Study never started

Although intrapartum epidural analgesia is frequently implicated in adverse breast-feeding outcomes, many previous studies feature major design limitations that preclude widespread applicability of findings. Some fail to control for the precise pharmacologic composition of the epidural infusion, including whether or not an opioid, such as fentanyl, is even used at all in addition to local anesthetic or whether a combined spinal/epidural or purely epidural technique is used. The drugs used in epidural infusions not only have different mechanisms of action and lipophilicities but are also transferred across the placenta in varying proportions, with one study identifying an umbilical vein/maternal vein ratio of 0.94 for epidural fentanyl and 0.30 for bupivacaine, and another finding significantly different umbilical cord fentanyl concentrations among neonates whose mothers' epidural infusions contained >150 micrograms fentanyl, <150 micrograms, or none at all. It is also possible that the same total dose of epidural fentanyl could affect neonates differently depending upon the time course over which it was administered - namely, whether fentanyl is included in both the initial epidural bolus and the subsequent infusion or solely in the infusion. In two studies, mean umbilical vein concentrations of fentanyl did not correlate with total epidural infusion time, but both of these featured sample sizes fewer than 30, necessitating further research. Another limitation of some previous studies is defining success solely as the time to cessation of breast-feeding. Questionnaires mailed to mothers months or even years postpartum may generate unreliable data. If epidural medications truly mediate some physiologic effect upon breast-feeding, then the optimal study period is immediately post-delivery, specifically before the drugs are cleared from the maternal and neonatal circulations. After hospital discharge, many new factors - such as a mother's need to return to work or lack of social support - begin to confound the picture of breast-feeding success. Full-time employment outside the home has been significantly associated with decreased likelihood of breast-feeding at 6 months postpartum. Some studies also fail to control for intent to breast-feed at the time of hospital admission, number of infants previously breast-fed, or labor duration. Failure to account for oxytocin augmentation of labor is also problematic, as intravenous intrapartum oxytocin infusion has been shown to decrease a woman's endogenous serum oxytocin concentration on the second day postpartum in a dose-dependent fashion, which can subsequently impair milk release and, thus, decrease breast-feeding success.

Epidural analgesia may worsen breast-feeding outcomes by attenuating neonatal exhibition of neurobehaviors tied to feeding, such as sucking, rooting, and swallowing, during the immediate postpartum period. This critical period is when mother and baby make their first attempts at breast-feeding and set a precedent for subsequent interactions. Neonatal feeding behavior in the early postpartum period is an important predictor of long-term breast-feeding success; those babies who feed most vigorously during their first days of life are significantly more likely to still be breast-feeding at 3 or 6 months than those who exhibit any lesser degree of breast-feeding enthusiasm. Radzyminski et al. found no significant dose-response relationship for either epidural bupivacaine or fentanyl regarding neonatal feeding behaviors, and Porter et al. found no significant effect of epidural fentanyl, mean dose 184 micrograms, upon neonatal APGAR scores, incidence of respiratory depression, or NACS scores (Neurologic and Adaptive Capacity Scores) at 2 or 24 hours post-delivery. Beilin et al. found that neonatal NACS scores were significantly lower when mothers' epidural infusions contained greater than 150 micrograms total epidural fentanyl than when they contained only bupivacaine.

In this randomized, controlled, double-blinded study, we investigate whether intrapartum epidural fentanyl significantly decreases the likelihood of breast-feeding at hospital discharge and increases the incidence of neonatal deficits in latching on to the breast and audibly swallowing during the first three hours of life. We hypothesize that these effects will be dose-dependent but will have no relation to the time course over which the epidural fentanyl is administered. We also investigate whether oxytocin augmentation of labor and decreased amount of skin-to-skin contact during the first hour of life are associated with significantly decreased breast-feeding rates at hospital discharge.

NACS scores do not specifically measure feeding behaviors, but rather the five general areas of adaptive capacity, passive tone, active tone, primary reflexes, and alertness/crying/motor activity. Conversely, the Preterm Infant Breastfeeding Behavior Scale (PIBBS)specifically examines neonatal rooting, latching, sucking, and swallowing behaviors as well as general activity level, thus producing a comprehensive picture of breast-feeding interactions.

Beilin et al. previously performed a randomized, controlled, double-blinded study of multiparous women who were randomized to receive intrapartum epidural analgesia that contained, in addition to bupivacaine, either greater than 150 micrograms total fentanyl, less than 150 micrograms, or zero. This study had several limitations. Breast-feeding difficulties were not assessed until 24 hours postpartum, long after the neonate would have been expected to have fully metabolized the drugs. Like most previous studies, Beilin et al. also did not control for skin-to-skin mother-infant contact. According to a prospective cohort study of 21,842 mothers in 19 hospitals, the likelihood of exclusive breast-feeding throughout the hospital stay is positively, significantly correlated with the number of minutes of skin-to-skin contact during the first three hours postpartum, even after controlling for intention to breast-feed at the time of hospital admission. Another randomized, controlled, double-blinded study found no significant differences regarding breast-feeding initiation or breast-feeding rates at 12 months postpartum among women who received bupivacaine-only epidural analgesia, combined spinal-epidural analgesia with mean fentanyl dose 107.3 micrograms, or epidural infusion with mean fentanyl dose 162.8 micrograms. All fentanyl was administered only as part of a continuous infusion and never in an initial bolus, so the potentially different effects resulting from large loading doses of fentanyl compared to gradual infusion could not be studied. The authors did not objectively measure neonatal feeding behavior, relying solely upon maternal interviewing at 24-48 hours post-delivery, and they studied only nulliparous women.

In a pilot study of 310 mothers who delivered at University Hospitals Case Medical Center (Cleveland, Ohio) between September 2009 and February 2010, we found that duration of epidural infusion was significantly correlated with decreased neonatal ability to successfully latch onto the breast and audibly swallow during the first three hours of life. All infusions contained identical doses of fentanyl and bupivacaine.

Inclusion Criteria:

• Singleton pregnancy
• Maternal age 18+
• Intention to breast-feed at time of hospital admission

Exclusion Criteria:

• Delivery by Cesarean section
• Neonatal morbidity requiring admission to neonatal intensive care unit