU/L and/or TPO-Ab titers above 1,250 U/mL before 20 weeks, these markers were indicative of the propensity to develop hypothyroidism by the end of pregnancy. These observations are important, providing clinicians with simple tools to identify during early gestation those women who carry the highest risk. As a consequence, thyroid function can then be closely monitored and preventive thyroxine treatment administered, to avoid the potential deleterious effects of hypothyroxinemia on both maternal and fetal outcomes.

Recommendations and main "take home" messages

Pregnancy dampens the immune system, leading to a pattern where both arms of the immune responses (cell-mediated and humoral) are reduced. The rapid reduction in immune suppressor functions following delivery leads to the re-establishment and exacerbation of these conditions during the postpartum.

Even in the absence of evident thyroid dysfunction, there is good evidence to suggest that thyroid autoimmunity is associated with an increased risk of infertility. This constitutes an argument for the systematic screening of infertile women for the presence of mild thyroid underfunction, frequently associated with thyroid autoimmunity, particularly when these patients seek medical advice before in vitro fertilization procedures.

With regard to pregnancy loss, the vast majority of available studies have clearly established that thyroid autoimmunity is associated with a significant increase in the risk of miscarriage. Association does not imply causality and the etiology of this association is probably multi-factorial, including underlying dysregulation of the immune system, subtle forms of mild thyroid failure, and older age. Although there is a positive association between presence of thyroid antibodies & pregnancy loss, universal screening for thyroid antibodies (and possible treatment) can not be recommended at this time.

With regard to the repercussions of positive thyroid antibodies, the main risk is the occurrence of maternal hypothyroidism, with its potential deleterious effects for both the mother and fetus. This could be prevented by systematic screening for thyroid dysfunction and presence of thyroid antibodies during early gestation, followed by the administration of thyroxine treatment when required.