Hypothyroidism due to Hashimoto’s thyroiditis was diagnosed prior to the second pregnancy with TSH, FT4, and TPO antibodies. Effectiveness of therapy was evaluated by TSH values regularly during pregnancy.

Physiologic changes in thyroid function occur during normal pregnancy. hCG stimulates thyroid activity; as hCG increases during the first trimester, thyroid-stimulating hormone (TSH) decreases [1–3] and then gradually increases and plateaus. Free T4 (FT4) initially increases with the increase in hCG and then decreases after the first trimester. TBG and total T4 both increase during pregnancy. Gestation-specific reference intervals for TSH exist during pregnancy [4–6], which should alleviate the misinterpretation of thyroid function tests [6]. Dashe et al. constructed a nomogram in 13,599 singleton pregnancies and a separate nomogram in 132 twin pregnancies to estimate expected reference ranges throughout gestation [7]. However, in iodine-deficient populations, no gestation-specific levels have been established [8]. According to the American Thyroid Association guidelines, trimester-specific reference ranges for TSH as defined in populations with optimal iodine intake should be applied [9]. If not available, TSH in the first trimester should fall between 0.1–2.5 mIU/L, 0.2–3.0 mIU/L in the second trimester, and 0.3–3.0 mIU/L in the third trimester [9]. FT4 is optimally measured by T4 in dialysate or ultrafiltrate with online extraction, liquid chromatography, or tandem mass spectrometry; but if not available, TSH is a more reliable estimate of thyroid function [9].

New-onset hypothyroidism during pregnancy is a rare occurrence. Allan et al. screened 9403 healthy women between 15 and 18 weeks gestation, of which 172 women (1.8 %) had a TSH from 6 to 9.9 mIU/L and 37 women (0.4 %) had a TSH greater than 10 mIU/L [10]. Casey et al. screened 17,298 women before 20 weeks gestation; similarly, 404 women (2.3 %) had elevated TSH and 32 women (0.2 %) had both an elevated TSH and a suppressed FT4 [4].

Both overt and subclinical hypothyroidism are associated with adverse pregnancy outcomes. Casey et al. found a significant (p < 0.05) increase in both placental abruption and delivery at less than 34 weeks gestation in women with subclinical hypothyroidism compared to age-matched euthyroid women [4]. In the Generation R Study, thyroid function tests and birth outcomes were studied in 5971 women with singleton pregnancies and no current or past thyroid disease except for a low (3.4 %) incidence of subclinical hypothyroidism. Five percent of women had a preterm delivery at less than 37 weeks gestation, 4.4 % had a spontaneous preterm delivery at less than 37 weeks gestation, and 1.4 % delivered at less than 34 weeks gestation (defined as very premature delivery) [11]. Using the ATA guideline of normal gestation-specific TSH values, the authors found that women with elevated TSH had an increased incidence of overall preterm delivery but not of spontaneous preterm delivery. Women with hypothyroxinemia had a 2.5-fold increased risk of preterm delivery, a 3.4-fold increased risk of spontaneous preterm delivery, and a 3.6-fold increased risk of very preterm delivery. Women with TPO antibody positivity had a 1.7-fold increased risk of preterm delivery, a 2.1-fold increased risk of spontaneous preterm delivery, and a 3.6-fold increased risk of very preterm delivery [11]. LaFranchi et al. found greater rates of maternal complications including gestational hypertension and c-section in women with overt and subclinical hypothyroidism [12]. Higher c-section rates may be in part due to the associated increased incidence of breech presentation at term in mothers with elevated TSH [13].

In addition to preterm delivery, there are serious outcomes on fetal morbidity and mortality in hypothyroid women. In a study of 9403 women with singleton pregnancies, the fetal death rate after 16–18 weeks gestational age was fourfold higher in the 2.2 % of women with a TSH greater than 6 mIU/L (3.8 % vs 0.9 %), which rose to 8.1 % among women with a TSH greater than 10 mIU/L [10]. In women with a TSH greater than the 97.5th percentile compared to euthyroid women, birth weight less than 1500 g, NICU admissions, and respiratory distress were all significantly increased [4]. Another prospective population-based study of 1017 women with singleton pregnancies found that those with overt hypothyroidism had an increased risk of fetal death, fetal loss, circulatory system malformations, and low birth weight [14]. Those with subclinical hypothyroidism also had increased risk of fetal distress, preterm delivery, poor vision, and neurodevelopmental delay [14].

Much research has focused on maternal hypothyroidism and fetal neurodevelopmental effects. During the first 14 weeks of gestation when brainstem and cerebral neurogenesis occurs, the fetus’s source of free T4 is entirely maternal. From 14 weeks onward, during neuronal maturation and synaptic development, the fetus begins producing and supplying its own thyroid hormones [15]. Man et al. studied the offspring of 365 otherwise uncomplicated pregnancies. Hypothyroxinemic mothers were defined as women with two or more values below the normally elevated levels or who received no thyroid hormone replacement or inadequate therapy. IQ levels in the 4-year-old and 7-year-old offspring of these mothers were significantly less (p < 0.05) compared to euthyroid mothers [16]. In another study, children of women with untreated hypothyroidism during pregnancy averaged 7 points lower on IQ testing and had a significant percentage (19 % vs 5 % of controls) of IQ less than or equal to 85 [17]. The Generation R Study looked at expressive language at 18 months and nonverbal cognitive function at 30 months in 3659 children, of which all mothers had normal TSH. The authors found a significant increased odds ratio of expressive language delay in mildly and severely depressed FT4 levels compared to controls. The odds ratio was also significantly increased for nonverbal cognitive delay but only with severely depressed FT4 levels [18]. Li et al. compared women with subclinical hypothyroidism and low T4 and TPO antibodies with control groups matched for gestational age, gender, birth condition (Apgars), birth weight, and other factors. IQ and motor function score of offspring at 25–30 months old were significantly lower (p < 0.01) in all three groups compared to their matched controls [19]. A study of over 5700 children found that only girls of mothers with elevated TSH had a significant increased odds ratio of inattention and ADHD symptoms at age 8 as measured by a teacher-administered test [20].

Infertile women who are euthyroid but thyroid peroxidase antibody (TPO-Ab) positive have also been shown to have poor pregnancy outcomes. A prospective study of 484 euthyroid women undergoing assisted reproductive technology, of which 72 (15 %) were TPO-Ab positive, found that TPO-Ab-positive women treated with levothyroxine had no difference in pregnancy rate compared to placebo-treated TPO-Ab-positive women and the TPO-Ab-negative control group but did have an increase in miscarriage rate. Delivery rate was lower in both groups of TPO-Ab-positive women compared to the negative control group, but there was no difference in delivery rates between the two TPO-Ab-positive groups [21].

Euthyroid women with autoimmune thyroid disease are at risk for obstetrical complications. Negro et al. randomized 115 TPO-Ab-positive women with a normal TSH to treatment with levothyroxine (57 women) versus placebo (58 women). The control group consisted of 869 TPO-Ab-negative women. Primary outcome measure was the rate of obstetrical complications. Levothyroxine was dosed by initial TSH value obtained in the first trimester. Both groups of TPO-Ab-positive patients had higher initial TSH values compared to the control group. Untreated TPO-Ab-positive patients had significantly elevated TSH values throughout gestation and significantly lower free T4 values after 30 weeks gestation and a significant increase in miscarriage and preterm delivery compared to the other two groups (p < 0.05 for all). These results suggest that treatment with levothyroxine in euthyroid women with TPO antibody positivity can reduce the risk of miscarriage and preterm delivery to control levels. The authors recommended levothyroxine treatment for all pregnant women with TSH greater than 2.0 mIU/L or high TPO antibody titers [22].

Given the above risks, should all pregnant patients be screened for hypothyroidism? Vaidya et al. screened 1560 pregnant women during their first obstetric visit using TSH, FT4, and FT3. TPO antibodies were checked in 85 % of patients. 26.5 % of patients were considered high risk, defined as a personal history of thyroid disease or autoimmune disease or a family history of thyroid disease. Of the patients, forty (2.6 %) had an elevated TSH, and the prevalence was significantly higher in the high-risk group (6.8 %) compared to the low-risk group (1 %). However, twelve of the forty low-risk patients (30 %) had an elevated TSH. The authors concluded that targeted screening of only the high-risk group would miss approximately one third of pregnant women with overt or subclinical hypothyroidism [23]. However, universal screening versus case finding has not been shown to decrease adverse outcomes. Negro et al. randomly assigned 4562 pregnant women to a universal screening group or a case-finding group. Women in both groups were labeled as high risk or low risk depending on their history. High-risk women in both groups and low-risk women in the universal screening group were tested for TSH, FT4, and TPO antibody; low-risk women in the case-finding group were tested during pregnancy, and results were reviewed postpartum. Patients were treated with levothyroxine if TSH was greater than 2.5 mIU/L. Euthyroid TPO-antibody-positive patients were retested in the second and third trimesters. Similarly, hyperthyroid women were screened and treated as indicated. In the case-finding group, 20 (4.4 %) of patients versus 34 (1.9 %) of low-risk patients had hypothyroidism. Similarly, in the universal screening group, 4.0 % of high-risk patients and 2.5 % of low-risk patients screened positive. Adverse outcomes examined included miscarriage, cesarean delivery, preterm delivery, and NICU admission. No differences were seen in adverse outcomes between the universal screening and the case-finding groups, but screening and treating the low-risk patients resulted in a significantly decreased rate of adverse outcomes compared to the untreated low-risk patients not identified during pregnancy [24].

The Controlled Antenatal Thyroid Screening study (CATS) was a prospective randomized controlled trial whose aim was twofold: to identify subclinical hypothyroidism (SCH) in early gestation and to evaluate levothyroxine (LT4) intervention therapy in SCH. Women were initially randomized into screening (sample measured immediately) and control groups (sample stored and measured after delivery). Women identified in the screening group with SCH were started on LT4 150mc daily in the late first to early second trimester, at a median gestational age of 13 weeks and 3 days. The primary outcome was the IQ of children at age 3 years, which was not significantly different between the control and the screening groups. Two other tests, the child behavior checklist and the behavior rating preschool, also showed no difference between the two groups. This trial showed no benefit for antenatal screening and treatment for SCH, as it appears to have been implemented too late to be impactful in early childhood cognitive function [25].

However, challenges in screening may occur. In euthyroid women with TPO antibody positivity who are at risk for miscarriage and preterm delivery, LT4 therapy is associated with a decreased risk of these complications [22]. Another prospective study of 1560 pregnant women concluded that screening only pregnancies with risk factor for thyroid disorders would miss approximately one third of pregnant women with either overt or subclinical hypothyroidism [23]. No national organization recommends screening of low-risk pregnant women.

In patients with known hypothyroidism, thyroid function can be optimized in pregnancy by paying close attention to those with known thyroid disease. Mandel et al. reviewed 12 women with primary hypothyroidism on LT4 before, during, and after pregnancy. All patients experienced an increase in their TSH, and 9 of 12 patients (75 %) required an increase in their LT4 dose during pregnancy [26]. A prospective study of 63 pregnant women taking LT4, 83 % of whom evaluated were in the first trimester, showed that 49 % of initial TSH values were outside the normal laboratory reference range. The rate of fetal loss was significantly higher (29 % versus 6 %) in women with abnormal initial TSH compared to initial TSH within the reference range [27].

Thyroid dysfunction in early pregnancy has been associated with miscarriage. Of 1013 women in whom LT4 had been started at least 6 months prior to conception and whose TSH was measured in the first trimester, 62.8 % had a TSH greater than 2.5 mIU/L, with 29.1 % greater than 4.5mIU/L and 7.4 % greater than 10mIU/L. Miscarriage risk was increased in women with TSH greater than 4.5mIU/L (odds ratio 1.8) which further increased in women with TSH greater than 10mIU/L (odds ratio 3.95) [28].

In patients with planned pregnancies, preconception TSH levels should be optimized. Of 53 pregnant women with hypothyroidism in whom preconception TSH was less than 2.5 and in the normal range 6 months prior to pregnancy, 17 had to increase their levothyroxine dose due to an elevated TSH at the first prenatal visit. 50 % of women with a preconception TSH of 1.2–2.4 mIU/L required an increase in their levothyroxine versus only 17.2 % of women with a preconception TSH of less than 1.2 (p < 0.02). The authors concluded that preconception TSH levels should not only be in the normal range prior to conception but should not be greater than about 1.2 mIU/L; therefore, fewer dose adjustments during pregnancy would be needed [29]. In another study of 25 hypothyroid women planning pregnancy, 14 were assigned to partially suppressive treatment and 11 continued their current levothyroxine dose. Women who received partially suppressive therapy showed higher (normal) FT4 and lower TSH at the first postconception evaluation. Women who continued their current dose were twice as likely to have an increased TSH later in pregnancy. The authors noted that a partially suppressive dose may be worthwhile in light of the effects of maternal hypothyroidism on future offspring [30]. Alexander et al. studied 19 hypothyroid women on preconception levothyroxine, of whom 17 resulted in term births. A mean levothyroxine dose increase of 47 % was required, with the median onset of the required increase at 8 weeks gestation. However, the requirements increased as early as 5 weeks gestation [31].

To minimize the risks of untreated hypothyroidism, the American Thyroid Association recommends that treated hypothyroid patients who are newly pregnant should self-increase their levothyroxine dose by 25–30 % if they have a missed menstrual cycle or have a positive home pregnancy test and promptly notify their caregiver. One way this can be accomplished is to increase their weekly dose of 7 pills to 9 pills [9].

This patient presented with signs and symptoms of thyrotoxicosis. Laboratory evaluation was consistent with hyperthyroidism. The etiology of her hyperthyroidism was suspected to be Graves’ disease due to the physical findings and confirmed with the positive thyroid receptor antibodies. A normal intrauterine pregnancy was confirmed by a pelvic ultrasound.

Maternal complications of uncontrolled or untreated hyperthyroidism include hypertensive disorders of pregnancy, preterm labor, congestive heart failure, placental abruption, and thyroid storm. Fetal and neonatal complications include intrauterine growth restriction, small for gestational age, prematurity, stillbirth, fetal and neonatal hyperthyroidism, and congenital malformations.

Laboratory diagnosis of hyperthyroidism in pregnancy can be challenging due to the physiologic changes in thyroid function during pregnancy. Total T4, free T4 (FT4), total triiodothyronine (T3), and thyroid-binding globulin (TBG) are elevated in normal pregnancy. Additionally, FT4 analogue assays may underestimate the actual FT4 value. TSH suppression and/or FT4 elevation appear transiently in approximately 10 % of pregnancies in the first trimester [3]. TRAb may be useful in establishing a diagnosis. Radioactive iodine is contraindicated in pregnancy.

It is important to differentiate the etiology of thyrotoxicosis as gestational versus clinical. The majority of patients with thyrotoxicosis have gestational thyrotoxicosis. Gestational thyrotoxicosis, which is hCG mediated, appears in the late first and early second trimesters and resolves as hCG decreases. Patients may be asymptomatic with a suppressed TSH only or may have hyperemesis gravidarum with suppressed TSH and mildly elevated FT4. No goiter is usually present on physical exam in areas of iodine sufficiency and patients are TRAb negative. The course is self-limited and no antithyroid medication is indicated. A serum TSH in the first trimester less than 0.1 mIU/L should prompt a history and physical exam, and FT4 measurements should be obtained. T3 and TRAb may be helpful to establish a diagnosis of hyperthyroidism [9]. In women with gestational hyperthyroidism and hyperemesis gravidarum, management includes supportive therapy, treating dehydration and hospitalization if indicated. Antithyroid drugs are not recommended [9].

Approximately 0.1–4 % of pregnant women with thyrotoxicosis have a clinical, rather than a gestational etiology for their symptoms. Of these, Graves’ disease is the most frequent. Other etiologies include thyroiditis, toxic nodule(s), and hydatidiform mole.

Thyroiditis is a rare diagnosis in pregnancy, the etiology of which is postinfectious or autoimmune mediated. Because so few cases have been reported, the percentage of patients who have complications is unknown. Clinical findings include a goiter, which may be painful (De Quervain’s or postinfectious etiology) or painless (autoimmune etiology). Patients are TRAb negative and may be TPO negative or positive in those with an autoimmune etiology. Similar to gestational thyrotoxicosis, the course is self-limited and no antithyroid treatment is indicated [32].

Hydatidiform mole is another hCG-mediated form of thyrotoxicosis in pregnancy. Onset appears following mole formation and resolves if hCG decreases after treatment of the molar pregnancy. Like gestational thyrotoxicosis, patients may be asymptomatic with a suppressed TSH alone or may have hyperemesis gravidarum with a suppressed TSH and elevated FT4. Patients do not have a goiter and are TRAb negative [32].

Graves’ and toxic nodular goiter, whose etiology is autoimmune or of nodular autonomy, have a variable onset whose course is persistent until addressed. Like other etiologies of hyperthyroidism, patients may be asymptomatic with a suppressed TSH alone or may have hyperemesis gravidarum and symptoms of thyrotoxicosis with a suppressed TSH and elevated FT4. Clinical findings include a goiter, TRAb positivity, or positive thyroid ultrasound findings. Antithyroid treatment is indicated due to the complications described at the beginning of this segment [32]. Evidence is lacking to recommend for or against a thyroid ultrasound to differentiate the cause of hyperthyroidism in pregnancy. Radioactive iodine scanning or uptake should not be performed in pregnancy [9]. Additionally, thyrotoxic women who are not pregnant but desire to conceive should be euthyroid prior to attempting pregnancy [9].

The use of the antithyroid drugs PTU and MMI during pregnancy and their potential effects on the developing fetus have been evaluated in multiple studies. MMI may have increased transplacental passage compared to PTU [33], and its use in the first trimester has been associated with aplasia cutis and choanal atresia [34]. Cord serum PTU levels at term are higher than maternal levels, suggesting the fetus has a slower clearance of PTU [35]. In a case-affected control analysis of 18,131 cases of congenital malformations and first-trimester medication use, 127 infants were born to mothers who took antithyroid medication; 47 took PTU and 80 took MMI. 52 groups of malformations were analyzed. Significant associations were found between PTU and situs inversus, unilateral renal agenesis/dysgenesis, and cardiac outflow tract anomalies and between MMI and situs inversus, choanal atresia, and omphalocele [36]. Another study examined 6,744 women with Graves’ disease, of whom 1426 took MMI, 1578 took PTU, and 2065 took no antithyroid medication during the first trimester. Overall, there was a 4.1 % rate of major anomalies with MMI compared with 2.1 % in the control group and 1.9 % in the PTU group. Anomalies included aplasia cutis, omphalocele, and a symptomatic omphalomesenteric duct anomaly [37]. A Danish nationwide cohort study of 817,093 children born from 1996 to 2008 found that MMI/carbimazole (CMZ) and PTU were associated with an increased risk of birth defects diagnosed before age 2. The prevalence was an excess of cases per live birth (5.7 % for nonexposed and 5.4 % for history of antithyroid drug use prior to pregnancy compared with 8 % for PTU alone, 9.1 % for MMI/CMZ alone, and 10.1 % for MMI/CMZ with a switch to PTU in early pregnancy). The odds of developing a birth defect were 1.41 for PTU, 1.66 for MMI/CMZ, and 1.82 in MMI/CMZ with a switch to PTU in early pregnancy. Specific birth defect associations were seen in the different groups. Children exposed to MMI/CMZ had a combined odds ratio of 21.8 and exhibited aplasia cutis, eye and circulatory anomalies, choanal and esophageal atresia, omphalocele, and omphalocele-mesenteric duct anomalies. Malformations involving the face, neck, and preauricular sinus or cyst with fistula were seen in PTU-exposed children. Those receiving both medications had an increase in urinary tract anomalies [38]. A subanalysis of the same cohort examined the severity of birth defects associated with PTU exposure in early pregnancy. Of 14 cases identified, 11 children were exposed to PTU only and 3 to MMI/carbimazole with a switch to PTU in early pregnancy. The prevalence of face and neck defects and urinary system anomalies were both increased, with hazard ratios of 4.92 and 2.73, respectively [39]. Although any antithyroid drug use in pregnancy may pose a risk, both hyper- and hypothyroidism threaten pregnancy outcomes more than antithyroid drugs. Ideally, patients with preexisting Graves’ disease should be permanently treated with surgery or 131-I prior to pregnancy [40].