Breastfeeding after immediate vs delayed postpartum contraceptive implant placement: a noninferiority randomized controlled trial.

Abstract

Breastfeeding and access to contraception are important to patient and public health. The etonogestrel contraceptive implant is a popular and highly effective long-acting progestin contraceptive with few medical contraindications. Despite theoretical concern about early exogenous progestin exposure inhibiting lactogenesis, previous studies showed reassuring lactogenesis and long-term breastfeeding with immediate placement. However, no studies compare placement within 24 hours postpartum to delayed placement at a postpartum visit. We studied the association between etonogestrel contraceptive implant placement within 24 hours vs at least 2 weeks postpartum and breastfeeding continuation at 8 weeks postpartum. We conducted a noninferiority randomized controlled trial of postpartum participants who planned to breastfeed and use the etonogestrel implant at 2 university hospitals in the Mountain West. Participants were at least 13 years old, fluent in English or Spanish, had term deliveries, and lacked contraindications to implant use. They were randomized 1:1 to implant placement within 24 hours or at least 2 weeks postpartum. We collected baseline participant characteristics from chart review and questionnaires. Participants reported breastfeeding status and implant use via electronic questionnaires at 2, 4, 8, 12, and 24 weeks postpartum. The study was powered to assess a 15% noninferiority margin between groups of the primary outcome, any breastfeeding at 8 weeks in the per protocol population. We enrolled 150 participants (n=78 immediate, n=72 delayed). After removing participants who declined implant placement, withdrew, or were ineligible (n=8) or lacked primary outcome data (n=16), the modified intention-to-treat analysis included 126 participants (n=69 immediate, n=57 delayed). The per protocol analysis included 115 participants (n=62 immediate, n=53 delayed) after excluding additional participants who received the implant outside of the placement timing windows (n=11). Participants were similar in age, race, and ethnicity, as well as previous breastfeeding experience, delivery mode, and epidural use; the delayed group reported prior implant use less frequently (29% vs 43%). In the per protocol analysis, 77.4% (48/62) of participants in the immediate group and 81.1% (43/53) in the delayed group reported any breastfeeding at 8 weeks; the lower limit of the one-sided 95% confidence interval around this -3.7% difference was -16%, exceeding our predefined noninferiority margin. The difference between groups in the modified intention-to-treat group was smaller (-0.6%) and within the noninferiority margin (-12.8% lower limit of the 1-sided 95% confidence interval): 78.3% (54/69) immediate and 78.9% (45/57) delayed. Implant continuation at 24 weeks trended toward favoring immediate placement (96%, 52/54) compared to in those assigned to delayed placement, (85%, 39/46, P=.08). Exclusive or any breastfeeding and implant continuation through 24 weeks were similar between groups. The results demonstrated no clinically important differences in breastfeeding and implant continuation outcomes though our primary outcome exceeded our predefined noninferiority margin. These reassuring data add to the evidence base for supporting etonogestrel implant initiation for breastfeeding people whenever they desire it postpartum.