Antithyroid Drug Treatment in Pregnancy

Abstract

Antithyroid (thionamide) drugs, discovered after the work of the Mackenzies (1) and introduced into clinical medicine by Astwood and colleagues, have been in widespread use for the treatment of thyrotoxicosis for nearly 70 yr (2). The drugs methimazole (MMI), carbimazole (CBZ), the prodrug of MMI, and propylthiouracil (PTU) have been administered to pregnant women during this time; for example, an early study reported outcomes in 19 patients (22 pregnancies) studied over 6 yr (3). However, it was not until 1972 that the first reports of scalp defects in babies whose mothers had received MMI during pregnancy (4) were recorded. Since then, a number of congenital anomalies including aplasia cutis congenita, choanal atresia, and omphalomesenteric duct anomalies have been ascribed to antenatal exposure to MMI/CBZ, but less so or not at all to PTU. The term MMI embryopathy has been employed as an umbrella description of this group of anomalies (5). The incidence of these malformations is very low (prevalence of aplasia cutis in babies not exposed to teratogens is 0.03%, and that of patent vitellointestinal duct is 0.0053%), so that studies of association with a particular drug require a very large sample size as well as careful design to confirm the association with antithyroid drugs. In this issue of the JCEM, Yoshihara et al. (6) investigated whether in utero exposure to antithyroid drugs led to a higher rate of major malformations compared with infants born to a control group of women. The study was a retrospective review of the pregnancy outcomes of 6744 pregnant women with Graves’ disease in relation to all observed congenital anomalies. Of these, 1426 had received MMI alone, 1578 received PTU alone, and 2065 served as the control group because they had received no medication (total, 5069). Another 1675 women had received multiple therapy and were excluded. Approximately 11.5% of the women did not have live births, consistent with a recent analysis of miscarriage rates in the general population (7). The results showed a significantly higher rate of major anomalies in the MMI group of babies (4.1%) compared with those seen in the PTU group (2.1%) (P 0.002). Furthermore, the rate of abnormalities observed in the PTU group was not significantly different from the 1.9% found in the control group (P 0.709). The odds ratio (95% confidence interval) for the number of malformed infants in the MMI group was 2.28 (1.54–3.33), and that for the PTU group was 0.66 (0.41– 1.03). The authors concluded that MMI exposure in the first trimester increases the risk of congenital anomalies. However, inspection of the detailed results seems to show that only the previously noted MMI-associated abnormalities (omphalocele, omphalomesenteric duct anomalies, and aplasia cutis congenita) were significantly increased compared with the PTU group or the controls. Based on this study, 1.62% of infants developed these three anomalies when exposed to MMI, compared with 0% in those exposed to PTU or controls. The question that should be asked is whether the data of Yoshihara et al. (6) are consistent with previous studies in this area. In a prospective study of 241 women exposed to MMI, no increase in the overall rate of abnormalities was found in the MMI group compared with controls. However, two of eight infants with abnormalities in the MMI group had MMI embryopathy (8). Clementi et al. (9) conducted a case-affected control analysis in which 18,131 cases with malformations with first-trimester exposure to medication were noted. Of these, 127 were born to mothers with known antithyroid drug treatment during the first trimester. Prenatal exposure to CBZ/MMI was significantly (P 0.01) associated with choanal atresia, omphalocele, and situs inversus, whereas PTU exposure was linked to a small number of cases of cardiac and renal