So, as I wrote a couple of posts back, I said I would take a look at the study summarized under the Medpage Today headline “Normal Vaginal Births Cause Infant Brain Hemorrhages.” I have pasted in the study’s abstract and the URL for it from PubMed for those of you not familiar with the study’s details.

Intracranial hemorrhage in asymptomatic neonates: prevalence on MR images and relationship to obstetric and neonatal risk factors.

Looney CB, Smith JK, Merck LH, Wolfe HM, Chescheir NC, Hamer RM, Gilmore JH.

Department of Psychiatry, CB No. 7160, 7025A Neurosciences Hospital, University of North Carolina School of Medicine, Chapel Hill, NC 27599-7160, USA.

PURPOSE: To retrospectively evaluate the prevalence of neonatal intracranial hemorrhage (ICH) and its relationship to obstetric and neonatal risk factors. MATERIALS AND METHODS: Pregnant women were recruited for a prospective study of neonatal brain development; the study was approved by the institutional review board and complied with HIPAA regulations. After informed consent was obtained from a parent, neonates were imaged with 3.0-T magnetic resonance (MR) imaging without sedation. The images were reviewed by a neuroradiologist with 12 years of experience for the presence of ICH. Medical records were prospectively and retrospectively reviewed for selected risk factors, which included method of delivery, duration of labor, and evidence of maternal or neonatal birth trauma. Risk factors were assessed for relationship to ICH by using Fisher exact test statistics. RESULTS: Ninety-seven neonates (mean age at MR imaging, 20.8 days +/- 6.9 [standard deviation]) underwent MR imaging between the ages of 1 and 5 weeks. Eighty-eight (44 male and 44 female) neonates (65 with vaginal delivery and 23 with cesarean delivery) completed the MR imaging evaluation. Seventeen neonates with ICHs (16 subdural, two subarachnoid, and six parenchymal hemorrhages) were identified. Seven infants had two or more types of hemorrhages. All neonates with ICH were delivered vaginally, with a prevalence of 26% in vaginal births. ICH was significantly associated with vaginal birth (P < .005) but not with prolonged duration of labor or with traumatic or assisted vaginal birth. CONCLUSION: Asymptomatic ICH following vaginal birth in full-term neonates appears to be common, with a prevalence of 26% in this study. (c) RSNA, 2007.

http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD=Display&DB=pubmed

Here are the problems and weaknesses that jumped out at me:

• The study reports that instrumental vaginal delivery was not associated with these “small bleeds,” as they are described in the article (don't you love them using the word "hemorrhage" in the title and in the take-home message to be given to parents?), that caused no symptoms; however, studies consistently find an association between instrumental vaginal delivery and larger, symptomatic bleeds. With only 65 vaginal births, of which 13 were instrumental vaginal deliveries, the study lacks the statistical power to detect a difference, so this conclusion is misleading.
• These births were not “normal” by Lamaze International’s standards. Twenty out of every 100 women having a vaginal birth in this population had an instrumental delivery (13 out of 65). Fourteen out of every 100 babies (9 out of 65, assuming all trauma occurred during vaginal delivery) experienced delivery trauma, defined as “cephalohematoma [a fluid-filled swelling under the scalp], scalp laceration [cuts], or bruising associated with the use of forceps.” Given this rate of vaginal instrumental delivery and delivery trauma (can you imagine these stats in a home birth practice or at a free-standing birth center even including transfers to hospital?), it’s a safe bet that pushing position and technique were also not physiologic. “But,” as they say on the late night commercials, “there’s still more.” Investigators note that most bleeds were at the base of the skull. As a retired doula, I can tell you that conventional obstetricians often don’t just “catch” babies. The head emerges, and then they tug, and maneuver, and twist the baby’s head to “deliver” the shoulders, as if the shoulders wouldn’t come out on their own. It’s also a safe bet that a high proportion of the women were given drugs to start or intensify labor. Overly strong contractions in second stage might have a harmful effect too. In summary, like the accusations that vaginal birth weakens the pelvic floor, we don’t actually know to what degree these bleeds are inherent to vaginal birth and to what degree they are iatrogenic (caused by doctors) injuries. It would be interesting to do a study comparing MRIs of babies whose mothers pushed in positions other than on their backs, who pushed and breathed according to their inner urges, and who birthed their babies truly spontaneously with babies whose mothers had typical obstetric management.
• Amy Romano, my boss at Lamaze International, and author of Lamaze’s quarterly Research Roundup, points out another confounding factor: Babies with bleeds had the MRI at a mean of a week earlier than babies who didn’t have bleeds. Why is this important? The investigators note that they excluded infants older than 5 wks from the study because other studies had found that “all hemorrhages identified at birth had resolved by that age.” In other words, a week could make a big difference as to whether a bleed was found or not. If women having cesarean surgery tended to have their babies scanned later, this could explain some of the difference between groups.
• Even if vaginal birth is inherently associated with small bleeds, what the radiologists measured is a surrogate outcome, that is, an outcome believed to be in the causal pathway to a clinical outcome but not one itself. Surrogate outcomes often turn out to be misleading. For example, hormone replacement therapy improves blood lipid profiles in menopausal women but doesn't reduce heart attacks. It is likely that head molding can cause a bit of bleeding at the suture lines even when care during labor is optimal--which it certainly was not here--but it would be faulty logic to assume that this is anything but innocuous. Remember that no adverse effects were seen in these babies. In point of fact, we have no evidence that the liberal use of cesarean surgery improves neurologic outcomes compared with spontaneous labor and spontaneous vaginal birth. On the contrary, we have ample evidence of its potential harms for babies, not to mention women, in both the current and future pregnancies.

It is, to put it mildly, disingenuous to recommend telling parents that vaginal birth causes “brain hemorrhage” while cesareans do not, and then add “but we think it’s probably nothing to worry about.” There is, however, a real take-home message that was noted by the investigators but didn’t make it into the news or the study abstract: Should a baby in the first weeks of life undergo neurologic imaging, finding a bit of blood under the skull does not necessarily mean the baby has been physically abused—at least, considering the number of instrumental deliveries and the amount of delivery trauma reported in this study, not by the baby’s family.

-- Henci