The Use of Antithyroid Drugs in Pregnancy and Lactation

The management of hyperthyroidism in the pregnant woman has been the topic of several recent reviews (1–6) and case series (7, 8). These papers have discussed maternal-fetal thyroid relationships, the epidemiology of maternal hyperthyroidism, causes of gestational thyrotoxicosis, the clinical and laboratory diagnosis of hyperthyroidism in pregnancy, therapy with ATDs, surgery, β adrenergic blocking drugs, and iodides, as well as the consequences of inadvertent radioiodine therapy and the implications of breast feeding. The purpose of this discussion is to provide an in depth examination of the use of ATDs in the pregnant hyperthyroid woman, rather than to present comprehensive guidelines for the treatment of hyperthyroidism during pregnancy furnished in other reviews (1–6). The subject continues to engender controversy because of the lack of prospective clinical trials and the relative rarity of the condition, occurring in only 1 of 1000–2000 pregnancies (7, 8). With more women receiving radioiodine before conception, the frequency may be even lower in the future. When one discusses the risks and benefits of ATD therapy in the management of the hyperthyroid pregnant women, the risks to the mother and fetus of untreated hyperthyroidism must also be considered. These risks appear to be directly related to the control or, more likely, the severity of the maternal hyperthyroidism. In a 1989 retrospective analysis of 181 hyperthyroid pregnant women (8), the subjects were separated into 3 groups: women who were euthyroid at presentation and remained that way (group 1, n = 34), women who were hyperthyroid at presentation but who were controlled during the pregnancy (group 2, n = 90), and women whose hyperthyroidism was not controlled (group 3, n = 57). The odds ratio (OR) for low birth weight (<2500 g) infants was not elevated in group 1 compared with normal control women. Group 2 women had an OR for low birth weight infants of 2.4 [95% confidence interval (CI) 1.4–4.1], and the OR was 9.2 (95% CI 5.5–16) in group 3 women. Similarly, prematurity was more common in groups 2 and 3 [OR in group 2, 2.8 (CI 0.33–23.5); OR in group 3, 16.5 (CI 2.1–130)] compared with group 1. Eclampsia was also more common in group 3 women (OR 4.7, CI 1.1–19.7). In this study there was no increase in the frequency of infants who were small for gestational age, although Mitsuda et al. (9) found a higher frequency of small-for-gestational-age infants in women who were thyrotoxic for more than 30 weeks during the pregnancy (26.7% vs. 7.7% among women who were euthyroid throughout pregnancy). Davis, et al. (7) found that 50% (4 of 8) of the infants born to untreated hyperthyroid women were stillborn vs. no stillbirths in the 36 treated euthyroid women. Also, 62% (5 of 8) of untreated hyperthyroid women had congestive heart failure vs. 3% (1 of 36) of treated euthyroid women. Interestingly, another study found that the altered hemodynamics seen in pregnant hyperthyroid women with congestive heart failure improved but did not normalize with restoration of a euthyroid state (10). This may be one explanation for the development of preeclampsia even in women whose hyperthyroidism is controlled during pregnancy (8). It is uncertain whether untreated Graves’ disease is associated with a higher frequency of congenital abnormalities. Mitsuda et al. (9) described congenital abnormalities in 6 of 230 (2.6%) infants born to mothers with Graves’ disease (3 with ventricular septal defect, 1 with cleft lip and palate, 1 with polydactyly, and 1 with diastasis recti abdominis). This rate was similar to the malformation rate in normal pregnancies, but it is not specified whether the mothers’ hyperthyroidism was controlled or not. In another report from Japan examining the relationship of fetal malformations to thyroid status in the first trimester (11), the rate of fetal malformations was 6% (3 of 50; imperforate anus, anencephaly, and cleft lip: 1 each) in untreated hyperthyroid women, 1.7% (2 of 117; ear lobe malformation and omphalocele) in treated hyperthyroid women, and 0% (0 of 126) in treated euthyroid women. In a report from the U.S. (12), the overall rate of fetal malformations was 3% (6 of 185): 3 of 99 (3%) of babies were born to women who were hyperthyroid during the first trimester (pulmonic stenosis, ventricular septal defect, patent ductus arteriosus: 1 each), and 3 of 96 (3%) born to women who were euthyroid in the first trimester(inguinal hernia in 2, hydrocephalus in 1). These data suggest that there may be a small increase in the rate of congenital anomalies in women with untreated or incompletely treated hyperthyroidism in the first trimester, but more information is clearly needed. ATD therapy is obviously indicated for hyperthyroidism when it is moderate or severe. Although there is no generally accepted definition of moderate or severe hyperthyroidism, it would be reasonable to treat women whose free T4 levels are more than 2.5 ng/dL if the upper limit of normal is 2.0 ng/mL. However, for an asymptomatic woman whose pregnancy is progressing satisfactorily and whose serum free T4 is minimally above 2.5 ng/dL, careful clinical follow-up without antithyroid therapy may be appropriate. It should be noted that all prior studies linking hyperthryoidism to adverse pregnancy outcomes have reported a correlation with elevations of maternal thyroid hormone levels rather than the presence or absence of maternal symptomatology. Certainly, in milder cases, when free T4 and/or free T3 serum levels are only slightly elevated, it may be appropriate to follow the patient expectantly, given that hyperthyroidism tends to spontaneously improve over the course of the gestation (1–6), and that neonatal hypothyroidism is seen more often if the mother is euthyroid rather than mildly hyperthyroid (see below) (13). It might also be reasonable to measure free T4 by equilibrium dialysis in borderline cases, because some free T4 assays may report erroneously high free T4 levels in situations of increased T4 binding such as pregnancy (14). Throughout gestation, communication with the patient’s obstetrician is important for assessment of the health of the pregnancy, which must be correlated with maternal thyroid status for optimal ATD dose titration. Both propylthiouracil (PTU) and methimazole (MMI) have been used in pregnancy and are equally effective in the management of hyperthyroidism in this setting (12, 13, 15). In the U.S., PTU has traditionally been preferred because of MMI’s purported increased passage across the placenta and breast epithelium, and because of a reported association between MMI use and aplasia cutis as well as other rare annomalies (see below). However, MMI or its precursor carbimazole are used widely throughout the world to treat pregnant women. Transplacental passage.In the only in vivo human study to formally examine the possible placental transfer PTU and MMI, 9 pregnant women ingested 35S-labeled compounds 2 h before the elective termination of pregnancies that ranged in gestational age from 8 to 20 weeks (16). Only seven of the nine women had complete data. For MMI (n = 2) or carbimazole (n = 3), the ratio of fetal serum or cord blood MMI levels to maternal drug levels ranged from 0.72 to 1, indicating a high rate of drug transfer. For PTU, the ratio ranged from 0.27 to 0.35 (n = 2), indicating a substantially lower transfer rate. These investigators also obtained drug transfer data from pregnant rats that supported these observations. The differences in placental passage were mainly attributed to the known disparity in drug binding to albumin (PTU≫MMI) and in lipid solubility, as well as possible differences in maternal/fetal volumes of distribution, excretion, and metabolism of each compound (16). More recently, a study using isolated perfused human placentae found no difference in the rate or extent of transplacental passage between PTU and MMI (17). Even though PTU is highly protein bound[ 61% to human serum albumin (17)], transfer rates across the placenta were independent of the perfusate protein concentration, suggesting that there is highly efficient placental extraction of unbound drug. Failure to achieve a steady-state with a single dose of drug was offered as the likely explanation for the differences that were observed between this study and the earlier report by Marchant et al. (16). Although it is uncertain whether this in vitro model is completely representative of in vivo events, a lack of a difference in placental transfer of PTU and MMI is consistent with clinical observations showing similar fetal outcomes with either drug, in terms of thyroid function and congenital anomalies (12, 15), and with data showing that cord blood PTU levels were similar to or higher than simultaneously obtained maternal serum PTU levels (18). Possible drug-related congenital anomalies.Aplasia cutis is a congenital localized absence of skin, typically present as a 0.5–3 cm isolated, punched out, midline lesion at the vertex or occipital area of the scalp (19). The lesions may close spontaneously or skin grafting may be required in more severe cases. It occurs spontaneously in approximately 1 in 2000 births (20), is often familial, and can occur alone or in association with other anomalies. In 1972, Milham and Elledge (21) reported that of 12 cases of aplasia cutis that occurred in Washington state in a 6-month period, 2 mothers had taken MMI. Subsequently, additional cases were reported in association with MMI (20, 22). Although Van Dijke et al. presented circumstantial evidence that there may be no association between MMI use and aplasia cutis (20), and Momotani et al. (11) found no cases among the offspring of 243 MMI-treated mothers, cases continue to be reported (23–28). Furthermore, Martinez-Frias et al. (29) reported a 3-fold increase in the frequency of this anomaly in regions of Spain where there was illegal use of MMI in animal feed as a fattening agent, providing epidemiological evidence for an association. Perhaps of greater concern than aplasia cutis are recent descriptions of a MMI embryopathy (8 cases), which includes aplasia cutis, choanal atresia, tracheal-esophageal fistulae, hypoplastic nipples, facial anomalies, and psychomotor delay (reviewed in Refs. 30, 31). When considering the significance of all of these reports, it must be emphasized that there are no case reports of aplasia cutis or other congenital anomalies in association with PTU exposure. This is particularly significant because PTU is the preferred drug therapy of hyperthyroidism in pregnancy in the U.S. The data suggesting that PTU crosses the placenta more easily than previously thought would seem to remove PTU’s advantage over MMI in pregnancy. On the other hand, the rare but probable real association between MMI usage and fetal anomalies makes MMI a less attractive first line alternative. All things being equal, PTU is probably still the first choice for the treatment of hyperthyroidism in pregnancy, although MMI remains a reasonable second-line agent in the case of an allergic reaction, intolerance, or poor response to PTU. When the diagnosis of moderate to severe hyperthyroidism is made, therapy with 100 mg PTU every 8 h should be initiated. Lower doses should be used in milder cases, because baseline thyroid function is a major factor in the rate of drug response. If the patient has been taking MMI before she becomes pregnant, switching to PTU would be reasonable before a planned conception or if the pregnancy is diagnosed in the first trimester during the period of organogenesis. The patient’s thyroid function should be monitored monthly, and the drug dose adjusted to maintain the serum free T4 at the very upper limit of the normal range or slightly above normal. Indeed, the ATD dose should probably be tapered if the TSH rises into the normal range. On the other hand, if the severe hyperthyroidism persists, there should be no hesitation to increase the dose in an effort to control the hyperthyroidism to the appropriate pregnancy target. Doses in the range of 600–800 mg/day are needed occasionally, possibly related to poor compliance (32) or altered PTU pharmacodynamics in pregnancy; in one study, the pharmacokinetics of PTU in pregnancy appeared to be normal (33), whereas others suggested that peak serum PTU levels and area under the curve were lower during pregnancy compared with levels in the same women when they were postpartum (18) or compared with nonpregnant controls (34). Limited data for MMI in pregnancy also suggest that drug clearance may be more rapid than in nonpregnant women (35). Despite these caveats, Wing et al. (12) found that the median maximal and minimum doses of PTU (450 mg and 100 mg daily, respectively) and MMI (40 mg and 10 mg daily, respectively), were similar to what might be required in nonpregnant patients. However, if the requirement for high doses of PTU or MMI persists during pregnancy, thyroidectomy may be warranted (1–6). As thyroid function improves, the dose of PTU may be gradually tapered, and when it is in the range of 50–100 mg daily, it is often possible to discontinue the drug altogether. As alluded to above, and as proposed by Daniels (2) low doses of ATDs should be discontinued if serum TSH levels become normal (2), given the relationship between maternal and fetal thyroid function noted above. In one retrospective report of 30 PTU-treated women, 10 were able to discontinue it before the end of pregnancy (36). It is clear that overzealous treatment of maternal hyperthyroidism can lead to fetal hypothyroidism (37) and goiter (38, 39), although most cases of hypothyroidism are really transient episodes of subclinical hypothyroidism, with mild hyperthyrotropinemia and normal serum free T4 and T3 levels (13, 15, 40). Momotani et al. (13, 15) have reported the most extensive data documenting the relationship between maternal and neonatal thyroid function. In one report (13), no cases of neonatal hypothyroxinemia or elevated serum TSH were seen in the offspring of treated women whose serum free T4 levels remained elevated. There was a strong relationship between maternal and fetal thyroid function, illustrated by rates of transient neonatal hypothyroxinemia of 10, 36, and 100% when the mother’s free T4 was maintained in the upper one third of the normal range, the lower two thirds of the normal range, or the normal range, there were no differences in cord blood free T4 and TSH in infants to PTU (n = vs. MMI (n = and there was no relationship between the dose of either ATD and neonatal thyroid function. This and the close relationship between maternal and fetal thyroid function might be by the that fetal thyroid function may be under the of maternal thyroid as suggested by et al. and by Momotani et al. (13, 15). This may also provide a for using higher ATD doses to control maternal thyroid function, if rather than to control the disease for of fetal On the other hand, data from et al. suggest that even mild maternal hypothyroidism may have an the of the which is with However, it not appear that of the women in the study by et al. were because of ATD usage during pregnancy. data from infants to either PTU or MMI in have not differences in when compared with controls or but maternal thyroid function during the pregnancy is not reported in these it is to whether of the were born to mothers who were during there a for and ATD therapy to maternal hypothyroidism or fetal of maternal ATD In a of the the frequency of neonatal goiter was lower when women therapy or than when they ATDs alone or In another report there were no differences in cord blood or TSH in born to mothers (n = who had been treated with PTU and thyroid hormone PTU dose or PTU alone (n = PTU dose This small study that therapy not neonatal but also that PTU doses are not higher when thyroid hormone therapy is ATD and therapy in pregnancy is because it lead to inadvertent of ATDs However, this might if it be that maternal hypothyroidism, a strong of fetal hypothyroxinemia (13), be In one report for of women treated with ATDs alone at some during pregnancy in in another the frequency of hypothyroidism by a low free T4 at was (13). In women with either hyperthyroidism or such mild disease that ATDs be to ATD therapy would not be However, in women who drug doses above the minimum more than mg/day for PTU or more than mg/day for the of to hypothyroidism has some in the first However, only prospective studies with follow data would be able to this The of the of during ATD therapy in several clinical The of Graves’ disease pregnancy, with in the third trimester and or of ATD only to be by of the disease in the postpartum period with of ATD women with Graves’ disease may ATD therapy and would to may have previously ATD therapy in pregnancy and are but an of hyperthyroidism ATD they to continue In Graves’ disease may first occur in the postpartum period, either as the of hyperthyroidism or postpartum if such women are they often to continue even if ATD therapy is In the women taking ATD therapy have been because of the that the of these in breast were to the This from studies of one of the first compounds used for treatment of In et al. reported that the in breast was 3 higher than that in serum 2 h of 1 to two women. Although treatment of hyperthyroidism was discontinued for other no studies of the used ATD were 20 In et al. a with the in of a documenting that PTU was not in breast women without thyroid disease and two with Graves’ treated with were given mg PTU, and the serum and PTU were for The of PTU in breast over that period was of the with the being only approximately of simultaneously serum in with an earlier study these the that a mother treated with mg PTU a would of PTU a to which for a is to a dose of 3 mg for a studies the transfer of MMI or into have that a higher of an dose of these to the mother in the The to serum ratio of MMI is with a in breast over an period to of an dose This is to seven higher than that for PTU. a single MMI dose to the an may which for a is to mg for a to the higher of MMI, this dose the However, in these studies were by of a single dose of MMI to women who did not have thyroid disease were not in a very recent study serum MMI levels in infants who were by thyrotoxic mothers receiving mg of MMI MMI levels in the babies were less than 2 h the mother ingested the dose the range a higher passage of MMI or vs. PTU into breast be from the of the MMI is minimally to serum whereas PTU is more in to albumin In MMI is not in serum whereas PTU, a is more in serum than in the more breast This would its transfer from serum into the breast the data in the several that if a woman to PTU rather than MMI or should be However, these studies passage of ATD in breast but did not the of during ATD As noted by one is important from a as well as a it important to study thyroid function of infants during when mothers are given antithyroid treatment for the 2 several such studies have been which infants of thyrotoxic mothers taking PTU, MMI, or Refs. The of these reports with some a of thyroid function of the and a more In other independent of maternal ATD may thyroid function of the The most important of these is the transplacental passage of maternal that can the thyroid in the weeks of of and development of infants of thyrotoxic mothers treated with ATD function This study has been to a group of of and development of infants of thyrotoxic mothers treated with ATD function This study has been to a group of There are several that can be from a of the studies in 1. maternal use of either MMI to 20 or PTU to during not the thyroid hormone levels The only reported infants who had high serum TSH or low serum T4 levels were during the first of and were born to mothers who had been treated with PTU during gestation and therapy postpartum In at two of these infants had even more of thyroid function, as by lower levels of T4 and higher TSH in cord However, thyroid function to normal the first the of maternal ATD during this transient of neonatal thyroid function most likely the required for clearance of PTU transplacental rather than by breast In even if the mother becomes because of ATD the thyroid function of the remains normal as by et al. 57). These reported women taking MMI mg daily, 1 20 mg daily, who elevated serum TSH levels and serum T4 levels 1 the serum TSH and T4 levels were normal. the and development of whose mothers MMI has been although no such reports have been for PTU. the and of and et al. and and in of thyrotoxic mothers and controls between the of The and did not between the two groups Although has these studies to a group of whose mothers MMI during and found no difference in or This still a small that may be to However, given the normal thyroid function in these it reasonable to that thyroid hormone levels were for normal allergic associated with ATD such as and have not been reported in infants who mothers were treated with However, because most of these adverse are rare and the of reported is only approximately a remains a In maternal ATD use during to be whether it is gestation or in the postpartum For MMI, doses of to 20 mg have been not to thyroid function 57). For PTU, the of reported infants is and thyroid function was in only infants whose mothers were taking high doses mg daily, 1 and mg daily, 2 to be PTU doses used during should be mg or mother should ATD dose which should provide a h interval before she Although maternal thyroid hormone levels must be monitored with appropriate ATD it that the thyroid function not to be as as and development are progressing