Vol. 9, No. 10, December 1998
Reviews
The Systematic Screening and Management of Hypothyroidism and Hyperthyroidism During Pregnancy
Daniel Glinoer
Trends in Endocrinology and Metabolism 1998, 9:403-411
Department of Internal Medicine, Thyroid Investigation Clinic, University Hospital Saint-Pierre, 322 Rue Haute, B-1000 Brussels, Belgium. |
|
Article Outline
- Abstract
- Hypothyroidism and Pregnancy
- Hypothyroidism (HO) in the General Population
- Fertility and Pregnancy Outcome in Hypothyroid Women
- AITD and the Risk of Miscarriage
- AITD and the Risk of Hypothyroidism in the Pregnant State
- Subclinical and Overt Hypothyroidism During Pregnancy
- Systematic Screening for AITD and Hypothyroidism During Pregnancy
- Hyperthyroidism and Pregnancy
- Hyperthyroidism (HR) in the Community and in Pregnancy
- GD in Pregnancy
- The Management of GD in Pregnancy
- Gestational Transient Thyrotoxicosis (GTT)
- Systematic Screening for Hyperthyroidism During Pregnancy
- Conclusions
- References
- Copyright
Abstract
Altogether, thyroid function abnormalities during pregnancy can affect up to 10% of all women. The high prevalence of both hypo- and hyperthyroidism, the obstetrical repercussions associated with thyroid dysfunction in the mothers, as well as the potential role of maternal thyroid dysfunction as an influence on fetal development constitute solid arguments for a further increase of our knowledge of the pathophysiological processes underlying the alterations of thyroid function related to the pregnant state. In this review, the focus will be on the most clinically relevant aspects associated with hypothyroidism [autoimmune thyroid disorders (AITDs), subfertility, risk of miscarriage, risk of hypothyroidism in women with AITD and treatment of hypothyroid women] and with hyperthyroidism (clinical presentations during pregnancy, Graves' disease and its management, fetal hyperthyroidism in women with antithyroid-stimulating hormone receptor antibodies and gestational transient thyrotoxicosis associated with human chorionic gonadotropin stimulation of the maternal thyroid gland). I also propose a global strategy for the systematic screening of hypo- and hyperthyroidism in the pregnant state.
Hypothyroidism and Pregnancy
Hypothyroidism (HO) in the General Population
Twenty years ago, a representative sample of an entire adult community was surveyed for thyroid disease, and in particular for thyroid insufficiency, in the northeastern part of England (the Whickham study) (
Tunbridge et al., 1977). Even today the Whickham study remains a unique example, and other surveys carried out with different methodological approaches and in different parts of the world have largely corroborated the main findings of the original survey (Wang and Crapo, 1997). Globally, these studies have shown that the prevalence of HO is high in the population (1–10%), that it remains frequently undiagnosed (subclinical), and that it is mainly a female disorder. The main cause of HO is chronic autoimmune thyroid disease (AITD), and the prevalence of antithyroid antibodies is closely associated with aging in the female population (6% below 45 years of age; 10% between 45 and 75 years; >15% over 75 years); in parallel, the prevalence of HO increases with aging. Recently revisited in a follow-up study, the original Whickham survey has yielded another major finding, namely, that the incidence of overt HO is at least four cases per 1000 women per year, with a close association between the rise in thyroid-stimulating hormone (TSH) and positive antithyroid antibodies (Vanderpump et al., 1995). Hence, because of its high frequency in the young population and its female preponderance, it is logical to consider whether HO might both affect pregnancy and its course be affected by pregnancy.
Fertility and Pregnancy Outcome in Hypothyroid Women
There is a known association between HO and decreased fecundity which, in most instances, is primarily associated with ovulatory disturbances and not with miscarriage: women who require treatment with thyroid hormone have a twofold greater risk of primary ovulatory infertility (Garber, 1997).
When hypothyroid women do become pregnant, they have an increased risk of obstetrical complications (such as intrauterine fetal demise, gestational hypertension, placental abruption and poorer perinatal outcome). Several studies have indicated that adequate thyroid hormone administration greatly improves, although it does not entirely suppress, the frequency of these abnormalities (Montoro et al., 1981). In general, infants of hypothyroid mothers appear healthy, without evidence of thyroid dysfunction, provided that no iodine deficiency was present in utero. Some studies, although not confirmed by all, have indicated that infants born to hypothyroid mothers might have an increased risk of a higher perinatal mortality and congenital malformations, and of a lower birth weight. The causal relationship is, however, difficult to ascertain, as other medical problems might be associated with HO, such as anemia or nutritional inadequacy. Finally, severe HO during pregnancy raises evident concern about potential long-lasting, psychoneurological consequences for the progeny, related to insufficient transplacental transfer of maternal thyroid hormones to the developing fetus during the first half of gestation, before the fetal thyroid gland becomes functional (Porterfield and Hendrich, 1993).
AITD and the Risk of Miscarriage
Since the early 1990s, several studies have shown an increased prevalence of spontaneous abortion in apparently euthyroid pregnant women who have antithyroid antibodies (Stagnaro-Green et al., 1990) (Fig 1). Altogether, these studies have indicated: (1) that the risk occurs primarily in the first trimester of gestation; (2) that the miscarriage rate is increased two- to fourfold compared with women without detectable thyroid immunity; (3) that the risk of miscarriage is further amplified in women who have a history of consecutive abortions (three or more); and (d) that these women present no clear evidence of hypothyroidism, even though in some instances borderline high serum TSH levels have been demonstrated. The presence of thyroid immunity is an independent marker of an at-risk pregnancy, and is not linked directly to other autoimmune disturbances, such as positive anticardiolipin, antinuclear or antiphospholipid antibodies (Bakimer et al., 1994).
| View Image |
| Figure 1. Between 1988 and 1990 in our institution, a large group of consecutive euthyroid pregnant women was investigated prospectively and systematically for the presence of antithyroid autoantibodies and the occurrence of spontaneous abortion before 14 weeks of gestation. While the overall rate of miscarriage was 6% in the cohort, the prevalence of autoimmune thyroid disorders (AITD) reached 24% among those women with a miscarriage. Furthermore, among women with AITD, the prevalence of a miscarriage reached 22%, which was significantly increased compared with the 5% prevalence of spontaneous abortion in the antithyroid antibody negative group (p<0.005) (adapted from Glinoer et al., 1991). Ab+, positive for antithyroperoxidase and/or antithyroglobulin autoantibodies; Ab-, negative for both autoantibodies. |
Can the association between thyroid immunity and an increased risk of fetal wastage owing to early miscarriage be explained? The easiest explanation is that the presence of thyroid antibodies reveals a more generalized underlying abnormal stimulation of the immune system (Geva et al., 1997). However, other explanations cannot at present be ruled out, and the question clearly requires further investigation. For example, it could be speculated that mild degrees of thyroid insufficiency (difficult or impossible to detect by serum hormone measurements) might play a role locally at the level of the female genital tract. In addition, it has been our experience that women with AITD are two to three years older on the average when becoming pregnant, compared with control subjects from the same cohort, and the age difference is statistically significant (Glinoer et al., 1991). Therefore, it is possible that thyroid immunity acts by hindering fecundity, hence delaying conception. This hypothesis, even though not proven to date, might have clinical relevance, as it has been established clearly that the risk of having a spontaneous miscarriage increases significantly with age (Knudsen et al., 1991).
AITD and the Risk of Hypothyroidism in the Pregnant State
Between June 1990 and December 1992, 1660 consecutive pregnancies, with no previous history of thyroid disorder, were investigated systematically for the presence of antithyroid antibodies, and for the occurrence of alterations in serum TSH and free T4 concentrations (Fig 2). The prevalence of AITD was 6.5%, which is the expected prevalence among female subjects of child-bearing age. Among these patients, 16 women were biochemically hypothyroid (serum TSH >4 mU l-1) and four were hyperthyroidic (with both suppressed TSH and supranormal free T4). These 20 cases will be discussed later. Eighty-seven women (5.2%) showed the presence of antithyroid antibodies at the time of the initial visit, but free T4 and TSH concentrations were in the normal range. Thyroid function was monitored sequentially during gestation. Despite the expected decrease in the titers of antithyroid antibodies during gestation, the parameters of thyroid function showed a gradual deterioration towards HO in a significant fraction of women. Already during the first trimester, the distribution curve of serum TSH was shifted significantly towards higher (albeit still normal) values when compared with normal pregnant controls from the same cohort. At the time of delivery, 40% of women with AITD had a serum TSH >3 mU l-1, with almost half of them >4 mU l-1. Hence, in the early stages of pregnancy, women with AITD were able to maintain normal thyroid function owing to the sustained thyrotropic stimulation. Three days after delivery, however, their serum-free T4 concentrations were markedly lower than those of controls: on average, there was a 30% reduction in serum-free T4, with almost a half of AITD women in the HO range, confirming that these women have a reduced functional thyroid reserve. An important observation was that it was possible, at the individual level, to predict progression to HO on the basis of serum TSH levels and thyroid antibody titers in the first trimester (Glinoer et al., 1994).
| View Image |
| Figure 2. Prospective assessment of the prevalence of thyroid dysfunction and autoimmune thyroid disorders (AITD) in a population of consecutive healthy pregnant women (with no previous history of thyroid disease or treatment), based on the determination of thyroid function parameters and antibody detection at the time of the initial visit. Antithyroid antibodies were found in 6.5% of the cohort, with 5.2% being clinically and biochemically euthyroid, and 1.3% presenting thyroid dysfunction, associated with thyroid immunity, in the early stages of gestation. |
Subclinical and Overt Hypothyroidism During Pregnancy
When HO does not result from the previous treatment of hyperthyroidism with radioiodine or surgery, the most common cause of primary HO in women of child-bearing age is chronic autoimmune thyroiditis, which occurs in both goitrous and atrophic forms. Population-based studies have indicated that as many as 2.5% of all pregnancies might have undiagnosed subclinical HO (Klein et al., 1991). In our studies, a similar prevalence of 2.2% undiagnosed HO was found, with serum TSH ranging between 4 and 20 mU l-1; in addition, free T4 concentrations, even though not systematically subnormal, tended to cluster near the lower limits of normal. These women, diagnosed in the early stages of gestation by systematic screening, were treated with T4 (50–125 µg day-1) throughout gestation, resulting in a normalization of thyroid function. In 16 of the 41 women diagnosed with subclinical HO, the etiology was clearly related to AITD, with antithyroperoxidase antibody (TPO-Ab) titers between 400 and 5000 U ml-1. In the remaining cases, the etiology could not be attributed to AITD, in the absence of detectable antithyroid antibodies or a family history of goiter or HO (Glinoer, 1997).
In relation to thyroid hormone replacement during pregnancy for women with a previously established diagnosis of primary HO, several studies carried out during the past ten years have indicated the need for a systematic adjustment of the T4 replacement dose. The most comprehensive study was reported by Kaplan (1992). In a retrospective analysis of thyroid hormone requirements in a group of 65 women who were hypothyroid because of Hashimoto's thyroiditis or previous thyroid ablation for hyperthyroidism, the author showed that serum TSH rose markedly when thyroxine replacement doses were maintained at prepregnancy levels. Moreover, free T4 concentrations decreased (on the average by 40%) and became subnormal in 13% of the women. In contrast, raising the daily thyroxine dosage by 40–100 µg day-1 resulted in a normalization of serum TSH. After parturition, the T4 requirements returned approximately to the prepregnancy dosage. The study also showed that the increment in T4 dosage required to maintain euthyroidism during pregnancy depended upon the etiology of HO: women with AITD had a residual thyroid secretory capacity allowing them to adapt (although insufficiently) to the changes in hormone requirements associated with pregnancy, while women with a previous thyroid ablation were incapable of adaptation to those changes, hence requiring a greater increment in the thyroxine dosage to remain euthyroid.
The following consensus guidelines have been proposed for the care and management of hypothyroid women during pregnancy (Glinoer, 1997). First, the daily dose of T4 should be increased in at least 80% of hypothyroid women. Pregnant women who do not require such an increase in dosage were probably overtreated before becoming pregnant. Second, the increased need for T4 is already apparent in the first trimester of gestation, concomitant with major changes in thyroidal economy associated with pregnancy. Hence, the adjustment of T4 dosage should be accomplished without delay, in the early stages of gestation. Third, the required hormone increments vary widely (between 10% and 150%), with a median dosage increment of 40–50% over the pregestational replacement T4 dosage. Therefore, each case needs to be assessed individually. Fourth, a regular clinical and laboratory follow-up is essential, with periodic determinations of TSH and free T4 concentrations, indicating the mandatory need for a close collaboration between the endocrinologist and the obstetrician.
Systematic Screening for AITD and Hypothyroidism During Pregnancy
It is justifiable to propose a systematic screening for HO during pregnancy for the following reasons: (1) both AITD and hypothyroidism are common in young female subjects, (2) subclinical HO often remains undiagnosed, and (3) potentially specific obstetrical risks are associated with HO. In addition, there is a rationale for proposing the systematic detection of AITD during pregnancy, based on the following arguments: (1) the increased risk of spontaneous abortion in women with AITD; (2) the risk of progressive HO in euthyroidic women with antithyroid antibodies; (3) the risk of postpartum thyroiditis in the year following pregnancy (50% of AITD women display biochemical evidence of thyroid dysfunction); and (4) the well-known, long-term risk of developing definitive HO later on in life in women with chronic autoimmune thyroiditis.
The following scheme could be proposed (Fig 3). As a first step, serum TSH and thyroid antibodies are measured as early as possible, preferably between 12 and 20 weeks of gestation. Ideally, both antithyroglobulin antibody (TG-Ab) and TPO-Ab titers should be measured. However, if for economical reasons only one thyroid antibody can be assessed, screening with TPO-Ab is to be preferred, because it yields the best diagnostic score, TPO-Ab being positive in 75–80% of women with AITD. If serum TSH is below 2 mU l-1 and thyroid antibodies are negative, nothing more needs to be done. If serum TSH concentrations are above 4 mU l-1, irrespective of the positivity or negativity of antithyroid antibodies, the subject should be considered hypothyroid. We propose determination of free T4 concentrations, ultrasonography of the thyroid gland (in search of glandular hypotrophy) and, possibly, in specific cases, a thyrotropin-releasing hormone (TRH) test. Based on these results, women should be treated with T4 throughout pregnancy and thyroid function parameters should be monitored every two to three months. Obviously, such women should also be followed during the postpartum period.
| View Image |
| Figure 3. A proposed two-step algorithm for the systematic screening of autoimmune thyroid disorders (AITD) and hypothyroidism during pregnancy, based on the determination of antithyroperoxidase antibodies (TPO-Ab), antithyroglobulin antibodies (TG-Ab), and serum thyroid-stimulating hormone (TSH) concentrations in the first half of gestation. |
The second step in the algorithm concerns the women with positive antithyroid antibodies. In this setting, we propose that treatment should be based on serum TSH levels during early gestation. When serum TSH is below 2 mU l-1 (most often associated with low titers of antithyroid antibodies), systematic treatment with T4 is not warranted. We propose that serum TSH should be monitored at six months gestation, and that women should be followed during the postpartum period. In contrast, for women with thyroid immunity and serum TSH concentrations within the normal range, but already between 2–4 mU l-1 during early gestation (most often associated with higher thyroid antibody titers), we propose the measurement of free T4 concentrations and, if they are low or low–normal for the gestational age, treatment with T4 for the remainder of the pregnancy (usually, between 50 and 100 µg day-1) will be sufficient to maintain normal thyroid function. Obviously, such women should also be followed during the postpartum period.
Future prospective studies are necessary to demonstrate the relevance of the proposed scheme, as there is not yet direct evidence of the advantages of treating subclinical hypothyroidism during pregnancy. However, indirect arguments suggest strongly that no harm can be done and treatment with thyroid hormone can only be beneficial.
Hyperthyroidism and Pregnancy
Hyperthyroidism (HR) in the Community and in Pregnancy
The epidemiological characteristics of HR in the adult population present marked differences from those of HO. First, while HO is a relatively common disorder, HR occurs much less frequently, with an overall prevalence rate of 1–2% and estimated incidence rates of 0.1–0.8 per 1000 women per year (Wang and Crapo, 1997). Second, while the risk of developing HO clearly increases with age, this is not the case for HR. The main age-related change for HR is the type of medical condition that causes the metabolic disorder: specifically, HR is most often of autoimmune origin in young patients, while it is more often related to toxic nodular disease in older patients. Third, as alluded to above, there is no clear association between aging, positive antithyroid antibodies and hyperthyroidism.
During pregnancy, HR is usually thought to occur in ~2 of every 1000 women (Becks and Burrow, 1991). The causes of hyperthyroidism include the classic causes found in the general population, but also two others that are specific for the pregnant state (1). Clinical entities such as toxic adenoma, multinodular toxic goiter, subacute or silent thyroiditis and thyrotoxicosis factitia are extremely uncommon. Molar disease should always be considered and can potentially lead to fulminant hyperthyroidism, particularly in women with a pre-existing autonomous nodular goiter. However, uncomplicated hydatidiform mole is now easily diagnosed in the early stages of gestation, and will therefore rarely lead to severe hyperthyroidism, because it lasts only for a few weeks or months and is cured rapidly by the removal of the pathological trophoblast (Hershman, 1972). The major cause of hyperthyroidism in women of child-bearing age is Graves' disease (GD). In recent years, another cause has been characterized, resulting from the direct stimulation of the thyroid gland by human chorionic gonadotropin (hCG), which induces a transient—although also potentially severe—form of hyperthyroidism, and is observed in the first half of gestation. The syndrome, referred to as `gestational transient thyrotoxicosis', has to be differentiated from GD, as the course of both conditions, the fetal risks associated with them, and the management and follow-up of both entities are different (Glinoer et al., 1993, Yoshimura and Hershman, 1995).
-adrenergic blocking agents for a short period. In rare cases, the severity of the clinical presentation may lead to treatment with propylthiouracil (usually for only a few weeks). In all the cases that we have seen, GTT was transient and the normalization of free T4 concentrations paralleled the decrease in hCG concentrations. Also, GTT was not associated with a less favorable outcome of pregnancy.
and 
text only | 
