Sleep-Related Eating Disorder Associated With Mirtazapine

et al, our associated sample size was not sufficient to enable adequate statistical power to justify their investigation. We do, however, appreciate the significance of including data on infant mortality to more completely describe the effects of prenatal SSRI exposure on extreme clinical manifestations and the infant life prognosis. In accordance with this, we have included data on stillbirths and neonatal death later. Our original analysis included the investigation of neonatal outcomes among live-born singletons. After including data on infants previously excluded from the study (ie, fetal deaths), there was only 1 (0.61%; 6.1 per 1000 births) stillbirth in the group of women who received a dispensing for an SSRI, 14 (0.89%; 8.9 per 1000 births) stillbirths in the group of women who had a documented psychiatric illness but did not receive a dispensing for an SSRI, and 198 (0.45%; 4.5 per 1000 births) stillbirths in the group of women who did not have a psychiatric illness and did not receive a dispensing for an SSRI. With recognized limitations of sample size set aside, these differences were not statistically significant. From the available data, we can confirm that there were 2 (0.90%; 9.0 per 1000 births) neonatal deaths in the group of women who received a dispensing for an SSRI, 5 (0.3%; 3.2 per 1000 births) neonatal deaths in the group of women who had a documented psychiatric illness but did not receive a dispensing for an SSRI, and 96 (0.30%; 3.0 per 1000 births) neonatal deaths in the group of women who did not have a psychiatric illness and did not receive a dispensing for an SSRI. Again, with recognized limitations of sample size set aside, these differences were not statistically significant. Furthermore, no data were available to us in relation to the cause of death. Given the relatively small number of outcomes, we feel that these results should be interpreted with caution, making it difficult to draw accurate comparisons with previous studies, such as that published by Stephansson et al. An important note is that our overall rates of fatal outcomes in our cohort are higher than what was identified by Stephansson et al. This could be a manifestation of our cohort being derived from a single specialist tertiary level teaching hospital that is likely to attract high-risk pregnancies, as opposed to the populationbased approach undertaken by Stephansson et al. We also acknowledge the limitation of not having data available to assess the severity of underlying maternal illness. Further studies with adequate sample size are required to clarify these findings, and we look forward to this evolving literature.