Human Milk Bioactive Components and Child Growth and Body Composition in the First 2 Years: A Systematic Review

Human milk (HM) contains macronutrients, micronutrients, and a multitude of other bioactive factors, which can have a long-term impact on infant growth and development. We systematically searched MEDLINE, EMBASE, Cochrane Library, Scopus, and Web of Science to synthesize evidence published between 1980 and 2022 on HM components and anthropometry through 2 y of age among term-born infants. From 9992 abstracts screened, 141 articles were included and categorized based on their reporting of HM micronutrients, macronutrients, or bioactive components. Bioactives including hormones, HM oligosaccharides (HMOs), and immunomodulatory components are reported here, based on 75 articles from 69 unique studies reporting observations from 9980 dyads. Research designs, milk collection strategies, sampling times, geographic and socioeconomic settings, reporting practices, and outcomes varied considerably. Meta-analyses were not possible because data collection times and reporting were inconsistent among the studies included. Few measured infant HM intake, adjusted for confounders, precisely captured breastfeeding exclusivity, or adequately described HM collection protocols. Only 5 studies (6%) had high overall quality scores. Hormones were the most extensively examined bioactive with 46 articles (n = 6773 dyads), compared with 13 (n = 2640 dyads) for HMOs and 12 (n = 1422 dyads) for immunomodulatory components. Two studies conducted untargeted metabolomics. Leptin and adiponectin demonstrated inverse associations with infant growth, although several studies found no associations. No consistent associations were found between individual HMOs and infant growth outcomes. Among immunomodulatory components in HM, IL-6 demonstrated inverse relationships with infant growth. Current research on HM bioactives is largely inconclusive and is insufficient to address the complex composition of HM. Future research should ideally capture HM intake, use biologically relevant anthropometrics, and integrate components across categories, embracing a systems biology approach to better understand how HM components work independently and synergistically to influence infant growth.


Introduction
Beyond providing a custom-made source of micronutrients and macronutrients, human milk (HM) contains a multitude of other bioactive factors [1].Together, these components create a biologically active system to meet the health and nutritional needs of infants and young children.The WHO recommends exclusive breastfeeding for the first 6 mo and continued HM feeding as a component of the diet through 2 y and beyond [2].HM bioactives can have a prolonged impact on the infant microbiome, growth and development, as well as immune function [3].Although infant formula provides a safe nutrient substitute for HM, it is deficient in its capacity to replace HM bioactives.Despite decades of research, we still have a limited understanding of how many HM components inform the most fundamental of infant outcomes, including growth and development.However, as technology improves, there has been an increasing push to expand the breadth and scope of HM composition research, moving beyond nutrient analysis to investigate diverse components and clinical outcomes in the infant.
Anthropometry is a primary indicator of health for physicians, care providers, and parents.In high-resourced countries, a key goal is to prevent childhood obesity and identify risk factors or predictors of obesity in early life [4].Conversely, lower resourced settings often have a greater need to understand and mitigate concerns around child under-nutrition and stunting [5].Investigating how HM components contribute to infant anthropometry in healthy full-term infants, will provide a broad group of researchers and clinicians with an enhanced understanding of the role that HM feeding plays in child growth.This in turn will help to provide improved evidence to inform practice recommendations, and health promotion strategies to support breastfeeding and optimal infant growth and will assist the industry to better design HM alternatives when HM is not available.
The aim of this systematic review was to assess and synthesize evidence on the associations between HM components and child anthropometry measured in the first 2 y.Because of a large number of articles retrieved, results were organized into 3 manuscripts encompassing the following categories: micronutrients (vitamins and minerals [6]), macronutrients (lipids, proteins, and digestible carbohydrates [7]), and the current manuscript which examines bioactive components (for example, cytokines, hormones, and non-digestible carbohydrates).Maternal cells, ribonucleic acid, and microbiota were excluded from this review.

Bioactive components in HM
Bioactive components of HM are defined as components that "affect biological processes or substrates and hence have an impact on body function or condition and ultimately health" [1].Bioactive components in HM include lactoferrin, growth factors, hormones, nucleotides, human milk oligosaccharides (HMOs), immunoglobulins, and cytokines [8,9].Although this list is not exhaustive, for the purposes of this review, we have classified bioactives in HM into 3 categories: hormones, HMOs, and immunomodulatory components.
Hormones enter milk from the maternal bloodstream and are produced endogenously in the maternal epithelium of the mammary gland [10].Insulin, ghrelin, adiponectin, and leptin are commonly examined for their appetite-regulating functions [10] and are among the most extensively studied bioactive components in HM in relation to infant growth [10].Previous research examining the link between hormones and infant growth has reported conflicting results [10].
HMOs are the third most abundant component in HM.HMOs are complex carbohydrates that are indigestible to the infant yet serve as prebiotics for commensal bacteria in the infant's gut [11], thus impacting the infant microbiome.HMOs can also have anti-adhesive functions, sequester pathogens, and directly interact with the gut epithelium and immune cells [12].Furthermore, HMOs are minimally absorbed into the infant circulation where they can have systemic effects [13].The most influential predictor of HMO composition is maternal secretor status, which is determined by a single nucleotide polymorphism on the fucosyltransferase 2 (FUT2) gene [14].HM produced by secretors contains HMOs that have α1,2-fucosylated oligosaccharides, whereas milk produced by nonsecretors is deficient in this class of HMOs.Although relatively little research has addressed the impact of HMOs or maternal secretor status on infant growth, some commercial formulas are now adding HMOs to their products, often including or limited to α1,2-fucosylated HMOs that are not produced by non-secretor mothers [11,15].HM contains multiple components that impact the infant's immune system.These include cytokines, growth factors, lactoferrin, lysozyme, and immunoglobulins [16].Although each of these factors has a different pathway of impact, they all influence the development and function of the immune system and help provide immunity for the breastfed infant during a critical period when the infant's own immune system is developing [17].Although they are best known for their immunomodulating

Statement of Significance
Our work comprehensively synthesizes evidence regarding associations between individual human milk bioactives and child anthropometrics among healthy, term-born infants.This manuscript is part of a larger three-part systematic review (PROSPERO: CRD42020187350).
properties, it is conceivable that these factors could also influence infant growth-either by supporting optimal immune health or via immune-independent mechanisms such as energy spared by preventing illness in the infant.

Methods
This review was registered with PROSPERO: CRD42020187350 [6] and is reported according to the PRISMA [18].Nine reviewers (SMR, JMM, DC, MG, KS, SM, PPP, RR, and MB) independently participated in abstract and full-text screening, quality assessment, and data extraction.Covidence Systematic Review Software (2020) was used to manage screening and data extraction.

Search strategy and screening
In consultation with the review team, a health sciences librarian (CM) developed and tested the search strategy.Using a combination of controlled vocabulary and keywords to create search concepts for HM, growth and development, macronutrients, micronutrients, and bioactive components.We also included an infant search filter adapted from the Pediatric Search Filter from the Cochrane Childhood Group to limit 24 mo of age [19].The search was peer-reviewed by another health sciences librarian using the Peer Review of Electronic Search Strategies method [20].The original search strategy was created in MEDLINE (Ovid) and translated to the other databases.The MEDLINE (Ovid) strategy is available in Appendix A. All other strategies are available upon request.
We searched the following databases in March 2020: MED-LINE (Ovid; MEDLINE® All 1946-2020), EMBASE (Ovid; 1974-2020), the Cochrane Library (Wiley; CENTRAL and Cochrane Database of Systematic Reviews), Scopus , and Web of Science Core Collection (Clarivate, 1900(Clarivate, -2020)).To locate grey literature we searched Agricola, Practice-based Evidence in Nutrition (PEN®), OpenSIGLE, Google Advanced, and PROSPERO.These resources were selected to ensure the retrieval of materials relevant to nutrition, food science, and technology.Finally, we conducted reverse snowballing (using the reference list of a paper to identify additional papers; [21] on review articles retrieved with our search strategy.The search was updated in March 2022 revisiting all the original databases and grey literature sources.The records were exported into Endnote (version x9; Clarivate Analytics) and duplicates were removed [22].Inclusion criteria were 1) References published in English and 2) after 1980.All records were screened in duplicate in Covidence (Veritas Health Innovation).

Selection criteria
Search results were screened in duplicate.Any randomized controlled trial (RCT) or observational study was eligible for inclusion if it reported associations between HM components and infant anthropometrics.Data from RCTs were evaluated as observational studies because, in all cases, associations between HM composition and infant anthropometrics were secondary trial outcomes.We required that studies reported on healthy, term, HM-fed infants (aged 0-24 mo).Healthy was defined as term birth (37 wk, 0 d of gestation, or later) with no congenital or other morbidities and no admission in the neonatal intensive care unit, as described by study authors.Studies that included preterm infants were excluded unless it was possible to extract data for the term infants separately.Although breastfeeding exclusivity was not an inclusion criterion, it was recorded when reported by authors.Our main outcomes were weight-for-age Zscore (WAZ), length-for-age Z-score (LAZ), weight-for-length Zscore (WLZ), BMI or BMI-for-age Z-score, and growth velocity.Reference populations used to calculate Z-scores varied across studies, and some studies reported percentiles rather than Zscores.To simplify the synthesis of results only Z-scores were summarized in heatmaps.
We also included articles that reported other infant anthropometrics, including but not limited to weight, length, rapid weight gain (as reported by study authors), total adiposity (percent body fat by DEXA [dual-energy x-ray absorptiometry] or skinfold thickness), body composition (fat mass, fat-free mass, percent fat mass by bioelectrical impedance spectroscopy or skinfold thickness), stunting, wasting, under-or overweight, and head circumference.

Quality assessment
Articles were assessed for quality using a modified Newcastle-Ottawa scale [23] (Supplemental Table 1).On the basis of previous research [24], and in collaboration with multiple subject matter experts, we created a 17-point evaluation scale.We designated 8 points for HM exposure assessment, including HM collection and handling protocol (3 points), HM sample preparation (1 point), analytical method used to measure HM analyte (2 points), longitudinal HM sampling strategy (1 point), and accounting for infants' HM intake (1 point); 5 points for confounders considered, including infant diet (2 points), birth anthropometrics (1 point), baseline characteristics of mothers and infants (2 points); and 4 points for infant anthropometry outcome assessment, including whether infant anthropometrics were measured by trained staff (1 point), using technical replicates (1 point), longitudinally (1 point), and with all infants measured within 1 wk of each other at each time point (1 point).Quality assessment for each article was conducted in duplicate by independent reviewers, with conflicts addressed through consensus.Overall quality scores between >13 and 17 were considered high; 7 and 13 moderate; and <7 low.Quality scores were also evaluated individually for exposure assessment (high: >6-8, moderate: 3-6, and low: <3), confounders considered (high: >4-5, moderate: 3-4; and low: <3), and outcome assessment (high: >3-4, moderate: 2-3, and low: <2).

Data extraction
Data extraction was conducted using a standardized form that was developed and piloted in collaboration with subject matter experts.Data extracted included publication year, location, design, baseline characteristics of mothers and infants, HM sampling times, and timing of infant anthropometric measurements, whether HM values were reported as concentrations or estimated intakes, outcomes, associations reported (correlations and unadjusted and adjusted ß-estimates, as reported by study authors), and major confounders considered (via study design or statistical analyses) including maternal age, parity, maternal BMI, ethnicity, time postpartum, breastfeeding exclusivity (as stated by study authors-usually at time of enrollment), birth anthropometrics, infant age and sex, and any others reported.Study authors were contacted to request data in instances they were missing or presented in non-extractable formats.Each article was extracted in duplicate, and conflicts were addressed through consensus.

Analytical strategies
Data were summarized in tables as reported by study authors, and directional associations reported for HM concentrations were visualized in heat maps with colors determined by vote counting based on the mean direction of significant associations [25].Because of minimal studies reporting calculated daily intakes (CDIs), these were summarized narratively.Only studies reporting HM concentrations were reflected in the heatmaps.The color gradient was determined by 1) assigning a score to each outcome (þ1 for positive associations, 0 for no/assumed no association, and À1 for inverse associations) and 2) taking the mean direction of association (range: À1 to þ1) for all articles reviewed.Some papers only reported outcomes that were statistically significant; for these studies, we considered the unreported associations as "assumed no association."Components investigated in just 1 study were excluded from heatmaps.When studies reported results from both HM concentration and estimated daily intake, only associations using concentrations were reported in heatmaps.
Narrative synthesis was conducted according to the synthesis without meta-analyses (SWiM) reporting guidelines [25].Bioactives were divided into 3 categories of hormones, HMOs, and immunomodulatory components, based on their distinct roles within infant physiology.Bioactives that did not fall into these 3 categories were discussed individually.

Description of included studies
In total, 9992 unique abstracts were identified and 1001 full texts were screened (Figure 1).The main reasons for excluding articles were no infant anthropometrics or only birth anthropometrics were reported (n ¼ 510); no associations between HM analytes and infant anthropometrics were reported (n ¼ 165); or no HM analytes of interest were reported (n ¼ 89).Together, these 3 reasons accounted for 76% (764/1001) of the articles excluded during full-text screening.Overall, 141 articles were included for the broader systematic review, of which 75 articles examining bioactives in HM were included in this review.
Among the 75 included bioactive articles (Table 1) , none were published before 2000.The 75 articles represented 69 unique studies involving 9980 dyads.Thirty-two studies were conducted in low-and middle-income countries (LMICs, according to World Bank criteria) and 36 in high-income countries (HICs).One study examined milk components from 2 cohorts, 1 in the USA (HIC) and 1 in Mexico (LMIC [47]).Fifty-seven studies were longitudinal, including 5 RCTs and 1 case-control study.The remaining 12 studies were cross-sectional.

Study quality
The majority of studies (n ¼ 65) were rated as moderate (8.5-12.75score on the modified Newcastle-Ottawa scale; maximum 17 points), with 5 studies being rated as low quality (<8) and only 5 studies rated as high quality (>13) (Figure 2, Supplemental Table 2).The most common quality issue across studies was failing to adjust for confounders, such as breastfeeding exclusivity, maternal BMI, or maternal age.
A variety of infant anthropometrics were reported across studies.Weight and length were the most common anthropometrics reported for each of the bioactive categories.Among studies examining hormones, 24 different anthropometrics were reported, compared with 22 for HMOs and 26 for immunomodulatory components.
Milk collection strategies and time points varied considerably across studies.Thirty articles reported analyte concentrations in milk from a single collection time point.Only 10 articles reported intakes and many of these were from the same research group [57,60,71].In addition, there was substantial variation in how 24-h milk intake was assessed.Most studies used pre-and post-feed weights; however, some did this for each feed over 24 h [60,83] whereas others only weighed at 1 feed and multiplied it by the number of feeds [84].Milk sampling times varied from birth (colostrum) to 4 y (which was outside the scope of this study).The most common time points for milk sampling were 1 mo (30 articles), 2 mo (18 articles), 3 mo (19 articles), and 6 mo (21 articles).Twelve studies reported "varied" milk collection time points which ranged from 10 d to 4 y postpartum.Two studies [57,77] did not report when milk samples were collected.
Heterogeneity in infant anthropometrics and variation in milk sampling procedure and collection time limited the ability to compare across a wide range of studies.Because of heterogeneity in study designs, sampling times, and reporting practices, metaanalyses and pre-planned subgroup analyses (for example, by study setting, mode of HM feeding, and nutritional status of mothers) were not feasible.The full list of pre-specified subgroups is available in the review protocol [6].

Hormones
Hormones were the most extensively examined bioactive in the body of literature with 46 articles (n ¼ 6773 dyads) exploring how hormone concentrations were related to infant growth (Figure 3, Supplemental Table 3).Overall, 13 hormones were examined with leptin being the most common (n ¼ 35 articles; 5857 dyads), followed by adiponectin (n ¼ 18 articles; 3479 dyads).Other hormones studied included cortisol, ghrelin, glucagon-like peptide-1 (GLP-1), insulin-like growth factor 1 (IGF-1), insulin peptide tyrosine tyrosine (PYY), and resistin.Notably, of the 46 included articles, only 15 accounted for maternal pre-pregnancy BMI.This is an important consideration because maternal BMI appears to impact several appetiteregulating milk hormone concentrations, including leptin and adiponectin [105].However, BMI is subject to bias as it is often self-reported and thus inaccurate and the relevance to HM composition is relevant at milk sampling time rather than before FIGURE 1. Systematic review of associations between HM bioactives and infant growth in the first 2 y: PRISMA flow diagram.Reasons for study exclusion were recorded in the order listed in the figure.Though some studies had more than 1 reason for exclusion, each study was only counted once (for example, if a study reported no HM analytes of interest and was not in English, it was recorded as the former).Bioactive studies are reported in the current paper; Macronutrient and Micronutrient studies are reported separately.HM, human milk.pregnancy [106].Finally, there were inconsistencies across studies regarding milk preparation before analyzing hormone levels.Using skim milk compared with homogenized milk is an important consideration as lipids may interfere with certain hormones and assays, especially when examining leptin and adiponectin levels in HM [107].We identified inconsistencies between studies with some using whole milk samples [70,65] and other studies using skim milk samples [68,55] to test for adiponectin and leptin levels.

Leptin.
Overall, HM leptin concentrations were negatively associated with infant anthropometry in more than half of the comparisons (11/20 comparisons), including for weight, weight gain, length, BMIZ and percent fat (Figure 3).Most remarkably, 3 of 4 studies examining the association between leptin and percent fat mass found significant inverse associations (Figure 3) [65,68,108], whereas 2 studies observed positive associations for lean body mass [70,55].The 1 exception to this trend was a positive association with change in BMI over the first 2 y in infants born with an average BMI [85]; however, only 1 study evaluated this association.Two studies consistently showed no association between milk leptin and triceps skinfold thickness [40,83].Four studies [96,60,83,84] examined CDI of HM leptin and none found meaningful associations with infant growth outcomes despite 3 of these studies [60,83,84] using the same technique to calculate CDI.

Adiponectin.
Similar to leptin, HM adiponectin concentrations were overall inversely associated with infant growth, with weight and length being reported most frequently.Multiple studies found inverse associations between milk adiponectin concentrations and infant weight [77,54,88,99] and length [99,54], and weight-for-age Z-Score (WAZ) [47,99,91].Only 3 of 21 studies [71,54,104] found positive associations between milk adiponectin and infant growth outcomes.However, Yu et al. [104] stratified by healthy mothers and mothers with gestational diabetes, and only found a positive association with infants from healthy mothers.Gridneva et al. [70], found an inverse association with infant abdominal adiposity; however, this was a time-dependent relationship calculated from daily intake rather than a cross-sectional 1-time adiponectin measurement.

Other hormones.
Milk insulin concentrations were examined in 9 studies (10 articles) with 1189 dyads.All studies, except Chan et al. [59] and Yu et al. [104] excluded mothers with diabetes; both of which found that HM insulin concentrations were significantly higher in mothers who were diagnosed with diabetes.Overall, inverse relationships were observed between milk insulin and infant growth; specifically with infant weight [108,51], head circumference [108,104], BMIZ [59], and percent fat [108].However, 8 of 10 studies found no relationship between milk insulin and weight and no studies found associations between milk insulin and infant length.Conversely, Sims et al. [96] and Ellsworth et al. [64] found positive relationships between insulin and fat mass index (FMI), WAZ, and head circumference-for-age z-score.However, both of these studies stratified mothers by maternal BMI examining outcomes of infants consuming milk from normal Alternative versions organized by component are available in Supplemental Tables 3-5.
weight compared with overweight mothers and neither study controlled for exclusive HM feeding.Milk ghrelin was examined in 9 studies (701 dyads) and demonstrated minimal associations with infant growth outcomes.An inverse association was observed with infant weight in 1 study [77], whereas 5 other studies [104,51,72,103,58] found no association.Positive associations were observed for weight gain in 2 [58,78] of the 5 studies that reported an increased rate of weight gain when higher concentrations of milk ghrelin were present.
Milk IGF-1 was analyzed in 5 studies (820 dyads).No consistent associations were found with infant growth across all studies.Two studies demonstrated opposite trends between milk IGF-1 and infant weight gain [78,69].However, Kon et al. [78], categorized infant weight gain into low weight gain (<500 g/mo), normal weight gain (between 500 and 1000 g/mo), and high weight gain (>1000 g/mo), whereas Galante et al. [69] treated weight gain as a continuous variable.In addition, the small molecule metabolite cyclic glycine-proline, which is derived from IGF-1, demonstrated a positive association with WAZ [73].
Milk cortisol and resistin each were analyzed in 2 studies, respectively.Pundir et al. [90] found positive associations between milk cortisol concentrations and infant head circumference and infant adiposity (measured by percent fat mass), whereas Hollanders et al. [73] found no associations between milk cortisol concentrations and any infant growth outcomes.
Both studies controlled for percent breastfeeding at the time of milk collection.Milk resistin was also examined by [51,94] and no associations were found with infant growth outcomes in either study.
PYY and GLP-1 were both assessed by [95] and [81].Although Schueler et al. [95] reported on a subset of the larger Larson-Meyer et al. cohort [81], they examined different infant growth outcomes.Both Schueler et al. [95] and Larson-Meyer et al. [81] found that GLP-1 concentrations were inversely associated with infant growth (weight, weight gain, and WAZ), but found no associations between PYY and infant growth outcomes.
Only 1 study explored sex hormones and infant growth [33].Inverse associations were observed between infant weight at 6 mo and progesterone, luteinizing hormone, and follicle-stimulating hormone, but not estradiol.Finally, 1 study [66] explored the relationship between irisin and infant growth, observing a positive association between milk irisin and infant weight at 6 wk of age.
Overall, our systematic review indicates that milk leptin and milk adiponectin are the only hormones that have been consistently associated with infant growth, with higher concentrations of leptin typically being associated with fat deposition and lean body mass, and adiponectin being associated with reduced growth outcomes.

HMOs
In total, 13 studies (n ¼ 2640 dyads) examined the association between HMO composition and infant growth (Supplemental FIGURE 2. Summary of quality assessments of included articles.Quality scores are awarded based on the number of points assigned according to the criteria in Supplemental Table 1.Detailed numeric scores are presented in Supplemental Table 2. HM.HM, human milk.*Indicates the same study, but separate articles.Table 4, Figure 4).Some studies stratified all analyses on maternal secretor status [80,97,98,100], a major genetic determinant of HMO composition, whereas others did not stratify [48,53,62,86,92], and some performed both stratified and combined analyses [61,76,82,87].Classification and identification of HMOs were inconsistent among studies, and there was considerable variation in which HMOs authors examined.All studies used some form of liquid chromatography to analyze HMOs.Lagstrom et al. [80], Alderete et al. [48], and Larsson et al. [82] used HPLC after fluorescent derivatization, resulting in a uniform comparison of individual HMOs and consistent measures of composite HMOs, whereas Jorgensen et al. [76] and Davis et al. [62] used an LC-MS (nano-LC-chip/time-of-flight mass spectrometry) approach resulting in the identification of several previously unnamed and unstudied HMOs.This difference in methodology made comparison between studies difficult.For example, Tonon et al. [98] analyzed 16 HMOs which they indicated were most the abundant, whereas Davis et al. [62] analyzed 39 HMOs.Jorgensen et al. [76] used the same analytical techniques as Davis et al. [62]; however, the infant anthropometrics examined were not comparable with other studies (change in LAZ, HAZ, WLZ, and HCZ).FIGURE 3. Mean directions of associations between HM hormones and infant growth in the first 2 y.Significant associations reflect results as reported by individual study authors (for example, using HM concentrations as the predictor variable, see Table 1).Value in cells indicates the number of studies examining each comparison.Red squares indicate mean positive associations, blue squares indicate mean inverse associations, white squares indicate a mean association of 0, and black squares indicate that association was not assessed.HCA, head circumference-for-age; HM, human milk; LFA, length-for-age; WFA, weight-for-age; WFL, weight-for-length; BFA, BMI-for-age; FFM, fat-free mass; FMI, fat mass index.*Indicates that equal numbers of positive and negative associations were observed, resulting in a gradient of zero (0).

Maternal secretor status
All 13 studies measured maternal secretor status in their analysis of HMO profiles and infant growth; however, only 8 studies reported outcomes stratified by secretor status [80,97,98,100,61,76,82,87].Tonon et al. [98] , Binia et al. [53], and Wang et al. [100] further stratified their analyses by Lewis status, which is a functional dominant allele that is coded on the FUT3 gene on chromosome 19 and has a lesser known impact on HMO patterns in HM [109].Presence or near-absence of the HMO 2'Fucosyllactose (2'FL) was used across all studies to determine secretor status, with milk containing 2'FL being classified as secretor milk.Sprenger et al. [97] and Menzel et al. [87] compared infant growth between infants receiving secretor or non-secretor milk.Although Sprenger et al. [97] found no association between any infant growth parameters and secretor and non-secretor milk, Menzel et al. [87] found that infants who consumed milk from non-secretor mothers had significantly higher BMI at 3 mo and 6 mo.In addition, Menzel et al. [87] found statistically higher head circumference in infants consuming non-secretor milk at 3, 6, and 12 mo and that this difference was sustained until 7 y of age.These findings indicate a need to consider the secretor status of the mother when looking at growth outcomes overall.Notably, only Binia et al. [53] accounted for infant secretor status, which could be an important consideration because the infant's own secretor status can also impact microbiome development and other aspects of physiology [76].

Combined secretor and non-secretor milk-individual HMOs
Nine studies (n ¼ 1594 dyads) examined the association between individual HMO concentrations and infant growth with secretor and non-secretor samples combined (Figure 4A) [48,53,61,62,76,82,86,87,92].There were no consistent associations between individual HMOs and infant growth outcomes observed across studies.Although lacto-N-fucopentaose I (LNFPI) demonstrated an inverse association with infant weight, total fat mass, and lean body mass, this was only observed in 1 of 6 studies reporting on these associations [48].Conversely, Davis et al. [62], found a positive association with BMIZ, but only when looking at LNFP I and LNFP III in concert.Binia et al. [53] concluded that fat mass accretion and fat mass index did not appear to be related to a specific HMO and highlighted the importance of examining HMOs in composite groups rather than individually.
The majority of studies only analyzed HM component concentrations.However, Cheema et al. [61] also assessed the daily intake of HMOs, finding positive associations between 2'FL, 3'FL, difucosyllactose (DFLac), Sialyl-lacto-N-tetraose b (LSTb), and difucosyllacto-N-tetrose and infant growth, whereas these associations were not observed when simply measuring concentrations.

Secretor milk-individual HMOs
Eight studies (n ¼ 1928 dyads) reported results for individual HMO concentrations in secretor milk and resultant associations, with lacto-N-neotetraose (LNnT) and lacto-N-tetrose (LNT) demonstrating the most consistent (inverse) relationships with FIGURE 4. Mean directions of associations between HMOs and infant growth in the first 2 y, regardless of secretor status (A), or separately for secretor (B) and non-secretor (C) milk.Significant associations between HMOs and infant anthropometrics reflect results as reported by individual study authors (for example, using HM concentrations as the predictor variable, see Table 1).Value in cells indicates the number of studies examining each comparison.Red squares indicate mean positive associations, blue squares indicate mean inverse associations, white squares indicate a mean association of 0, and black squares indicate that association was not assessed.Note that some studies did not stratify on secretor status while others only presented stratified data (i.e. the studies in A are not all represented in B&C, and vice-versa).Despite using the 'near absence' of 2'FL to define non-secretor status, some studies still reported on 2'FL in this group, though no associations with infant growth were observed.HCA, head circumference-for-age; HM, human milk; HMO, human milk oligosaccharide; LFA, length-for-age; WFA, weight-for-age; WFL, weight-for-length; BFA, BMI-for-age; FFM, fat-free mass; FMI, fat mass index.*Indicates that equal numbers of positive and negative associations were observed, resulting in a gradient of zero (0).infant growth (Figure 4B: Secretor HM).In 3 of 6 studies [80,98,48], LNnT demonstrated inverse associations with infant length and in 2 of 6 studies [92], LNnT demonstrated inverse associations with infant weight.Furthermore, inverse associations were observed between LNnT and weight velocity [82], WAZ and weight-for-height Z-Score (WHZ) [109], and fat mass index [82].LSTb also demonstrated a consistent inverse association with infant weight across 2 of 3 studies reporting on this relationship [48].Finally, 6'sialyllactose (6'SL) demonstrated a potential negative association with infant growth, though only 1 of 5 studies observed this inverse association with weight, weight gain, and length [61].
Several positive associations were identified between secretor milk 2'FL, 3'FL, DFLac, lacto-N-hexaose, and difucosyllacto-Nhexaose concentrations and infant growth outcomes (Figure 4B).However, these positive associations were only identified in 1 study each and were not replicated in other studies.2'FL and DFLac were positively associated with infant weight [76], length [80,82], and weight velocity [82].Larsson et al. [82] also found an additional positive association between 2'FL and fat mass index.Finally, using the untargeted method of HMO identification Jorgensen et al. [76] found positive associations between unnamed HMOs 5130b and HMO 4240a and change in LAZ as well as a positive association between the composite of 5230a þ DFLNnO I/DFLNO II (Difucosyllacto-N-neooctaose I/difucosyllacto-N-octaose II) and change in WAZ.Wang et al. [100] identified unique positive associations between monofucosylpara-lacto-N-hexaose IV, fucosyl-para-lacto-N-hexaose I, trifucos yllacto-N-hexaose and infant length gain.
Mixed associations were observed between 3'SL and infant weight and length, with Lagstrom et al. and Larsson et al. [80,82] finding positive associations and Tonon et al. [98] finding negative associations.Although stratification among infants receiving only milk from secretor mothers showed increased significant associations with infant growth compared with combined analyses, these associations were only found in single studies for most individual HMOs.

Non-secretor milk-individual HMOs
Seven studies (n ¼ 1782 dyads) reported results for individual HMOs from infants receiving milk from non-secretor mothers (Figure 4C) [60,76,80,87,97,98,100], which make up a relatively smaller proportion of the population.Typically, 6%-24% of populations in European and North and Latin American Countries and as many as 29%-51% of populations in some South Asian countries are nonsecretors [111].Because of the lower proportion of nonsecretors in the population, many of these studies may not be sufficiently powered to detect statistical differences in non-secretor participants only.Again, associations between individual HMOs and infant growth outcomes were typically found in single studies and consistent associations between studies were not observed.Lagstrom et al. [80] and Jorgensen et al. [76] did not find any associations between individual HMOs in non-secretor milk and infant anthropometrics.Tonon et al. [98] and Lagstrom et al. [80] found that a concentration of 6'SL in HM had an inverse association with infant weight and Menzel et al. [87] found an inverse association with BMI.However, when Cheema et al. [60] accounted for intakes (not shown in heatmaps) instead of milk concentration, they found a positive association between 6'SL and infant weight.As such, it is hard to draw meaningful conclusions about the association between 6'SL and infant growth outcomes.

Composite HMO measures
Composite measures of HMOs, such as acidic or fucolsylated HMOs, tended to demonstrate inconsistent directional associations with infant growth across studies.HMO diversity showed inverse associations with several infant growth outcomes for both secretor-only and combined secretor and non-secretor milk (Figure 4A and B).Inverse associations were observed for HMO diversity and BMIZ [82], total fat mass [48], fat mass index [82], and percent fat [48] in analyses combining secretors and nonsecretors; whereas weight and length [76] demonstrated inverse associations when the analysis was restricted to just secretor milk.Acidic HMOs demonstrated positive associations when analyses combined secretors and nonsecretors, but inverse associations when analyses were restricted to only secretor milk.
Composite HMO measures also revealed an inverse association between total HMOs and infant weight in non-secretor milk [111,112].No other meaningful associations were observed between composite HMOs in non-secretor milk and infant growth outcomes across studies.Finally, Jorgensen et al. [76] found inverse associations between composite HMO measures of LNT þ LNnT and LNFP I þ III and change in WLZ in infants consuming secretor milk.
Jiang et al. [75] explored infant growth patterns by combining proteome, lipidome, and HMO data using principal component analysis (PCA) rather than looking at individual HMOs.Using this method, they found that the group high in LNnH (lacto-N-neothexaose)and LNDFH II (lacto-N-difucohexaose) and low in 3'SL was inversely associated with infant length-for-age (LFA) scores when adjusted for infant age, infant sex, birth weight, birth length, and maternal age.They also found that the factor high in LNnH , LSTa, LSTb, LSTc, 3'FL, and 2'FL and low in disialyllacto-N-tetraose was inversely associated with infant BMIZ.Noting that these factors also included proteomics and lipidomics, Jiang et al. [75] was the only study to look at multiple aspects of HM composition and infant growth outcomes using PCA.
The studies examining HMOs and infant growth included in this review had considerable variation in analytic techniques, study settings, and infant feeding status.For example, some studies [82] examined outcomes in exclusively breastfed infants whereas others did not account for breastfeeding exclusivity [48], which is an important predictor of HMO composition [113].In addition, the lactation stage when samples were collected can also impact HMO composition, with 6'SL concentrations dropping quickly and 3'FL increasing over time.Sample collection ranged from 2 d postpartum [75] to 9 mo postpartum [82], which could result in important variability in HMO concentrations.Finally, the setting is an important influential factor in HMO composition [76] and HMO studies ranged across all 6 habitable continents in this review.As such, it is difficult to draw clear conclusions about the role of HMOs on infant growth because of the likely and unaccounted variability in HMO composition.

Immunomodulatory components
Twelve studies (n ¼ 1422 dyads) were included in the immunomodulatory component category (Supplemental Table 5, Figure 5).Thirty-one anthropometrics were examined across the 12 studies, with weight and length being the most common.Bioactive components in the immunomodulatory category included cytokines, immunoglobulins, lactoferrin, lysozyme, and malondialdehyde.The area of investigation surrounding many of these HM components is still highly exploratory with studies using varied assays and minimal consensus on quality control.IL-6 (n ¼ 7; n ¼ 501 dyads) and tumor necrosis factor-alpha (TNF-α; n ¼ 5, n ¼ 441 dyads) were the most investigated bioactive components in this category.Interestingly, 10 of 12 studies were published after 2017, indicating that this is a relatively new area of exploration.
Nuss et al. [108] and Durilova et al. [63] both observed inverse associations between milk IL-6 concentrations and infant weight.Furthermore, Nuss et al. [108] found additional inverse associations between IL-6 and infant length, head circumference, and percent body fat.WAZ was the only growth outcome to demonstrate a positive association with milk IL-6 concentrations [93].Among the ILs, IL-6 was the only component where >1 study found similar directional (inverse) associations between milk concentrations and infant growth outcomes (weight) [108,63].Associations were only observed in one study each for all other IL milk components.
Six studies explored the relationship between milk TNF-α concentrations and infant growth.Inverse associations were observed between milk TNF-α and infant WFA [93] and infant head circumference [108]; whereas positive associations were observed for weight and percent fat mass [108].
Lactoferrin demonstrated mixed directional associations with infant growth within the same study.Gridneva et al. [71], observed a positive association with trunk fat mass and an inverse association with fat-free mass index, indicating that lactoferrin may have a positive association with overall infant body fat composition.Significant positive associations were found between infant growth outcomes and IL-1 β (BMIZ) [65], IL-4 (weight) [63], transforming growth factorbeta (weight) [49], and lysozyme (Fat-free mass, fat mass index, and total fat-free mass) [56].However, these significant associations were only observed in 1 study each and findings were not replicated elsewhere.

Metabolomics-untargeted analysis
Untargeted metabolomics analyses facilitate the observation of a broader range of compounds and report a relative abundance of molecules.Two studies [74,102] conducted untargeted metabolomics on bioactive components in HM and examined how these metabolites related to infant growth.In their case-control study of Chinese females with normal and gestational diabetic pregnancies, Wu et al. [102] identified 620 metabolites.Ten of these compounds (phosphocreatine, creatine, D-glutamic acid, N-methyl-D-aspartic acid, L-serine, phosphocholine, iditol, sorbitol, galactitol, and cytarabine) had significant inverse associations with infant weight gain and 2 (the lipids eicosatrienoic acid and lysophosphatidylcholine) had positive associations.Direct comparisons between Wu et al. [102], which looked at infant body weight gain as their infant growth metric, and Isganitis et al. [74], which examined infant fat mass percentage and fat accrual and excluded mothers with gestational diabetes, are obfuscated by differing growth outcomes as well as different chromatographic separation methods (liquid compare with gas chromatography).Indeed, when comparing the compounds in Wu et al. [102] to the compounds found to have significant associations with infant growth in Isganaitis et al. [74], there was no overlap of specific metabolites, which is not unexpected given the different data collection methods.However, at a broader metabolic level, there is evidence for similar biological effects in the creatine degradation pathway (negatively associated in both studies; phosphocreatine, creatine, and creatinine), and a general role for bioactive lipids-observed in both studies.A key component in fatty acid metabolism, carnitine, was negatively associated with infant fat accrual and infant fat mass percentage; however, it was positively associated with infant percent body fat.

Key findings
The findings from this systematic review revealed inconsistent associations between HM bioactive components and infant body composition in the first 2 y, highlighted inconsistent data collection methods and identified many knowledge gaps for future research.Among bioactive constituents of HM, the largest body of evidence (spanning >20 y, hundreds of studies, and over 5000 dyads) exists for leptin and adiponectin, showing consistent inverse associations with infant growth across studies, although several studies found no association.In contrast, research on other hormones, HMOs, and immunomodulators has primarily emerged in the last decade, with data often limited to a few hundred infants.There were no consistent associations found between individual HMOs and infant growth outcomes across studies, partially because of differences in laboratory methodology and data analysis strategies.However, examining HMOs in concert with each other revealed more consistent associations across studies.Among all immunomodulatory components, IL-6 appeared to demonstrate the most consistent association with infant growth, showing an inverse relationship.In addition, metabolomic analysis of HM is also a new area of exploration and could yield many important relationships between HM metabolites and infant growth.Targeted analyses of metabolites in milk indicate that bioactive lipids are a broad group of molecules with diverse physiological effects, which remain to be elucidated in HM.However, consistency of assays and methodological quality control is still of concern in these emerging areas when considering the validity of outcomes.As such, it is important to acknowledge these limitations when examining immunomodulatory components in HM as they are likely impacted by the different methodological strategies and may not demonstrate consensus among studies.
Moving forward, researchers aiming to examine the association between components in HM should continue to focus on appetite-regulating hormones while expanding exploration to other hormones and composite analysis of HMOs as the number of observations for these components are still quite small.Furthermore, the relationship between immunomodulatory components in HM and infant growth is still preliminary and warrants further research through both targeted and untargeted explorations.In addition, it is important to think beyond examining components within the silos of macronutrient, micronutrient, and non-nutrient categories and consider how these components work in combination with each other.

Improving the quality of HM research
Our ability to synthesize and interpret research findings was tempered by substantial variation in HM sampling times and collection strategies.Inconsistent findings related to HMO composition and infant growth were similarly related to FIGURE 5. Mean directions of associations between HM immunomodulators and infant growth in the first 2 y.Significant associations between immunomodulators and infant anthropometrics reflect results as reported by individual study authors (for example, using HM concentrations as the predictor variable, see Table 1).Value in cells indicates the number of studies examining each comparison.Red squares indicate mean positive associations, blue squares indicate mean inverse associations, white squares indicate a mean association of 0, and black squares indicate that association was not assessed.HCA, Head circumference-for-age; HM, human milk; LFA, Length-for-age; WFA, weight-for-age; WFL, weight-forlength; BFA, BMI-for-age; FFM, fat-free mass; FMI, fat mass index.*Indicates that equal numbers of positive and negative associations were observed, resulting in a gradient of zero (0).inconsistent sampling time points and reporting in a recent review of microbiome-related products in HM and infant growth [115].It is well established that HM composition changes throughout the day and over the course of lactation for some components [116].Previous research examining the impact of circadian rhythms on bioactive components in HM has been inconsistent; likely because of methodological issues with milk collection and sampling strategies [116].The impact of temporal variance in milk composition should be considered when developing milk collection strategies.Consistent collection times across studies that align longitudinally with the lactation stage and temporally with circadian rhythms will allow for more consistent comparisons of findings.
Extensive variation in the anthropometric outcomes measured across studies further complicated this review.Most studies examined length and weight; however, an additional 24 different anthropometric measures were also reported across studies, making it difficult to synthesize results.Reliability and reproducibility have long plagued infant growth research [117] and our results demonstrate that this issue persists with extensive variation even among many studies published within the last 5 y.Clinicians and researchers currently have over 100 growth or size charts from which to reference based on their location and local guidelines [118] and there are no universally accepted guidelines to identify clinically important growth trajectories [117].This lack of consistency and guidance complicates the identification of meaningful research outcomes and the resultant translation to clinical practice.In our systematic review, only 4 studies [59,64,81,97] explicitly stated they were following WHO Child Growth Standards [113], to guide their anthropometric outcomes which are considered to be the gold standard for assessing child growth [117,119].Over 25 different anthropometric measures were reported across all included studies which challenged our ability to conduct a meta-analysis and compare outcomes across studies.Aris et al. [117] highlights the importance of using longitudinal growth trajectories, specifically BMIZ, rather than cross-sectional assessments to identify clinically meaningful trajectory patterns, risk for obesity, and adverse health outcomes.Consistent reporting of anthropometric measures, such as BMIZ will align research outcomes in a more understandable and meaningful way so that researchers and clinicians can guide their practice based on consistent outcomes.
Specific to studies examining HMOs, stratification by secretor status is an important consideration necessary to identify certain associations between HMOs and infant growth.Larsson et al. [82] and Jorgensen et al. [76] presented results stratified by secretor status and with combined secretor status; both identified associations for secretor-positive milk that were not present when combining secretors with nonsecretors.These differences indicate the vital importance of stratifying results by secretor status and to adequately power studies for stratified analysis.The extent to which secretor status impacts HM composition is still being explored and may extend beyond HMOs.As such, stratification by secretor status in future analyses may be warranted to better understand this relationship and the subsequent impact on child growth.Furthermore, Larsson et al. [82] were not able to examine the relationships between HMO composition and infant growth in non-secretor mothers because of the small sample size.Although nonsecretors only represent less than half of the world population [120], limiting research findings to only combined and secretor milk impacts the generalizability of HM composition research.Consequently, a greater emphasis needs to be placed on recruiting an adequate number of nonsecretor mothers to conduct sufficient analyses and make recommendations for this portion of the population.
All studies used the presence of 2'FL as an indicator of secretor status as a phenotypic indicator of secretor status.Although identifying 2'FL in HM is feasible and the required technology is becoming more accessible through point-of-care testing [121], there remain many scenarios where genetic information is used to determine secretor status.In these instances, there may be added value to distinguishing between heterozygous and homozygous secretors because differences may exist for some HMOs between heterozygous and homozygous secretors.Finally, findings from this review indicate that infant secretor status is rarely considered when examining HM composition and child growth.Only 1 study included in this review [98] accounted for infant secretor status and there is a general paucity of research examining the role of infant secretor status in child health and developmental outcomes.Considerations of both maternal and infant physiology (including secretor status) and the interplay with HM as a triadic relationship will provide better insight toward recommended practices and therapeutic concepts related to HM composition [120].
Moving research forward: the need to study HM from a systems perspective Although the full impact of cytokines in HM is yet to be fully understood, it is believed that cytokines mediate infant immune responses through interactions with epithelial receptors [122].In addition, cytokines are believed to act in concert with other cytokines, rather than in isolation.Similarly, it is believed that HMOs act in combination with each other [122] as well as individually.Furthermore, different types of bioactive constituents (for example, HMOs and cytokines) might also interact.In addition, some HM constituents are consumed or modified by the infant's gut microbiota [123].This underscores the importance of examining HM bioactive components as a network of interactions and pathways, from an ecological or biological systems perspective to develop a more comprehensive understanding of how HM components interact with each other to inform growth and subsequent health outcomes [124,125].Yet very few studies took this approach.Jiang et al. [75] examined the combined role that the proteome, lipidome, and HMO composition in HM has on infant growth using PCA.Only 2 studies attempted integrated analyses of different HMOs [48,76] while 2 other studies performed an untargeted metabolomic analysis [74,102].With the exception of Jiang et al. [75], Alderete et al. [48] who included HM insulin into their HMO model and Jorgensen et al. [76] who explored HMOs and bioactive proteins in concert with each other, no studies integrated different classes of milk bioactives.Conducting these comprehensive analyses and looking at HM from a systems perspective requires investment from scientists and funders to develop assays to optimize milk volume, to collect samples using consistent and validated strategies, and to include large sample sizes from diverse populations across the globe.Echoing Reyes et al. [6], there is a need for large high-quality studies to better understand how HM components work independently and together to influence infant growth and health outcomes.

Strengths and limitations
Across 3 studies [6,7], we have comprehensively synthesized the available evidence for HM components and child anthropometric measures in the first 2 y.The main limitation of our review was the inability to conduct meta-analyses because of incomplete reporting practices and variability in sampling times and outcome measurements.Although meta-analyses are considered to be the gold standard, using SWiM [25] as a reporting guideline is an acceptable synthesis method in the absence of meta-analyses and allows us to synthesize and present study findings that would otherwise be left unreported.Individual studies included in this review also had limitations; only 5 of 75 achieved a high-quality score.Most did not adequately control for confounding (maternal BMI, birth anthropometrics, time postpartum, and HM exclusivity) and many did not provide results for all examined outcomes.We assumed that no relationship existed when an outcome was not reported, which limited our understanding and ability to incorporate these findings into the overall review.Finally, most studies measured bioactive concentrations, rather than calculated intakes.Assessing bioactives in milk using concentrations from 1 feed does not appropriately indicate the overall intake of these components by an infant over time and can lead to measurement bias.

Conclusions
Bioactive components in HM are increasingly being examined as important influencers of child growth.Although our findings in this review were largely inconsistent, general trends were observed for HM adiponectin and leptin, demonstrating inverse relationships with infant growth.In addition, in secretor mothers, the HMO LNnT consistently demonstrated inverse trends with child growth outcomes, whereas 2'FL demonstrated positive associations.No immunomodulatory components demonstrated consistent directional relationships with infant growth outcomes.
Our synthesis of this literature was limited by methodological issues with milk collection strategies, timing of milk collection, inconsistent anthropometric measures that were not aligned with WHO guidelines, and insufficient reporting of findings.Moving forward, HM researchers need to be cognizant of aligning research strategies with clinically meaningful growth outcomes, developing HM sampling strategies that reflect the temporal fluctuations in composition, capturing daily HM intake, and applying analysis strategies that acknowledge and investigate HM as a complex biological system.

), growth velocity between 3 mo and 1 y
Nonsecretors: (À) Association for height and LNT at 2 y (No) Association for height and 2'FL, 3'-FL, 3 0 SL, 6 0 SL, LNnFP, or LNFP-V.(À) Association for growth velocity and LNnT at 3 mo-1 y and 1 y -2 y (À) Association for BMI-SDS and LNFP-V at 3 mo, 6 mo, 1 y and 2 y. (À) Association for BMI-SDS and 6'SL at 3 mo, and 1 y (À) Association for LNFP-V and HC at 3 mo, 1 y and 2 y. (þ) Association for LNnFP and HC at 2 y Secretors: (þ) Association for height and LNFP I at 3 mo, 6 mo and 12 mo (No) Association for height and 2'FL, 3'FL, 3 0 SL, 6 0 SL, LNnFP, or LNFP-V.(À) Association for growth velocity and LNFP I at 1 y -2 y (No) Association for growth velocity and 2'FL, (À) Association for BMI-SDS and LNT and LNFP-V at 2 y Secretor status, maternal prepregnancy weight and height and infant birth parameters (continued on next page) M.(M.Brockway et al.Advances in Nutrition 15 (2024) 100127 , HAZ, HC (No) Significant association was found for weight for Z-score of infants and TAC and MDA levels in HM Infant birth weight, Association for leptin and weight (À) Association for leptin and HC (À) Association for leptin and % fat mass (No) Association for leptin and length (À) Association for Insulin and weight (À) Association for Insulin and HC (À) Association for Insulin and % fat mass (No) Association for Insulin and length (þ) Association for TNF-α and weight (À) Association for TNF-α and HC (þ) Association for TNF-α and % fat mass (No) Association for TNF-α and length Infant age at visit (continued on next page) M.(M.Brockway et al.Advances in Nutrition 15 (2024) 100127

TABLE 1
Detailed characteristics and results of included studies reporting on human milk bioactives and infant anthropometrics (organized alphabetically by first author)

TABLE 1
Association for weight gain during first 15 d and leptin at 15 d (À) Association for weight gain during first mo and leptin at 15 d IL-4, IL-6, IL-10, IL-17, IL-18, IL-23, interferongamma (IFN-gamma) and transforming growth factor beta 1 (TGF-β1) (concentrations) Unclear body weight (þ) Association for IL-4 and body weight (À) Association for IL-6 and body weight (þ) Association for milk insulin and WFA from 2 wk to 6 mo and HCAZ change from 2 wk to 2 mo in infants receiving any type of nutrition (þ) Association for IL-1beta and LAZ in infants at V4 (þ) Association for IL-8 and BMIZ at V7 (þ) Association for IL-1beta and BMIZ at V7 (þ) Association for leptin and lean mass at V4 (À) Association for leptin and % fat mass at V4 (À) Association for leptin and BMIZ at V1 (þ) Association for leptin and % lean mass at V1 and V4 (þ) Association for cytokines and WAZ at birth (À) Association for cytokines and WAZ for 1 mo-(continued on next page) M.(M.Brockway et al.Advances in Nutrition 15 (2024) 100127

TABLE 1
Association for mo 1 leptin levels and mo 6 body length, % fat, total fat mass, and trunk fat mass (No) Association for mo 1 leptin levels and total fat-free mass (No) Association for mo 1 insulin levels and mo 6 body length, % fat, total fat mass, total fat-free mass and trunk fat mass (No) Association for mo 1 TNF levels and mo 6 body length, % fat, total fat mass, total fat-free mass and trunk fat mass (No) Association for mo 1 IL6 levels and mo 6 body length, % fat, total fat mass, total fat-free mass and trunk fat mass (À) Association for cGP and weight Z at 13 mo (þ) Association for IGF-1:cGP ration and weight Z at 13 mo (À) Association for IGF-1:cGP ratio and BMIZ at 3 and 5 y (þ) Association for cGP and BMIZ at 5 y (þ) Association for lysozyme CDI at 12 mo and decrease in FFMI for 5 and 12 mo (À) Association for lactoferrin CDI and FFMI (ultrasound) at 12 mo (þ) Association for lysozyme CDI and Fat mass (ultrasound) and FMI (ultrasound) at 12 mo (À) Association for CDI adiponectin and lean body mass.(þ) Association for CDI adiponectin and adiposity (þ) Association for CDI skim milk leptin and adiposity (No) Association for HM adiponectin and weight gain.Higher CDI of skim milk leptin was associated with a lower accrual of FFM over 12 mo (À) Association for lactoferrin concentration and visceral depth (þ) Association for LActose, total carbohydrates and total protein (intakes) and subcutaneous abdominal fat area (no other significant associations)

TABLE 1
endopeptidase activity, signal transduction, immune response and low in β-casein; high in Phospholipids including PE, PC, Ceramide, SM, TG-SU2, TG-UUU and low in TG-SSS; high in LNnH, LNDFH II and low in 3'SL) (À) Association for Factor 1 and LAZ Factor 2: (high in Vitronectin, CD81 molecule, complement C4A, fibroblast growth factor binding protein 1, milk fat globule-EGF Factor 8 protein, immunoglobulin heavy constant gamma 4 and low in as1casein; high in TG-SSS and low in Phospholipids including PE, PC, SM; high in LNnH, LSTa, LSTb, LSTc, 3'FL, 2'FL and low in DSLNT) (À) Association for Factor 2 and BAZ and WAZ Factor 3: (high in no proteins and low in 27 proteins related to platelet degranulation, endopeptidase activity, innate immune response; high in Phospholipids including PE, PC, PI, LPC, LPE, LPI, SM, Ceramide and low in TG-S2U; high in LSTc, LNFP II and low in 6'Sl and 2'FL) (þ) Association for Factor 3 and LAZ.For secretors only: In primary analyses: (þ) Association for absolute abundance of all HMOs and delta LAZ.In exploratory analyses: (À) Association for LNT þ LNnT and delta WLZ.For nonsecretors only: In primary analyses: (No) Significant associations for bioactive proteins or groups of HMOs and growth.In exploratory analyses: (À) Association for LNFP II and delta WLZ.(À) Association for unnamed HMO 4120a and delta WLZ In secretors: (À) Association for HMO diversity and LAZ and WAZ during first y (À) Association for HMO diversity and LAZ at 1 y and 5 y (þ) Association for 2'FL and LAZ for 3 mo and 12 mo, and 1 y and 5 y (þ) Association for 2'FL and WAZ for 3 mo and 12 mo (À) Association for LNnT and weight and LAZ throughout first 5 y Maternal secretor status, maternal prepregnancy BMI, infant sex, birthweight zscore, time point (that (continued on next page) M.(M.Brockway et al.Advances in Nutrition 15 (2024) 100127