To assess potential long-term effects of prenatal SSRI exposure, we administered clinical doses of fluvoxamine and fluoxetine to pregnant mice. Embryos were exposed to SSRIs between E8 and E18, when the 5-HTT is present 
. After prenatal fluoxetine exposure, we found a dose-dependent increase in neonatal mortality in the offspring, a permanent decreased 5-HTT binding and significant long-term alterations in depression- and anxiety related behavior.
In order to ensure the relevancy of this study for the human situation, we made a comparison of the fetal exposure of fluvoxamine and fluoxetine between mouse and human in terms of placental transfer. The results showed that the placental passage is similar between mouse and human, with respect to fluvoxamine and fluoxetine. Fluvoxamine displayed a relatively low placental transfer, while fluoxetine showed a relatively high placental transfer. Thus, fluoxetine-treated fetuses are exposed to higher levels compared to fluvoxamine-treated fetuses. Therefore, we hypothesized that offspring exposed to fluoxetine would have more possible long-term consequences compared to offspring exposed to fluvoxamine.
In the present study, no alterations were found in body weight between E18 and adulthood of both the fluvoxamine and fluoxetine exposed offspring, which is consistent with others who administered a therapeutic dose of SSRIs in mice 
. However, our findings conflict with a report which describes that rats treated with high doses of fluoxetine during pregnancy have smaller pups with poorer weight gain 
. In human, many studies have described lower birth weight after prenatal SSRI exposure 
. Although, Chambers et al. 
described a link between prenatal fluoxetine exposure and low birth weight that disappeared when maternal weight gain was controlled.
Surprisingly, the offspring exposed to high doses of fluoxetine showed a dramatic high level of mortality during the postnatal period, whereas mortality was low in the saline- and fluvoxamine-treated group. The cross-fostering experiments demonstrated that the effects of prenatal fluoxetine treatment on mortality were due to fetal aspects, which excludes maternal effects on the survival rate of the offspring. By histopathological examination and measuring the wall thickness/radius ratio, we observed that the majority of the fluoxetine-treated offspring died of heart failure, due to dilated cardiomyopathy. Dilated cardiomyopathy was also observed in mice lacking the serotonin 2B receptor, which is required for heart development 
. Nebigil et al. 
also showed that the serotonin 2B receptor is required for 5-HT to regulate cardiovascular functions. This suggests that the 5-HT system is involved in heart development and that fluoxetine treatment during fetal development affects the heart, resulting in dilated cardiomyopathy. Major cardiac malformations have also been associated with prenatal SSRIs in human practice, although most of the cardiac malformations were observed after prenatal paroxetine treatment 
. Fortunately, no association has been found between prenatal SSRI exposure and neonatal death in human practice. Differences in neonatal mortality between mouse and human can be explained by the neonatal care, which is excellent in humans but poor in mice. An alternative explanation is that, due to species differences in development, neonatal death in mice is similar to prenatal death in humans. Chambers et al. 
observed that mothers exposed to fluoxetine had an increase in the incidence of miscarriages. Moreover, in a meta-analysis of clinical trials it was shown that prenatal SSRIs significantly increased the risk for spontaneous abortion 
, which is also observed in mice 
. Since we found no effects of prenatal fluvoxamine exposure on the survival rate and the hearts of the offspring, fluvoxamine may be a safer SSRI during pregnancy compared to fluoxetine. Nevertheless, more research is necessary to exclude the effects of prenatal fluoxetine exposure on heart development in humans.
The observed changes in 5-HTT binding in the raphe nucleus at P20 and adulthood demonstrates that fluoxetine exposure during fetal development permanently changes the serotonin homeostasis, whereas prenatal fluvoxamine exposure did not. This data is in agreement with other reports, which showed that chronic treatment with fluoxetine downregulates the 5-HTT in adult rodents 
. However, these animals were treated with SSRIs at adulthood. To our knowledge, this is the first study which describes the effects of prenatal SSRI treatment on 5-HTT density. Since 5-HTT appeared to be a critical regulator of emotional function, we have investigated whether these 5-HTT alterations resulted in changes of behavior. All behavioral experiments showed that fluoxetine exposed mice demonstrated depression- and anxiety-related behavior, whereas fluvoxamine exposed offspring showed no changes in behavior at adulthood compared to saline-treated offspring. These results are comparable to the behavioral data of the 5-HTT null mutant mice, which demonstrated a range of behavioral and neurophysiological abnormalities that resemble symptoms of mood and anxiety disorders 
. To date, there have been insufficient long-term follow-up studies in human to demonstrate effects of prenatal SSRI on the risk of developing affective disorders in the offspring. Further investigation of the long-term consequences of fetal exposure to SSRIs, as well as the mechanisms involved, are required for a better understanding of the impact of SSRIs on development of the offspring.
Decreasing the dose of fluoxetine in a dose-response experiment, with therapeutic doses ranging from 0.3 till 0.8 mg/kg/day, resulted in higher survival rates and less dramatic effects on the long-term behavior in the offspring. Interestingly, the effects observed after administration of 0.8 mg/kg/day were less severe when a lower dose (0.6 mg/kg/day) was used, and completely diminished at the lowest dose administered (0.3 mg/kg/day). Noteworthy, the sub-maximal clinical dose of fluoxetine (0.6 mg/kg/day) still resulted in heart failure and behavioral pathology in the offspring. These results indicate that the effects of prenatal fluoxetine treatment on the fetus are dose-dependent.
The present findings demonstrate that prenatal fluoxetine treatment has dramatic effects on the survival rate of the offspring, alter the 5-HT system homeostasis at adulthood and indicate that fluoxetine exposed mice are more vulnerable to anxiety disorders at adulthood. Ultimately, clinical studies will be required to determine whether our findings have applicability to the risks for anxiety or affective disorders in humans. Two recent papers on this subject argue that it is safe to use antidepressants during pregnancy 
. Interestingly, the selection of patients was very restricted. Only pregnant women were included that used SSRIs in the first trimester. Although this is fine for general teratogenic effects of SSRIs, it is not appropriate to evaluate the effects of the pharmacological activity of SSRIs during development. It has been shown that the human serotonin system is developing in week 14–16 
. This would indicate that pharmacological intervention of the serotonin system by SSRIs in the infant is only effective after the first trimester. Taken together, we would argue that a clear study in humans about the relationship between SSRI intake and exposure of the infant, through placenta transfer efficiency parameters, is needed together with a broad behavioral assessment of the infant after the withdrawal effects.
Based on our findings in mice, prenatal fluvoxamine has no long-term consequences on the offspring in the clinical dose used, suggesting that fluvoxamine may be a safer antidepressant drug during pregnancy compared to fluoxetine. However, our data indicate that more detailed and specific follow-up studies in humans are required, since irreversible long-term adverse effects of SSRI treatment may be observed later in life. Our findings stress that it is important to be restrictive with prenatal fluoxetine administration, and that fluvoxamine may be the preferred SSRI during pregnancy.