CHAPTER 58: Endocrine Disorders

In 1903, little was known of many endocrine disorders. Still, endocrinopathies seem particularly closely related to pregnancy because of its gestational proclivity for prodigious hormone secretion. This is best illustrated by placental lactogen in diabetes, the most common endocrinopathy encountered in pregnancy (Chap. 57, Types of Diabetes). Pregnancy is also interrelated with some endocrinopathies that are at least partially due to autoimmune dysregulation. Clinical manifestations of this result from complex interplay among genetic, environmental, and endogenous factors that activate the immune system against targeted cells within endocrine organs. An extraordinary example of these interactions comes from studies that implicate maternal organ engraftment by fetal cells that were transferred during pregnancy. These cells later provoke antibody production, tissue destruction, and autoimmune endocrinopathies.

Taken in aggregate, disorders of the thyroid gland are common in young women and thus frequently encountered in pregnancy. Maternal and fetal thyroid function are intimately related, and drugs that affect the maternal thyroid also affect the fetal gland. Moreover, thyroid autoantibodies have been associated with increased rates of early pregnancy wastage. Also, uncontrolled thyrotoxicosis and untreated hypothyroidism are both associated with adverse pregnancy outcomes. Finally, evidence suggests that the severity of some autoimmune thyroid disorders may be ameliorated during pregnancy, only to be exacerbated postpartum.

Thyroid Physiology and Pregnancy

Maternal thyroid changes are substantial, and normally altered gland structure and function are sometimes confused with thyroid abnormalities. These alterations are discussed in detail in Chapter 4 (Thyroid Gland), and normal serum hormone level values are found in the Appendix (Serum and Blood Constituents). First, maternal serum concentrations of thyroid binding globulin are increased concomitantly with total or bound thyroid hormone levels (Fig. 4-16). Second, thyrotropin, also called thyroid-stimulating hormone (TSH), currently plays a central role in screening and diagnosis of many thyroid disorders. Notably, TSH receptors are cross stimulated, albeit weakly, by massive quantities of human chorionic gonadotropin (hCG) secreted by placental trophoblast. Because TSH does not cross the placenta, it has no direct fetal effects. During the first 12 weeks of gestation, when maternal hCG serum levels are maximal, thyroid hormone secretion is stimulated. The resulting greater serum free thyroxine (T₄) levels act to suppress hypothalamic thyrotropin-releasing hormone (TRH) and in turn limit pituitary TSH secretion (Fig. 58-1). Accordingly, TRH is undetectable in maternal serum. Conversely, in fetal serum, beginning at midpregnancy, TRH becomes detectable, but levels are static and do not increase.

Throughout pregnancy, maternal thyroxine is transferred to the fetus (American College of Obstetricians and Gynecologists, 2017). Maternal thyroxine is important for normal fetal brain development, especially before the onset of fetal thyroid gland function (Bernal, 2007; Korevaar, 2016). And, even though the fetal gland begins concentrating iodine and synthesizing thyroid hormone after 12 weeks’ gestation, maternal thyroxine contribution remains important. In fact, maternal sources account for 30 percent of thyroxine in fetal serum at term (Thorpe-Beeston, 1991). Still, developmental risks associated with maternal hypothyroidism after midpregnancy remain poorly understood (Morreale de Escobar, 2004; Sarkhail, 2016).

Autoimmunity and Thyroid Disease

Most thyroid disorders are inextricably linked to autoantibodies against nearly 200 thyrocyte components. These antibodies variably stimulate thyroid function, block function, or cause thyroid inflammation that may lead to follicular cell destruction. Often, these effects overlap or even coexist.

Thyroid-stimulating autoantibodies, also called thyroid-stimulating immunoglobulins (TSIs), bind to the TSH receptor and activate it, causing thyroid hyperfunction and growth. Although these antibodies are identified in most patients with classic Graves disease, simultaneous production of thyroid-stimulating blocking antibodies may blunt this effect (Jameson, 2015). Thyroid peroxidase (TPO) is a thyroid gland enzyme that normally functions in the production of thyroid hormones. Thyroid peroxidase antibodies, previously called thyroid microsomal autoantibodies, are directed against TPO and, as shown in Figure 58-2, have been identified in 5 to 15 percent of all pregnant women (Abbassi-Ghanavati, 2010; Sarkhail, 2016). These antibodies have been associated in some studies with early pregnancy loss and preterm birth (Negro, 2006; Korevaar, 2013; Plowden, 2017; Thangaratinam, 2011). In another study with more than 1000 TPO antibody-positive pregnant women, the risk for preterm birth was not elevated, however, the risk for placental abruption was greater (Abbassi-Ghanavati, 2010). These women are also at high risk for postpartum thyroid dysfunction and at lifelong risk for permanent thyroid failure (Andersen, 2016; Jameson, 2015).

Fetal Microchimerism

Autoimmune thyroid disease is much more common in women than in men. One intriguing explanation for this disparity is fetal-to-maternal cell trafficking (Greer, 2011). Fetal cells are known to enter maternal circulation during pregnancy. When fetal lymphocytes enter maternal circulation, they can live for more than 20 years. Stem cell interchange can lead to engraftment in several maternal tissues and is termed fetal microchimerism. In some cases, this may involve the thyroid gland (Bianchi, 2003; Boddy, 2015; Khosrotehrani, 2004). A high prevalence of Y-chromosome-positive cells has been identified using fluorescence in situ hybridization (FISH) in the thyroid glands of women with Hashimoto thyroiditis—60 percent, or with Graves disease—40 percent (Renné, 2004). In another study of women giving birth to a male fetus, Lepez and colleagues (2011) identified significantly more circulating male mononuclear cells in those with Hashimoto thyroiditis. Ironically, such microchimerism may have a protective role for autoimmune thyroid disorders (Cirello, 2015).

Hyperthyroidism

The incidence of thyrotoxicosis or hyperthyroidism in pregnancy is varied and complicates between 2 and 17 per 1000 births when gestational-age appropriate TSH threshold values are used (Table 58-1). Because normal pregnancy simulates some clinical findings similar to thyroxine excess, clinically mild thyrotoxicosis may be difficult to diagnose. Suggestive findings include tachycardia that exceeds that usually seen with normal pregnancy, thyromegaly, exophthalmos, and failure to gain weight despite adequate food intake. Laboratory testing is confirmatory. TSH levels are markedly depressed, while serum free T₄ (fT₄) levels are elevated (Jameson, 2015). Rarely, hyperthyroidism is caused by abnormally high serum triiodothyronine (T₃) levels—so-called T₃-toxicosis.

Thyrotoxicosis and Pregnancy

The overwhelming cause of thyrotoxicosis in pregnancy is Graves disease, an organ-specific autoimmune process associated with thyroid-stimulating TSH-receptor antibodies as previously discussed (De Leo, 2016). Because these antibodies are specific to Graves hyperthyroidism, such assays have been proposed for diagnosis, management, and prognosis in pregnancies complicated by hyperthyroidism (Barbesino, 2013). At Parkland Hospital, these receptor antibody assays are generally reserved for cases in which fetal thyrotoxicosis is suspected. With Graves disease, during the course of pregnancy, hyperthyroid symptoms may initially worsen because of hCG stimulation but then subsequently diminish with drops in receptor antibody titers in the second half of pregnancy (Mestman, 2012; Sarkhail, 2016). Amino and coworkers (2003) have found that levels of blocking antibodies also decline during pregnancy.

Treatment

Thyrotoxicosis during pregnancy can nearly always be controlled by thionamide drugs. Propylthiouracil (PTU) has been historically preferred because it partially inhibits the conversion of T₄ to T₃ and crosses the placenta less readily than methimazole. The latter has also been associated with a rare methimazole embryopathy, characterized by esophageal or choanal atresia as well as aplasia cutis, a congenital skin defect. Yoshihara and associates (2012, 2015) analyzed outcomes in Japanese women with first-trimester hyperthyroidism and found a twofold increased risk of major fetal malformations in pregnancies exposed to methimazole compared with either PTU or potassium iodide. Specifically, seven of nine cases with aplasia cutis and the only case of esophageal atresia were in the group of methimazole-exposed fetuses. There have also been reports of a PTU-associated embryopathy (Andersen, 2014).

In 2009, the Food and Drug Administration issued a safety alert on PTU-associated hepatotoxicity. This warning prompted the American Thyroid Association and the American Association of Clinical Endocrinologists (2011) to recommend PTU therapy during the first trimester followed by methimazole beginning in the second trimester. The obvious disadvantage is that this might lead to poorly controlled thyroid function. Accordingly, at Parkland Hospital, we continue to prescribe PTU treatment throughout pregnancy.

Transient leukopenia can be documented in up to 10 percent of women taking antithyroid drugs, but this does not require therapy cessation (American College of Obstetricians and Gynecologists, 2017). In approximately 0.3 percent, however, agranulocytosis develops suddenly and mandates drug discontinuance (Thomas, 2013). It is not dose related, and because of its acute onset, serial leukocyte counts during therapy are not helpful. Thus, if fever or sore throat develops, women are instructed to discontinue medication immediately and report for a complete blood count.

Therapy may have other side effects. First, as noted, hepatotoxicity is a possibility and develops in approximately 0.1 percent of treated women. Serial measurement of hepatic enzyme levels does not prevent fulminant PTU-related hepatotoxicity. Second, approximately 20 percent of patients treated with PTU develop antineutrophil cytoplasmic antibodies (ANCA). Despite this, only a small percentage of these subsequently develops serious vasculitis (Kimura, 2013). Finally, although thionamides have the potential to cause fetal complications, these are uncommon. In some cases, thionamides may even be therapeutic for the fetus, because TSH-receptor antibodies cross the placenta and can stimulate the fetal thyroid gland to cause thyrotoxicosis and goiter.

The initial thionamide dose is empirical. For nonpregnant patients, the American Thyroid Association recommends that methimazole be used at an initial higher daily dose of 10 to 20 mg orally followed by a lower maintenance dose of 5 to 10 mg. If PTU is selected, a dose of 50 to 150 mg orally three times daily may be initiated depending on clinical severity (Bahn, 2011). At Parkland Hospital, we usually initially give 300 or 450 mg of PTU daily in three divided doses for pregnant women. Occasionally, daily doses of 600 mg or higher are necessary. As discussed, we generally do not transition women to methimazole during the second trimester. The goal is treatment with the lowest possible thionamide dose to maintain thyroid hormone levels slightly above or in the high normal range, while TSH levels remains suppressed (Bahn, 2011). Serum free T₄ concentrations are measured every 4 to 6 weeks.

Subtotal thyroidectomy can be performed after thyrotoxicosis is medically controlled. This seldom is done during pregnancy but may be appropriate for the very few women who cannot adhere to medical treatment or in whom drug therapy proves toxic (Stagnaro-Green, 2012a). Surgery is best accomplished in the second trimester. Potential drawbacks of thyroidectomy include inadvertent resection of parathyroid glands and injury to the recurrent laryngeal nerve.

Thyroid ablation with therapeutic radioactive iodine is contraindicated during pregnancy. The necessary doses may also cause fetal thyroid gland destruction. Thus, when radioactive iodine is given unintentionally, many clinicians recommend abortion. Any exposed fetus must be carefully evaluated, and the incidence of fetal hypothyroidism depends on gestational age and radioiodine dose (Berlin, 2001). There is no evidence that radioiodine given before pregnancy causes fetal anomalies if enough time has passed to allow radiation effects to dissipate and if the woman is euthyroid (Ayala, 1998). The International Commission on Radiological Protection has recommended that women avoid pregnancy for 6 months after radioablative therapy (Brent, 2008). Moreover, during lactation, the breast also concentrates a substantial amount of iodine. This may pose neonatal risk due to ¹³¹I-containing milk ingestion and maternal risk from significant breast irradiation. To limit the latter, a delay of 3 months after breastfeeding cessation will more reliably ensure complete breast involution.

Pregnancy Outcome

Women with thyrotoxicosis have pregnancy outcomes that largely depend on whether metabolic control is achieved. For example, excess thyroxine may cause miscarriage or preterm birth (Andersen, 2014; Sheehan, 2015). In untreated women or in those who remain hyperthyroid despite therapy, incidences of preeclampsia, heart failure, and adverse perinatal outcomes are higher (Table 58-2). A prospective cohort study from China showed that women with clinical hyperthyroidism had a 12-fold greater risk of delivering an infant with hearing loss (Su, 2011).

Fetal and Neonatal Effects

In most cases, the perinate is euthyroid. In some, however, hyper- or hypothyroidism can develop with or without a goiter (Fig. 58-3). Clinical hyperthyroidism develops in up to 1 percent of neonates born to women with Graves disease (Barbesino, 2013; Fitzpatrick, 2010). If fetal thyroid disease is suspected, nomograms are available for sonographically measured thyroid volume (Gietka-Czernel, 2012).

The fetus or neonate who was exposed to excessive maternal thyroxine may have any of several clinical presentations. First, goitrous thyrotoxicosis is caused by placental transfer of thyroid-stimulating immunoglobulins. Nonimmune hydrops and fetal demise have been reported with fetal thyrotoxicosis (Nachum, 2003; Stulberg, 2000). The best predictor of perinatal thyrotoxicosis is presence of thyroid-stimulating TSH-receptor antibodies in women with Graves disease (Nathan, 2014). This is especially true if their levels are more than threefold higher than the upper normal limit (Barbesino, 2013). In a study of 72 pregnant women with Graves disease, Luton and associates (2005) reported that none of the fetuses in 31 low-risk mothers had a goiter, and all were euthyroid at delivery. Low risk was defined as no requirement for antithyroid medications during the third trimester or an absence of antithyroid antibodies. Conversely, in a group of 41 women who either were taking antithyroid medication at delivery or had thyroid receptor antibodies, 11 fetuses—27 percent—had sonographic evidence of a goiter at 32 weeks’ gestation. Seven of these 11 fetuses were determined to be hypothyroid, and the remaining fetuses were hyperthyroid. In response to these results, the American Thyroid Association and American Association of Clinical Endocrinologists (2011) recommend routine evaluation of TSH-receptor antibodies between 22 and 26 weeks’ gestation in women with Graves disease. The American College of Obstetricians and Gynecologists (2017), however, does not recommend such testing. If the fetus is thyrotoxic, maternal thionamide drugs are adjusted even though maternal thyroid function may be within the targeted range (Mestman, 2012). Although usually short-lived, neonatal thyrotoxicosis may require short-course antithyroid drug treatment (Levy-Shraga, 2014; Nathan, 2014).

A second presentation is goitrous hypothyroidism caused by fetal exposure to maternally administered thionamides (see Fig. 58-3). Although there are theoretical neurological implications, reports of adverse fetal effects seem to have been exaggerated. Available data indicate that thionamides carry an extremely small risk for causing neonatal hypothyroidism (Momotani, 1997; O’Doherty, 1999). For example, in at least 239 treated thyrotoxic women shown in Table 58-1, evidence of hypothyroidism was found in only four newborns. Furthermore, at least four long-term studies report no abnormal intellectual and physical development of these children (Mestman, 1998). If maternal hypothyroidism developed, the fetus can be treated by a reduced maternal antithyroid medication dose and injections of intraamnionic thyroxine if necessary.

A third presentation, nongoitrous hypothyroidism, may develop from transplacental passage of maternal TSH-receptor blocking antibodies (Fitzpatrick, 2010; Gallagher, 2001). And finally, fetal thyrotoxicosis after maternal thyroid gland ablation, usually with ¹³¹I radioiodine, may result from transplacental thyroid-stimulating antibodies. In one report of early fetal exposure to radioiodine, neonatal thyroid studies indicated transient hyperthyroidism from maternal transfer of stimulating antibodies (Tran, 2010).

Fetal Diagnosis

Evaluation of fetal thyroid function is somewhat controversial. Although the fetal thyroid volume can be measured sonographically in women taking thionamide drugs or in those with thyroid-stimulating antibodies, most investigators do not currently recommend this routinely (Cohen, 2003; Luton, 2005). Kilpatrick (2003) recommends umbilical cord blood sampling and fetal antibody testing only if the mother has previously undergone radioiodine ablation. Because fetal hyper- or hypothyroidism may cause hydrops, growth restriction, goiter, or tachycardia, fetal blood sampling may be appropriate if these are identified (Brand, 2005). The Endocrine Society clinical practice guidelines recommend umbilical cord blood sampling only when the diagnosis of fetal thyroid disease cannot be reasonably ascertained based on clinical and sonographic data (Garber, 2012). Diagnosis and treatment are considered further in Chapter 16 (Surgical Therapy).

Thyroid Storm and Heart Failure

Both of these are acute and life-threatening in pregnancy. Thyroid storm is a hypermetabolic state and is rare in pregnancy. In contrast, pulmonary hypertension and heart failure from cardiomyopathy caused by the profound myocardial effects of thyroxine are common in pregnant women (Sheffield, 2004). As shown in Table 58-2, heart failure developed in 8 percent of 90 women with uncontrolled thyrotoxicosis. In these women, cardiomyopathy is characterized by a high-output state, which may lead to a dilated cardiomyopathy (Fadel, 2000; Klein, 1998). The pregnant woman with thyrotoxicosis has minimal cardiac reserve, and decompensation is usually precipitated by preeclampsia, anemia, sepsis, or a combination of these. Fortunately, thyroxine-induced cardiomyopathy and pulmonary hypertension are frequently reversible (Sheffield, 2004; Siu, 2007; Vydt, 2006).

Management

Treatment is similar for thyroid storm and heart failure and should be carried out in an intensive care area that may include special-care units within labor and delivery (American College of Obstetricians and Gynecologists, 2017). Shown in Figure 58-4 is our stepwise approach to medical management of thyroid storm or thyrotoxic heart failure. An hour or two after initial thionamide administration, iodide is given to inhibit thyroidal release of T₃ and T₄. It can be given intravenously as sodium iodide or orally as either saturated solution of potassium iodide (SSKI) or Lugol solution. With a history of iodine-induced anaphylaxis, lithium carbonate, 300 mg every 6 hours, is given instead. Most authorities recommend dexamethasone, 2 mg intravenously every 6 hours for four doses, to further block peripheral conversion of T₄ to T₃. If a β-blocker drug is given to control tachycardia, its effect on heart failure must be considered. Propranolol, labetalol, and esmolol have all been used successfully. Coexisting severe preeclampsia, infection, or anemia should be aggressively managed before delivery is considered.

Hyperemesis Gravidarum and Gestational Transient Thyrotoxicosis

Transient biochemical features of hyperthyroidism may be observed in 2 to 15 percent of women in early pregnancy (Fitzpatrick, 2010). Many women with hyperemesis gravidarum have abnormally high serum thyroxine levels and low TSH levels (Chap. 54, Management). This results from TSH-receptor stimulation from massive—but normal for pregnancy— concentrations of hCG. This transient condition is also termed gestational transient thyrotoxicosis. Even if associated with hyperemesis, antithyroid drugs are not warranted (American College of Obstetricians and Gynecologists, 2017). The degree of hCG level elevation does not correlate with thyroxine and TSH values, which become more normal by midpregnancy (Nathan, 2014; Yoshihara, 2015).

Thyrotoxicosis and Gestational Trophoblastic Disease

The prevalence of increased thyroxine levels in women with a molar pregnancy ranges between 25 and 65 percent (Hershman, 2004). As discussed, abnormally high hCG levels lead to overstimulation of the TSH receptor. Because these tumors are now usually diagnosed early, clinically apparent hyperthyroidism has become less common. With molar evacuation, serum free T₄ levels usually normalize rapidly in parallel with declining hCG concentrations. This is discussed further in Chapter 20 (Diagnosis).

Subclinical Hyperthyroidism

Third-generation TSH assays with an analytical sensitivity of 0.002 mU/mL permit identification of subclinical thyroid disorders. These biochemically defined extremes usually represent normal biological variations but may herald the earliest stages of thyroid dysfunction. Subclinical hyperthyroidism is characterized by an abnormally low serum TSH concentration in concert with normal thyroxine hormone levels (Surks, 2004). Long-term effects of persistent subclinical thyrotoxicosis include osteoporosis, cardiovascular morbidity, and progression to overt thyrotoxicosis or thyroid failure. Casey and Leveno (2006) reported that subclinical hyperthyroidism was found in 1.7 percent of pregnant women. Importantly, subclinical hyperthyroidism was not associated with adverse pregnancy outcomes. In separate retrospective analyses of almost 25,000 women who underwent thyroid screening throughout pregnancy, Wilson and colleagues (2012) and Tudela and coworkers (2012) also found no relationship between subclinical hyperthyroidism and preeclampsia or gestational diabetes.

Treatment of subclinical hyperthyroidism is unwarranted in pregnancy because antithyroid drugs may affect the fetus. These women may benefit from periodic surveillance, and approximately half eventually have normal TSH concentrations.

Hypothyroidism

Overt or symptomatic hypothyroidism, as shown in Table 58-3, has been reported to complicate between 2 and 12 per 1000 pregnancies. It is characterized by insidious nonspecific clinical findings that include fatigue, constipation, cold intolerance, muscle cramps, and weight gain. A pathologically enlarged thyroid gland depends on the etiology of hypothyroidism and is more likely in women in areas of endemic iodine deficiency or those with Hashimoto thyroiditis. Other findings include edema, dry skin, hair loss, and prolonged relaxation phase of deep tendon reflexes. Clinical or overt hypothyroidism is confirmed when an abnormally high serum TSH level is accompanied by an abnormally low thyroxine level. Subclinical hypothyroidism, discussed later, is defined by an elevated serum TSH level and normal serum thyroxine concentration (Jameson, 2015). Sometimes included in the spectrum of subclinical thyroid disease are asymptomatic individuals with high levels of anti-TPO or antithyroglobulin antibodies. Autoimmune euthyroid disease represents a new investigative frontier in screening and treatment of thyroid dysfunction during pregnancy.

Overt Hypothyroidism and Pregnancy

The most common cause of hypothyroidism in pregnancy is Hashimoto thyroiditis, characterized by glandular destruction from autoantibodies, particularly anti-TPO antibodies. Another cause is postablative Graves disease. Clinical identification of hypothyroidism is especially difficult during pregnancy because many of the signs or symptoms are also common to pregnancy itself. Thyroid analyte testing should be performed on symptomatic women or those with a history of thyroid disease (American College of Obstetricians and Gynecologists, 2017). Severe hypothyroidism during pregnancy is uncommon, probably because it is often associated with infertility and higher spontaneous abortion rates (De Groot, 2012). Even women with treated hypothyroidism undergoing in vitro fertilization have a significantly lower chance of achieving pregnancy (Scoccia, 2012).

Treatment

The American Thyroid Association and American Association of Clinical Endocrinologists (2011) recommend replacement therapy for overt hypothyroidism beginning with levothyroxine in doses of 1 to 2 μg/kg/d or approximately 100 μg daily. Women who are athyreotic after thyroidectomy or radioiodine therapy may require higher doses. Surveillance is with TSH levels measured at 4- to 6-week intervals, and the thyroxine dose is adjusted by 25- to 50-μg increments until TSH values become normal. Pregnancy is associated with an increased thyroxine requirement in approximately a third of supplemented women (Abalovich, 2010; Alexander, 2004). The increased demand in pregnancy is believed to be related to augmented estrogen production (Arafah, 2001).

Greater thyroxine requirements begin as early as 5 weeks’ gestation. In a randomized trial that provided an increased levothyroxine dose at pregnancy confirmation in 60 mothers, Yassa and coworkers (2010) found that a 29- to 43-percent increase in the weekly dose maintained serum TSH values <5.0 mU/L during the first trimester in all women. Importantly, however, this increase caused TSH suppression in more than a third of women. Significant hypothyroidism may develop early in women without thyroid reserve, such as those with a previous thyroidectomy, those with prior radioiodine ablation, or those undergoing assisted reproductive techniques (Alexander, 2004; Loh, 2009). Anticipatory 25-percent increases in thyroxine replacement at pregnancy confirmation will reduce this likelihood. All other women with hypothyroidism should instead undergo TSH testing at initiation of prenatal care.

Pregnancy Outcome with Overt Hypothyroidism

Observational studies, although limited, indicate that excessive adverse perinatal outcomes are associated with overt thyroxine deficiency (Table 58-4). Preterm birth rates, for example, are higher (Sheehan, 2015). With appropriate replacement therapy, however, rates of adverse effects are not increased in most reports (Bryant, 2015; Matalon, 2006; Tan, 2006). In one dissenting study, however, risks for some pregnancy complications were greater even in women taking replacement therapy (Wikner, 2008). Most experts agree that adequate hormone replacement during pregnancy minimizes the risk of adverse outcomes and most complications.

Fetal and Neonatal Effects

Undoubtedly, maternal and fetal thyroid abnormalities are related. In both, thyroid function is dependent on adequate iodide intake, and its deficiency early in pregnancy can cause both maternal and fetal hypothyroidism. And, as discussed, maternal TSH-receptor-blocking antibodies can cross the placenta and cause fetal thyroid dysfunction. Rovelli and colleagues (2010) evaluated 129 neonates born to women with autoimmune thyroiditis. They found that 28 percent had an elevated TSH level on the third or fourth day of life, and 47 percent of these had TPO antibodies on day 15. Still, autoantibodies were undetectable at 6 months of age. It seems paradoxical that despite these transient laboratory findings in the neonate, TPO and antithyroglobulin antibodies have little or no effect on fetal thyroid function (Fisher, 1997). Indeed, prevalence of fetal hypothyroidism in women with Hashimoto thyroiditis is estimated to be only 1 in 180,000 newborns (Brown, 1996).

Subclinical Hypothyroidism

Although common in women, the incidence of subclinical hypothyroidism varies depending on age, race, dietary iodine intake, and serum TSH thresholds used to establish the diagnosis (Jameson, 2015). In two large studies totaling more than 25,000 pregnant women screened in the first half of pregnancy, subclinical hypothyroidism was identified in 2.3 percent (Casey, 2005; Cleary-Goldman, 2008). The rate of progression to overt thyroid failure is affected by TSH level, age, other disorders such as diabetes, and presence and concentration of antithyroid antibodies.

Diez and Iglesias (2004) prospectively followed 93 nonpregnant women with subclinical hypothyroidism for 5 years and reported that in a third, TSH values became normal. In the other two thirds, those women whose TSH levels were 10 to 15 mU/L developed overt disease at a rate of 19 per 100 patient years. Those women whose TSH levels were <10 mU/L developed overt hypothyroidism at a rate of 2 per 100 patient years.

Regarding screening for subclinical hypothyroidism in nonpregnant individuals, the U.S. Preventative Services Task Force also reports that nearly all patients who develop overt hypothyroidism within 5 years have an initial TSH level >10 mU/L (Helfand, 2004; Karmisholt, 2008).

For gravidas, in a 20-year follow-up study of 5805 women who were screened in early pregnancy, only 3 percent developed thyroid disease. Of the 224 women identified with subclinical hypothyroidism during pregnancy, 17 percent developed thyroid disease in the next 20 years, and most of these had either TPO or thyroglobulin antibodies during pregnancy (Männistö, 2010). Thus, the likelihood of progression to overt hypothyroidism during pregnancy in otherwise healthy women with subclinical hypothyroidism seems remote.

Subclinical Hypothyroidism and Pregnancy

Earlier studies were suggestive that subclinical hypothyroidism might be associated with adverse pregnancy outcomes. In 1999, interest was heightened by two studies indicating that undiagnosed maternal thyroid hypofunction may impair fetal neuropsychological development. In one study, Pop and associates (1999) described 22 women with free T₄ levels <10th percentile whose offspring were at higher risk for impaired psychomotor development. In the other study, Haddow and coworkers (1999) retrospectively evaluated children born to 48 untreated women whose serum TSH values were >98th percentile. Some had diminished school performance, reading recognition, and intelligent quotient (IQ) scores. Although described as “subclinically hypothyroid,” these women had an abnormally low mean serum free thyroxine level, and thus, many had overt hypothyroidism.

To further evaluate any adverse effects, Casey and colleagues (2005) identified subclinical hypothyroidism in 2.3 percent of 17,298 women screened at Parkland Hospital before midpregnancy. These women had small but significantly higher incidences of preterm birth, placental abruption, and neonates admitted to the intensive care nursery compared with euthyroid women. In another study of 10,990 similar women, however, Cleary-Goldman and associates (2008) did not find such associations.

Other studies subsequently confirmed a link between subclinical thyroid function and adverse outcomes (Chen, 2017; Maraka, 2016). One included 24,883 women screened throughout pregnancy and showed an almost twofold greater risk of severe preeclampsia (Wilson, 2012). In an analysis of the same cohort, a consistent relationship was shown between rising TSH levels and the risk for gestational diabetes (Tudela, 2012). Finally, Nelson and colleagues (2014) found an elevated risk for diabetes and stillbirth.

Lazarus and colleagues (2012) reported the findings of the international multicenter Controlled Antenatal Thyroid Screening (CATS) study. This study evaluated prenatal thyroid screening and randomized treatment of both subclinical hypothyroidism and isolated maternal hypothyroxinemia. They reported that offspring IQ scores at age 3 years were not superior in the treated pregnancies.

Despite these findings, the unanswered question concerned whether treatment of subclinical hypothyroidism would mitigate any or all of these reported adverse outcomes. To address this, the Maternal-Fetal Medicine Units Network screened more than 97,000 pregnant women for thyroid disorders and reported that 3.3 percent had subclinical hypothyroidism. These 677 women were randomly assigned to thyroxine replacement therapy or placebo. As reported by Casey and colleagues (2017), and shown in Table 58-5, maternal adverse pregnancy outcomes or cognitive development in the offspring at 5 years did not differ between groups. Annual developmental testing scores and behavioral and attention-deficit hyperactivity disorder results also did not differ.

Screening in Pregnancy

Because of the findings in the studies from 1999 cited above, some professional organizations began to recommend routine prenatal screening and treatment for subclinical hypothyroidism. Consequent to the Lazarus study, however, clinical practice guidelines from the Endocrine Society, the American Thyroid Association, and the American Association of Clinical Endocrinologists uniformly recommended screening only those at greater risk during pregnancy (De Groot, 2012; Garber, 2012). This has been and still is the recommendation of the American College of Obstetricians and Gynecologists (2017). The findings of Casey and colleagues (2017) further buttress these recommendations.

Isolated Maternal Hypothyroxinemia

Women with low serum free T₄ values but a normal-range TSH level are considered to have isolated maternal hypothyroxinemia. Its incidence in two large trials was 1.3 to 2.1 percent (Casey, 2007; Cleary-Goldman, 2008). As shown in Figure 58-2, unlike in subclinical hypothyroidism, these women had a low prevalence of antithyroid antibodies.

Evolution of the knowledge of this thyroid disorder was similar to that seen with subclinical hypothyroidism. Initial studies reported that offspring of women with isolated hypothyroxinemia had neurodevelopmental difficulties (Kooistra, 2006; Pop, 1999, 2003). In another study, Casey and colleagues (2007) found no higher risks for other adverse perinatal outcomes compared with those of euthyroid women. Also, the aforementioned CATS study did not find improved neurodevelopmental outcomes in women with isolated hypothyroxinemia who were then treated with thyroxine (Lazarus, 2012).

The randomized trial conducted by the Maternal-Fetal Medicine Units Network also provided data to settle this question. Casey and colleagues (2017) noted no higher rates of adverse outcomes between groups and found that early thyroxine treatment offered no benefits (see Table 58-5).

Euthyroid Autoimmune Thyroid Disease

Autoantibodies to TPO and thyroglobulin have been identified in 6 to 20 percent of reproductive-aged women (Thangaratinam, 2011). Most who test positive for such antibodies, however, are euthyroid. That said, such women carry a two- to fivefold increased risk for early pregnancy loss (Stagnaro-Green, 2004; Thangaratinam, 2011). The presence of thyroid antibodies has also been associated with preterm birth (Stagnaro-Green, 2009). In a randomized treatment trial of 115 euthyroid women with TPO antibodies, Negro and coworkers (2006) reported that treatment with levothyroxine astoundingly reduced the preterm birth rate from 22 to 7 percent. Contrarily, Abbassi-Ghanavati and associates (2010) evaluated pregnancy outcomes in more than 1000 untreated women with TPO antibodies and did not find an increased risk for preterm birth compared with the risk in 16,000 euthyroid women without antibodies. These investigators, however, did find a threefold greater risk of placental abruption in these women.

As with nonpregnant subjects with TPO antibodies, these women are also at increased risk for progression of thyroid disease and postpartum thyroiditis (Jameson, 2015; Stagnaro-Green, 2012a). Currently, universal screening for the thyroid autoantibodies is not recommended by any professional organization (De Groot, 2012; Stagnaro-Green, 2011a, 2012a).

Iodine Deficiency

Decreasing iodide fortification of table salt and bread products in the United States during the past 25 years has led to occasional iodide deficiency (Caldwell, 2005; Hollowell, 1998). Importantly, the most recent National Health and Nutrition Examination Survey indicated that, overall, the United States population remains iodine sufficient (Caldwell, 2011). Even so, experts agree that iodine nutrition in vulnerable populations, such as pregnant women, requires continued monitoring. In 2011, the Office of Dietary Supplements of the National Institutes of Health sponsored a workshop to prioritize iodine research. Participants emphasized the decline in median urinary iodine levels to 125 μg/L in pregnant women and the serious potential effects on developing fetuses (Swanson, 2012).

Dietary iodine requirements are higher during pregnancy due to augmented thyroid hormone production, increased renal losses, and fetal iodine requirements. Adequate iodine is requisite for fetal neurological development beginning soon after conception, and abnormalities are dependent on the degree of deficiency. The World Health Organization (WHO) has estimated that 38 million children are born every year at risk of lifelong brain damage associated with iodine deficiency (Alipui, 2008).

Although it is doubtful that mild deficiency causes intellectual impairment, supplementation does prevent fetal goiter (Stagnaro-Green, 2012b). Severe deficiency, on the other hand, is frequently associated with damage typically encountered with endemic cretinism (Delange, 2001). It is presumed that moderate deficiency has intermediate and variable effects. Berbel and associates (2009) began daily supplementation in more than 300 pregnant women with moderate deficiency at three time periods—4 to 6 weeks, 12 to 14 weeks, and after delivery. They found improved neurobehavioral development scores in offspring of women supplemented with 200 μg potassium iodide very early in pregnancy. Similarly, Velasco and coworkers (2009) found improved Bayley Psychomotor Development scores in offspring of women supplemented with 300 μg of iodine daily in the first trimester. In contrast, Murcia and colleagues (2011) identified lower psychomotor scores in 1-year-old infants whose mothers reported daily supplementation of more than 150 μg. To address this, randomized controlled trial of iodine supplementation in mildly to moderately iodine-deficient pregnant women in India and Thailand is nearing completion (Pearce, 2016).

Regarding daily iodine intake, the Institute of Medicine (2001) recommends 220 μg/d during pregnancy and 290 μg/d during lactation (Chap. 9, Minerals). The Endocrine Society recommends an average iodine intake of 150 μg/d in reproductive-aged women, and this should be increased to 250 μg during pregnancy and breastfeeding (De Groot, 2012). The American Thyroid Association has recommended that 150 μg of iodine be added to prenatal vitamins to achieve this average daily intake (Becker, 2006). According to Leung and coworkers (2011), however, only 51 percent of the prenatal multivitamins in the United States contain iodine. It has even been suggested that because most cases of maternal hypothyroxinemia worldwide are related to relative iodine deficiency, supplementation may obviate the need to consider thyroxine treatment in such women (Gyamfi, 2009). However, without evidence of benefit, it is hard to justify the cost of iodine supplementation of large numbers of pregnant women in areas with mild iodine deficiency (Pearce, 2016). Importantly, experts caution against oversupplementation. Teng and associates (2006) contend that excessive iodine intake—defined as >300 μg/d—may lead to subclinical hypothyroidism and autoimmune thyroiditis. The Endocrine Society, in accordance with the WHO, advises against exceeding twice the daily recommended intake of iodine, or 500 μg/d (De Groot, 2012; Leung, 2011).

Congenital Hypothyroidism

Universal newborn screening for neonatal hypothyroidism was introduced in 1974 and is now required by law in all states (Chap. 32, Routine Newborn Care). This develops in approximately 1 in 3000 newborns and is one of the most preventable causes of mental retardation (LaFranchi, 2011). Developmental disorders of the thyroid gland such as agenesis and hypoplasia account for 80 to 90 percent of these cases. The remainder is caused by hereditary defects in thyroid hormone production (Moreno, 2008).

Early and aggressive thyroxine replacement is critical for newborns with congenital hypothyroidism. Still, some neonates identified by screening programs who were treated promptly will exhibit cognitive deficits into adolescence (Song, 2001). Therefore, in addition to timing of treatment, the severity of congenital hypothyroidism is an important factor in long-term cognitive outcomes. Olivieri and colleagues (2002) reported that 8 percent of 1420 newborns with congenital hypothyroidism also had other major congenital malformations.

Postpartum Thyroiditis

Transient autoimmune thyroiditis is consistently found in approximately 5 to 10 percent of women during the first year after childbirth (Nathan, 2014; Stagnaro-Green, 2011b, 2012a). Postpartum thyroid dysfunction with an onset within 12 months includes hyperthyroidism, hypothyroidism, or both. The propensity for thyroiditis antedates pregnancy and is directly related to increasing serum levels of thyroid autoantibodies. Up to 50 percent of women who are thyroid-antibody positive in the first trimester will develop postpartum thyroiditis (Stagnaro-Green, 2012a). In a Dutch study of 82 women with type 1 diabetes, postpartum thyroiditis developed in 16 percent and was threefold higher than in the general population (Gallas, 2002). Importantly, 46 percent of those identified with overt postpartum thyroiditis had TPO antibodies in the first trimester.

Clinical Manifestations

In clinical practice, postpartum thyroiditis is diagnosed infrequently because it typically develops months after delivery and causes vague and nonspecific symptoms (Stagnaro-Green, 2004). The clinical presentation varies, and classically two clinical phases that may develop in succession are recognized. The first and earliest is destruction-induced thyrotoxicosis with symptoms from excessive release of hormone from glandular disruption. The onset is abrupt, and a small, painless goiter is common. Although there may be many symptoms, only fatigue and palpitations are more frequent in thyrotoxic women compared with normal controls. This thyrotoxic phase usually lasts only a few months. Thionamides are ineffective, and if symptoms are severe, a β-blocking agent may be given. The second and usually later phase between 4 and 8 months postpartum is hypothyroidism from thyroiditis. Thyromegaly and other symptoms are common and more prominent than during the thyrotoxic phase. Thyroxine replacement at doses of 25 to 75 μg/d is typically given for 6 to 12 months.

Stagnaro-Green and associates (2011b) reported postpartum surveillance results in 4562 Italian gravidas who had been screened for thyroid disease in pregnancy. Serum TSH and anti-TPO antibody levels were measured again at 6 and 12 months. Overall, two thirds of 169 women (3.9 percent) with postpartum thyroiditis were identified to have hypothyroidism only. The other third were diagnosed with hyperthyroidism. Only 14 percent of all women demonstrated the “classic” biphasic progression described above. These findings are consistent with data compiled from 20 other studies between 1982 and 2008 (Stagnaro-Green, 2012a).

Importantly, women who experience either type of postpartum thyroiditis have a 20- to 30-percent risk of eventually developing permanent hypothyroidism, and the annual progression rate is 3.6 percent (Nathan, 2014). Women at greater risk for developing hypothyroidism are those with higher titers of thyroid antibodies and higher TSH levels during the initial hypothyroid phase. Others may develop subclinical disease, but half of those with thyroiditis who are positive for TPO antibodies develop permanent hypothyroidism by 6 to 7 years (Stagnaro-Green, 2012a).

An association between postpartum thyroiditis and postpartum depression has been proposed but remains unresolved. Lucas and coworkers (2001) found a 1.7-percent incidence of postpartum depression at 6 months in women with thyroiditis as well as in controls. Pederson and colleagues (2007) found a significant correlation between abnormal scores on the Edinburgh Postnatal Depression Scale and total thyroxine values in the low normal range during pregnancy in 31 women. Similarly unsettled is the link between depression and thyroid antibodies. Kuijpens and associates (2001) reported that TPO antibodies were a marker for postpartum depression in euthyroid women. In a randomized trial, however, Harris and coworkers (2002) reported no difference in postpartum depression in 342 women with TPO antibodies who were given either levothyroxine or placebo.

Nodular Thyroid Disease

Thyroid nodules can be found in 1 to 2 percent of reproductive-aged women (Fitzpatrick, 2010). Management of a palpable thyroid nodule during pregnancy depends on gestational age and mass size. Small nodules detected by sensitive sonographic methods are more common during pregnancy in some populations. Kung and associates (2002) used high-resolution sonography and found that 15 percent of Chinese women had nodules larger than 2 mm in diameter. Almost half were multiple, and the nodules usually enlarged modestly across pregnancy and did not regress postpartum. Biopsy of those >5 mm³ that persisted at 3 months usually showed nodular hyperplasia, and none were malignant. In most studies, 90 to 95 percent of solitary nodules are benign (Burch, 2016).

Evaluation of thyroid nodules during pregnancy should be similar to that for nonpregnant patients. As discussed in Chapter 46 (Radiographic Contrast Agents), radioiodine scanning in pregnancy is usually not recommended (American College of Obstetricians and Gynecologists, 2017). Sonographic examination reliably detects nodules >5 mm, and their solid or cystic structure also is determined. According to the American Association of Clinical Endocrinologists, sonographic characteristics associated with malignancy include hypoechogenic pattern, irregular margins, and microcalcifications (Gharib, 2005). Fine-needle aspiration (FNA) is an excellent assessment method, and histological tumor markers and immunostaining are reliable to evaluate for malignancy (Hegedüs, 2004). If the FNA biopsy shows a follicular lesion, surgery may be deferred until after delivery.

Evaluation of thyroid cancer involves a multidisciplinary approach (Fagin, 2016). Most thyroid carcinomas are well differentiated and pursue an indolent course. Messuti and coworkers (2014) provided evidence that persistence or recurrence of these tumors may be more common in pregnant women. When thyroid malignancy is diagnosed during the first or second trimester, thyroidectomy may be performed before the third trimester (Chap. 63, Thyroid Cancer). In women without evidence of an aggressive thyroid cancer or in those diagnosed in the third trimester, surgical treatment can be deferred to the immediate puerperium (Gharib, 2010).

The function of parathyroid hormone (PTH) is to maintain extracellular fluid calcium concentration. This 84-amino acid hormone acts directly on bone and kidney and indirectly on small intestine through its effects on synthesis of vitamin D (1,25-(OH)₂D) to increase serum calcium (Potts, 2015). Secretion is regulated by serum ionized calcium concentration through a negative feedback system. Calcitonin is a potent parathyroid hormone that acts as a physiological parathyroid hormone antagonist. The interrelationships between these hormones, calcium metabolism, and PTH-related protein produced by fetal tissue are discussed in Chapter 4 (Parathyroid Glands).

Of fetal demands, calcium requirements reach 300 mg/d in late pregnancy and 30 g for the entire gestation. These needs and greater renal calcium loss from augmented glomerular filtration substantively raise maternal calcium demands. Pregnancy is associated with a twofold rise in serum concentrations of 1,25-dihydroxyvitamin D, which increases gastrointestinal calcium absorption. The effectuating hormone is probably of placental and decidual origin because maternal PTH levels are low normal or decreased during pregnancy (Cooper, 2011; Molitch, 2000). Total serum calcium levels decline with serum albumin concentrations, but ionized calcium levels remain unchanged. Vargas Zapata and colleagues (2004) have suggested a role for insulin-like growth factor-1 (IGF-1) in maternal calcium homeostasis and bone turnover.

Hyperparathyroidism

Hypercalcemia is caused by hyperparathyroidism or cancer in 90 percent of cases (Potts, 2015). Because many automated laboratory systems include serum calcium measurement, hyperparathyroidism has changed from being a condition defined by symptoms to one that is discovered on routine screening (Pallan, 2012). It has a reported prevalence of 2 to 3 per 1000 women, but some have estimated the rate to be as high as 14 per 1000 when asymptomatic cases are included. Almost 80 percent are caused by a solitary adenoma, and another 15 percent by hyperfunctioning of all four glands. In the remainder, a malignancy as the cause of increased serum calcium levels is usually obvious. Of note, PTH produced by tumors is not identical to the natural hormone and may not be detected by routine assays.

In most patients, the serum calcium level is elevated to within only 1 to 1.5 mg/dL above the upper normal limit. This may help to explain why only 20 percent of those who have abnormally elevated levels are symptomatic (Bilezikian, 2004). In a fourth, however, symptoms become apparent when the serum calcium level continues to rise. Hypercalcemic crisis manifests as stupor, nausea, vomiting, weakness, fatigue, and dehydration.

All women with symptomatic hyperparathyroidism should be surgically treated (Potts, 2015). Indications for parathyroidectomy include a serum calcium level 1.0 mg/dL above the upper normal range, a calculated creatinine clearance <60 mL/min, reduced bone density, or age >50 years (Bilezikian, 2009). Those not meeting these criteria should undergo annual serum calcium and creatinine level measurement and bone density assessment every 1 to 2 years (Pallan, 2012).

Hyperparathyroidism in Pregnancy

In their review, Schnatz and Thaxton (2005) found fewer than 200 reported cases of hyperparathyroidism complicating pregnancy. As in nonpregnant patients, parathyroid adenoma is the most common etiology. Ectopic parathyroid hormone production and rare cases of parathyroid carcinoma have been reported in pregnancy (Montoro, 2000; Saad, 2014). Symptoms include hyperemesis, generalized weakness, renal calculi, and psychiatric disorders. Occasionally, pancreatitis is the presenting disorder (Cooper, 2011; Hirsch, 2015).

Pregnancy theoretically improves hyperparathyroidism because of significant calcium shunting to the fetus and augmented renal excretion (Power, 1999). When the “protective effects” of pregnancy are withdrawn, however, postpartum hypercalcemic crisis is a significant danger. This life-threatening complication can be seen with serum calcium levels greater than 14 mg/dL and is characterized by nausea, vomiting, tremors, dehydration, and mental status changes (Malekar-Raikar, 2011).

Early reports described excessive stillbirths and preterm deliveries in pregnancies complicated by hyperparathyroidism. More recent reports, however, described lower rates of stillbirth, neonatal death, and neonatal tetany (Kovacs, 2011). Other fetal complications include miscarriage, fetal-growth restriction, and low birthweight (Chamarthi, 2011). Schnatz (2005) reported a 25-percent incidence of preeclampsia.

Management in Pregnancy

Surgical removal of a symptomatic parathyroid adenoma is preferable. This should prevent fetal and neonatal morbidities and postpartum parathyroid crises (Kovacs, 2011). Elective neck exploration during pregnancy is usually well tolerated, even in the third trimester (Hirsch, 2015; Schnatz, 2005; Stringer, 2017). In at least two cases, a mediastinal adenoma was removed at midpregnancy (Rooney, 1998; Saad, 2014).

Medical management may be appropriate in asymptomatic pregnant women with mild hypercalcemia (Hirsch, 2015). If so, patients are carefully monitored in the puerperium for hypercalcemic crisis. Initial medical management might include calcitonin to decrease skeletal calcium release, or oral phosphate, 1 to 1.5 g daily in divided doses, to bind excess calcium. For women with dangerously elevated serum calcium levels or those who are mentally obtunded with hypercalcemic crisis, emergency treatment is instituted. Diuresis with intravenous normal saline is begun so that urine flow exceeds 150 mL/hr. Furosemide is given in conventional doses to block tubular calcium reabsorption. Importantly, hypokalemia and hypomagnesemia should be prevented. Adjunctive therapy includes mithramycin, which inhibits bone resorption.

Neonatal Effects

Normally, cord blood calcium levels are higher than maternal levels (Chap. 7, Placental Role in Embryofetal Development). With maternal hyperparathyroidism, abnormally elevated maternal and thus fetal levels further suppress fetal parathyroid function. Because of this, newborn calcium levels rapidly drop after birth, and 15 to 25 percent of these neonates develop severe hypocalcemia with or without tetany (Molitch, 2000). Neonatal hypoparathyroidism caused by maternal hyperparathyroidism is usually transient and is treated with calcium and 1,25-dihydroxyvitamin D₃ (calcitriol). The latter will not be effective in preterm infants, however, because the intestinal vitamin D receptor is not sufficiently expressed (Kovacs, 2011). Neonatal tetany or seizures should stimulate an evaluation for maternal hyperparathyroidism (Beattie, 2000; Ip, 2003).

Hypoparathyroidism

The most common cause of hypocalcemia is hypoparathyroidism that usually follows parathyroid or thyroid surgery. Hypoparathyroidism is estimated to follow up to 7 percent of total thyroidectomies (Shoback, 2008). It is characterized by facial muscle spasms, muscle cramps, and paresthesias of the lips, tongue, fingers, and feet. This can progress to tetany and seizures (Potts, 2015). Chronically, hypocalcemic pregnant women may also have a fetus with skeletal demineralization resulting in multiple bone fractures in the neonatal period (Alikasifoglu, 2005).

Maternal treatment includes calcitriol, dihydrotachysterol, or large vitamin D doses of 50,000 to 150,000 U/d; calcium gluconate or calcium lactate in doses of 3 to 5 g/d; and a low-phosphate diet. Fetal risks from large doses of vitamin D have not been established. During treatment, the therapeutic challenge in women with known hypoparathyroidism is management of blood calcium levels. It is possible that the greater calcium absorption typical of pregnancy will result in lower calcium requirements or that the fetal demand for calcium will result in greater need. The goal during pregnancy is to maintain a corrected calcium level in the low normal range.

Pregnancy-Associated Osteoporosis

In most gravidas, even with their remarkably increased calcium requirements, it is uncertain whether pregnancy causes osteopenia (Kaur, 2003; To, 2003). In one study of 200 pregnant women in which bone mass was measured, Kraemer and colleagues (2011) demonstrated a decline in bone density during pregnancy. Women who breastfed, carried twin pregnancies, or had a low body mass index were at higher risk of bone loss. From their review, Thomas and Weisman (2006) cite a 3- to 4-percent average reduction in bone mineral density during pregnancy. Lactation also represents a period of negative calcium balance that is corrected through maternal skeletal resorption. Feigenberg and coworkers (2008) found cortical bone mass reductions using ultrasound in young primiparas in the puerperium compared with nulligravid controls. Rarely, some women develop idiopathic osteoporosis while pregnant or lactating (Hellmeyer, 2007).

The most common symptom of osteoporosis is back pain in late pregnancy or postpartum. Other symptoms are hip pain, either unilateral or bilateral, and difficulty in weight bearing until the woman is nearly immobilized (Maliha, 2012). In more than half of women, no apparent reason for osteopenia is found. Some known causes include heparin (unfractionated only), prolonged bed rest, and corticosteroid therapy (Cunningham, 2005; Galambosi, 2016). In a few cases, overt hyperparathyroidism or thyrotoxicosis eventually develops.

Treatment is problematic and includes calcium and vitamin D supplementation and standard pain management. Shown in Figure 58-5 is a hip radiograph from a woman treated at Parkland Hospital during the third trimester for transient osteoporosis of pregnancy. For women with pregnancy-associated osteopenia, long-term surveillance indicates that although bone density improves, these women and their offspring may have chronic osteopenia (Carbone, 1995). Related, prenatal supplementation of normal women with cholecalciferol, 1000 IU/d, did not increase offspring bone mineral content, although it did ensure maternal vitamin D repletion (Cooper, 2016).

Pregnancy has profound effects on adrenal cortical secretion and its control or stimulation. These interrelationships were reviewed by Lekarev and New (2011) and are discussed in detail in Chapter 4 (Parathyroid Glands).

Pheochromocytoma

Pheochromocytomas are chromaffin tumors that secrete catecholamines and usually are located in the adrenal medulla, although 10 percent are located in sympathetic ganglia. They are called the 10-percent tumor because approximately 10 percent are bilateral, 10 percent are extraadrenal, and 10 percent are malignant. These tumors can be associated with medullary thyroid carcinoma and hyperparathyroidism in some of the autosomally dominant or recessive multiple endocrine neoplasia syndromes, as well as in neurofibromatosis and von Hippel-Lindau disease (Neumann, 2015).

These tumors complicate approximately 1 per 50,000 pregnancies (Quartermaine, 2017). Notably, they are found in 0.1 percent of hypertensive patients (Abdelmannan, 2011). However, they are more commonly found at autopsy but with infrequent clinical recognition. Symptoms are usually paroxysmal and manifest as hypertensive crisis, seizure disorders, or anxiety attacks. Hypertension is sustained in 60 percent of patients, but half of these also have paroxysmal crises. Other symptoms during paroxysmal attacks are headaches, profuse sweating, palpitations, chest pain, nausea and vomiting, and pallor or flushing.

The standard screening test is quantification of metanephrines and catecholamine metabolites in a 24-hour urine specimen (Neumann, 2015). Diagnosis is established by measurement of a 24-hour urine collection with at least two of three assays for free catecholamines, metanephrines, or vanillylmandelic acid (VMA). Determination of plasma catecholamine levels is the most sensitive test. In nonpregnant patients, adrenal localization is usually successful with either computed tomography (CT) or magnetic resonance (MR) imaging. For most cases, preferred treatment is laparoscopic adrenalectomy (Neumann, 2015).

Pheochromocytoma Complicating Pregnancy

These tumors are rare but result in dangerous pregnancy complications. Geelhoed (1983) provided an earlier review of 89 cases in which 43 mothers died. Maternal death was much more common if the tumor was not diagnosed antepartum—58 versus 18 percent. As seen in Table 58-6, maternal mortality rates are now lower but still formidable. In their review of 77 cases, Biggar and Lennard (2013) reported that antepartum diagnosis is the most important determinant of maternal mortality risk. That said, Salazar-Vega and colleagues (2014) described good outcomes in women diagnosed after delivery.

Diagnosis of pheochromocytoma in pregnancy is similar to that for nonpregnant patients. MR imaging is the preferred technique because it almost always locates adrenal and extraadrenal pheochromocytomas (Fig. 58-6). In many cases, the principal challenge is to differentiate preeclampsia from the hypertensive crisis caused by pheochromocytoma. Grimbert and colleagues (1999) diagnosed two pheochromocytomas during 56 pregnancies in 30 women with von Hippel-Lindau disease.

Management

Immediate control of hypertension and symptoms with an α-adrenergic blocker such as phenoxybenzamine is imperative. The dose is 10 to 30 mg, two to four times daily. After α-blockade is achieved, β-blockers may be given for tachycardia. In many cases, surgical exploration and tumor removal are performed during pregnancy, preferably during the second trimester (Biggar, 2013; Dong, 2014). Successful laparoscopic removal of adrenal tumors has become the norm (Miller, 2012; Zuluaga-Gómez, 2012). If diagnosed later in pregnancy, either planned cesarean delivery with tumor excision or postpartum resection is appropriate.

Recurrent tumors are troublesome, and even with good blood pressure control, dangerous peripartum hypertension may develop. We have cared for three women in whom recurrent pheochromocytoma was identified during pregnancy. Hypertension was managed with phenoxybenzamine in all three. Two newborns were healthy, but a third was stillborn in a mother with a massive tumor burden who was receiving phenoxybenzamine, 100 mg daily. In all three women, tumor was resected postpartum.

Cushing Syndrome

This syndrome is rare and the female:male ratio is 3:1 (Arit, 2015). Most cases are iatrogenic from long-term corticosteroid treatment. Cushing disease refers to bilateral adrenal hyperplasia stimulated by corticotropin-producing pituitary adenomas. Corticotropin is also called adrenocorticotropic hormone (ACTH). Most adenomas are microadenomas measuring <1 cm, and half measure <5 mm. Rarely, abnormal secretion of hypothalamic corticotropin-releasing factor may cause corticotropic hyperplasia. Such hyperplasia may also be caused by nonendocrine tumors that produce polypeptides similar to either corticotropin-releasing factor or corticotropin. Less than a fourth of cases of Cushing syndrome are corticotropin independent, and most of these are caused by an adrenal adenoma. Tumors are usually bilateral, and half are malignant. Occasionally, associated androgen excess may lead to severe virilization.

The typical cushingoid body habitus is caused by adipose tissue deposition that characteristically results in moon facies, a buffalo hump, and truncal obesity. Fatigability and weakness, hypertension, hirsutism, and amenorrhea are each encountered in 75 to 85 percent of nonpregnant patients (Hatipoglu, 2012). Personality changes, easy bruisability, and cutaneous striae are common. Up to 60 percent may have impaired glucose tolerance. Diagnosis can be difficult and is suggested by elevated plasma cortisol levels that cannot be suppressed by dexamethasone or by elevated 24-hour urine free cortisol excretion (Arit, 2015; Loriaux, 2017). Neither test is totally accurate, and each is more difficult to interpret in obese patients. Serum corticotropin levels and CT and MR imaging are used to localize pituitary and adrenal tumors or hyperplasia.

Cushing Syndrome and Pregnancy

Because most women have corticotropin-dependent Cushing syndrome, associated androgen excess may cause anovulation, and pregnancy is rare. In their review, Lekarev and New (2011) identified fewer than 140 reported cases of Cushing syndrome in pregnancy. These differ compared with nonpregnant women in that half are caused by corticotropin-independent adrenal adenomas (Kamoun, 2014; Lacroix, 2015). Approximately 30 percent of cases are from a pituitary adenoma, and 10 percent from adrenal carcinomas. All reports stress difficulties in diagnosis because of pregnancy-induced increases in plasma cortisol, corticotropin, and corticotropin-releasing factor levels. Measurement of 24-hour urinary free cortisol excretion is recommended, with consideration for the normal elevation seen in pregnancy.

Pregnancy outcomes in women with Cushing syndrome are listed in Table 58-7. Heart failure is common during pregnancy and is a major cause of maternal mortality (Buescher, 1992). Hypercortisolism in pregnancy may also cause poor wound healing, osteoporotic fracture, and psychiatric complications (Kamoun, 2014).

Long-term medical therapy for Cushing syndrome usually is ineffective, and definitive therapy is resection of the pituitary or adrenal adenoma or bilateral adrenalectomy for hyperplasia (Lacroix, 2015; Motivala, 2011). During pregnancy, management of hypertension in mild cases may suffice until delivery. In their review, Lindsay and associates (2005) described primary medical therapy in 20 women with Cushing syndrome. Most were successfully treated with metyrapone as an interim treatment until definitive surgery after delivery. A few cases were treated with oral ketoconazole. However, because this drug also blocks testicular steroidogenesis, treatment during pregnancy with a male fetus is worrisome. Mifepristone, the norethindrone derivative used for abortion and labor induction, has shown promise for treating Cushing disease but should not be used in pregnancy for obvious reasons. If necessary, pituitary adenomas can be treated by transsphenoidal resection (Boscaro, 2001; Lindsay, 2005). Unilateral adrenalectomy has been safely performed in the early third trimester and can also be curative (Abdelmannan, 2011).

Adrenal Insufficiency—Addison Disease

Primary adrenocortical insufficiency is rare because more than 90 percent of total gland volume must be destroyed for symptoms to develop. Autoimmune adrenalitis is the most common cause in the developed world, but tuberculosis is a more frequent etiology in resource-poor countries (Arit, 2015; Kamoun, 2014). The incidence has been cited as being as high as 1 in 3000 births in Norway (Lekarev, 2011). In the United States, the prevalence was 1 per 10,000 to 20,000 pregnancies (Schneiderman, 2017). There is an increased incidence of concurrent Hashimoto thyroiditis, primary ovarian insufficiency, type 1 diabetes, and Graves disease. These polyglandular autoimmune syndromes also include pernicious anemia, vitiligo, alopecia, nontropical sprue, and myasthenia gravis.

Untreated adrenal hypofunction frequently causes infertility, but with replacement therapy, ovulation is restored. If untreated, symptoms often include weakness, fatigue, nausea and vomiting, and weight loss. Because serum cortisol levels are increased during pregnancy, the finding of a low value should prompt a cosyntropin test to document the lack of response to infused corticotropin (Salvatori, 2005).

In a large Swedish cohort study, 1188 women with Addison disease were compared with more than 11,000 age-matched controls who delivered between 1973 and 2006 (Björnsdottir, 2010). Women diagnosed with adrenal insufficiency within 3 years of delivery were significantly more likely to deliver preterm, to deliver a low-birthweight newborn, and to undergo cesarean delivery. Others have reported similar adverse outcomes (Quartermaine, 2017). Most pregnant women with Addison disease are already taking glucocorticoid and mineralocorticoid replacement drugs. These should be continued and women observed for evidence of either inadequate or excessive corticosteroid replacement (Lebbe, 2013). During labor, delivery, and postpartum, or after a surgical procedure, corticosteroid replacement must be increased appreciably to approximate the normal adrenal response—the so-called stress dose. Hydrocortisone, 100 mg, is usually given intravenously every 8 hours for 48 hours. It is important that shock from causes other than adrenocortical insufficiency—for example, hemorrhage or sepsis—be recognized and treated promptly.

Primary Aldosteronism

Hyperaldosteronism is caused by an adrenal adenoma—Conn syndrome—in approximately 75 percent of cases. Idiopathic bilateral adrenal hyperplasia makes up the remainder, except for rare cases of adrenal carcinoma (Abdelmannan, 2011; Eschler, 2015). Findings include hypertension, hypokalemia, and muscle weakness. High serum or urine levels of aldosterone confirm the diagnosis.

In normal pregnancy, as discussed in Chapter 4 (Musculoskeletal System), progesterone blocks aldosterone action, and thus there are very high aldosterone levels (Appendix, Serum and Blood Constituents). Accordingly, the diagnosis of hyperaldosteronism during pregnancy can be difficult. Since renin levels are suppressed in pregnant women with hyperaldosteronism, a plasma aldosterone-to-renin activity ratio may be helpful for diagnosis (Kamoun, 2014). Hypertension worsens as pregnancy progresses, and medical management includes potassium supplementation and antihypertensive therapy. In many cases, hypertension responds to spironolactone, but β-blockers or calcium-channel blockers may be preferred because of the potential fetal antiandrogenic effects of spironolactone. Mascetti and coworkers (2011) reported successful use of amiloride in a pregnant woman. Use of eplerenone, a selective aldosterone-receptor antagonist, has also been reported (Cabassi, 2012). Laparoscopic tumor resection is curative (Eschler, 2015; Miller, 2012).

The pituitary enlarges impressively during pregnancy, predominately from lactotrophic cellular hyperplasia induced by estrogen stimulation (Chap. 4, Gastrointestinal Tract). Several pituitary disorders can also complicate pregnancy.

Prolactinomas

These adenomas are found often in nonpregnant women since the advent of widely available serum prolactin assays. Serum levels <25 pg/mL are considered normal in nonpregnant women (Motivala, 2011). Adenoma symptoms and findings include amenorrhea, galactorrhea, and hyperprolactinemia. Tumors are classified arbitrarily by their size measured by CT or MR imaging. A microadenoma is ≤10 mm, and a macroadenoma is >10 mm. Treatment for microadenomas is usually with bromocriptine, a dopamine agonist and powerful prolactin inhibitor, which frequently restores ovulation. For suprasellar macroadenomas, most recommend surgical resection before pregnancy is attempted (Araujo, 2015).

In a pooled analysis of more than 750 pregnant women with prolactinomas, only 2.4 percent with microadenomas developed symptomatic enlargement during pregnancy (Molitch, 2015). Symptomatic enlargement of macroadenomas, however, is more frequent and was found in 21 percent of 238 pregnant women. Schlechte (2007) also reported that 15 to 35 percent of suprasellar macroadenomas have tumor enlargement that causes visual disturbances, headaches, and diabetes insipidus. Nonfunctioning adenomas can also cause symptoms of pituitary expansion in pregnancy (Lambert, 2017).

Pregnant women with microadenomas should be queried regularly for headaches and visual symptoms. Those with macroadenomas are followed more closely and have visual field testing during each trimester. CT or MR imaging is recommended only if symptoms develop (Fig. 58-7). Serial serum prolactin levels serve little use because of normal rises during pregnancy (Appendix, Serum and Blood Constituents). Symptomatic tumor enlargement should be treated immediately with a dopamine antagonist. The safety of bromocriptine in pregnancy is well established. The safety profile is less well known for cabergoline, which is increasingly used in nonpregnant women because it is better tolerated and more effective. Cabergoline is generally considered safe for use in pregnancy (Araujo, 2015; Auriemma, 2013). Lebbe and colleagues (2010) described 100 pregnancies exposed to cabergoline and found no adverse effects. Similar findings were reported in 85 exposed Japanese pregnant women (Ono, 2010). Surgery is recommended for women with no response.

Acromegaly

This is caused by excessive growth hormone, usually from an acidophilic or a chromophobic pituitary adenoma. In normal pregnancy, pituitary growth hormone levels decrease as placental epitopes are secreted. Diagnosis is confirmed by elevated IGF-1 serum levels (Katznelson, 2014). Fewer than 100 cases of acromegaly have been reported during pregnancy (Cheng, 2012; Dias, 2013; Motivala, 2011). Pregnancy is probably rare in women with acromegaly because half are hyperprolactinemic and anovulatory. During pregnancy, affected women are at marginally greater risk for gestational diabetes and hypertension (Caron, 2010; Dias, 2013).

Management is similar to that for prolactinomas, with close monitoring for symptoms of tumor enlargement. Dopamine agonist therapy is less effective than for prolactinomas. And, transsphenoidal resection, generally considered first-line treatment outside of pregnancy, may be necessary for symptomatic tumor enlargement during pregnancy (Motivala, 2011). Guven and associates (2006) reported a case of pituitary apoplexy necessitating emergent transsphenoidal adenoma resection and cesarean delivery at 34 weeks. Successful treatment of pregnant women with the somatostatin-receptor ligand octreotide and with the GH analogue pegvisomant has also been reported (Dias, 2013; Fleseriu, 2015).

Diabetes Insipidus

The vasopressin deficiency evident in diabetes insipidus is usually due to agenesis or destruction of the neurophypophysis (Robertson, 2015). True diabetes insipidus is a rare complication of pregnancy.

Diabetes insipidus therapy is intranasal administration of a synthetic analogue of vasopressin, desmopressin, which is 1-deamino- 8-d-arginine vasopressin (DDAVP). Ray (1998) reviewed 53 cases in which DDAVP was used during pregnancy with no adverse sequelae. Most women require increased doses during pregnancy because of an increased metabolic clearance rate stimulated by placental vasopressinase (Lindheimer, 1994). By this same mechanism, subclinical diabetes insipidus may become symptomatic or cases of transient diabetes insipidus may be encountered during pregnancy (Bellastella, 2012; Robertson, 2015). The prevalence of vasopressinase-induced diabetes insipidus is estimated at 2 to 4 per 100,000 pregnancies (Wallia, 2013).

In our experiences, as described in Chapter 55 (Acute Viral Hepatitis), transient secondary diabetes insipidus is more likely encountered with acute fatty liver of pregnancy (Nelson, 2013). This probably is due to altered vasopressinase clearance because of hepatic dysfunction.

Sheehan Syndrome

Sheehan (1937) reported that pituitary ischemia and necrosis associated with obstetrical blood loss could result in hypopituitarism. With modern methods of hemorrhagic shock treatment, Sheehan syndrome is now seldom encountered (Feinberg, 2005; Pappachan, 2015; Robalo, 2012). Affected women may have persistent hypotension, tachycardia, hypoglycemia, and lactation failure. Because deficiencies of some or all pituitary-responsive hormones may develop after the initial insult, Sheehan syndrome can be heterogenous and may not be identified for years (Tessnow, 2010). In one cohort study of 60 women from Costa Rica with Sheehan syndrome, the average time to diagnosis was 13 years (Gei-Guardia, 2011). Because adrenal insufficiency is the most life-threatening complication, adrenal function should be immediately assessed in any woman suspected of having Sheehan syndrome. After glucocorticoid replacement, subsequent analyses and replacement of thyroid, gonadal, and growth hormones is considered.

Lymphocytic Hypophysitis

This rare autoimmune pituitary disorder is characterized by massive infiltration by lymphocytes and plasma cells with parenchymal destruction of the gland. Many cases are temporally linked to pregnancy (Foyouzi, 2011; Honegger, 2015; Melmed, 2015). There are varying degrees of hypopituitarism or symptoms of mass effect, including headaches and visual field defects. A sellar mass is seen with CT or MR imaging. A mass accompanied by a modestly elevated serum prolactin level—usually <100 pg/mL—suggests lymphocytic hypophysitis. In contrast, levels >200 pg/mL are encountered with a prolactinoma. The etiology is unknown, but nearly 30 percent have a history of coexisting autoimmune diseases including Hashimoto thyroiditis, Addison disease, type 1 diabetes, or pernicious anemia. Treatment is with glucocorticoids and pituitary hormone replacement. The disease may be self-limited, and a careful withdrawal of hormone replacement is attempted after inflammation subsides (Foyouzi, 2011; Melmed, 2015).