Introduction
The World Health Organization recommends exclusive human milk feeding for the first six months of life, followed by continued human milk feeding alongside appropriate complementary foods for up to two years or longer(1). During this critical developmental period, human milk provides a uniquely tailored source of nutrition that supports infant growth, immune development, and long-term health(Reference Andreas, Kampmann and Mehring Le-Doare2). Human milk is not a static substance; it is a dynamic, bioactive fluid that evolves in composition across lactation stages to meet the changing physiological needs of the infant(Reference Ballard and Morrow3).
The progression of human milk follows three phases: colostrum, transitional milk, and mature milk. Colostrum, produced from the second trimester through the first few postpartum days, is rich in immunological components such as secretory Immunoglobulin A (sIgA), lactoferrin, and leukocytes, along with growth factors and proteins essential for gut and immune system maturation(Reference Ballard and Morrow3). Transitional milk follows between days 5 and 14 postpartum, with increasing volume and changing nutrient composition, eventually giving way to mature milk, which constitutes the predominant form of milk from approximately day 15 onward(Reference Ballard and Morrow3).
The health benefits of human milk feeding have been well documented for both infants and mothers. For infants, human milk may offer protection against infections, obesity, and non-communicable diseases such as hypertension and type 2 diabetes mellitus, with evidence of dose-dependent effects(Reference Kelishadi and Farajian4). Mothers also experience reduced risks of postpartum obesity, breast cancer, and ovarian cancer(Reference Alotiby5). While infant formula provides an alternative feeding option, it lacks the dynamic and individualised qualities of human milk, which cannot be fully replicated(Reference Martin, Ling and Blackburn6).
Human milk composition is influenced by a wide range of factors, including maternal diet, mode of delivery, lactation stage, infant sex, and metabolic conditions such as gestational diabetes mellitus (GDM)(Reference Mantziari and Rautava7). GDM is a common pregnancy complication characterised by glucose intolerance with onset of first recognition during pregnancy, affecting up to 18% of pregnancies globally depending on the diagnostic criteria used(Reference Choudhury and Devi Rajeswari8). Beyond its well-known implications on maternal and foetal outcomes during pregnancy(Reference Choudhury and Devi Rajeswari8), GDM may have lasting effects on maternal metabolism and milk biosynthesis.
The mammary gland is a highly dynamic endocrine-metabolic organ that undergoes extensive structural and functional maturation across pregnancy and lactation in response to coordinated hormonal and metabolic signalling. Metabolic hormones, particularly insulin, interact with classical lactogenic hormones to support mammary differentiation and nutrient partitioning, while glucose metabolism is central to lactation, as glucose serves as the primary precursor for lactose synthesis and mammary glucose uptake increases markedly from late pregnancy into established lactation via upregulation of glucose transporters such as GLUT1(Reference Neville, Anderson, McManaman, Badger, Bunik and Contractor9,Reference Mueckler and Thorens10) .
GDM is characterised by systemic insulin resistance and altered glucose metabolism during the critical window of secretory activation. Because lactogenesis requires tightly coordinated endocrine-metabolic signalling, the insulin resistance and hyperglycaemia of GDM may disrupt insulin-dependent pathways within mammary epithelium, contributing to delayed differentiation and secretory activation(Reference Neville, Anderson, McManaman, Badger, Bunik and Contractor9,Reference Nommsen-Rivers, Dolan and Huang11) . Consistent with this physiology, women with GDM have an increased risk of delayed onset of lactogenesis II and reduced colostrum supply compared with women without GDM(Reference Nommsen-Rivers, Dolan and Huang11,Reference Wu, Pang, Jiang, Zheng, Han and Zhang12) . Together, these findings support a model in which GDM alters mammary adaptations by perturbing hormonal and metabolic networks essential for efficient milk production.
Existing reviews have largely examined the effects of GDM on specific components of human milk, such as macronutrients and metabolic hormones(Reference Suwaydi, Zhou, Wlodek, Lai, Gridneva and Geddes13,Reference Qin, Li and Wang14) . Over the past several years, there has been a rapid expansion of research examining a broader spectrum of milk constituents – including macronutrients, bioactive lipids, oligosaccharides, immunological factors, and the milk microbiome – in relation to GDM. The objective of this systematic review and meta-analysis is to evaluate and meta-analyse the current evidence on how GDM affects human milk composition across all stages of lactation. By synthesizing findings from a diverse body of studies, this review aims to provide a comprehensive overview of whether and how GDM affects human milk, and to identify consistent patterns or gaps in the current literature. Given the critical role of human milk in shaping early-life development and long-term health, understanding the links between maternal metabolic status and human milk composition is essential for informing targeted nutritional strategies, clinical guidelines, and public health interventions.
Methods
This systematic review and meta-analysis were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis Protocols (PRISMA-P) and registered in the PROSPERO database (CRD420250596663).
Exposure and outcome measurements
The exposure is GDM diagnosis, while the primary outcome of interest is any human milk component, including macronutrients, lipids, oligosaccharides, hormones, and bioactive peptides, inflammatory and immune factors including cytokines and growth factors, micronutrients and antioxidants, metabolomic and proteomic profiles, microbiomes, and micro ribonucleic acids (microRNAs/miRNAs).
Search strategy
A search of five databases was performed on 25 October 2024, with no prior date limits. Databases included CINAHL, EMBASE, Medline, Scopus, and Dissertations & Theses (ProQuest). The specific search string of the search terms was modified as necessary for each database. The terms for the diabetes concept were combined as follows: MH (‘Diabetes Mellitus, Gestational’) OR TI,AB (pregnant* or gestation* or maternal) N3 (glucose intolerance) OR (TI,AB pregnant* N3 diabetes*) OR TI,AB (‘gestational diabetes’ or ‘GDM’ or ‘gestational diabetes mellitus’ or ‘diabetes in pregnancy’ or ‘pregnancy diabetes’ or ‘pregnancy-related diabetes’). These terms were added, using the Boolean ‘AND’, to the following terms for the human milk concept: TI, AB (breastmilk or milk or breastfeeding or breastfed or ‘breast fed’ or lactate* or colostrum or ‘liquid gold’ or foremilk or hindmilk). Two reviewers (SB, AG) independently screened study titles, abstracts, and full texts using Covidence. Any discrepancies were resolved by discussion, and arbitration by a third assessor (MK) as required.
Eligibility criteria
Inclusion criteria were human studies, published in English, examining differences in human milk composition based on glycaemic state among adults (18+ years of age) who delivered full-term. Exclusion criteria were adolescents (<18 years of age), pre-term delivery, complications other than GDM that may affect human milk composition, types of diabetes other than GDM, and studies published in a language other than English.
Data extraction
Data extraction was completed independently by four reviewers (SB, AG, JD, MK). Any discrepancies were resolved by discussion, and arbitration by the fourth assessor (MK) as required. Variables included:
-
1. Author(s) and year of publication;
-
2. Study location;
-
3. Inclusion and exclusion criteria;
-
4. Sample size (total and by group (with and without GDM));
-
5. Timing of GDM diagnosis;
-
6. Timing of milk collection;
-
7. Lactation stage;
-
8. Method of milk collection;
-
9. Primary outcome(s);
-
10. Main finding(s); and
-
11. Quality scoring.
Quality assessment
Each study was independently assessed by three reviewers (SB, AG, JD) using the Newcastle-Ottawa Scale(15,16) . Quality assessment scales were used for two different types of studies: cross-sectional studies and cohort studies. Any discrepancies were resolved by discussion, and arbitration by a fourth assessor (MK) as required.
Data synthesis and analysis
Findings from full text papers were synthesised narratively using a descriptive and comparative summary. Basic descriptive statistics were calculated to summarise study characteristics. A set of meta-analyses was performed. Only studies with requisite data were included in the meta-analyses. Meta-analyses using a fixed-effects model were completed if heterogeneity was not too high (I 2 < 75%); a random-effects model was employed if heterogeneity was ≥75%. Standardised mean differences and standard deviations were employed as a summary statistic when the trials assessed the same outcome using different measurements; otherwise, raw mean differences and standard deviations were used with 95% confidence intervals (CIs).
Results
Characteristics of selected studies
Following the removal of duplicates, 4787 titles and abstracts were eligible for screening, of which 71 underwent full-text review. A total of 52 full-text articles with 3233 participants were eligible for inclusion. Of the 52 articles, nine had data for inclusion in the meta-analyses(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17–Reference Azulay Chertok, Haile, Shuisong and Kennedy25). A flowchart of the article selection process is shown in Figure 1.
PRISMA 2025 flow diagram illustrating published studies and abstracts included in the systematic review.
Fig. 1. Long description
The flowchart begins with the identification stage, where studies are sourced from databases and registers, including Scopus, Embase, MEDLINE, CINAHL, and citation searching, totaling 7675 studies. No references are added from other sources. 2888 references are removed due to duplicates identified manually and by Covidence. In the screening stage, 4787 studies are screened, with 4716 studies excluded. 71 studies are sought for retrieval, and all are retrieved. These 71 studies are assessed for eligibility, with 19 studies excluded for various reasons such as wrong setting, outcomes, comparator, indication, intervention, study design, and patient population. Finally, 52 studies are included in the review.
Of the 52 studies, 41 (78.85%) studies had a longitudinal cohort design(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17,Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Wang, Liu, Li, Xu, Wang and Lu22–Reference Dou, Luo, Xing, Liu, Chen and Zhu59) , six (11.54%) studies had a cross-sectional design(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21,Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60–Reference Resende, Clemente, Bezerra, Grilo, de Melo and Bellot64) , and five (9.62%) studies included grey literature (abstracts)(Reference Nteka, Liosis, Kastanidou, Papgaroufalis, Mastorakos and Lykeridou65–Reference Kotilnska68) and theses(Reference Salguero19). Geographically, the majority of studies were conducted in China (n = 17; 32.69%)(Reference Wang, Liu, Li, Xu, Wang and Lu22–Reference Zhang, Zhang, Yan, Duan, Wang and Szeto24,Reference Tao, Wang, Xiao, Li, Wang and Mao40–Reference Zhong, Zhang, Xia, Zhu, Chen and Shan47,Reference Ma, Lu, Huang, Guo, Xiang and Gao52,Reference Wen, Wu, Yang, Han, Wang and Fu55,Reference Wu, Yu, Liu, Wang, Chen and Qiao56,Reference Yu, Rong, Sun, Ding, Wan and Zou58,Reference Dou, Luo, Xing, Liu, Chen and Zhu59,Reference Li, Ning, Rui, Wang, Lei and Yu63) , followed by North America (n = 10; 19.23%)(Reference Salguero19,Reference Azulay Chertok, Haile, Shuisong and Kennedy25,Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26,Reference Churchill, Zawawi, Elisia, Seider, Noseworthy and Thompson29,Reference Gámez-Valdez, García-Mazcorro, Montoya-Rincón, Rodríguez-Reyes, Jiménez-Blanco and Rodríguez31,Reference Grapov, Lemay, Weber, Phinney, Azulay Chertok and Gho35,Reference Shah, Fields, Pezant, Kharoud, Gulati and Jacobs38,Reference Smilowitz, Sarah, Jincui, Dmitry, Holiday and Carol39,Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48,Reference Fradet, Castonguay-Paradis, Dugas, Perron, St-Arnaud and Marc61) , Europe (n = 9; 17.31%)(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21,Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz32,Reference Metallinou, Lykeridou, Karampas, Liosis, Skevaki and Rizou36,Reference Rold, Guldbæk, Lindegaard, Kirk, Nygaard and Bundgaard-Nielsen37,Reference Klein, Bancher-Todesca, Graf, Garo, Roth and Kautzky-Willer49,Reference Peila, Riboldi, Spada, Coscia, Barbagallo and Volti53,Reference Nteka, Liosis, Kastanidou, Papgaroufalis, Mastorakos and Lykeridou65,Reference Kotilnska68) , Turkey (n = 7; 13.46%)(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17,Reference Aydin, Kuloglu and Aydin28,Reference Aydin33,Reference Aydin, Celik, Gurates, Sahin, Ulas and Yilmaz50,Reference Ustebay, Baykus, Deniz, Ugur, Yavuzkir and Yardim54,Reference Yavuz, Aydin, Ugur, Aydin, Senol and Baykus57,Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60) , Greater Middle East (n = 4; 7.69%)(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Fatima, Rehman, Muhammad, Martins, Mohammed and Khan30,Reference Fatima, Khalid, Ladak and Ali34,Reference Azulay Chertok, Haile, Eventov-Friedman, Silanikove and Argov-Argaman51) , Brazil (n = 4; 7.69%)(Reference Avellar, Oliveira, Caixeta, Souza, Teixeira and Faria27,Reference Gomes Fagundes, França, da Silva Fernandes, Hara, Morceli and Honorio-França62,Reference Resende, Clemente, Bezerra, Grilo, de Melo and Bellot64,Reference Cunha67) , and Africa (n = 1; 1.92%)(Reference Li66).
Quality assessment scores were completed based on the type of study; that is, cohort or cross-sectional studies. For the cohort studies, 30 were rated as high quality(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17,Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Wang, Liu, Li, Xu, Wang and Lu22,Reference Zhang, Zhang, Yan, Duan, Wang and Szeto24,Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26–Reference Aydin, Kuloglu and Aydin28,Reference Fatima, Rehman, Muhammad, Martins, Mohammed and Khan30,Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz32–Reference Fatima, Khalid, Ladak and Ali34,Reference Metallinou, Lykeridou, Karampas, Liosis, Skevaki and Rizou36–Reference Wang, Li, Li, Li, Liu and Lu41,Reference Yao, Shen, Zhang, Tang, Yu and Tu43,Reference Zhang, Zhang, Han, Chen, Liu and Sun45–Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48,Reference Aydin, Celik, Gurates, Sahin, Ulas and Yilmaz50–Reference Ma, Lu, Huang, Guo, Xiang and Gao52,Reference Ustebay, Baykus, Deniz, Ugur, Yavuzkir and Yardim54–Reference Wu, Yu, Liu, Wang, Chen and Qiao56,Reference Yu, Rong, Sun, Ding, Wan and Zou58,Reference Dou, Luo, Xing, Liu, Chen and Zhu59) , and 11 as moderate(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Azulay Chertok, Haile, Shuisong and Kennedy25,Reference Churchill, Zawawi, Elisia, Seider, Noseworthy and Thompson29,Reference Gámez-Valdez, García-Mazcorro, Montoya-Rincón, Rodríguez-Reyes, Jiménez-Blanco and Rodríguez31,Reference Grapov, Lemay, Weber, Phinney, Azulay Chertok and Gho35,Reference Yao, Shen, Yu, Tang, Zhang and Xu42,Reference Yu, Wu, Shen, Yao, Tang and Song44,Reference Klein, Bancher-Todesca, Graf, Garo, Roth and Kautzky-Willer49,Reference Peila, Riboldi, Spada, Coscia, Barbagallo and Volti53,Reference Yavuz, Aydin, Ugur, Aydin, Senol and Baykus57) . Of the six cross-sectional studies, one was rated as very good quality(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21), four as good quality(Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60–Reference Gomes Fagundes, França, da Silva Fernandes, Hara, Morceli and Honorio-França62,Reference Resende, Clemente, Bezerra, Grilo, de Melo and Bellot64) , and one as satisfactory quality(Reference Li, Ning, Rui, Wang, Lei and Yu63). Excluded from the quality assessment were the four grey literature abstracts or posters, due to insufficient information for assessment(Reference Nteka, Liosis, Kastanidou, Papgaroufalis, Mastorakos and Lykeridou65–Reference Kotilnska68). However, Salguero (2009), a thesis with a longitudinal cohort design, was assessed based on its study design and rated as high quality(Reference Salguero19). Further details are provided in Table 1.
Characteristics of included studies
Table 1. Long description
A table detailing the quality assessment scores of included studies. The table has 10 rows and 3 columns. Column headers are Study Type, High Quality, Moderate Quality, Good Quality, Satisfactory Quality, and Very Good Quality. Row labels include Cohort Studies, Cross-Sectional Studies, and Grey Literature. Row 1: Cohort Studies, High Quality, 30, Moderate Quality, 11. Row 2: Cross-Sectional Studies, Very Good Quality, 1, Good Quality, 4, Satisfactory Quality, 1. Row 3: Grey Literature, Excluded, 4. Row 4: Thesis, High Quality, 1. The table summarizes the quality ratings of different types of studies, indicating the number of studies rated as high, moderate, good, satisfactory, or very good quality.
Only data relevant to the research question of this systematic review and meta-analysis are provided. Abbreviations: 2′FL: 2′-fucosyllactose; 2-AG: 2-arachidonoylglycerol; 2-DHG: 2-docosahexaenoyl-glycerol; 2-DPG: 2-docosapentaenoyl-glycerol; 2-EPG: 2-eicosapentaenoyl-glycerol; 2-LG: 2-linoleoyl-glycerol; 2-OG: 2-oleoyl-glycerol; 3′FL: 3′-fucosyllactose; 3′GL: 3′-galactosyllactose; 3′SL: 3′-sialyllactose; 6′SL: 6′-sialyllactose; AEA: N-arachidonoylethanolamine (anandamide); BMI: body mass index; CRP: C-reactive protein; DFLNH: difucosyl-lacto-N-hexaose; DHA: docosahexaenoic acid; DHEA: N-docosahexaenoyl-ethanolamine; DPEA: N-docosapentaenoyl-ethanolamine; DSL: disialyllacto-N-tetraose; DSLNT: disialyl-lacto-N-tetraose; EPA: eicosapentaenoic acid; EPEA: N-eicosapentaenoyl-ethanolamine; GDM: gestational diabetes mellitus; HMO(s): human milk oligosaccharide(s); IFN-γ: interferon-gamma; IGF-1: insulin-like growth factor 1; IGFBP-2: insulin-like growth factor–binding protein 2; IgA: immunoglobulin A; IgG: immunoglobulin G; IgM: immunoglobulin M; IL: interleukin; LEA: N-linoleoyl-ethanolamine; LNDFH-I: lacto-N-difucohexaose I; LNFP: lacto-N-fucopentaose; LNnT: lacto-N-neotetraose; LNT: lacto-N-tetraose; LPG: lysophosphatidylglycerol; LPI: lysophosphatidylinositol; LST: sialyl-lacto-N-tetraose; MFGM: milk fat globule membrane; MFLNH: monofucosyl-lacto-N-hexaose; MFLNnH: monofucosyl-lacto-N-neohexaose; miRNAs: microRNAs; MMP-9: matrix metalloproteinase-9; MUFAs: monounsaturated fatty acids; NGAL: neutrophil gelatinase-associated lipocalin; OEA: N-oleoyl-ethanolamine; OPL: 1-oleoyl-2-palmitoyl-3-linoleoylglycerol; PA: phosphatidic acid; PC: phosphatidylcholine; PE: phosphatidylethanolamine; PG: phosphatidylglycerol; PS: phosphatidylserine; PUFAs: polyunsaturated fatty acids; SFA: saturated fatty acid; SIgA: secretory immunoglobulin A; SREBP-1c: sterol regulatory element–binding protein 1c; TGF-β: transforming growth factor-beta; Th: T helper; TNF-α: tumor necrosis factor-alpha.
Meta-analyses
Colostrum
Lactose in colostrum: Differences in lactose concentrations in colostrum from mothers with and without GDM were examined in two studies(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Azulay Chertok, Haile, Shuisong and Kennedy25) . Both studies reported significantly lower lactose concentrations in the GDM group compared with controls. The pooled analysis across both studies (n = 195) revealed a statistically significant overall mean difference of −31.41 mmol/L (95% CI: −54.67 to −8.15), favouring lower values in the GDM group (Z = 2.65, P = 0.008, I 2 = 0%) (Figure 2).
Meta-analysis of the effects of GDM on lactose in colostrum.
Ghrelin in colostrum: Differences in ghrelin levels in colostrum from mothers with and without GDM were examined in three studies(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17,Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21) . The pooled analysis across all three studies (n = 97) yielded an overall mean difference of −0.03 ng/ml (95% CI: −0.05 to −0.00), indicating no statistically significant differences in the colostrum of mothers with and without GDM (Z = 2.02, P = 0.04) (Figure 3). Nonetheless, very high heterogeneity was observed (I 2 = 97%).
Meta-analysis of the effects GDM on ghrelin in colostrum.
Fig. 3. Long description
A table comparing the effects of GDM on ghrelin levels in colostrum across different studies. The table has 4 rows and 10 columns. Column headers are: Study or Subgroup, GDM Mean, GDM SD, GDM Total, Non-GDM Mean, Non-GDM SD, Non-GDM Total, Weight, Mean difference IV, Random, 95% CI. Row 1: Aydin 2010, 0.0277, 0.002, 10, 0.0392, 0.002, 10, 49.6%, -0.01 [-0.01, -0.01]. Row 2: Aydin et al. 2007, 0.0261, 0.007, 12, 0.0639, 0.009, 14, 48.3%, -0.04 [-0.04, -0.03]. Row 3: Lis-Kuberka et al. 2024, 0.33, 0.31, 30, 0.44, 0.31, 21, 2.1%, -0.11 [-0.28, 0.06]. Row 4: Total (Wald), 52,, , 45,, , 100.0%, -0.03 [-0.05, -0.00]. The table also includes footnotes about the methods used for calculations.
Human milk oligosaccharides (HMOs) in colostrum – LNT, 2′FL, 3′SL, and 3FL: Two studies examined the relationship between levels of HMOs in colostrum in mothers with and without GDM(Reference Wang, Liu, Li, Xu, Wang and Lu22,Reference Zhang, Zhang, Yan, Duan, Wang and Szeto24) . The biomarkers that demonstrated significant differences were HMOs 2′FL and 6′SL(Reference Wang, Liu, Li, Xu, Wang and Lu22,Reference Zhang, Zhang, Yan, Duan, Wang and Szeto24) . The concentration of 2′FL found in colostrum of mothers with GDM was statistically lower than in mothers without GDM (mean difference: −478.92 μg/ml, 95% CI: −827.05 to −130.80, P = 0.007) with low heterogeneity (I² = 26%) (Figure 4). Similarly, 6′SL concentrations found in mothers with GDM were statistically lower than in mothers without GDM (mean difference: −285.05 μg/ml, 95% CI: −490.79 to −79.31, P = 0.007) with no heterogeneity (I² = 0%) (Figure 5). In contrast, no significant differences were observed for LNT, 3′SL, and 3FL in the colostrum of mothers with and without GDM (Figures 6–8)(Reference Wang, Liu, Li, Xu, Wang and Lu22,Reference Zhang, Zhang, Yan, Duan, Wang and Szeto24) .
Meta-analysis of the effects of GDM on 2′FL in colostrum.
Meta-analysis of the effects of GDM on 6′SL in colostrum.
Meta-analysis of the effects of GDM on LNT in colostrum.
Meta-analysis of the effects of GDM on 3′SL in colostrum.
Fig. 7. Long description
A table comparing the effects of GDM on 3’SL levels in colostrum across different studies. The table has 3 rows and 10 columns. Column headers are: Study or Subgroup, GDM Mean [ug/ml], GDM SD [ug/ml], GDM Total, Non-GDM Mean [ug/ml], Non-GDM SD [ug/ml], Non-GDM Total, Weight, Mean difference IV, Random, 95% CI [ug/ml], and Mean difference IV, Random, 95% CI [ug/ml]. Row 1: Wang et al. 2022, 144, 161, 13, 252, 181, 51, 49.0%, -108.00 [-208.63, -7.37], -108.00 [-208.63, -7.37]. Row 2: Zhang et al. 2024, 286.4, 128.9, 9, 237.9, 51.09, 9, 51.0%, 48.50 [-42.09, 139.09], 48.50 [-42.09, 139.09]. Row 3: Total (Wald), 22,, , 60,, , 100.0%, -28.15 [-181.49, 125.18], -28.15 [-181.49, 125.18].
Meta-analysis of the effects of GDM on 3′FL in colostrum.
Mature milk
Ghrelin in mature milk: Mature milk ghrelin concentrations were examined in mothers with and without GDM in two studies(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17,Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18) . The pooled analysis across both studies (n = 46) yielded no overall mean difference of −0.00 ng/ml (95% CI: −0.01 to 0.01) in ghrelin concentrations in mature milk between mothers with and without GDM (Z = 0.44, P = 0.66, I 2 = 0%).
Fatty acids C8:0, C10:0, C12:0, C14:0. C16:0, C18:0, C20:0, and C24:0 in mature milk: Differences in the levels of fatty acids in mature milk from mothers with and without GDM were examined in two studies(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21,Reference Yu, Wu, Yao, Song, Zhang and Xu23) . The pooled analysis across both studies (n = 125) did not demonstrate a significant overall difference in any of the fatty acids in mature milk of mothers with and without GDM (Figures 9–17).
Meta-analysis of the effects of GDM on ghrelin in mature milk.
Meta-analysis of the effects of GDM on C8:0 fatty acid in mature milk.
Meta-analysis of the effects of GDM on C10:0 fatty acid in mature milk.
Fig. 11. Long description
A table comparing the effects of GDM on C10:0 fatty acid in mature milk. The table has 3 rows and 11 columns. Column headers are: Study or Subgroup, GDM Mean [% of total FA], GDM SD [% of total FA], GDM Total, Non-GDM Mean [% of total FA], Non-GDM SD [% of total FA], Non-GDM Total, Weight, Mean difference IV, Random, 95% CI [% of total FA], Mean difference IV, Random, 95% CI [% of total FA]. Row 1: Sarkadi et al. 2022, 0.84, 0.21, 30, 1.05, 0.62, 39, 44.1%, -0.21 [-0.42, -0.00]. Row 2: Yu et al. 2023, 1.17, 0.05, 12, 1.1, 0.29, 44, 55.9%, 0.07 [-0.02, 0.16]. Row 3: Total (Wald), 42, 83, 100.0%, -0.05 [-0.33, 0.22].
Meta-analysis of the effects of GDM on C12:0 fatty acid in mature milk.
Meta-analysis of the effects of GDM on C14:0 fatty acid in mature milk.
Fig. 13. Long description
A table comparing the effects of GDM on C14:0 fatty acid in mature milk across different studies. The table has 4 rows and 10 columns. Column headers are: Study or Subgroup, GDM Mean [% of total FA], GDM SD [% of total FA], GDM Total, Non-GDM Mean [% of total FA], Non-GDM SD [% of total FA], Non-GDM Total, Weight, Mean difference IV, Fixed, 95% CI [% of total FA], and Mean difference IV, Fixed, 95% CI [% of total FA]. Row 1: Sarkadi et al. 2022, 7.14, 3.44, 30, 7.74, 2.54, 39, 28.8%, -0.60 [-2.07, 0.87]. Row 2: Yu et al. 2023, 4, 1.3, 12, 3.85, 1.94, 44, 71.2%, 0.15 [-0.78, 1.08]. Row 3: Total, 42, 83, 100.0%, -0.07 [-0.85, 0.72]. Row 4: Test for overall effect: Z = 0.16 (P = 0.87), Test for subgroup differences: Not applicable, Heterogeneity: Chi² = 0.72, df = 1 (P = 0.40); I² = 0%.
Meta-analysis of the effects of GDM on C16:0 fatty acid in mature milk.
Meta-analysis of the effects of GDM on C18:0 fatty acid in mature milk.
Meta-analysis of the effects of GDM on C20:0 fatty acid in mature milk.
Meta-analysis of the effects of GDM on C24:0 fatty acid in mature milk.
Fig. 17. Long description
A table comparing mean percentages of total fatty acids between GDM and Non-GDM groups across two studies. The table has 4 rows and 11 columns. Column headers are: Study or Subgroup, GDM Mean [% of total FA], GDM SD [% of total FA], GDM Total, Non-GDM Mean [% of total FA], Non-GDM SD [% of total FA], Non-GDM Total, Weight, Mean difference IV, Random, 95% CI [% of total FA], and Mean difference IV, Random, 95% CI [% of total FA]. Row 1: Sarkadi et al. 2022, 0.06, 0.13, 30, 0.33, 0.69, 39, 40.8%, -0.27 [-0.49, -0.05]. Row 2: Yu et al. 2023, 0.19, 0.04, 12, 0.2, 0.09, 44, 59.2%, -0.01 [-0.04, 0.02]. Row 3: Total (Wald), 42, 83, 100.0%, -0.12 [-0.37, 0.13]. Row 4: Test for overall effect: Z = 0.91 (P = 0.36), Test for subgroup differences: Not applicable, Heterogeneity: Tau² (REML) = 0.03; Chi² = 5.16, df = 1 (P = 0.02); I² = 81%. Footnotes: CI calculated by Wald-type method, Tau² calculated by Restricted Maximum-Likelihood method.
Narrative results
Of the 52 articles included in this review, 43 were not included in the meta-analyses. Reasons for exclusion varied, including differences in the lactation stages examined for specific components, the use of non-comparable reporting formats (e.g., medians versus mean differences), the limited reporting of outcomes (i.e., a human milk component examined in only one study, precluding a meta-analysis). A narrative synthesis for these studies, including those presented in meta-analyses, is presented below.
Energy and macronutrients
Energy
Five studies (n = 31–92) examined associations between GDM and energy content in human milk(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Yao, Shen, Zhang, Tang, Yu and Tu43,Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48,Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60) . Overall, findings were mixed: two studies reported no differences in energy content between mothers with and without GDM, with similar values observed in colostrum (n = 62 and 92)(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60) , transitional (n = 62)(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20), and mature milk (n = 53)(Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48). In contrast, three studies reported higher energy content in mature milk from mothers with GDM(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Yao, Shen, Yu, Tang, Zhang and Xu42) .
Carbohydrates
Across seven studies (n = 31–133), findings on carbohydrate and lactose concentrations in human milk from mothers with GDM varied by lactation stage. In colostrum, higher carbohydrate concentrations were reported in GDM milk between days 2–5 postpartum (n = 92)(Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60), while lower lactose concentrations in colostrum (days 2–7) were reported in two studies (n = 133, 130, and 46)(Reference Azulay Chertok, Haile, Shuisong and Kennedy25,Reference Li, Ning, Rui, Wang, Lei and Yu63,Reference Li66) . In contrast, one longitudinal study (n = 130) spanning days 3 to 28 postpartum found no effect of GDM on carbohydrates at any time point(Reference Kotilnska68).
In mature milk, carbohydrates were reported to be similar between groups after day 10 postpartum (n = 56)(Reference Yu, Wu, Yao, Song, Zhang and Xu23). At approximately six weeks postpartum (day 42), lactose concentrations were higher in mature milk from mothers with GDM compared with controls (n = 130)(Reference Li, Ning, Rui, Wang, Lei and Yu63), although this difference was no longer present by day 90. Similarly, no differences in lactose were observed at month 2 postpartum in another study (n = 53)(Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48).
In later lactation, one study (n = 20) examining milk collected beyond three months postpartum reported that carbohydrates were higher in mothers with GDM compared with controls(Reference Yao, Shen, Yu, Tang, Zhang and Xu42).
Proteins
Eight macronutrient studies (n = 24–130) examining bulk protein content reported no differences between mothers with and without GDM across colostrum(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60,Reference Cunha67,Reference Kotilnska68) , transitional milk(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Smilowitz, Sarah, Jincui, Dmitry, Holiday and Carol39,Reference Kotilnska68) , or mature milk(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Yao, Shen, Yu, Tang, Zhang and Xu42,Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48,Reference Kotilnska68) . Similarly, one study of free amino acids (n = 68) found no differences in colostrum between GDM and control mothers(Reference Klein, Bancher-Todesca, Graf, Garo, Roth and Kautzky-Willer49).
Two proteomic studies (n = 18–90) reported differential expression of specific proteins in colostrum from mothers with GDM(Reference Grapov, Lemay, Weber, Phinney, Azulay Chertok and Gho35,Reference Zhang, Zhang, Han, Chen, Liu and Sun45) . Grapov et al. (2015) (n = 18) identified 27 proteins distinguishing GDM from non-GDM, with elevations in apolipoprotein D, Ig heavy chain V-II region ARH-77, and prostasin, and reductions in alpha-2-HS-glycoprotein, apolipoproteins A1 and E, 14-3-3 protein zeta-delta, protein DJ-1, and FAM3B(Reference Grapov, Lemay, Weber, Phinney, Azulay Chertok and Gho35). Zhang et al. (2021) (n = 90) similarly found 192 proteins differing between groups, with glycated peptides upregulated and glycosylated peptides downregulated, primarily involving carbohydrate and lipid metabolism and immune function(Reference Zhang, Zhang, Han, Chen, Liu and Sun45).
In a glycoprotein study, Wang et al. (2023) (n = 12) reported that 12 of 56 N-glycans and 11 of 25 O-glycans in colostrum were different between GDM and control mothers. Overall, N- and O-glycan concentrations were lower in GDM milk, with fucosylated N-glycans higher and sialylated O-glycans lower compared with controls(Reference Wang, Li, Li, Li, Liu and Lu41).
Examining milk fat globule membrane (MFGM) proteins specifically (n = 31–48), Tao et al. (2024) (n = 48) reported four upregulated and 24 downregulated proteins in mature milk from mothers with GDM, including upregulated carbonyl reductase [NADPH] 3 (CBR3), cysteine-rich protein 2 (CRIP2), ras-related C3 botulinum toxin substrate 1 (RAC1), and GTPase KRas (KRAS), and downregulated neurabin-2 (PPP1R9B), gamma-aminobutyric acid receptor-associated protein-like 2 (GABARAPL2), major histocompatibility complex class I-related gene protein (MR1), methylmalonic aciduria type A protein, mitochondrial (MMAA), and ribosyldihy-dronicotinamide dehydrogenase [quinone] (NQO2)(Reference Tao, Wang, Xiao, Li, Wang and Mao40). Another study (n = 31) of mature milk reported 174 differentially expressed MFGM proteins, with albumin and transthyretin specifically upregulated in mothers with GDM(Reference Yao, Shen, Yu, Tang, Zhang and Xu42).
Fats
Five studies (n = 31–130) examined associations between GDM and fat content in human milk, reporting mixed findings(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Yao, Shen, Yu, Tang, Zhang and Xu42,Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60,Reference Kotilnska68) . Three studies observed no differences in fat levels between mothers with and without GDM, across colostrum or mature milk (n = 62, 92, and 130)(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20,Reference Korkut, Köse Çetinkaya, Işık, Özel, Gökay and Şahin60,Reference Kotilnska68) , whereas two studies reported higher fat levels in mature milk from mothers with GDM compared with controls (n = 56 and 31)(Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Yao, Shen, Yu, Tang, Zhang and Xu42) .
Lipids
Triglycerides: Three studies (n = 53–74) examined associations between GDM and triglyceride content in human milk, reporting mixed findings(Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48,Reference Cunha67) . One study (n = 53) observed no differences in triglyceride levels in mature milk between mothers with and without GDM(Reference Dugas, Laberee, Perron, St-Arnaud, Richard and Perreault48), whereas another study (n = 56) reported lower OPL triacylglycerol and higher CaOL triacylglycerol in mature milk(Reference Yu, Wu, Yao, Song, Zhang and Xu23). One study (n = 130) examined triglycerides content in colostrum, which reported higher triglycerides 48 hours after delivery in women with GDM compared with controls(Reference Cunha67).
Fatty acids: Across three studies (n = 40–69), global milk fatty acid class distribution, including saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs), and polyunsaturated fatty acids (PUFAs), was similar between mothers with and without GDM(Reference Salguero19,Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21,Reference Yu, Wu, Yao, Song, Zhang and Xu23) , with no consistent differences in total PUFA content, total unsaturation, or phospholipid classes beyond the early postpartum period (n = 56)(Reference Yu, Wu, Yao, Song, Zhang and Xu23). Instead, GDM-related differences were predominantly specific to individual fatty acids, varied by lactation stage, and were influenced by maternal diet and body mass index (BMI).
Among fatty acid outcomes, n-3 PUFAs, particularly docosahexaenoic acid (DHA), were the most consistently altered. In a study of transitional and mature milk (n = 18), mothers with GDM who did not consume phosphatidylserine (PS)/DHA supplements had lower DHA concentrations than unsupplemented controls, particularly at six to ten weeks postpartum(Reference Salguero19). In contrast, PS/DHA supplementation increased milk DHA concentrations in both groups, resulting in no between-group differences(Reference Salguero19). Lipidomic profiling of colostrum (n = 40) similarly showed lower levels of DHA (22:6) in milk from mothers with GDM compared with controls(Reference Zhong, Zhang, Xia, Zhu, Chen and Shan69).
Findings for n-6 PUFAs were less consistent and appeared stage-dependent. Higher concentrations of γ-linolenic acid, eicosatrienoic acid, arachidonic acid, and DHA were reported in colostrum from GDM mothers compared with controls (n = 63)(Reference Azulay Chertok, Haile, Eventov-Friedman, Silanikove and Argov-Argaman51), whereas later in lactation, linoleic acid was lower in GDM than non-GDM milk among normal-weight mothers (n = 69)(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21). The linoleic acid to α-linolenic acid ratio was not different in GDM milk(Reference Salguero19).
For MUFAs, total MUFA content did not differ by GDM status(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21,Reference Yu, Wu, Yao, Song, Zhang and Xu23) ; however, specific MUFAs were altered, including lower palmitoleic acid (C16:1) in GDM milk beyond the early postpartum period(Reference Yu, Wu, Yao, Song, Zhang and Xu23) and higher levels of very-long-chain MUFAs (21:1 and 22:1) and GDM in lipidomic analyses(Reference Zhong, Zhang, Xia, Zhu, Chen and Shan69).
SFA differences were selective and BMI-dependent. Caproic acid (C6:0) was consistently lower in GDM mothers regardless of BMI, while lauric acid (C12:0) was lower only among mothers with obesity and GDM(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21). In contrast, palmitic acid (C16:0) was higher in normal-weight mothers with GDM(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21). Despite these individual differences, total SFA content did not differ by GDM status(Reference Simon Sarkadi, Zhang, Muránszky, Vass, Matsyura and Benes21).
Polar lipids: Three studies (n = 40–56) examined associations between GDM and polar lipid content in human milk(Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Yu, Wu, Shen, Yao, Tang and Song44,Reference Zhong, Zhang, Xia, Zhu, Chen and Shan47) . One study (n = 40) identified 60 lipid species that differed between mothers with and without GDM, with phosphatidylethanolamine (PE) and phosphatidylcholine (PC) being the most abundant polar lipids in colostrum milk from mothers with GDM(Reference Zhong, Zhang, Xia, Zhu, Chen and Shan47). In contrast, two studies reported no differences between GDM and controls in mature milk glycerophospholipids (including PC, PS, PE, PI, PG, PA, LPE, LPC, LPI, LPG), sphingomyelins, and plasmalogens (n = 56 each)(Reference Yu, Wu, Yao, Song, Zhang and Xu23,Reference Yu, Wu, Shen, Yao, Tang and Song44) .
Cholesterol: In one study (n = 74), cholesterol was higher in mature milk of mothers with GDM compared with controls(Reference Cunha67).
Endocannabinoidome: A cross-sectional study (n = 53) examined associations between GDM and components of the endocannabinoidome in human milk(Reference Fradet, Castonguay-Paradis, Dugas, Perron, St-Arnaud and Marc61). Mothers with GDM had higher concentrations of several N-acylethanolamine bioactive lipids compared with controls, including N-palmitoylethanolamide, 2-arachidonoylglycerol, 2-docosapentaenoyl-glycerol, and 2-docosahexaenoyl-glycerol, while levels of Noleoyl–ethanolamine, Nlinoleoyl–ethanolamine, N-arachidonoylethanolamine, N-eicosapentaenoyl-ethanolamine, N-docosapentaenoyl-ethanolamine, N-docosahexaenoyl-ethanolamine, 2-oleoyl-glycerol, 2-linoleoyl-glycerol, and 2-eicosapentaenoyl-glycerol did not differ between groups(Reference Fradet, Castonguay-Paradis, Dugas, Perron, St-Arnaud and Marc61).
Human milk oligosaccharides
Six studies (n = 18–130) examined associations between maternal GDM and HMOs across colostrum, transitional, and mature milk(Reference Wang, Liu, Li, Xu, Wang and Lu22,Reference Zhang, Zhang, Yan, Duan, Wang and Szeto24,Reference Smilowitz, Sarah, Jincui, Dmitry, Holiday and Carol39,Reference Ma, Lu, Huang, Guo, Xiang and Gao52,Reference Dou, Luo, Xing, Liu, Chen and Zhu59,Reference Li, Ning, Rui, Wang, Lei and Yu63) . Colostrum from mothers with GDM generally showed lower concentrations of total HMOs, fucosylated HMOs, sialylated HMOs (including 3′-SL and 6′-SL), and specific individual HMOs such as LNT, LNnT, LNFP-I, LNFP-II, DSLNT, LSTc, and LNDFH-I (n = 99,18, 127, and 130)(Reference Wang, Liu, Li, Xu, Wang and Lu22,Reference Zhang, Zhang, Yan, Duan, Wang and Szeto24,Reference Ma, Lu, Huang, Guo, Xiang and Gao52,Reference Li, Ning, Rui, Wang, Lei and Yu63) . Some structure-specific differences were also reported, with H4N2F2 elevated and several fucosylated HMOs (H5N3F2, H5N3F3, H6N4F1, H6N4F2, H6N4F3) reduced in GDM colostrum (n = 127)(Reference Ma, Lu, Huang, Guo, Xiang and Gao52).
For transitional and mature milk, findings were more variable. In two studies, total HMOs and individual HMOs (e.g., 2′-FL, 3′-SL, 6′-SL, LNT, LNFP-I, LSTc) were higher in GDM milk at later stages, often surpassing control levels by day 42, though differences tended to normalise by day 90 (n = 22 and 130)(Reference Dou, Luo, Xing, Liu, Chen and Zhu59,Reference Li, Ning, Rui, Wang, Lei and Yu63) . Another study (n = 24) found no changes in free oligosaccharides during transitional milk(Reference Smilowitz, Sarah, Jincui, Dmitry, Holiday and Carol39).
Hormones and metabolic regulatory factors
Ghrelin
Across studies (n = 20–96), ghrelin concentrations in human milk were generally lower in mothers with GDM, with some variation by lactation stage and treatment. In colostrum, lower ghrelin levels in GDM milk compared with controls were reported in three studies(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17,Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Yu, Rong, Sun, Ding, Wan and Zou58) . One study (n = 51) further showed that colostral ghrelin was lowest in diet-treated GDM mothers and higher in insulin-treated GDM mothers, where mothers without GDM showed the highest levels(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18). In mature milk, two studies reported lower ghrelin concentrations in mothers with GDM compared with controls (n = 26 and 96)(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17,Reference Yu, Rong, Sun, Ding, Wan and Zou58) , whereas one study (n = 26) observed no between-group differences in mature milk despite lower levels in colostrum(Reference Aydin, Geckil, Karatas, Donder, Kumru and Kavak17).
Leptin
Three studies (n = 51–190) consistently reported no differences in leptin concentrations in human milk between mothers with and without GDM, across colostrum and mature milk(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26,Reference Yu, Rong, Sun, Ding, Wan and Zou58) .
Adiponectin
Five studies (n = 46–190) in this review examined associations between GDM and adiponectin concentrations in human milk(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26,Reference Yu, Rong, Sun, Ding, Wan and Zou58,Reference Nteka, Liosis, Kastanidou, Papgaroufalis, Mastorakos and Lykeridou65,Reference Li66) . Overall, while one study (n = 96) suggests that GDM may be associated with lower adiponectin levels in colostrum and mature milk(Reference Yu, Rong, Sun, Ding, Wan and Zou58), the majority of evidence – particularly from colostrum samples – does not support a consistent difference in adiponectin concentrations between mothers with and without GDM (n = 51, 190, 50, and 46)(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26,Reference Nteka, Liosis, Kastanidou, Papgaroufalis, Mastorakos and Lykeridou65,Reference Li66) .
Glucose
Findings on glucose concentrations in human milk differed by lactation stage (n = 28–190). In colostrum, higher glucose concentrations in GDM milk were reported in two studies (n = 130 and 74)(Reference Li, Ning, Rui, Wang, Lei and Yu63,Reference Cunha67) , while one study (n = 28) found no between-group differences(Reference Gomes Fagundes, França, da Silva Fernandes, Hara, Morceli and Honorio-França62). In mature milk, no differences in glucose were observed in one study (n = 130)(Reference Li, Ning, Rui, Wang, Lei and Yu63), whereas lower glucose concentrations at one and three months postpartum were reported in mothers with GDM compared with controls in a larger cohort (n = 190)(Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26).
Insulin
Two studies (n = 96–190) examined associations between GDM and insulin concentrations in human milk, with inconsistent findings(Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26,Reference Yu, Rong, Sun, Ding, Wan and Zou58) . In one study (n = 190), insulin concentrations in mature milk were lower at one and three months postpartum in mothers with GDM compared with controls(Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26). In contrast, a second study (n = 96) reported higher insulin concentrations in both colostrum and mature milk among mothers with GDM across multiple postpartum time points (days 3, 42, and 90)(Reference Yu, Rong, Sun, Ding, Wan and Zou58).
Chemerin
Two cohort studies examined associations between GDM and chemerin concentrations in human milk(Reference Fatima, Rehman, Muhammad, Martins, Mohammed and Khan30,Reference Ustebay, Baykus, Deniz, Ugur, Yavuzkir and Yardim54) . In one study (n = 66), chemerin levels were higher in the colostrum of mothers with GDM, with elevations persisting at six weeks postpartum(Reference Fatima, Rehman, Muhammad, Martins, Mohammed and Khan30). Similarly, a second study (n = 53) reported that chemerin levels were higher in colostrum, transitional, and mature milk from mothers with GDM compared with controls(Reference Ustebay, Baykus, Deniz, Ugur, Yavuzkir and Yardim54).
Irisin
Two studies (n = 30–66) examined associations between GDM and irisin concentrations in colostrum, transitional, and mature milk, both reporting lower levels in colostrum and mature milk of mothers with GDM(Reference Aydin, Kuloglu and Aydin28,Reference Fatima, Khalid, Ladak and Ali34) .
Apelin
One study (n = 20) examined associations between GDM and the bioactive peptides apelin-36 and apelin-12 in colostrum and mature milk, reporting lower concentrations of both peptides in mothers with GDM compared with controls(Reference Aydin33).
Nesfatin-1
One cohort study (n = 20) examined associations between GDM and the bioactive peptide nesfatin-1 in colostrum and mature milk, reporting lower concentrations in mothers with GDM compared with controls(Reference Aydin33).
Resistin
In one study (n = 51), resistin was not associated with maternal glycaemic state in colostrum(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18).
Other peptides
Three studies examined associations between GDM and other bioactive regulatory peptides in human milk(Reference Aydin, Kuloglu and Aydin28,Reference Aydin, Celik, Gurates, Sahin, Ulas and Yilmaz50,Reference Yavuz, Aydin, Ugur, Aydin, Senol and Baykus57) . Overall, mothers with GDM showed higher concentrations of some peptides – betatrophin, asprosin, glucagon, subfatin, and copeptin – across all lactation stages, with levels typically highest in colostrum (n = 24, 30 and 26)(Reference Aydin, Kuloglu and Aydin28,Reference Aydin, Celik, Gurates, Sahin, Ulas and Yilmaz50,Reference Yavuz, Aydin, Ugur, Aydin, Senol and Baykus57) . Other peptides, including elabela, preptin, salusin-α, salusin-β, hepcidins, and adropin, were lower in mothers with GDM, though these differences were often restricted to specific stages such as colostrum or transitional milk(Reference Aydin, Kuloglu and Aydin28,Reference Aydin, Celik, Gurates, Sahin, Ulas and Yilmaz50,Reference Yavuz, Aydin, Ugur, Aydin, Senol and Baykus57) .
Immune and inflammatory factors
Immune factors
Across six studies (n = 24–190), maternal GDM was associated with alterations in specific immune factors in human milk, though the patterns varied by analyte and lactation stage. In colostrum, mothers with GDM exhibited a mixed profile of immune changes: some studies reported higher pro-inflammatory cytokines, including IFN-γ, interleukin (IL)-6, and IL-15, alongside lower IL-1 receptor antagonist (n = 25)(Reference Avellar, Oliveira, Caixeta, Souza, Teixeira and Faria27), whereas others found increases in anti-inflammatory markers, such as IL-10 and anti-insulin IgA, accompanied by reductions in IFN-γ and TGF-β, with no changes in total IgA, TNF-α, or IL-17 (n = 74)(Reference Cunha67). Cellular composition was also affected; although total CD3+ and memory T cells were unchanged, GDM colostrum contained a higher proportion of CD4+ T cells, including naive CD4+CD45RA+ cells, while Th1, Th2, and Th17 cytokines levels were similar (n = 28)(Reference Gomes Fagundes, França, da Silva Fernandes, Hara, Morceli and Honorio-França62). Early colostral IgG (days 1–3) was higher in GDM mothers, particularly in those receiving insulin treatment, whereas sIgA, IgM, and total IgA were largely stable across lactation stages (n = 102)(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz32).
In transitional milk, GDM was associated with substantial reductions in sIgA, including decreases in mannose-, fucose-, and sialic acid–containing N-glycans (n = 24)(Reference Smilowitz, Sarah, Jincui, Dmitry, Holiday and Carol39), indicating that post-translational modifications of immune proteins may also be affected.
In mature milk, systemic markers of inflammation such as C-reactive protein (CRP) remained elevated at one and three months postpartum in mothers with GDM, while IL-6 showed no differences (n = 190)(Reference Choi, Nagel, Kharoud, Johnson, Gallagher and Duncan26), suggesting that some immune or inflammatory changes persist beyond colostrum even when specific milk cytokines normalise.
Growth factors
In studies on growth factors (n = 25–51), no differences in IGF-1 or IGFBP-2 were observed in colostrum between mothers with and without GDM(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18,Reference Nteka, Liosis, Kastanidou, Papgaroufalis, Mastorakos and Lykeridou65) . However, mothers with GDM exhibited a decrease in the growth factor granulocyte-macrophage colony-stimulating factor (GM-CSF) in their colostrum compared with controls (n = 25)(Reference Avellar, Oliveira, Caixeta, Souza, Teixeira and Faria27).
Lactoferrin
Two studies (n = 24–102) examined associations between GDM and lactoferrin concentrations in human milk, reporting mixed findings(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz32,Reference Smilowitz, Sarah, Jincui, Dmitry, Holiday and Carol39) . One study (n = 102) found no differences in lactoferrin concentrations in early colostrum (days 1–3), colostrum (days 4–7), or transitional milk (days 8–15) between mothers with and without GDM(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz32), whereas another (n = 24) reported higher levels of lactoferrin N-glycans and fucosylated and sialylated lactoferrin N-glycans in transitional (two-week) milk from mothers with GDM(Reference Smilowitz, Sarah, Jincui, Dmitry, Holiday and Carol39).
Dermcidin
In one study (n = 53), dermcidin concentrations tended to be higher in milk from mothers with GDM compared with controls across lactation stages; however, the figures and text do not clearly indicate whether these between-group differences in milk reached statistical significance at individual lactation stages(Reference Ustebay, Baykus, Deniz, Ugur, Yavuzkir and Yardim54).
NGAL/MMP
One study (n = 35) examined reported that the mean MMP-9/NGAL complex concentration was higher in day 3 milk, but not day 1, from the GDM group compared with controls(Reference Metallinou, Lykeridou, Karampas, Liosis, Skevaki and Rizou36).
Micronutrients and antioxidants
Alpha-tocopherol
Two studies (n = 18–51) examined associations between GDM and alpha-tocopherol concentrations in human milk, finding no differences between groups in colostrum (n = 51)(Reference Resende, Clemente, Bezerra, Grilo, de Melo and Bellot64) or mature milk (n = 18)(Reference Churchill, Zawawi, Elisia, Seider, Noseworthy and Thompson29).
Beta-carotene
One study (n = 18) examined associations between GDM and beta-carotene in mature milk, finding no differences between groups(Reference Churchill, Zawawi, Elisia, Seider, Noseworthy and Thompson29).
Ascorbic acid
One study (n = 18) examined associations between GDM and ascorbic acid concentrations in mature milk, finding no differences between groups(Reference Churchill, Zawawi, Elisia, Seider, Noseworthy and Thompson29).
Calcium
One study (n = 74) examined associations between GDM and calcium in mature milk, reporting higher calcium concentrations in mothers with GDM compared with controls(Reference Cunha67).
Citrate
In one study (n = 133), citrate concentrations were lower in the colostrum of mothers with GDM compared with controls(Reference Azulay Chertok, Haile, Shuisong and Kennedy25).
Glutathione
One study (n = 120) examined associations between GDM and glutathione in human milk, reporting lower concentrations in colostrum in mothers with GDM compared with controls, with no sustained differences in transitional or mature milk(Reference Peila, Riboldi, Spada, Coscia, Barbagallo and Volti53).
Oxygen radical absorbance capacity
In one study (n = 18), there was no difference in concentrations of oxygen radical absorbance capacity (ORAC) between mature milk of mothers with and without GDM(Reference Churchill, Zawawi, Elisia, Seider, Noseworthy and Thompson29).
Multiomics
Metabolome
Across metabolomics studies (n = 31–184), GDM was associated with widespread but stage-dependent alterations in the human milk metabolome, with differences most pronounced in transitional and mature milk and spanning lipids, amino acids, organic acids, and energy-related metabolites(Reference Yao, Shen, Zhang, Tang, Yu and Tu43,Reference Wen, Wu, Yang, Han, Wang and Fu55,Reference Wu, Yu, Liu, Wang, Chen and Qiao56) . Rather than a single consistent metabolite signature, findings indicate dynamic, lactation-stage-specific shifts, with some overlap in affected metabolic pathways. In colostrum, two studies reported fewer differentially abundant metabolites compared with later stages, with a predominance of reduced amino acids and organic acids in milk from mothers with GDM compared to controls(Reference Wen, Wu, Yang, Han, Wang and Fu55,Reference Wu, Yu, Liu, Wang, Chen and Qiao56) . Specifically, colostrum from mothers with GDM had lower levels of glycine-related metabolites, glutamine, oxaloacetic acid, γ-aminobutyric acid, alongside other lipids and amino acids (n = 184)(Reference Wen, Wu, Yang, Han, Wang and Fu55), with only partial compensation from increases in select amino acids, lipids, and nucleotides (n = 100)(Reference Wu, Yu, Liu, Wang, Chen and Qiao56).
In transitional milk, the metabolomic impact of GDM was most extensive. Wen et al. (2019) identified reductions in amino acid derivatives, organic acids, and one unsaturated fatty acid (9-heptadecenoic acid), with concurrent increases in asparagine and the citric acid cycle intermediate malic acid. Similarly, Wu et al. (2021) reported that the majority of metabolites differing by GDM status in transitional milk were lipids – particularly glycerophospholipids, most of which were reduced, alongside mixed changes in amino acids, fatty acids, and nucleotides(Reference Wu, Yu, Liu, Wang, Chen and Qiao56). Notably, both studies highlight energy metabolism and lipid remodelling as key features of GDM-related differences during this stage(Reference Wen, Wu, Yang, Han, Wang and Fu55,Reference Wu, Yu, Liu, Wang, Chen and Qiao56) .
In mature milk, Wen et al. (2019) observed a predominance of downregulated metabolites in GDM, including amino acids and derivatives (e.g., glycine, pyroglutamic acid, β-alanine), organic acids linked to amino acid and citric acid cycle metabolism, and several saturated and unsaturated fatty acids, alongside selective increases in cysteine, lignoceric acid, malic acid, and heneicosane(Reference Wen, Wu, Yang, Han, Wang and Fu55). Wu et al. (2021) similarly reported mixed alterations in mature milk, with both decreased and increased lipids, organic acids, nucleotides, and vitamins, suggesting continued metabolic divergence rather than normalisation over time(Reference Wu, Yu, Liu, Wang, Chen and Qiao56).
Despite heterogeneity, a small number of metabolites showed reproducible patterns across stages. 4-Pyridoxic acid was consistently elevated in GDM milk across all lactation stages(Reference Wu, Yu, Liu, Wang, Chen and Qiao56), while malic acid showed repeated stage-specific alterations, being elevated in transitional milk but reduced in colostrum and mature milk(Reference Wu, Yu, Liu, Wang, Chen and Qiao56), underscoring dysregulation of energy metabolism. In contrast, most metabolites differed in a stage-restricted manner, with the majority unique to transitional or mature milk(Reference Wu, Yu, Liu, Wang, Chen and Qiao56).
At later lactation (>3 months), Yao et al. (2024) (n = 20) reported a substantially expanded set of differentially expressed metabolites, including higher linoleic acid, arachidonic acid, oxidised lipid derivatives, and L-glutamic acid in GDM milk, alongside lower levels of a specific oxidised fatty acid metabolite, suggesting that lipid and amino acid pathway perturbations persist beyond early lactation(Reference Yao, Shen, Zhang, Tang, Yu and Tu43).
Microbiome
Two studies (n = 31–45) examined associations between GDM and the human milk microbiome, reporting varied findings(Reference Gámez-Valdez, García-Mazcorro, Montoya-Rincón, Rodríguez-Reyes, Jiménez-Blanco and Rodríguez31,Reference Rold, Guldbæk, Lindegaard, Kirk, Nygaard and Bundgaard-Nielsen37) . One study (n = 31) observed increased microbial diversity and higher Staphylococcus abundance in colostrum from mothers with GDM(Reference Gámez-Valdez, García-Mazcorro, Montoya-Rincón, Rodríguez-Reyes, Jiménez-Blanco and Rodríguez31), whereas another study (n = 45) found no differences in alpha- or beta-diversity, nor in Staphylococcus in transitional/mature milk between groups(Reference Rold, Guldbæk, Lindegaard, Kirk, Nygaard and Bundgaard-Nielsen37). The first study did not observe between-group differences in Corynebacterium 1 and Prevotella in colostrum(Reference Gámez-Valdez, García-Mazcorro, Montoya-Rincón, Rodríguez-Reyes, Jiménez-Blanco and Rodríguez31), while the second study did not observe between-group differences in other specific taxa (see Appendix A) in transitional/mature milk(Reference Rold, Guldbæk, Lindegaard, Kirk, Nygaard and Bundgaard-Nielsen37).
MircoRNA
Two studies (n = 50–94) examined micro ribonucleic acid (microRNA/miRNA) profiles in human milk from mothers with GDM, reporting heterogeneous findings(Reference Shah, Fields, Pezant, Kharoud, Gulati and Jacobs38,Reference Zheng, Mo, Lin, Wang, Chen and Luo46) . In one study (n = 50), numerous miRNAs were differentially expressed in GDM milk across lactation stages(Reference Zheng, Mo, Lin, Wang, Chen and Luo46). Specifically, in colostrum exosomes, 124 miRNAs were differentially expressed (101 upregulated, 23 downregulated); in transitional milk exosomes, 144 miRNAs (112 upregulated, 32 downregulated); and in mature milk exosomes, 121 miRNAs (98 upregulated, 23 downregulated) were altered, compared with stage-matched exosomes from controls. Gene Ontology analysis indicated that these miRNAs were enriched in cellular regulatory pathways, biological regulation, and regulation of biological process categories(Reference Zheng, Mo, Lin, Wang, Chen and Luo46). The second study (n = 94) focused on mature milk and reported lower levels of specific miRNAs in mothers with GDM, including miRNA-148a, miRNA-30b, and miRNA-let-7a/d, compared with controls(Reference Shah, Fields, Pezant, Kharoud, Gulati and Jacobs38) .
Discussion
This systematic review provides a comprehensive synthesis of evidence examining the association between gestational diabetes mellitus and human milk composition across lactation. Overall, findings indicate that core macronutrient content (protein, total fat, and total carbohydrate) is largely preserved in human milk regardless of maternal GDM status. In contrast, select bioactive, regulatory, and structural components – particularly in colostrum and transitional milk – show stage-specific alterations, suggesting that GDM influences the functional properties of milk rather than its gross nutritional adequacy. These differences probably reflect the maternal metabolic environment and its downstream effects on mammary gland activity during early lactation. Note that although 52 studies were included in this review, meta-analyses were possible for only nine outcomes, and many milk components (e.g., apelin, ascorbic acid, β-carotene, calcium, cholesterol, citrate, dermcidin, endocannabinoidome markers, glutathione, nesfatin-1, NGAL/MMPs, ORAC, and resistin) were evaluated in a single study. Therefore, these findings should be interpreted with caution and require confirmation in future research.
Meta-analytic and narrative evidence indicates that colostrum macronutrient components are particularly sensitive to GDM, with reductions observed in colostrum lactose. Additionally, total protein content is preserved in mothers with GDM, while specific proteins – particularly those involved in lipid transport, carbohydrate metabolism, immune regulation, and MFGM structure – are expressed differentially. Increased glycated peptides further suggest that maternal hyperglycaemia may influence post-translational protein modification within the mammary gland.
While total fat content and global fatty acid class distributions (SFA, MUFA, PUFA) were generally similar between groups, GDM was associated with selective, potentially biologically meaningful differences in individual lipid species. Most notably, lower DHA concentrations in transitional and mature milk among unsupplemented mothers with GDM may suggest reduced n-3 PUFA availability, which appeared modifiable by maternal supplementation. In contrast, higher concentrations of certain n-6 PUFAs (e.g., arachidonic acid) in colostrum and a higher linoleic-to-α-linolenic acid ratio point to a modest shift toward a more n-6-dominant profile. Additional alterations included higher energy content, cholesterol, triglycerides, polar lipids (PE, PC), and endocannabinoid-related lipids in some studies of mature milk, indicating lipid remodelling rather than increased lipid quantity. Elevated concentrations of palmitoylethanolamide and 2-arachidonoylglycerol suggest potential modulation of infant energy balance or immune signalling, although infant outcomes were not assessed.
Beyond macronutrient and lipid alterations, reductions in colostrum HMOs, including 2′-fucosyllactose (2′FL) and 6′-sialyllactose (6′SL), were observed among mothers with GDM. Because these HMOs play established roles in shaping the infant gut microbiome and immune development, early-life reductions may have functional relevance(Reference Kotilnska68). Importantly, many HMO differences attenuated by mature milk, suggesting either postpartum normalisation of mammary glycosylation or compensatory upregulation over time. However, the persistence of altered concentrations of specific HMOs (e.g., LNT, LNFP-I, LSTc) in some cohorts indicates that recovery may be incomplete or heterogeneous, potentially influenced by maternal metabolic control, BMI, or secretor status.
Human milk contains numerous bioactive peptides involved in appetite regulation and metabolic signalling(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18). Evidence suggests that GDM is associated with selective alterations in these peptides, particularly in colostrum. Lower concentrations of irisin, nesfatin-1, and apelin were consistently observed in colostrum and, in some cases, mature milk from mothers with GDM. In contrast, peptides such as betatrophin, asprosin, glucagon, and copeptin were elevated, often peaking in colostrum, suggesting a milk hormonal profile that may reflect maternal insulin resistance rather than impaired mammary synthesis.
For ghrelin, the observed reduction in colostrum among mothers with GDM was small in magnitude, with no consistent differences reported in mature milk, providing limited evidence for clinical relevance. Similarly, leptin concentrations were comparable between GDM and non-GDM milk across lactation, suggesting no clear GDM-driven effect on milk leptin exposure. Findings for insulin were inconsistent, with reports of both higher and lower concentrations in GDM milk, probably reflecting differences in GDM severity, treatment modality, postpartum glycaemic control, and timing of sampling. These inconsistencies highlight the need for standardised study designs and careful adjustment for maternal metabolic factors.
Adiponectin concentrations showed minimal differences by GDM status, with no consistent differences in colostrum and one report of lower concentrations in mature milk. Given adiponectin’s role in glucose metabolism and inflammation(Reference Lis-Kuberka, Berghausen-Mazur and Orczyk-Pawiłowicz18), reduced levels in later lactation may reflect persistent maternal metabolic dysregulation, although infant-level consequences remain speculative.
GDM was associated with complex, stage-dependent alterations in immune-related milk components, rather than a uniform pro-inflammatory profile. Colostrum exhibited both pro- and anti-inflammatory shifts, including altered cytokines (e.g., IL-6, IL-15, IL-10), changes in immune cell proportions, and higher early IgG concentrations. While total SIgA levels were largely unchanged, structural modifications – particularly altered N- and O-glycosylation patterns – were evident, including increased fucosylation and reduced sialylation. Reductions in sIgA N-glycans in transitional milk suggest that post-translational modification of immune proteins, rather than concentration alone, may be a potential key mechanism by which GDM influences milk immune function. Persistent elevations in maternal CRP and altered insulin levels at later lactation may indicate that systemic metabolic inflammation may extend into the postpartum period.
Alterations in the human milk microbiome represent another potential molecular pathway through which GDM may influence milk composition. One study reported higher Staphylococcus abundance in colostrum from mothers with GDM, an effect not observed in transitional or mature milk. While Staphylococcus includes opportunistic species(Reference Grapov, Lemay, Weber, Phinney, Azulay Chertok and Gho35), the clinical significance of this finding remains unclear. Although not directly examined in this review, GDM-related shifts in milk microbial composition could theoretically influence early infant gut colonisation, given the documented co-occurrence between human milk and infant gut microbes(Reference Shapira, Mandel, Mimouni, Moran-Lev, Marom and Mangel20). However, the limited number of microbiome studies and inconsistent findings preclude firm conclusions.
Metabolomic studies identified transitional milk as the most metabolically altered stage, with disruptions in amino acids, organic acids, and lipid-related metabolites reflecting altered energy metabolism. Reproducible elevations in 4-pyridoxic acid across lactation and recurrent changes in malic acid support the presence of sustained metabolic perturbations rather than transient early effects. At later lactation stages, continued divergence in lipid and amino acid pathways suggests that metabolic remodelling may persist beyond early postpartum.
Finally, miRNA profiling revealed widespread, stage-spanning alterations in milk-derived miRNAs in mothers with GDM, with enrichment in pathways related to cellular regulation and metabolism. These findings raise the possibility that epigenetic signalling via milk may represent an additional pathway linking maternal metabolic status with infant development, although functional consequences remain to be elucidated.
This review represents the most comprehensive synthesis to date examining the effects of GDM on human milk composition. A key strength lies in the systematic inclusion of studies spanning colostrum, transitional, and mature milk, allowing for the identification of temporal patterns in milk component alterations. The inclusion of meta-analyses for selected outcomes – such as lactose, ghrelin, and HMOs – strengthens the quantitative interpretation of pooled findings and highlights consistent disruptions in specific bioactive components. Additionally, this review integrates findings across multiple milk analyte classes (e.g., hormones, oligosaccharides, lipids, proteins, inflammatory and immune factors, multiomics), providing a multidimensional understanding of how GDM may influence milk quality and infant exposure to bioactive signals. Moreover, the overall quality of included studies was favourable, with the majority receiving high-quality ratings. A smaller proportion was rated as moderate, good/very good, or satisfactory, while grey literature sources were not formally assessed. This indicates that the evidence base is largely supported by methodologically robust studies.
Nonetheless, several limitations should be acknowledged. Many included studies were small, observational, and cross-sectional in nature, limiting causal inferences and generalisability. In addition, substantial heterogeneity existed across studies in the diagnostic criteria used for GDM, glycaemic control during pregnancy, adjustment for confounding variables such as diet, parity, mode of delivery, BMI, or intrapartum insulin use, and milk collection and handling protocols. For example, studies differed in the timing of milk collection (e.g., pre-feed, post-feed, morning, afternoon, fasting, or non-fasting conditions), the volume of milk collected, whether foremilk, hindmilk, or a complete breast expression was obtained, and the collection method (manual expression or manual, electric, or hospital-grade pumps). These differences in milk collection protocols may have contributed to the variability observed across studies. Although milk collection characteristics were extracted during data collection, inconsistent reporting and substantial variation in protocols precluded formal adjustment for these factors in the analyses.
Taken together, the evidence indicates that GDM is associated with selective, lactation-stage-specific remodelling of human milk bioactivity, particularly in colostrum and transitional milk, rather than wholesale disruption of milk nutrition. Many alterations appear dynamic and potentially modifiable by diet, supplementation, and metabolic control, supporting the concept of adaptive mammary responses to maternal metabolic stress. Future longitudinal studies integrating maternal metabolic measures, standardised milk analytics, and infant outcomes are essential to determine whether these milk compositional changes contribute to, compensate for, or mitigate GDM-related risks in offspring.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0954422426100523
Acknowledgements
N/A.
Financial support
None.
Competing interests
The authors declare no conflicts of interest.
Authorship
MK designed the research (project conception, development of overall research plan, and study oversight). AG, SB, and MK completed data collection and statistical analysis. BC conducted the literature search. AG, SB, JD, and MK wrote the manuscript. MK had primary responsibility for final content. All authors read and approved the final manuscript.
Open access
