Longitudinal Profiling of the Human Milk Microbiome from Birth to 12 Months Reveals Overall Stability and Selective Taxa-Level Variation

Abstract

Human milk bacteria contribute to gut microbiome establishment in breastfed infants. Although breastfeeding is recommended throughout infancy, temporal variation in the milk microbiome-particularly beyond solid food introduction-remains understudied. We analyzed 539 milk samples from 83 mother-infant dyads between 1 week and 12 months postpartum using full-length 16S rRNA gene sequencing. The microbiota was dominated by Streptococcus (34%), Cutibacterium (12%), and Staphylococcus (9%), with marked inter-individual variation. Microbiome profiles remained largely stable across lactation, with only six taxa showing temporal fluctuations, including increases in typical oral bacteria such as Streptococcus salivarius, Streptococcus lactarius, Rothia mucilaginosa, and Granulicatella adiacens. Richness and evenness were higher at 1 week compared to 1 month postpartum (p = 0.00003 and p = 0.007, respectively), then stabilized. Beta diversity also remained stable over time. Maternal pre-pregnancy BMI was positively associated with Gemella haemolysans (p = 0.016), while Haemophilus parainfluenzae was more abundant in milk from mothers with allergies (p = 0.003) and those who gave birth in autumn or winter (p = 0.006). The introduction of solid food was linked to minor taxonomic shifts. Overall, the milk microbiome remained robustly stable over the first year of lactation, with limited but notable fluctuations in specific taxa. This study supports the role of human milk as a consistent microbial source for infants and identifies maternal BMI, allergy status, and birth season as key variables warranting further investigation.

Keywords: 16S rRNA gene sequencing; breastfeeding; determinants; human milk microbiome.

Figure 1

Composition of the human milk microbiota in the first 12 months postpartum. OTUs with a mean relative abundance of ≥0.5% are displayed.

Figure 2

Genus-level interindividual variation in human milk microbiota profiles across the first 12 months postpartum. Genera that made up <2% mean relative abundance are grouped together as “others”.

Figure 3

Alpha diversity was significantly elevated at 1 week compared to 1 month postpartum. The box plots represent Shannon diversity/richness at each time point. The model-predicted means at each time point are represented by red dots. The model-predicted means for richness are not estimable at 9 and 12 months due to insufficient data (** p < 0.01, **** p < 0.0001).

Figure 4

Milk microbiome beta diversity remains largely stable over time. (A) PCoA plot of Bray–Curtis distances between different time points. The ellipses show the 95% confidence level. (B) Volatility of the milk microbiome between subsequent time point pairs. Each point represents one mother’s dissimilarity between two consecutive time points; the LOESS line shows overall trend with the shaded area representing the 95% confidence intervals.

Figure 5

Temporal variation in six OTUs within the human milk microbiota across the first year of lactation. One month postpartum was used as the reference level. Loess lines represent CLR transformed abundance at each time point, and the shaded areas represent the 95% confidence intervals. Model-predicted means at each time point are represented by dots (* p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001).

Figure 6

Pre-pregnancy BMI, birth season, and maternal allergy were associated with the human milk microbiome. Loess lines represent CLR transformed abundance at each time point, and the shaded areas represent the 95% confidence intervals.

Figure 7

Impact of solid food introduction on the human milk microbiota. Pre-solids: the sample collected immediately prior to the first introduction of solid foods. Post-solids (1): the first sample collected after the introduction of solid foods. Post-solids (2): the second sample collected at two sample time points after the introduction of solid foods. * p < 0.05, ** p < 0.01, *** p < 0.001.