Urine NMR metabolomics analysis of breastfeeding biomarkers during and after pregnancy in a large prospective cohort study

Abstract

Background. Modern metabolomic profiling has not yet been applied to human breastfeeding research. A common reason for breastfeeding cessation is perceived insufficient milk production. We investigated broad biochemical profiles in maternal urine collected during and after pregnancy to identify biomarkers related to reduced reported breastfeeding. Methods. Fasting urine was collected at three consultations (visit V1: gestational week 8–20; V2: week 28 ± 2; V3: 10–16 weeks postpartum) in the STORK Groruddalen program, a prospective, multiethnic cohort study of gestational diabetes involving healthy, pregnant women in Oslo, Norway, and analyzed using NMR spectroscopy. Breastfeeding at V3 was recorded in three categories: Exclusively breastfeeding (n = 326), partially breastfeeding (n = 156) and formula feeding (n = 67). Results. Five metabolites were relevant to breastfeeding. Lactose was detected at V1 and increased to 0.1 mM/mM creatinine at V2. Postpartum excretion at V3 was significantly higher in exclusively breastfeeding women than partially or non-breastfeeding (median = 0.29, 0.23 and 0.04 mM/mM creatine, respectively; ANOVA p-value = 2e–70). Glycine excretion at V3 (0.12, 0.10 and 0.06, respectively; p = 2e–5) and at V2 were associated with breastfeeding (0.34, 0.33 and 0.26, respectively; p = 4e–5). Creatine and two unidentified substances also correlated with breastfeeding. NMR metabolomics found no other metabolites differing between categories during pregnancy (V1, V2), and did not predict individual breastfeeding postpartum (V3). Conclusion. Decreased glycine excretion at V2 may indicate difficulties meeting the metabolic demands of the growing fetus, but urine profiles contained otherwise little indication of early adaptations during pregnancy towards reduced biological potential to breastfeed.

Key Words::

• Breastfeeding
• ethnic background
• gestational diabetes mellitus
• glycine
• lactation
• lactose
• metabolomics
• nuclear magnetic resonance spectroscopy
• partial least squares discriminant analysis
• urine

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

The STORK Groruddalen research program is funded by the Norwegian Research Council (project number 185790), The South-Eastern Norway Regional Health Authority, Norwegian Directorate of Health, and collaborative partners in The City of Oslo, Stovner, Grorud and Bjerke administrative districts, and the present study was supported by grants from the University of Oslo and the Oslo Diabetes Research Centre. The authors received their salaries from their respective primary academic affiliations. The funding bodies had no role in the design of the present or the original study, in the collection, analysis, and interpretation of data, in the writing of the manuscript, or in the decision to submit the manuscript for publication.