Neonatal Mortality for Low‐Risk Women by Method of Delivery

Abstract

Reply: We would like to thank Drs. Pettker and Funai, and Drs. Källén and Olausson for their thoughtful reading and comments on our article. We agree that, as with all studies, there were limitations to the data used for the analysis. These limitations were carefully described and the discussion section was appropriately cautious in its interpretations (see in particular the last 4 paragraphs) 1. We acknowledged the fact that, although we excluded infants with medical risks and complications to the maximum extent possible, it was still possible that some infants with medical risks remained in the study. This fact was why we were careful to use the terminology “no indicated risk” throughout the study. In particular, we never (in contrast to Drs. Pettker and Funai’s claim) referred to the study population as representing “elective” or “cesarean delivery on maternal request” (CDMR). Although underreporting of medical risk factors and complications of labor and/or delivery on the birth certificate has been documented, there may be better ascertainment of these risk factors and complications among women who undergo a major surgical procedure (such as cesarean section) than among women not undergoing surgery 2. In addition, the birth certificate employs a checkbox format where a “none” checkbox is checked to indicate that the women had none of the 16 listed medical risk factors or 15 listed complications of labor and/or delivery. The percent of records with unknown responses (i.e., no risk factors or complications reported and none box not checked) was extremely low (0.6% for medical risk factors and 0.9% for complications in 2001), and these cases were excluded from the analysis. Thus, it seems unlikely that the study population represents events with less complete reporting. Drs. Källén and Olausson note “the presence of 3,222 infants with Apgar scores < 4 . . . without any diagnosis of asphyxia,” and use this statement to demonstrate “substantial underreporting of medical diagnoses.” We are unsure where this mistaken impression originated. Of the 3,222 infants with Apgar scores < 4, approximately 13 percent were neonatal deaths and many of these had asphyxia (P20-P21) coded as the cause of death. We used Apgar score < 4 rather than a cause-of-death diagnosis of asphyxia to exclude cases from the multivariate models because the Apgar score identified asphyxia in both survivors and neonatal deaths, whereas the cause-of-death diagnosis of asphyxia only identified asphyxia in neonatal deaths. We understand Drs. Källén and Olausson’s suggestion to exclude intrapartum complications from the cesarean, but not the vaginal group, but think that at this point it is preferable to use the same exclusions for both the cesarean and the vaginal groups in order to create the most similar and comparable groups for analysis. We welcome Drs. Källén and Olausson’s analysis of the Swedish data, and hope that they will publish a more comprehensive study using those data, particularly when additional years of data are available to increase the study power. Currently the inclusion of only 11 neonatal deaths in the elective cesarean category yields considerable uncertainty in the resulting estimate. The health care system in Sweden is very different from that in the United States, with much greater participation of midwives in the birthing process and only 16 percent of births delivered by cesarean section in 2001, compared with 24 percent in the United States 3. Therefore, it is not surprising that an analysis of Swedish birth statistics would yield different results. We hope that our article will continue to stimulate further research and discussion into the risks and benefits of cesarean delivery for low-risk women.