Placental transfer of lithium has been studied by Newport and colleagues (2005). This work combined the results of a prospective sample of ten women with 32 cases of neonatal dosages identified in the literature. Their results suggest that placental transfer of lithium is complete and that the balance between maternal and foetal circulation (average child/mother plasma lithium ratio = 1.05) is as between the different fluid compartments of an individual. Another recent study (van der Lugt et al. 2012) found neonatal blood lithium levels above 0.8 mEq/L in 2 of 30 examined children, 1 of whom presented signs of neonatal distress. This work does not provide the blood levels of the other children included in the sample and does not take into account co-prescriptions of other psychotropic drugs.

The first studies on lithium teratogenicity are animal studies dating from the end of nineteenth century (Giles and Bannigan 2006). In animals, with doses used for treatment in human beings, the studies do not show an increase in teratogenic risk. With very high doses, some studies reveal different types of malformations (central nervous system, skeletal, craniofacial, etc.), while others did not find any link between prenatal exposure to lithium and birth defects (Giles and Bannigan 2006). The first data on lithium’s teratogenicity in humans come from the Register of Lithium Babies, which reported, retrospectively, the pregnancy outcomes of patients treated with lithium in Denmark, the United States and Canada. The first work reported 118 cases of Danish children whose mothers had taken lithium during the first trimester of pregnancy (Schou et al. 1973). Nine of them (7.6 %) presented with congenital malformations, six of which related to the cardiovascular system, including one Ebstein anomaly (severe malformation of the tricuspid valve). By adding data from the United States and Canada, the number of cases reported increased to 143. In this second sample, cardiovascular malformations accounted for 77 % of congenital defects, against 12.5 % in the general population. Ebstein’s anomaly was largely overrepresented: 40 % of babies from the registry against 1.25 % of cardiac malformations in the general population. The final publication (Weinstein and Goldfield 1975) included 225 observations (at least three children were also exposed to other treatments), among which 11 % (n = 25) of the children presented congenital malformations (against 2 % in the general population). Three quarters of these malformations were cardiovascular (n = 18) and 33 % (n = 6) were Ebstein anomalies. The retrospective collection of data in this register represents a major bias, since it is likely that pathological cases have been more frequently reported than normal births. Subsequent retrospective studies have used more strict data collection methods and by and large did not reveal any statistically significant relationship between the use of lithium during pregnancy and the occurrence of cardiac malformations in the newborn, nor links between Ebstein’s anomaly and prenatal exposure to lithium (Nora et al. 1974; Kozma 2005). In addition, a review of 59 cases of Ebstein diseases found no cases of children exposed to lithium in early pregnancy (Zalzstein et al. 1990). There are a few prospective studies (Jacobson et al. 1992; Bogen et al. 2012), some unpublished (Gentile 2012), that do not show a statistically significant link with birth defects, although some note the existence of sporadic cases of Ebstein’s anomaly (Jacobson et al. 1992; Gentile 2012). Case-control studies have also failed to demonstrate a significant association between birth defects and in utero exposure to lithium (Zalzstein et al. 1990; Gentile 2012). Case reports in the literature indicate both heart defects, sometimes with Ebstein’s anomaly, and other types of defects, including neural tube defects (Gentile 2012).