Drs. van Erp and Cannon Reply

To the Editor: In our article, we presented the finding of smaller hippocampal volumes among schizophrenic probands who experienced fetal hypoxia than those who did not, a difference that was not noted in their full siblings and/or in unrelated healthy volunteers. We interpreted this finding as potentially reflecting a genotype-environment interaction. Dr. Preti suggests that the result may actually reflect the significantly higher risk of unfavorable pregnancy outcome among schizophrenic patients’ mothers, which in turn may be under genetic control (a genotype-environment covariation), or more generally, a greater degree of hypoxic complications in the patients.

We interpreted our data as consistent with a genotype-environment interaction rather than a genotype-environment covariation because we did not find evidence of a greater number of hypoxia-associated obstetric complications in the probands or their full siblings in relation to the comparison subjects in this study group (1, 2), which would have been predicted in the case of genotype-environment covariation. We did, however, observe more hypoxia-related complications among the patients with an early age at onset.

Measures of fetal hypoxia based on standard records collected at the time of pregnancy and delivery, as used in our study, are likely to be less biased than measures taken from maternal interview or other retrospective scores. Nevertheless, the records-based measure of hypoxia used in this study is dichotomous, coding the presence or absence of hypoxia-associated obstetric complications. Ideally, fetal hypoxia would be assessed with direct quantitative measures, such as fetal blood oxygenation level.

We thus accept the possibility suggested by Dr. Preti that the magnitude of the hypoxic obstetric complications in the patients may have been larger than the magnitude in the siblings and comparison subjects, a possibility that awaits testing in studies with direct quantitative measures of fetal blood oxygenation.