Journal of Human Lactation | 2019
Reply to Letter to the Editor by Frank Nice
Abstract
Dear Editor, We appreciate the letter writer’s disappointment in the study results. However, we respectively disagree with their conclusion that not achieving our target enrollment in the randomized trial “invalidates the hypothesis and any conclusions.” Our purpose was not to determine if metformin increases milk production. Our primary aim was to determine the feasibility and acceptability of conducting a fullscale trial of metformin versus placebo. We received many referrals of mothers with insufficient milk production from breastfeeding medicine clinics and surrounding maternity hospital lactation follow-up clinics. We enrolled 56 mothers into baseline measurements, which was on target with our baseline goal. However, we did not meet our trial enrollment goal, which suggests feasibility concerns in scaling up to a larger randomized trial. This pilot study has highlighted the following next steps in this line of research: (a) investigation of the underlying drivers of the association between suboptimal metabolic health and low milk supply, (b) investigation of diagnostic methods that will enable earlier and more sensitive identification of women who are metabolically at risk for insufficient lactation, and (c) once identified, investigation of improved approaches to assessing and individualizing breastfeeding management plans to sustainably integrate breastfeeding with other new motherhood demands. Our second aim was to obtain estimates of variance and to determine if there was evidence of a positive trend toward greater improvement in milk production in the metforminassigned group. While our pilot study was never powered to draw definitive conclusions based on a significance level of p < .05, this does not preclude consideration of the clinical significance of the results. Our results uncover a critical question: Is a median improvement in milk output of 22 mL/24 hours (95% CI = −28 to 75 mL) clinically meaningful? Only 1 of the 10 mothers who were assigned to metformin had a maximum improvement in milk production across 28 days that was greater than 90 mL/24 hours, which was predetermined to be clinically relevant. We felt it was very important for the readers to know that none of the mothers assigned to metformin felt it was worthwhile. As lactation consultants, we view human milk as liquid gold and every drop worth it, but one of our aims was to determine acceptability from the mother’s perspective. We stand by the key message that only 2 of 10 mothers assigned to metformin sustained 4 weeks of improvement in milk production, which raises concerns about the clinical significance of scaling up to a fully powered randomized trial. Dismissing a pilot study due to null results is a major problem within academia, as it fuels publication bias toward studies with positive results. We applaud the Journal of Human Lactation for fighting against this bias by publishing our results, in which we present an unbiased description of the barriers that we encountered (and what others can expect to encounter) in conducting a rigorous pilot study following consensus guidelines for publication of randomized pilot and feasibility trials (SM Eldridge, BMJ 2016;355:i5239). While we deeply empathize with the frustration that mothers with low milk supply face, offering solutions lacking a strong evidence base only adds to their burden. In our opinion, the clear message from our pilot study results is that foundational upstream research is the next logical step to improve lactation for metabolically vulnerable women. Sincerely,