[S] How should hyperthyroidism in pregnancy be managed?

Hyperthyroidism due to GD is common in women in the reproductive age range and both the thyrotoxicosis and therapy of the disease may complicate the course and outcome of pregnancy. Further, normal pregnancy is accompanied by changes in thyroid physiology, and altered thyroid function testing will reflect this. In early pregnancy, physiological changes can mimic biochemical hyperthyroidism that does not require therapy. In these guidelines, we will address only the most common issues related to hyperthyroidism in pregnancy, pending full guidelines on thyroid disease and pregnancy currently being developed by the ATA.

[S1] Diagnosis of hyperthyroidism in pregnancy


The diagnosis of hyperthyroidism in pregnancy should be made using serum TSH values, and either total T4 and T3 with total T4 and T3 reference range adjusted at 1.5 times the nonpregnant range or free T4 and free T3 estimations with trimester-specific normal reference ranges. 1/+00

The diagnosis of hyperthyroidism in pregnancy can be challenging. In the vast majority of patients, the disease is caused by a primary thyroid abnormality, and the principal finding will be a suppressed serum TSH, with estimated serum free T4 and/or free T3 levels above the reference range (overt hyperthyroidism), or within the reference range (SH).A key point is that reference ranges for thyroid function tests are different during various stages of pregnancy, and for some types of assays, the change may be assay-dependent. GD is the most common cause of hyperthyroidism during pregnancy (278); nodular thyroid disease is less common. Hyperthyroidism caused by a human chorionic gonadotropin (hCG)-producing molar pregnancy or a choriocarcinoma presents with a diffuse hyperactive thyroid similar to GD, but without eye signs and without serum TRAb. In these patients, serum hCG will be higher than expected, and the cause can be identified by obstetrical investigation.

An understanding of pregnancy-related variations in thyroid function tests is important in making the diagnosis of hyperthyroidism in pregnancy. Serum TSH levels may be below the nonpregnant reference range in the first half of a normal-term pregnancy (279,280), presumably the result of stimulation of the normal thyroid by high levels of serum hCG (281). Therefore, low serum TSH levels with normal free T4 values in early pregnancy do not indicate abnormal thyroid function. During the second half of pregnancy, the lower limit for TSH in the nonpregnant population can be used (282).

Free T4 and T3 measured in an equilibrium dialysate or an ultrafiltrate of serum may be slightly higher (5%–10%) than nonpregnancy values around week 10 of pregnancy, corresponding to the period of high serum hCG and low serum TSH. From this normal or slightly high level, a gradual decrease occurs during pregnancy, and late third trimester reference values are 10%–30% below nonpregnancy values (283).

Serum total T4 and T3 increase in early pregnancy. From the late first trimester, they remain stable, with reference ranges close to 1.5 times nonpregnancy ranges during the second and third trimesters (283,284). Total T4 and T3 values may be combined with a T3 uptake test or measurements of TBG to adjust for pregnancy-associated variations in TBG. Such ”free T4 index” or ”TBG adjusted T4” values may be useful for diagnosing hyperthyroidism in pregnancy. However, trimester-specific normal reference ranges should be established for each individual test and assay used.

Technical remarks: The reliability of automated analog-based assays for free T4 and free T3 estimations has been questioned for more than 25 years (285), but these estimates are currently widely used; in many clinics, they are the standard of measurement in pregnancy. Because pregnancy may influence results of these assays from different manufacturers in different ways (286), method-specific reference ranges for each trimester of pregnancy should be employed by the manufacturer (287,288).

[S2] Management of hyperthyroidism in pregnancy


    Transient hCG-mediated thyrotropin suppression in early pregnancy should not be treated with antithyroid drug therapy. 1/+00

Once the diagnosis of hyperthyroidism is made in a pregnant woman, attention should be focused on determining the etiology of the disorder and whether it warrants treatment. Clinical features that may indicate the presence of significant hyperthyroidism include failure to gain weight, heat intolerance, excessive sweating, and tachycardia, beyond that normally associated with pregnancy.

The two most common types of biochemical hyperthyroidism that occur during pregnancy are gestational hyperthyroidism (e.g., hCG-mediated transient TSH suppression) and GD. Gestational hyperthyroidism is a generally asymptomatic, mild biochemical hyperthyroidism that may be observed in the first trimester of normal pregnancy. It is presumably caused by the high serum hCG of early pregnancy (281) and is not associated with adverse pregnancy outcomes (289). Pregnant women having gestational hyperthyroidism with emesis, and particularly hyperemesis, may develop more profound abnormalities in thyroid function, with biochemically overt hyperthyroidism and clinical symptoms and signs of hyperthyroidism. Complicated cases of gestational hyperthyroidism should be referred to medical centers with specific expertise in treating these patients.

Technical remarks: There is no evidence that treatment of gestational hyperthyroidism with ATDs is beneficial. In these patients, physical examination and repeat thyroid function tests at intervals of 3–4 weeks is recommended. If the differential diagnosis of the type of hyperthyroidism is unclear (i.e., if there is suspicion of GD) or in the case of very symptomatic disease, a trial of ATD therapy may be considered if significant clinical hyperthyroidism is evident.


    Antithyroid drug therapy should be used for hyperthyroidism due to GD that requires treatment during pregnancy. Propylthiouracil should be used when antithyroid drug therapy is started during the first trimester. Methimazole should be used when antithyroid drug therapy is started after the first trimester. 1/+00

Untreated or insufficiently treated hyperthyroidism may seriously complicate pregnancy (290–292), and patients with this disorder should be treated at centers with specific expertise in this area. GD as the cause of hyperthyroidism in pregnancy may be diagnosed from typical clinical findings, including the presence of GO and/or serum TRAb in a hyperthyroid patient. Approximately 5% of patients with newly diagnosed Graves’ hyperthyroidism are TRAb negative (43,293), especially those with milder disease.

A woman found to have GD before pregnancy and treated with ATD who goes into remission and is euthyroid off medication has a low risk of recurrent hyperthyroidism during pregnancy. However, her risk of relapse (as well as the risk of postpartum thyroiditis) during the postpartum period is relatively high (294). Antithyroid drugs have much the same effect on thyroid function in pregnant as in nonpregnant women. Both ATDs and TRAb pass the placenta and can affect fetal thyroid. On the other hand, T4 and T3 cross the placenta only in limited amounts.

PTU generally has been preferred in pregnancy because of concerns about rare but well-documented teratogenicity associated with MMI, namely, aplasia cutis and choanal or esophageal atresia (81). However, recent concerns about rare but potentially fatal PTU hepatotoxicity have led to a re-examination of the role of PTU in the management of hyperthyroidism in pregnancy (92). The U.S. Food and Drug Administration recently recommended that PTU be reserved for patients who are in their first trimester of pregnancy, or who are allergic to or intolerant of MMI (92,93).

MMI and PTU both appear in breast milk in small concentrations and studies of breast-fed infants of mothers taking ATDs have demonstrated normal thyroid function and subsequent intellectual development (81). However, because of the potential for hepatic necrosis in either mother or child from maternal PTU use, MMI is the preferred ATD in nursing mothers.


    We suggest that patients taking methimazole who decide to become pregnant obtain pregnancy testing at the earliest suggestion of pregnancy and be switched to propylthiouracil as soon as possible in the first trimester and changed back to methimazole at the beginning of the second trimester. Similarly, we suggest that patients started on propylthiouracil during the first trimester be switched to methimazole at the beginning of the second trimester. 2/+00

Concern is that changing back and forth between MMI and PTU might lead to poorly controlled thyroid function because of differences in pharmacokinetics and uncertainty about dose equivalency between the two drugs. This situation is complicated by the changing levels of TRAb in pregnancy. In general, a potency ratio of MMI to PTU of at least 20–30:1 is recommended when changing from one drug to another, although there are no studies that have examined this potency ratio directly. For example, 300 mg of PTU would be roughly equivalent to 10 to15mg of MMI (81). Alternatively, rather than switching to MMI at the end of the first trimester, the patient could remain on PTU during the second and third trimesters, and have hepatic enzymes measured every 4 weeks, at the same time that thyroid function is assessed. However, there are no prospective data that show that this type of monitoring is effective in preventing fulminant PTU related hepatotoxicity.


    GD during pregnancy should be treated with the lowest possible dose of antithyroid drugs needed to keep the mother’s thyroid hormone levels slightly above the normal range for total T4 and T3 values in pregnancy and the TSH suppressed. Free T4 estimates should be kept at or slightly above the upper limit of the nonpregnant reference range. Thyroid function should be assessed monthly, and the antithyroid drug dose adjusted as required. 1/+00

Even if the mother is euthyroid during ATD therapy, there is a risk of inducing fetal hypothyroidism during the second and third trimesters when the fetal thyroid has begun to function (295,296). Thus, the dose of ATD should be kept as low as possible. Block-replacement therapy consisting of ATD plus levothyroxine should not be used in pregnancy. If a woman receiving such therapy becomes pregnant, therapy should be changed to an ATD alone (278).

Technical remarks: Free T4 is the parameter that has been most closely correlated with good fetal outcome. Serum TSH may still be suppressed in these patients and should not be used as the sole guide in treatment, although normalization of maternal TSH during ATD therapy may indicate a need to reduce the dose of ATD (278).


    When thyroidectomy is necessary for the treatment of hyperthyroidism during pregnancy, the surgery should be performed if possible during the second trimester. 1/+00

Pregnancy is a relative contraindication to thyroidectomy and should only be used in this circumstance when aggressive medical management has not obviated the need for immediate treatment of the hyperthyroidism and antithyroid medications cannot be used. Thyroidectomy is best avoided in the first and third trimesters of pregnancy because of teratogenic effects associated with anesthetic agents and increased risk of fetal loss in the first trimester and increased risk of preterm labor in the third. Optimally, thyroidectomy would be performed in the latter portion of the second trimester. Although it is the safest time, it is not without risk (4.5%–5.5% risk of preterm labor) (47,48).

Evaluation by a high-risk obstetrician is advised along with counseling before surgery regarding risks involved (48). Thyroidectomy cures the hyperthyroid condition and is often followed by a gradual reduction in TRAb from the circulation (297). Until such remission takes place, TRAb produced by the mother may stimulate the thyroid of the fetus or newborn and induce hyperthyroidism. In the setting where the mother still harbours high levels of TRAb after thyroidectomy, close fetal monitoring for both cardiovascular and skeletal changes (fetal ultrasound) must be established.

There are no data concerning whether SSKI or iodine should be used to prepare pregnant patients for thyroidectomy. The risk of iodide therapy to the fetus is inhibition of iodine organification, the Wolff-Chaikoff effect. The fetal thyroid gland is particularly susceptible to the inhibitory effects of excess iodine at the end of gestation, and fetal goiter can occur with chronic therapy (298). However, there is no evidence that brief iodine preparation of the mother done preoperatively to reduce thyroid blood flow and control hyperthyroidism is harmful to the fetus.

Technical remarks: Preoperative preparation for thyroidectomy during the second trimester of pregnancy includes 10– 14 days of iodine, along with ATD therapy and beta-blockers to control hyperthyroidism (299–301).

[S3] The role of TRAb levels measurement in pregnancy


    TRAb levels should be measured when the etiology of hyperthyroidism in pregnancy is uncertain. 1/+00

The two best indicators of the activity of GD during pregnancy are thyroid function in the untreated patient and measurement of TRAb levels in the serum. TRAb measurement is useful in the diagnosis of GD in pregnant women with newly diagnosed hyperthyroidism who do not have clinical signs specific for GD, keeping in mind that the diagnostic sensitivity of good assays is around 95%, and the specificity is 99% (43).


    Patients who were treated with radioactive iodine or thyroidectomy for GD prior to pregnancy should have TRAb levels measured using a sensitive assay either initially at 22–26 weeks of gestation, or initially during the first trimester and, if elevated, again at 22–26 weeks of gestation. 1/+00

Measurement of TRAb levels can detect persistent TSH-receptor autoimmunity in a pregnant woman previously treated with ablative therapy (radioactive iodine or thyroidectomy) for GD who is now euthyroid with or without thyroid hormone replacement (297,302). If the mother still produces TRAb, they will cross the placenta and may affect fetal thyroid function in the last half of the pregnancy. Because of the slow clearance of maternal immunoglobulin G (IgG) from the neonatal circulation, thyroid dysfunction in the child may last for several months after birth. To evaluate the risk of such complications, a TRAb level should be measured in the pregnant woman either initially at 22–26 weeks of gestation, or initially during the first trimester and, if elevated, again at 22–24 weeks of gestation. If the level is high, a program of fetal and neonatal surveillance for thyroid dysfunction should be initiated (303). While measuring TRAb levels only at 22–26 weeks is more cost effective, the advantage to initial measurement during the first trimester is that this allows more time to initiate specialty consultation and, if the levels are found to be especially high at that time, intervention may be required before the third trimester. TRAb measurement is not necessary in a euthyroid pregnant patient previously found to have GD if she has an intact thyroid (i.e., not previously treated with surgery or radioactive iodine) and is not currently taking ATDs (295,297).


    Patients found to have GD during pregnancy should have TRAb levels measured at diagnosis using a sensitive assay and, if elevated, again at 22–26 weeks of gestation. 1/+00


  • RECOMMENDATION 77 TRAb levels measured at 22–26 weeks of gestation should be used to guide decisions regarding neonatal monitoring. 1/+00

TRAb (TBII or TSI) measurement is also useful to assist in the evaluation of disease activity in a woman being treated with ATDs for GD during pregnancy (297). In many patients, GD gradually remits during pregnancy. Disappearance of TRAb is an indication that ATD therapy may no longer be necessary, and that its continuation may put the fetus at risk for hypothyroidism. TRAb measurement also can be used during the third trimester to assess the risk of delayed neonatal hyperthyroidism when the mother continues to need MMI to control hyperthyroidism up to term. After delivery, MMI delivered to the fetus via placental passage is rapidly metabolized by the neonate, whereas the maternal TRAb disappears more slowly, with a half-life of around 3weeks. Thus, a high level of TRAb in the mother in late pregnancy is an indicator that the neonate may need to be monitored for the onset of neonatal hyperthyroidism starting a few days after birth.

Technical remarks: A sensitive TBII assay or TSI assay should be used to detect TRAb during pregnancy. A summary of TRAb measurement and management of hyperthyroidism caused by GD during pregnancy is presented in Table 9.


Timing of diagnosis

Specific circumstances


GD diagnosed during pregnancy

Diagnosed during first trimester

Begin propylthiouracila

Measure TRAb at diagnosis and, if elevated, repeat at 22–26 weeks of gestationb

If thyroidectomy is required, it is optimally performed during the second trimester

Diagnosed after first trimester

Begin methimazolec

Measure TRAb at diagnosis and, if elevated, repeat at 22–26 weeks of gestationb

If thyroidectomy is required, it is optimally performed during the second trimester

GD diagnosed and treated prior to pregnancy

Currently taking methimazole

Switch to propylthiouracil as soon as pregnancy is confirmed with early testinga

Measure TRAb either initially at 22–26 weeks of gestation, or initially during the first trimester and, if elevated, again at 22–26 weeks of gestationb

In remission after stopping antithyroid medication

TRAb measurement not necessary

Previous treatment with radioiodine or surgery

Measure TRAb either initially at 22–26 weeks of gestation, or initially during the first trimester and, if elevated, again at 22–26 weeks of gestationb

aSee remarks under Recommendation 71 for discussion regarding switching from one antithyroid drug to the other during pregnancy.

bIf a TRAb-positive woman becomes TRAb-negative during pregnancy, this may indicate a need to reduce or stop antithyroid drug therapy to avoid fetal hypothyroidism. If the antithyroid drug treated mother has high TRAb values in late pregnancy this indicates a risk of delayed neonatal hyperthyroidism (see remarks to Recommendation 77). If the mother has undergone some type of thyroid ablation (radioactive iodine or surgery) for GD and TRAb is high, evaluate fetus carefully for hyperthyroidism in second half of pregnancy and adjust or begin antithyroid drug therapy accordingly.

cAvoid fetal hypothyroidism, especially in second half of pregnancy (see recommendation 75 for details).

[S4] Postpartum thyroiditis


    In women with thyrotoxicosis after delivery, selective diagnostic studies should be performed to distinguish postpartum thyroiditis from postpartum GD. 1/+00

Postpartum thyroid dysfunction occurs in up to 10% of pregnancies in the United States. Postpartum thyroiditis is an autoimmune disorder unmasked in predisposed women as immune surveillance rebounds after pregnancy. The classic triphasic pattern is thyrotoxicosis at 1–6 months postpartum, followed by hypothyroidism and return to euthyroidism at 9– 12 months postpartum (304,305). However, this sequence is not observed in every patient. Among 371 cases in 13 studies, 25% of patients were found to have a triphasic pattern, 43% had hypothyroidism without preceding thyrotoxicosis, and 32% had thyrotoxicosis without subsequent hypothyroidism (305). In a prospective study of pregnant women, those with positive thyroperoxidase (TPO) antibodies in the first trimester were 27 times more likely to develop postpartum thyroiditis than were those with negative serology (306). In this study, tobacco smoking and bottle feeding (maybe because of higher exposure of the maternal thyroid to iodine, which is not excreted into breast milk) also increased the risk of developing thyroiditis.

Postpartum thyroiditis must be distinguished from GD to recommend proper therapy. Goiter is generally more pronounced in GD, and thyroid bruit or GO strongly suggest GD as well. TRAb may be measurable in patients with postpartum thyroiditis, but higher titers are suggestive of GD. When in vivo testing is required to make this distinction, 123I or technetium should be used rather than 131I in women who are nursing, since the shorter half-life of these agents will allow breast milk to be pumped and discarded for several days and nursing resumed, whereas breast-feeding should not be resumed if 131I is given as treatment for GD (307). Total T3 to T4 ratios (ng/dL:mcg/dL) tend to be higher (>20) in patients with GD than in those with postpartum thyroiditis.


    In women with symptomatic postpartum thyrotoxicosis, the judicious use of beta-adrenergic blocking agents is recommended. 1/+00

Treatment for postpartum thyroiditis is generally supportive in nature, with the use of beta-adrenergic blockers such as propranolol (lowest level in breast milk) (308) ormetroprolol to control pulse rate and hyperadrenergic symptoms during the thyrotoxic stage. Levothyroxine therapy may be beneficial, at least transiently, for women with symptomatic hypothyroidism or having TSH levels >10mU/L (305).

Technical remarks: Because beta blockers are secreted into breast milk in very low levels, no special monitoring is needed for breastfed infants of mothers on these medications (308).