INTRODUCTION
Advances in our understanding of the mechanisms responsible for autoimmune thyroid disease (AITD) have been developed in a few years, in large part by the application of novel molecular techniques. There have been a major comprenhension about self-tolerance mechanisms and how these may fail, antigen presentation, funcional T cell subsets and specific T cell receptor (TCRs) in autoimmunity. More gradually, these findings are being translated to thyroid autoimmunity where undoubtedly, the major advance in the last decade was the cloning and sequencing of the three major thyroid antigens: thyroglobulin (Tg), thyroidperoxidase (TPO) and the TSH receptor (TSH-R). As a result, detailed characterization of autoantibodies is now posible1 and its measurement useful in the clinical setting.
Genetic predisposition, together with ill-defined environmental and endogenous factors, determines susceptibility to AITD. Most attention has been focused on the role of HLA genes. A meta-analysis found evidence only for HLA-DR3 and -DR4 association in Hashimoto's thyroiditis, with a higher relative risk for the former2. However, subsequent studies have failed to substantiate this. Given this weak HLA association, it seems likely that one or more non-HLA genes makes a major contribution to susceptibility and novel approaches will be required to identify these3. On the other hand, we know more about the importance of environmental and endogenous (most probably hormonal) factors in thyroid autoimmunity4.
Another advance has been the recognition that the thyrocyte is not a helpless target of autoaggression, being capable of expressing a wide array of immunologically active molecules, which may exacerbate or diminish the autoimmune response5. Defining the contribution of thyrocytes to the intrathyroidal autoimmune response, whether from released cytokines or surface-bound molecules will be crucial to our future understanding as well as holding the promise that these thyroid derived production might be therapeutical targets.
AUTOIMMUNE THYROID DISEASE DURING PREGNANCY AND POSTPARTUM
Thyroid disorders are observed 4-to 5 -fold more frequently in women than in men, in particular during the childbearing period. It is therefore not unusual to encounter thyroid function abnormalities during a "routine" laboratory evaluation carried out for pregnant women. Pregnancy is accompanied by profound alterations in thyroidal economy, resulting from a complex combination of factors specific for the pregnant state: the rise in T4-binding globulin concentrations, the effect of hCG on the maternal thyroid, alterations in the requirement for iodine and modifications in autoimmune regulation.
Thyroid peroxidase (TPO) is now recognized as the specific antigen of thyroid microsoma. A new sensitive method for detection of antibodies to TPO (TPO-Ab), has been developed a few years ago 6. In our experience7, TPO-Ab determination using monoclonal antibody radioimmunoassay provides a clear improvement in the sensitivity of autoimmune thyroid disease diagnosis. We propose that in terms of systematic screening of thyroid autoimmunity TPO-Ab is clearly to be preferred than classical antimicrosomal antibodies (MIC-Ab) and antithyroglobulin (Tg-Ab)8.
In a cohort study of pregnant women with mild underlying abnormalities published in 1991, it was noted that women who are euthyroid but carry positive thyroid antibodies at the onset of pregnancy, have an increased risk of developing hypothyroidism during gestation9. We therefore investigated more systematically the role of autoimmune thyroid disease (AITD) on thyroid function during pregnancy. The systematic screening at the time of the initial visit (before 16 weeks) of thyroid autoantibodies among 437 healthy pregnant women, with no previous history of thyroid disorder, has showed a 9% prevalence of positive TPO-Ab titers, a 2.6% for positive Tg-Ab titers and a 6.5% for positive MIC-Ab. These percentages were not different from the nonpregnant female which is in the range of 5-10% reported by other authors10. We have found a highly significant positive correlation between TPO-Ab and MIC-Ab during early pregnancy in a subset of the sample. The same significant positive correlation, was found between TPO-Ab and Tg-Ab, in the complete cohort of pregnant women before 16 weeks (fig. 1).
Thyroid antibody titers were monitored sequentially during gestation and postpartum. A significant decline, both in TPO-Ab and Tg-Ab was observed during late gestation (fig. 2). These results provide evidence the immunosuppressive effect of pregnancy as a general phenomenon. We propose that if systematic screening of thyroid autoimmunity is to be organized during pregnancy, antibody determination should be carried out as early as possible during gestation.
Women with a positive test for TPO-Ab early in gestation showed a highly significant decrease in FT4 and increased TSH levels in the late trimester of gestation, despite a significant decline in TPO-Ab titers. These results are in agreement with Glinoer et al11, who recently showed that women with asymptomatic autoimmune thyroid disease who are euthyroid in early pregnancy carry a significant and progressive risk of becoming hypothyroid during gestation, despite a marked reduction in antibody titers. Progression to subclinical hypothyroidism was associated with and predicted by serum TSH levels and TPO-Ab titers in the first trimester. Hence, these parameters can be used to provide useful markers to identify those women who carry a higher risk and, therefore, initiate hormone substitution therapy.
Iodine and AITD during pregnancy
Some experimental studies12 have showed that iodine supplementation might be associated with an increase of thyroid autoantibodies. These results have been recently refuted13 in a group of women who received iodine supplementation (300 Fg KI/day) during pregnancy. In this study no mother developed autoantibodies during pregnancy and those mothers having already positive TPO-Ab at the beginning of the gestation showed no increase of the antibody titers.
Fetal loss and AITD
Another important question concerns the rate of spontaneous abortions in women with positive TPO-Ab. Our studies9 were the first to report a strong correlation between positive thyroid antibodies and the risk of spontaneous miscarriage in women who were clinically and biochemically euthyroid. Similar findings were reported by Stagnaro et al14 and Bussen et al15, emphasizing the notion that the risk of miscarriage occurs primarily in the first trimester and that women with a history of consecutive abortions carry an even greater risk. The most plausible hypothesis is that such women present an underlying more generalized defect in autoimmunity, leading to increased fetal loss16. Therefore, the early determination of thyroid auto-Ab, and in particular TPO-Ab, could serve as a sensitive marker to identify women at risk. TPO-Ab have been shown to cross the placenta9; however, the effect of these antibodies on fetal growth and subsequent development is not well known. It's noteworthy the study of Pop et al17, who concluded that children of pregnant women who have elevated TPO-Ab but normal thyroid function are at increased risk for decreased cognitive development. However, larger samples of women with TPO-Ab are needed to determine whether impaired childhood development is related to the titer of TPO-Ab or envirommental factors.
Smoking and AITD during pregnancy
There is an increasing suggestion of an effect of smoking on autoimmune thyroid disease. In particular, smoking has been shown to worsen Graves disease and Graves' ophthalmopathy18. The exact role of smoking on the autoimmune mechanism is unclear. However, it is known that smoking has an antithyroid effect19. We have shown a significant elevation in levels of TPO-Ab in pregnant women who smoked more than 10 cigarettes/day before gestation, compared with those who smoked less than 10 cigarettes/day or who were not smokers. A test for Tg-Ab showed similar results (fig. 3). We recommend systematic screening for TPO-Ab in women who smoke prior to or during early gestation in view of the effects of pregnancy on autoimmune thyroid disease20. Those with a positive TPO-Ab should be followed serially with testing for FT4 and TSH during gestation and the postpartum period as the risk of clinical thyroid disease in women with elevated TPO-Ab levels is big.
GRAVES DISEASE DURING PREGNANCY AND POSTPARTUM
Even though the immunological consequences of pregnancy are not yet fully elucidated, major influences on the course of autoimmune disease have been recognized for decades. In particular, pregnancy associated improvement in the hyperthyroidism of Graves, disease (GD) and exacerbation of hyperthyroidism during the postpartum period have been described21.
The prevalence of hyperthyroidism in pregnant women has been reported to be in the range 0.05-0.2%. Between 1974 and 1990 in the Los Angeles County (University of Southern of California) the prevalence of hyperthyroidism was 0.1%22. A screening of thyroid function in 9.453 early pregnant women residing in Japan revealed that 0.4% had hyperthyroidism (mainly untreated GD) and another 0.4% had a transient subclinical hyperthyroidism23. When the diagnosis of GD has not yet been established before the beginning of pregnancy, the disorder is not always readily suspected clinically, mainly because the symptoms and signs of mild to moderate hyperthyroidism may mimick the hypermetabolic state of normal pregnancy. Some findings such as a heart rate above 100 per minute, diffuse goiter and weight loss may be relevant and help in the diagnosis. Thyroid stimulating antibodies (TSAb) are considered to be the cause of hyperthyroidism in GD. Accurate diagnosis of GD is important, because untreated hyperthyroidism is associated with increased fetal loss, premature labor, and low birth weight24. The severity of hyperthyroidism frequently decreases during gestation in patients with Graves, disease, which often tends to recur after term. Due to the immune suppression associated with the pregnancy state, there is a progressive decrease in the titers of TSAb, as gestation progresses24. The reduced availability of iodine for the maternal thyroid may also play a role to improve the course of the disorder, at least when pregnancy occurs in women who reside in areas with a restricted iodine supply. It should, however, be noted that transient exacerbations of hyperthyroidism near the end of the first trimester (associated with peak hCG) are not exceptional25.
From 1988-1990 we analyzed in a prospective sequential study the changes in thyroid function and in the TSAb in patients with GD during pregnancy and postpartum. The patients constituted two different groups. One group was untreated before becoming pregnant and was named "remission group". The other group was taking antithyroid treatment, and was named "active group". Striking differences during pregnancy and postpartum period were shown when active and remission groups were compared. During early gestation some women with active GD, associated with a high rate of miscarriage (20%). However, throughout gestation these women improve, with a significant reduction of dosage of ATD. They started with mean dosage of 22 mg/day of methimazole and the mean dosage at delivery was 10 mg/day. The improvement observed during the la. te gestation, is likely to be related to the reduction in TSAb titers.
In the whole group, the levels of antibodies to TSH receptor decreased during pregnancy and increased after delivery, indicating that antibody production is regulated by events specific to pregnancy. As has been reported, immunosuppressive factors originating from the fetus, the placenta, or the mother (decreased circulating T-helper cells), could play a role in the level of antibodies to the TSH receptor26. The production of antibodies to the TSH receptor may be activated after delivery resulting in postpartum thyroid dysfunction. Fifty-five percent of the women in the active and 24% in the remission groups had a relapse of their hyperthyroidism, associated with an increase in the level of antibodies to TSH receptor. These findings suggest that the level of antibodies to TSH receptor is related to the frequency of relapse of hyperthyroidism. Hashizume et al27, provided evidence that the administration of L-thyroxine during the last trimester of pregnancy and the first year after delivery was effective in maintaining a lower level of TSAb and in decreasing the rate of postpartum hyperthyroidism in patients with GD. However, not everybody is in agreement with these results and the controversy still persists28.
Treatment of Graves' disease during pregnancy
Antithyroid drugs (ATD) are the therapy of choice in pregnant women, unless the severity of the condition justifies a more radical approach by surgery (which is then preferably carried out in the second trimester). Preference is usually given to propylthiouracil (PTU) over methimazole (or carbimazole), although this choice is not mandatory, as long as the minimal dose of ATD is implemented29. A controversial issue is whether the combined therapy (ATD associated to L-thyroxine) may be effective in controlling maternal hyperthyroidism and avoid fetal hypothyroidism and goiter. We do not recommend this combined therapy for the following reasons: first, transplacental passage of thyroid hormones is negligible and doesn't prevent the increase of fetal TSH; second, it may lead to the administration of larger doses of ATD which may cause fetal hypothyroidism and goiter.
In our personal experience, the combined therapy does not provide any help in the management of hyperthyroidism during pregnancy. We propose the use of the lowest dose ATD dose to maintain free thyroid hormones in the normal range or even in the upper limit of normality. It has been shown that such levels in maternal blood are associated with free hormone concentrations in the fetus that remain in the mid-range of normal values30.
Pregnancy complicated by hyperthyroidism is widely recognized as a cause of impaired fetal outcome. Momotani et al29, in a large number of hyperthyroid pregnant women observed that hyperthyroidism at the time of conception was associated to abortion in 25% and premature delivery in 15% of the women compared with 13 and 10% in euthyroid patients.
It is strongly recommended that thyroid stimulating antibodies (TSAb) titers are assayed in early pregnancy and the last trimester, because high TSAb levels predict the risk of neonatal hyperthyroidism and of recurrences of thyrotoxicosis during the postpartum period31.
Another controversial issue is breast feeding in mothers treated with ATD. For many years, breast-feeding was forbidden if ATD drugs were being used. Some studies have documented that, after a single dose of 400 mg of PTU the concentration of the drug in the milk was only 10% of that of serum and that only 90 µg of PTU were present in the milk collected over four hours32. Therefore mothers with GD who are receiving PTU have been advised that breast-feeding their infants is safe33,34.
In a recent study, serum thyroid hormones in newborns whose mothers were treated with maintenance doses of me thimazole were within the normal range limits35. Further studies should be conducted to evaluate the safety of long-term treatment before recommending breast feeding in the mother treated with antithyroid drugs.
Not autoimmune gestational hyperthyroidism or gestational transient thyrotoxicosis (GTT) has been also characterized36. This form of hyperthyroidism is different from GD, in that it occurs in women without a past history of GD and without detectable thyroid stimulating antibodies. GTT is not always clinically apparent, because it is most often transient. Its etiology is directly related to the thyrotropic stimulation of the thyroid gland associated with hCG.
Glinoer et al37 found an overall prevalence of 2.4% in a prospective cohort study undertaken between the 8th and 14th week of gestation. Symptoms compatible with hyperthyroidism were present, namely weight loss, absence of weight increase, tachycardia and fatigue. Patients with GTT manifested free T4 concentrations in the thyrotoxic range and circulating hCG abnormally elevated. Most women required no treatment with antithyroid drugs and were given beta-blocking agents for a short period, with a noticeable improvement of symptoms. In all cases, GTT was transient and the normalization of free T4 concentrations paralleled the decrease in hCG. GTT was not associated with a less favourable outcome of pregnancy. The presence of a variant hCG molecule with a potent thyrotropic activity is possible and has been advocated by several authors38, although this hypothesis is not absolutely required to explain the disorder. The cause of the anomaly in hCG regulation is presently unknown. Clinicians should be aware of the disorder and closely monitor thyroid function tests and hCG levels in women with symptoms suggesting hyperthyroidism and with early gestational emesis.
POSTPARTUM THYROIDITIS
Postpartum thyroiditis (PPT) is usually characterized by the development of transient hyperthyroidism and transient hypothyroidism three to six months after delivery. The condition is associated with the presence of circulating anti-TPO in the great majority of cases, and a pathogenetic similarity to Hashimoto's autoimmune thyroiditis has been suggested. Only half of the women identified as anti-TPO positive at around 12 to 16 weeks gestation will develop PPT. The other 50% will be euthyroid anti-TPO positive in the postpartum period39. The incidence of PPT varies from 5-9%40. In our area with mild iodine deficiency we found a 5.7% prevalence of PPT at 3 months postpartum (1.9% of patients had hypothyroidism and 3.8% had hyperthyroidism)41. In Catalunya, where the iodine intake is sufficient, the reported prevalence was 9.3%42. This variation may be due to the frequency of postpartum assessment, especially with regard to the recognition of postpartum hyperthyroidism, which may last only a few weeks. PPT is usually transient, although permanent hypothyroidism may ocur in up to 30% of patients43. Lazarus et al44, observed that there is about a 70% risk of developing PPT after a subsequent pregnancy if the condition has ocurred previously. Some authors have suggested that the titer of anti-TPO antibodies is a valid predictor of the severity of PPT and possibly of recurrent disease45. Although women with PPT after the first pregnancy had a greater incidence of thyroid enlargement than PPT negative TPO-Ab positive women, the presence or absence of goitre is not an accurate predictor of subsequent disease. These data reinforce the suggestion that women who are found to have thyroid dysfunction or circulating thyroid antibodies after pregnancy should be carefully assessed after a second birth.
TYPE 1 DIABETES MELLITUS AND AUTOIMMUNE THYROID DISEASE DURING PREGNANCY AND POSTPARTUM
Type 1 diabetes mellitus is commonly associated with other autoimmune endocrine disorders. Thyroid autoantibodies occur more frequently in these patients than in the normal population46. Furthermore, an increased prevalence of subclinical hypothyroidism has been reported in pregnant women with diabetes47. Several studies have found a wide range of prevalence of postpartum thyroid dysfunction (10 to 25%) in patients with type 1 diabetes mellitus48. This is a 3- fold increase compared to a similar study by the same group in a non-diabetic population.
We have shown that the prevalence of a positive test for TPO-Ab is increased in patients with type 1 diabetes mellitus and we reported a prevalence of 38% in our study population, with a higher frequency in women than in men. Our study in patients with type 1 diabetes mellitus showed a fourfold increase in thyroid antibody positivity in comparison with the autoantibody prevalence of 8.7% previously found by our group in healthy subjects7.
Thus, we studied the immunologic aspects and function of the thyroid gland in women with pregestational diabetes before conception, during pregnancy, and during the immediate postpartum period. Our goal was to evaluate the degree to which thyroid function was affected during these various periods and to determine whether glycemic control and neonatal complications correlated with thyroid status.
Pregestational diabetes and antithyroid antibodies
Seven our of 20 women (35%) had positive anti-TPO-Ab levels before pregnancy. The level of antibodies progressively declined during pregnancy, but they were never entirely negative. With the anticipated reactivation of the immune system during the postpartum period, 40% (8/20) of the women had positive antibody levels. The prevalence of positive anti-TPO of 38% in these women is equal to that previously described in subjects with type 1 diabetes mellitus7; and 3 to 4 times greater that found in the general population before, during, and after pregnancy10.
Antithyroid antibodies and thyroid function
During pregnancy, none of the women with negative antithyroid antibodies had a change in thyroid function; however, in three of the seven patients (43%), with positive anti TPO-Ab, hypothyroidism (TSH>10 mU/l) developed during pregnancy or the postpartum period and necessitated replacement therapy with thyroid hormone. Of the pregnant women with pregestational diabetes, 15% developed alterations in thyroid function during pregnancy - similar to the rate previously described by Jovanovic and Peterson47 and 2 to 3 times higher than that found by Gray et al49 in non-pregnant women with diabetes. In the immediate postpartum period, the prevalence of thyroid dysfunction was also 15% (3 of /20 patients). In the group with negative antithyroid antibody levels, the prevalence was 8%, significantly less than that seen in the women with positive antibody levels 29% (2 of 7). The global prevalence of thyroid dysfunction during pregnancy is reported to be between 10.5% found by Bech et al48 and 25% by Álvarez-Marfany et al50. These last two studies conducted postpartum follow-up of patients for 1 year, whereas our study had a 3-month postpartum follow-up50.
Our prevalence rate may indeed have been greater than 15% had follow-up of patients been continued for up to one year as well. In all studies, the development of thyroid dysfunction was greater in those women with increased anti-TPO-Ab.
Thyroid autoimmunity and glycemic control
Metabolic control was acceptable for both groups under study with glycosylated hemoglobin (HbA1c ) levels below 7% before and during pregnancy. Nevertheless, in the second and third trimesters, HbA1c levels in the group with positive anti-TPO-Ab were significantly higher than in the negative group (table 1). Another notable point is that, despite both groups having similar characteristics (table 2), the insulin dose per kilogram of body weight was also significantly higher51. These results are in agreement with those reported by Bech et al48 who also found higher HbA1c levels in pregnant patients with diabetes and positive anti-TPO-Ab than those with negative anti-TPO-Ab. In light of these results and considering the work of Vjvodic et al52 who suggest that hypothyroidism is a risk factor for gestational diabetes, we must ask whether a relationship exists between insulin resistance in pregnant patients with dia betes and elevated anti-TPO-Ab levels. If there is a relationship with insulin resistance, is it attributable to the hypothyroidism or are the actual anti-thyroid antibodies responsible? The answer continues to be evasive.
Thyroid status and fetal outcome
In contrast with previous investigations that have found a negative correlation between neonatal outcome and anti-TPO-Ab levels, our study failed to show a correlation between neonatal complications and the presence or absence of antithyroid antibodies53. This result is somewhat surprising, and we believe it may, in part, be due to small sample size. We have previously demonstrated increased fetal losses in patients with elevated antithyroid antibodies that have crossed the placenta9.
The increased prevalence of positive antithyroid antibodies in patients with type 1 diabetes mellitus as well as the relationship with thyroid dysfunction during pregnancy and the postpartum period, in addition to the possible repercussions on maternal glycemic control and ultimately the well-being of the neonate, makes it strongly advisable to screen for antithyroid antibodies routinely in all women with diabetes who are planning pregnancy. Additionally, those testing positively should be monitored with serial free T4 and TSH each trimester and in the postpartum period54,55.
GRATEFULNESS
This work was supported by grants from National Health Service (1986-1988), from Andalusian Health Service (1992-1995) and from Fondo de Investigación Sanitaria (1996).