As the gold standard of infant nutrition (1), human milk delivers a dynamic combination of nutrients and bioactive molecules influenced by both genetics and lifestyle factors, such as nutrition and smoking (2–4). Human milk consumption is also coterminous with the first 1,000 days of life, a critical time period to establish human metabolic mechanisms (5). The mother-to-child transmission of non-genetic factors influenced by lifestyle, which primes the infant’s development, can therefore be partly mediated by human milk composition (6). Human milk oligosaccharides (HMOs) are an important group of bioactive molecules in human milk. These non-nutritive milk components promote beneficial microbiota, protect against pathogens, and enhance immune function (7). HMOs are lactose-based molecules extended by the addition of galactose– N-acetylglucosamine disaccharides, fucose, or sialic acid residues. More than 200 distinct HMOs have been identified in human milk, with their concentrations varying across the postpartum period and between individuals (7, 8). Maternal characteristics, such as age, ethnicity, parity, and body composition affect HMOs concentration and composition (9, 10). Some studies indicate that the concentrations of some HMOs differ between mothers with obesity and those with a healthy weight, but findings are inconsistent (9–12).

The simplest sialylated oligosaccharides in human milk, 3′-sialyllactose (3′SL) and 6′-sialyllactose (6′SL), are generated in the Golgi apparatus of mammary epithelial cells when specific sialyltransferases attach sialic acid to lactose via α2-3 or α2-6 linkages. 3’SL and 6’SL promote infant gastrointestinal health and limit inflammation (13, 14). 3’SL and 6’SL supplementation also enhances exercise performance in mouse and Cænorhabditis elegans models, suggesting a role in cardiometabolic health (15–17). A recent Mexican study showed that mothers with overweight/obesity had lower 3’SL and 6’SL concentrations in colostrum, compared with mothers with a healthy body weight (10). The relationship between the level of these sialylated HMOs in human milk and infant growth is still unclear. Some studies report no association (18, 19), while others detected a positive correlation between 3’SL and/or 6’SL concentration and weight gain (20–22). Contrastingly, one study found a negative association between 3’SL and infant BMI (23).

The effect of maternal exercise on HMO concentrations is a largely unexplored area. In mice, exercise training during pregnancy caused increased milk concentrations of 3’SL but not 6’SL (24). The authors also showed that higher milk 3’SL concentrations mediated metabolic and cardiac health of the offspring in adulthood, with male offspring cross-fostered to trained mice displaying reduced body weight and fat percentage compared with offspring cross-fostered to sedentary mice. Based on the widespread beneficial effects of exercise on most organs and tissues in the body, it seems likely that exercise also benefits human milk composition (25). Indeed, endurance and strength training during pregnancy reduce the colostrum concentrations of two proinflammatory factors, IL-8 and TNF-α (26). Additionally, a single moderate-intensity exercise session in the postpartum period increased milk 12,13-diHOME, an activator of brown fat metabolism, while high-intensity endurance exercise increases milk lactate concentrations acutely after exercise (27, 28). Nevertheless, despite its major impact on metabolic health, exercise has undergone little investigation concerning its effect on human milk composition.

Our objective was to assess the acute effect of endurance exercise on 3’SL and 6’SL concentrations in human milk. Based on the noted effect of exercise on 3’SL (24), the main HMO in mice, we hypothesized that single sessions of both moderate-intensity continuous training (MICT) and high-intensity interval training (HIIT) would increase the concentration of sialylated HMOs in human milk, with the greatest effect seen after HIIT. To test our hypotheses, we analyzed 3’SL and 6’SL concentrations in human milk after a single session of MICT and HIIT in a randomized crossover study. Our results will contribute to the limited body of evidence on the impact of maternal exercise on human milk composition.

To our knowledge, this randomized crossover trial is the first study to investigate the acute effects of exercise on 3’SL and 6’SL concentrations in human milk. Contrary to our hypothesis, a single session of endurance exercise had no statistically significant effect on 3’SL or 6’SL concentrations. Future studies should investigate whether MICT transiently increases 3′SL and HIIT affects 6′SL concentrations.

Despite knowing that exercise affects the concentrations of several molecules in milk (26–28, 30), there is limited prior research on the impact of physical activity on HMO concentrations in human milk. Exercise induces profound systemic metabolic changes through increased skeletal-muscle energy demand, with outcomes depending on exercise intensity, duration, and frequency. A single bout of exercise triggers acute physiological responses, partly mediated by exercise-induced signaling molecules (exerkines) (34). Given that many metabolites, hormones, and bioactive molecules present in maternal circulation are transferred into milk, such exercise-induced changes may directly or indirectly influence human milk composition. Unlike the transient effects observed after a single exercise session, long-term adaptations arise from the cumulative impact of repeated bouts of exercise, including enhanced mitochondrial biogenesis and improved substrate utilization (35).

Although we did not observe statistically significant acute effects of endurance exercise on 3’SL or 6’SL concentrations within 4 h after exercise, Harris et al. (24) reported a positive association between physical activity during pregnancy and human milk concentrations of 3’SL at 2 months postpartum. Their findings suggest a chronic adaptation rather than an acute effect. For 6’SL, however, the relationship between human milk concentrations and physical activity is unclear. Milk 6’SL concentrations showed a weak, negative correlation with the number of daily steps in pregnancy, but no association with overall physical activity (24).

Both our findings and other studies (3, 24, 36, 37) show large inter-individual variations in human milk concentrations of 3’SL and 6’SL, making it difficult to ascertain the effect of exercise on these HMOs. We chose to define the REST condition as “sitting quietly” rather than “usual daily activities” to minimize individual variations. We also expected a higher contrast between “sitting quietly” and exercise than between “usual daily activities” and exercise. Combined with our small sample size (n = 20), the high inter-individual variations may explain the lack of observed significant effects in our study. We might be able to mitigate this challenge in the ongoing MILSHAKE trial, a larger randomized control trial (n ≈ 80) that aims to determine the chronic effects of endurance exercise training on human milk composition, including 3’SL and 6’SL (38).

Body composition is one of the factors that can impact HMO concentrations. Accordingly, another aspect of our investigation involved the association between sialylated HMOs and post-pregnancy BMI. The variations we observed in 3’SL and 6’SL concentrations between participants had an order of magnitude of 1.5 and 1, respectively. We observed a tendency of lower 3’SL concentration, but not 6’SL concentration, in mothers with overweight/obesity compared with mothers with a healthy postpartum weight. Also Harris et al. (24) found that those with a high BMI had lower 3’SL concentrations, while Urrutia-Baca et al. (10) found lower concentrations of both 3’SL and 6’SL in mothers with overweight/obesity than in those with a healthy weight.

In our cohort, 3’SL concentrations correlated significantly with adiponectin, whereas 6’SL did not show a statistically significant correlation (p = 0.06). On the other hand, neither 3’SL nor 6’SL correlated with insulin. The underlying mechanisms for the observed correlation between 3’SL and adiponectin should be investigated in future studies but may include links to maternal cardiorespiratory fitness. Exercise training has been shown to increase serum adiponectin concentrations in women (39) and 3’SL concentrations in human milk were found to be associated with the number of steps taken during pregnancy (24).

It is worth noting that the reported concentrations of 3’SL and 6’SL in human milk vary between studies. Some of this variation may be due to different methodologies used to determine these HMOs. Chromatography methods such as high-performance liquid chromatography, high-performance anion-exchange chromatography and liquid chromatography mass spectrometry are the most used to quantify HMOs because of their sensitivity, availability and ease-of-use. On the other hand, NMR allows for faster and more straightforward molecular quantification. The average concentrations of 3’SL (751 μmol/L) and 6’SL (219 μmol/L) in our study differed compared with what other studies covering the same lactation period have reported (36, 37, 40, 41). Two reviews including mostly studies which used chromatography methods to gather data on human milk HMO composition 31 to 100 and 15 to 90 days postpartum, calculated an average of 237 and 300 μmol/L for 3’SL and 347 and 631 μmol/L for 6’SL, respectively (40, 41). Fewer studies have used NMR to determine HMO concentrations.

Using NMR, the reported concentrations of 3’SL and 6’SL still differ between studies. One study found lower average concentrations than we did for 3’SL (110 μmol/L) and higher for 6’SL (280 μmol/L) over the first 3 months of lactation in 10 Chinese women (36). Also Smilowitz and colleagues reported lower concentrations of both HMOs 90 days postpartum compared with our results, but like us, they found higher average concentrations of 3’SL (144 μmol/L) than 6’SL (119 μmol/L) among 52 American women (37). These differences may reflect the influence of geographic location on human milk composition (42–44).

Genetic background, culture and lifestyle affect the amount and diversity of HMOs in human milk. As the concentration we found for 6’SL is similar to that found in other studies, it is possible that our Norwegian cohort presents a particularly high level of 3’SL due to genetics or lifestyle factors, such as diet. Pregnant Norwegian individuals might be less prone to consume a high-fat diet or sugar-sweetened beverages than in some other countries, two factors that have been negatively correlated with 3’SL concentration in milk (3, 45). We did not collect data on our participants’ general diet, which limits the interpretation of our results. However, the average BMI of our participants (26.2 kg/m² ± 3.5) was similar to that of the participants in the study by Smilowitz and colleagues (25.3 kg/m² ± 4.1), who also used NMR for HMO determination, indicating that they might share similar lifestyles (37). BMI was not reported by Kortesniemi et al. (36). Finally, the participants’ secretor status can impact at least 6’SL concentrations in human milk (37, 46). Contrary to most other studies, we did not investigate this parameter as we did not perform genetic analysis nor quantify milk fucosylation levels. Given that the Scandinavian populations (Denmark, Norway, Sweden) are genetically similar, we estimate the frequency of non-secretors to be 18% under the Hardy–Weinberg Equilibrium, based on the allele frequencies reported by NorthernSweden in the single nucleotide polymorphism database (47, 48). Genetic, cultural and lifestyle differences can impact human milk composition significantly; the results obtained from our healthy Norwegian cohort can therefore only be extrapolated to similar populations.

There is no gold-standard methodology for carrying out and analyzing NMR experiments in milk. Human milk is a complex biological fluid, explaining its poor characterization compared with other biofluids like serum or urine. Specifically, the overwhelming amount of lactose masks the signals of other milk metabolites, including peaks enabling the identification and quantification of 3’SL and 6’SL. A way to mitigate this problem is to identify all known metabolites individually and subtract the associated signals from the 1H NOESY spectra, leaving only the signals of less prominent metabolites and making it possible to quantify 3’SL and 6’SL reliably (36). This solution presents two major drawbacks: it is time-consuming and requires access to a comprehensive milk metabolite NMR library. The novel NMR sequence called DREAMTIME offers an efficient alternative to the problem of identifying specific molecules in a mixture (32). Nevertheless, a limitation of our study is that we did not analyze our samples with the more commonly used high-performance liquid chromatography coupled to mass spectrometry technique. We were thus not able to compare human milk 3’SL and 6’SL concentrations obtained with both methods. Future research should formally validate the use of DREAMTIME as an HMO quantification method, to ensure valid and reliable results.

To conclude, the present randomized crossover study demonstrated unaltered concentrations of 3’SL and 6’SL in human milk acutely after MICT and HIIT in a Norwegian cohort. The novel DREAMTIME sequence is a powerful NMR approach for analysis of complex fluids like human milk, even though further validation is required to confirm our results. This study informs health professionals and (future) parents on the relevance of a single moderate- or high-intensity exercise session in the postpartum period as a modifier for 3’SL and 6’SL milk concentrations. Further studies focusing on overweight and obese cohorts should be implemented to validate our results, as stronger effects of exercise are expected in this population compared with normal weight cohorts (49). Future research should also investigate chronic effects of exercise on human milk composition and the potential impact of such changes on infant health. Ideally, future trials will include data on habitual diet, longitudinal metabolite profiling, and infant development beyond 12 weeks postpartum.