Milk fat accounts for 45–55 % of the total energy provided by human milk(Reference German and Dillard1,Reference Mazzocchi, D’Oria and De Cosmi2) . It is the essential source of bioactive fatty acids and important delivery medium for fat-soluble antioxidants such as tocopherols and carotenoids(Reference German and Dillard1,Reference Mazzocchi, D’Oria and De Cosmi2) . The composition of human milk fluctuates with the progress of lactation and can be divided into three stages as colostrum, transitional milk and mature milk. As milk total fat content increases with lactation, long-chain PUFA (LC-PUFA) and fat-soluble antioxidants such as α-tocopherol and β-carotene indicated sharp declines(Reference Macias and Schweigert3–Reference Xue, Campos-Gimenez and Redeuil6). Specific mechanisms may exist in the transfer of different liphophilic components in order to satisfy the infant’s requirements(Reference Macias and Schweigert3), and a detailed analysis of human milk fatty acids and fat-soluble antioxidants over lactation is imperative for better estimations of the dynamic demands.
Linoleic acid (LA, C18 : 2n-6) and α-linolenic acid (ALA, C18 : 3n-3) are two essential PUFA in human body and compete for the same enzymes required for the conversion into the corresponding n-6 and n-3 long-chain derivatives, respectively(Reference Uauy, Mena and Rojas7). LC-PUFA exert a number of cognitive and immune benefits(Reference Weiser, Butt and Mohajeri8,Reference Waidyatillake, Stoney and Thien9) with tocopherols providing antioxidant protections(Reference Stimming, Mesch and Kersting10,Reference Raederstorff, Wyss and Calder11) . Recent studies have demonstrated the concern that the increasing maternal consumption of LA from vegetable oils may result in the predominance of proinflammatory arachidonic acid (ARA, C20 : 4n-6) in the infant’s body, which impairs Th1/Th2 immune balance and neocortical development(Reference Waidyatillake, Stoney and Thien9,Reference Kim, Kim and Lee12,Reference Zamberletti, Piscitelli and De Castro13) . Contrarily, n-3 LC-PUFA help in promoting psychomotor development and lowering the risk of lifelong allergic diseases and emotional problems(Reference Weiser, Butt and Mohajeri8,Reference Kim, Kim and Lee12,Reference Gibson, Muhlhausler and Makrides14,Reference Best, Gold and Kennedy15) . Hence, the evaluation of human milk n-6:n-3 PUFA or LA:ALA may outweigh the measurement of individual fatty acid alone. However, previous studies have rarely discussed these ratios in Chinese breast milks.
α-Tocopherol has been well documented, showing antioxidation, anti-lipid peroxidation and anti-inflammation functions. Other vitamin E isoforms were less investigated, but concerns have arisen that an increasing number of foods and formulas deliver much higher γ-tocopherol than α-tocopherol(Reference Martysiak-Zurowska, Szlagatys-Sidorkiewicz and Zagierski4,Reference Bai-fen, Ying, Jian-hua and Fei-fei16,Reference Wu, Xiang and Yang17) . γ-Tocopherol also decreases lipid peroxidation (weaker than α-tocopherol) and neutrophilic inflammation (stronger than α-tocopherol), but it can promote type 2 inflammation(Reference Berdnikovs, Abdala-Valencia and McCary18,Reference Stone, McEvoy and Aschner19) . The latest early life study in both Nigerian and American maternal–neonatal dyads reported the association between decreased circulating α-:γ-tocopherol and negative birth outcomes such as Caesarian sections(Reference Cave, Hanson and Schumacher20). The ratios of α-:γ-tocopherol could contribute to a more comprehensive understanding of the infant’s vitamin E requirements than α-tocopherol alone, but the ratios in human milk still lack investigation.
The majority of carotenoids can be classified into two groups: carotenes (α-carotene, β-carotene and lycopene) and their hydroxylated derivatives-xanthophylls (lutein, zeaxanthin and β-cryptoxanthin)(Reference Zielinska, Wesolowska and Pawlus21,Reference Eggersdorfer and Wyss22) . Although there are reports that high doses of carotenes might act as pro-oxidants and present cancer risks(Reference Moran, Mohn and Hason23,Reference Krinsky24) , carotenoids at physiological levels have been well recognised to exert antioxidant effect(Reference Zielinska, Wesolowska and Pawlus21,Reference Eggersdorfer and Wyss22) . Additionally, each carotenoid plays unique roles. For example, β-carotene mainly serves as vitamin A source in maternal–neonatal pairs(Reference Zielinska, Wesolowska and Pawlus21,Reference Eggersdorfer and Wyss22) ; Lutein and zeaxanthin contribute vitally to the macular and cognitive development(Reference Zielinska, Wesolowska and Pawlus21,Reference Eggersdorfer and Wyss22) . Consequently, the dominance of individual carotenoid may indicate different priorities in early life nutrition. The levels of human milk β-carotene have been investigated globally, but other carotenoids still warrant further investigations. Considering that the absorption of different carotenoids interferes substantially with each other and excessive consumption may pose undesirable risks(Reference Moran, Mohn and Hason23), it is important to elucidate the dynamic supply of human milk carotenoids to sustain a better estimate of the infant’s demand. However, to date, the available data are insufficient.
Given the insufficient data and the long-term neglect of the balance of lipid nutrients in human milk, the present study aimed to explore the profiles of fatty acids, tocopherols and carotenoids simultaneously in Chinese milks collected from healthy and well-nourished mothers. Nutrients were individually and contrastively analysed to report the concentrations and ratios. The lactational tendencies were specially focused to evaluate the dynamic nutrients supply in maternal–offspring dyads.
Materials and methods
Subject recruitment
The present study included lactating mothers aged between 18 and 40 years old with singleton and full-term delivery who were recruited from 1 July 2018 to 30 September 2018 at Xinhua Hospital affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China. Both mothers and their corresponding infants were medically certified as healthy (asymptomatic and with no clinical indications). Only mothers with family income beyond 150 000 CNY (21 513 USD) were included, which can surely be recognised as middle-income or above in China. All mothers were food-secured, well-nourished and had a good dietary habit as assessed by nutritionists. Anthropometric measurements were performed to collect the general characteristics including height, body weight, parity and educational background. All procedures of the present study were approved by the Ethics Committee of Xinhua Hospital (XHEC-2017-064), and informed consent was obtained from all the participants.
Human milk collection
Colostrum (1–5 d postpartum), transitional milk (11–15 d postpartum) and mature milk (41–45 d postpartum) were collected longitudinally from the same mothers. When maternal breasts were full of milk (around 09.00–11.00 hours), both sides were evacuated with an electric pump. The milk sample was carefully mixed to ensure homogenisation, from which a proportion (5–10 ml for colostrum; 20–50 ml for transitional or mature milk) was placed into a brown sterile conical tube and transported immediately to the freezer (−80°C) via cold chain within 5 h. The rest of the milk was fed to the infant. A total of ninety-five subjects participated in the study and provided 212 milk samples all together. Among them, 126 milk samples were collected longitudinally from the same forty-two mothers at three lactation stages and were chosen for the following analyses.
Fatty acid analysis
Lipids were extracted from 2·0 ml milk samples into dichloromethane–methanol–petroleum ether (30–60°C) (2:1:2, v/v/v; 20 ml) according to the previously described method(Reference Folch, Lees and Sloane Stanley25). The extraction solution was then evaporated to dryness under N2 flushing and re-dissolved in hexane. An accurate amount of tritridecanoin (C13 : 0; Nu-Chek-Prep) was added as internal standard (4·0 mg for each sample), and the mixture was then transesterified with sodium methoxide methanol solution (0·5 m, 2 ml). The generated fatty acid methyl ethers were then extracted with hexane and purified through a 0·22 μm nylon filter for the further determination.
The fatty acid methyl ethers were then separated in a capillary column (Supelco 2560, 100 m × 0·25 mm × 0·20 μm; Sigma-Aldrich) and quantified by Agilent 7890B GC with flame ionisation detector (Agilent Tech.)(Reference Wu, Gao and Tian26). The carrier gas was He with a flow rate of 0·8 ml/min and a split ratio of 1:20. Both the injector and detector temperature were 250°C. The oven programme was as follows: 100°C for the first 13 min, increased by 10°C/min until 180°C, maintained for 6 min, increased by 1°C/min until 200°C, maintained for 20 min, increased by 4°C/min until 230°C and maintained at this temperature for 13·5 min. A mixed fatty acid methyl ethers standard (GLC-746; Nu-Chek-Prep) was used to identify the fatty acid methyl ethers in milk samples. All of the twenty-eight kinds of fatty acids were quantified, and the presented results included eight kinds of major fatty acids (≥1 % of total fatty acids (TFA)) and seven kinds of LC-PUFA. All determinations were conducted by duplicate with a mixed mature milk control sample accompanying each batch.
Tocopherols and carotenoids analysis
Tocopherols and carotenoids were extracted and analysed according to the method described previously with some modifications(Reference Schimpf, Thompson and Pan27). Briefly, 4 ml water containing 0·5 g sodium ascorbate, 10 ml methanol and 1 ml tetrahydrofuran was added successively to 0·5 ml of milk sample to prevent oxidation. Saponification reaction was then operated with 1 ml aqueous solution of potassium hydroxide (45 %, w/w) under 65°C water bath for 15 min, and the tocopherols and carotenoids were extracted by 10·0 ml hexane–methyl tert–butyl ether (3:1, v/v). Then, 4·0 ml of the upper organic layer was evaporated to dryness under N2 flushing and re-dissolved in 400 μl dilution solution (10 % butylated hydroxytoluene in Methanol–methyl tert–butyl ether, 3:1, v/v) for further determination.
Tocopherols (α, γ-tocopherol) and carotenoids (lutein, zeaxanthin, β-cryptoxanthin, β-carotene, lycopene) were then analysed by reversed-phase Agilent 1260 HPLC (Agilent Tech.) on a C30 column (250 mm × 4·6 mm × 3 μm; YMC). The flow rate was set at 1·0 ml/min and mobile phase A was 100 % methanol and mobile phase B was 100 % methyl tert–butyl ether. The eluting gradient programme was: 0·0–10·0 min, 100 % A; 10·0–10·1 min, 90 % A; 10·1–20·0 min, 90 % A; 20·0–20·1 min, 62 % A; 20·1–29·0 min, 62 % A; 29·0–29·1 min, 30 % A; 29·1–40·0 min, 30 % A; 40·0–40·1 min, 100 % A; 40·1–46·0 min, 100 % A. Tocopherols were determined with fluorescence detection (Ex 290 nm, Em 327 nm; Agilent Tech.), and carotenoids were quantified with multiple wavelength detection (445 ± 4 nm; Agilent Tech.). External standards (analytical standards; Sigma-Aldrich) were used for quantification. All determinations were made by duplicate with a formula control sample (SRM 1849a, NIST) accompanying each batch.
Statistical analysis
Categorical data were expressed as percentages, and continuous data were expressed as medians and interquartile ranges (P25, P75) due to abnormal distribution assessed by the Shapiro–Wilk test. Data were log-transformed to meet normality, and indexes in human milk over different lactation stages were then compared using one-way ANOVA. Pairwise comparisons were then performed using Bonferroni correction if the results of ANOVA were significant. All of the analyses were carried out using SPSS Statistics 22.0 (IBM), and P < 0·05 was statistically significant. Figures were made by GraphPad Prism 6.0 (GraphPad Software).
Results
Basic characteristics of the lactating mothers and corresponding infants
As shown in Table 1, the forty-two lactating mothers were aged between 21 and 37 years with the median age 29 years. All mothers were Han Chinese and had completed high school education or above. The majority of the mothers had a normal pre-gestation BMI and appropriate weight gain during gestation. Most mothers were primiparas. Neonates all had an Apgar score of 10 and good growth parameters.
Table 1. Basic characteristics of the paired mothers and neonates (n 42)
(Medians and ranges (minimum, maximum) for continuous variables; numbers and percentages for categorical variables)
Fatty acids
As shown in Table 2, the amount of TFA doubled from colostrum (15·09 g/l) to transitional milk (29·95 g/l) and then remained relatively stable to mature milk (32·57 g/l). SFA (36·96–40·85 %) was the dominant fatty acids in human milk, followed by MUFA (32·53–34·95 %), while PUFA (24·85–26·86 %) occupied the least proportion. The percentage of LA remained stable over lactation, while ALA percentage increased. Consequently, the ratio of LA:ALA decreased from 19·37:1 in colostrum to 16·34:1 in transition, and 16·11:1 in mature milk. The ratio of n-6:n-3 PUFA remained constant as (10·80–11·65):1 over lactation. LC-PUFA including ARA, EPA (C20 : 5n-3) and DHA, only consisted a tiny part of milk TFA and showed significant decreases with lactation progress. The ratio of ARA:DHA was set at (1·27–1·37):1 among different lactation stages.
Table 2. Fatty acids in human milk over lactation (n 42)
(Medians and ranges (P25–P75))
TFA, total fatty acids; LA, linoleic acid; ALA, α-linolenic acid; ARA, arachidonic acid.
a,b,c Median values within a row with unlike superscript letters were significantly different (P < 0·05).
Concentrations of tocopherols and carotenoids
As shown in Table 3, milk tocopherols declined from 947·97 μg/100 ml in colostrum to 526·01 μg/100 ml in transitional milk and then declined to 361·01 μg/100 ml in mature stage. Both α-tocopherol and γ-tocopherol showed continuous decreases over different lactation stages. Compared with colostrum content, mature milk α-tocopherol declined by 65·10 % and γ-tocopherol declined by 42·11 %.
Table 3. Tocopherols and carotenoids (μg/100 ml) in human milk over lactation (n 42)
(Medians and ranges (P25–P75))
a,b,c Median values within a row with unlike superscript letters were significantly different (P < 0·05).
Similarly, total carotenoids in milk decreased from 34·42 μg/100 ml in colostrum to 20·36 μg/100 ml in transitional milk, and 10·43 μg/100 ml in mature milk. The concentrations of β-cryptoxanthin, β-carotene and lycopene all decreased continuously over different lactations. Lutein reached the highest level at transitional stage (7·12 μg/100 ml in colostrum; 9·49 μg/100 ml in transitional milk) and declined to the lowest level at mature stage (4·57 μg/100 ml). Zeaxanthin remained stable from colostrum to transitional stage and then significantly decreased at mature stage (2·15, 2·21 and 1·11 μg/100 ml, successively). Compared with colostrum content, mature milk lycopene indicated the biggest fall by 94·60 %, followed by β-carotene by 82·40 %, β-cryptoxanthin by 74·17 %, zeaxanthin by 39·47 % and lutein by 30·00 %. Carotenes decreased by 88·67 % and xanthophylls decreased by 35·92 %.
Ratio of α-:γ-tocopherol over lactation
As presented in Fig. 1, the ratio of α-:γ-tocopherol reached the highest level at 7·21:1 in colostrum, dropped to 5·29:1 in transitional milk, and further decreased to 4·21:1 in mature milk (P < 0·001).
Fig. 1. Ratio of α-:γ-tocopherol in human milk (n 42) decreased from colostrum (7:21:1) to mature milk (4·21:1). Data are medians and interquartile ranges.
Proportions of individual carotenoids over lactation
As shown in Fig. 2, the proportions of carotenoids in colostrum were listed as follows: lycopene 32·83 %, β-carotene 30·78 %, lutein 19·67 %, β-cryptoxanthin 10·78 % and zeaxanthin 5·94 %. Compared with colostrum, transitional milk and mature milk contained much higher percentages (P < 0·001) of lutein (51·64–52·49 %) and zeaxanthin (12·22–12·54 %) but much lower percentages (P < 0·001) of β-carotene (17·31–20·00 %) and lycopene (6·55–6·97 %). The percentages of β-cryptoxanthin in colostrum and the later lactation stages (9·27–11·01 %) were alike.
Fig. 2. Proportions of milk carotenoids over different lactations (n 42). Colostrum was predominated by lycopene (32·83 %) and β-carotene (30·78 %), while transitional milk and mature milk were mainly composed of lutein (51·64–52·49 %). 
, Lycopene; 
, β-carotene; 
, β-cryptoxanthin; 
, zeaxanthin; 
, lutein.
Correlations of milk fat and fat-soluble nutrients
Table 4 details the correlations between the milk fat profile (TFA content and fatty acid percentages) and tocopherols and carotenoids (absolute concentrations). After adjustment for lactation stages, milk TFA content correlated positively with tocopherols (r 0·275–0·349, P < 0·01) and xanthophylls (r 0·236–0·433, P < 0·01) but not with carotenes. The PUFA percentage showed positive correlations with tocopherols (r 0·178–0·200, P < 0·05) but not with carotenoids.
Table 4. Correlations between the milk fat profile and concentrations of tocopherols and carotenoids (n 42)*
TFA, total fatty acids; LA, linoleic acid; ALA, α-linolenic acid; ARA, arachidonic acid.
* Partial analysis adjusted for lactational stages.
‡ Correlations were significant at P < 0·05, two-tailed.
† Correlations were significant at P < 0·01, two-tailed.
Discussion
Fatty acids
Chinese human milk studies(Reference Giuffrida, Cruz-Hernandez and Bertschy5,Reference Jiang, Wu and Yu28,Reference Qi, Sun and Xia29) including our present one all reported a typically higher percentage of PUFA (22·40–30·04 %) than those (17·25–21·50 %) in the developed Asian countries(Reference Yuhas, Pramuk and Lien30,Reference Kim, Kang and Jung31) and those (14·68–19·71 %) in the Western countries(Reference Ribeiro, Balcao and Guimaraes32–Reference Zou, Guo and Huang34), which was mainly due to the higher levels of human milk LA (18·90–25·10 %) in China(Reference Giuffrida, Cruz-Hernandez and Bertschy5,Reference Jiang, Wu and Yu28,Reference Qi, Sun and Xia29) than in developed countries (11·48–16·59 %)(Reference Yuhas, Pramuk and Lien30–Reference Zou, Guo and Huang34). Consequently, a fairly high ratio of LA:ALA (16:1) was observed in the mature milk of our study, which was on the top of the range of (9·26–16·50):1 in the latest reports from China(Reference Giuffrida, Cruz-Hernandez and Bertschy5,Reference Jiang, Wu and Yu28,Reference Qi, Sun and Xia29) . This was much higher than the Chinese level of 7·62:1, dated back to 2006(Reference Yuhas, Pramuk and Lien30). Previous Chinese milk studies rarely emphasised the rising trend. According to literature(Reference Peng, Zhou and Wang35), the replacement of saturated animal fats with vegetable oils, such as sunflower oil, maize oil and soyabean oil in maternal diet, may be mainly responsible for the uptrend that we observed in the present study. Since n-6 series-derived eicosanoids are associated with proinflammatory properties, the predominance of n-6 PUFA is suggestive of higher risks of childhood allergic diseases, such as asthma and atopic eczema(Reference Waidyatillake, Stoney and Thien9,Reference Best, Gold and Kennedy15) . Also, the imbalance between n-6 and n-3 PUFA has been correlated with impaired cognitive and emotional development via epoxy metabolites and endocannabinoid system(Reference Kim, Kim and Lee12,Reference Zamberletti, Piscitelli and De Castro13,Reference Sakayori, Kikkawa and Tokuda36) . More investigations are warranted to confirm whether the increasing ratio of LA:ALA or n-6:n-3 PUFA in human milk is harmful to the short-term and long-term health of human infants.
We also found that LC-PUFA presented sharp decreases and mature milk showed the percentage of DHA as 0·44 % and ARA as 0·56 %, relatively comparable to the latest Chinese reports (DHA: 0·20–0·55 %; ARA: 0·50–0·89 %)(Reference Giuffrida, Cruz-Hernandez and Bertschy5,Reference Jiang, Wu and Yu28,Reference Qi, Sun and Xia29) and the Chinese report back to 2006 (DHA: 0·35 %; ARA: 0·49 %)(Reference Yuhas, Pramuk and Lien30). Since diets high in n-6 PUFA can result in lower endogenous conversion of ALA to n-3 LC-PUFA and impair tissue accumulation of n-3 LC-PUFA, there have been concerns that the rising ratio of LA:ALA may consequently have a negative influence on n-3 LC-PUFA levels in human milk(Reference Gibson, Muhlhausler and Makrides14,Reference Makrides, Simmer and Neumann37) . However, speculated from our findings, the human milk LC-PUFA may come from the maternal intake of preformed LC-PUFA more than the endogenous synthesis.
Tocopherols and carotenoids
The latest three Chinese studies demonstrated obviously lower levels of α-tocopherol (colostrum: 426·4–783·5 μg/100 ml; mature milk: 126·4–206·0 μg/100 ml)(Reference Xue, Campos-Gimenez and Redeuil6,Reference Jiang, Xiao and Wu38,Reference Wei, Yang and Xia39) than Western reports (colostrum: 886·0–2455·0 μg/100 ml; mature milk: 100·0–568·5 μg/100 ml)(Reference Lima, Dimenstein and Ribeiro40). However, dated back in 2002, the reported levels of α-tocopherol in Chinese milk were 898 μg/100 ml in colostrum and 331 μg/100 ml in mature milk(Reference Zhu, Zhang and Zhang41), which were used in the calculation of the Dietary Recommended Intakes for Chinese infants (the latest 2013 version)(42). Given the improving nutritional status of Chinese women, the lower levels in recent reports seemed to be illogical and potential confounding factors may exist. Nevertheless, our present data (colostrum: 840·44 μg/100 ml, mature milk: 290·65 μg/100 ml) were relatively close to the 2002 report(Reference Zhu, Zhang and Zhang41) and Western ranges(Reference Lima, Dimenstein and Ribeiro40). More high-quality studies are warranted. Human milk γ-tocopherol(Reference Martysiak-Zurowska, Szlagatys-Sidorkiewicz and Zagierski4,Reference Wei, Yang and Xia39) and carotenoids(Reference Macias and Schweigert3,Reference Schweigert, Bathe and Chen43–Reference Zielinska, Hamulka and Wesolowska45) were less investigated, and available data showed large disparities among global published data, which may be mainly due to different intakes of these nutrients.
In agreement with previous human milk studies worldwide(Reference Macias and Schweigert3,Reference Martysiak-Zurowska, Szlagatys-Sidorkiewicz and Zagierski4,Reference Wei, Yang and Xia39,Reference Lima, Dimenstein and Ribeiro40,Reference Schweigert, Bathe and Chen43–Reference Zielinska, Hamulka and Wesolowska45) , fat-soluble antioxidants in our study all indicated significant decreases, ranging between 30·00 and 94·60 % from colostrum to mature milk. Additionally, the present study demonstrated a down-regulated ratio of α-:γ-tocopherol from 7·21:1 to 4·21:1 with the progress of lactation, which is similar to previous milk studies (calculated by reported individual data)(Reference Martysiak-Zurowska, Szlagatys-Sidorkiewicz and Zagierski4,Reference Wei, Yang and Xia39) . Polarity may be involved because α-tocopherol possessing lower polarity indicated bigger decreases than γ-tocopherol over lactation. As positive roles have been attributed to α-tocopherol in bronchopulmonary dysplasia and childhood asthma, γ-tocopherol indicates more harmful effects through pro-inflammatory properties than its antioxidant benefits(Reference Stone, McEvoy and Aschner19). However, infant formulas, calculated by the published data, provide a much lower ratio of α-:γ-tocopherol as (0·68–2·51):1(Reference Martysiak-Zurowska, Szlagatys-Sidorkiewicz and Zagierski4,Reference Stimming, Mesch and Kersting10) . Whether the balance of α-:γ-tocopherol in human milk exerts special significance on the offspring health needs further investigations and the extensive additions of γ-tocopherol in formulas should also be reconsidered.
Among carotenoids, we observed that less polar carotenes were more subjected to changes, which was supported by other studies(Reference Macias and Schweigert3,Reference Zielinska, Wesolowska and Pawlus21) . Consequently, lycopene and β-carotene predominated in colostrum carotenoids, while mature milk was more enriched with lutein. From our perspective, none of the former studies figured out the special dominance mode of carotenoids during different lactation stages, but after calculation by the reported levels(Reference Macias and Schweigert3,Reference Zielinska, Wesolowska and Pawlus21,Reference Schweigert, Bathe and Chen43) , similar dominance was observed ignoring the distinct absolute concentrations from our study. Polarity has been found to influence the distribution of carotenoids among lipoprotein, carotenes are generally transported via LDL and enriched in LDL receptor-rich tissues, such as liver and adrenal gland, while xanthophylls indicating higher polarity tend to be more associated with HDL and enriched in the retina and nervous system(Reference Macias and Schweigert3,Reference Zielinska, Wesolowska and Pawlus21,Reference Schweigert, Bathe and Chen43) . Similar to the previous study(Reference Schweigert, Bathe and Chen43), colostrum carotenoids pattern resembled those of maternal plasma and the LDL fraction, whereas in mature milk, the pattern was similar to the HDL fraction. Supported by these findings, different transfer mechanisms may be involved during colostrogenesis and later lactation stages. Each fat-soluble nutrient has its own implications and may be of different priorities in the early life nutrition. As reported in literature(Reference Bohn46), infant food formulas contained limited levels of carotenoids (0–23·3 μg/100 ml) and the carotenoids profile indicated no obvious differences from the initial formulas to the follow-on formulas. Hence, bottle-fed infants may not benefit from the uniquely tailored human milk. Future formula design must pay close attention to the lipid patterns.
Correlations between lipid components in human milk
Based on the opposite altered trends of TFA content and the levels of carotenoids and tocopherols in human milk, it can be assumed that the transfer of lipophilic nutrients into milk is not a mere reflection of the transfer of TAG. However, we found that milk tocopherols and xanthophylls but not carotenes correlated positively with milk total fat content after adjustment for lactation stages. These findings supported our above speculation that milk fat might serve as the transfer vehicle for lipophilic nutrients via selective mechanisms, possibly the lipoprotein-associated processes involving polarity. We also found that tocopherol levels but not carotenoids were positively associated with PUFA percentages in human milk, which was consistent with the theory that vitamin E may protect PUFA from oxidation and the infant’s requirements of vitamin E are in relation to PUFA intake(Reference Stimming, Mesch and Kersting10,Reference Raederstorff, Wyss and Calder11) .
Limitations of the study
The limitations of the present study should be noted. Firstly, data on maternal dietary intakes were gathered but not calculated yet since the Chinese Food Composition Database lacks the data of γ-tocopherol and carotenoids. We are working on a comprehensive composition database using the US Department of Agriculture (USDA) Food and Nutrient Database, and further reports of dietary consumption can be expected. However, food security could not be an issue for the participants because the study inclusion criteria limited the study subjects to healthy, well-nourished women with balanced dietary habit and good socio-economic status. Secondly, only a single milk sample (09.00–11.00 hours) was collected and may not be sufficient to reflect the daily content. This limitation was shared by most milk studies due to the feasibility and compliance concerns(Reference Ruel, Dewey and Martinez47). The time interval was specially chosen because in our pre-survey, mothers were likely to undergo complete breast fullness during this period and there would be enough time for milk collection and intraday transportation.
Conclusions
The present study was a comprehensive report of human milk lipid profiles including fatty acids, tocopherols and carotenoids over three different lactation stages. Mature milk contained a distinct lipid profile from colostrum with higher total fat content but lower LC-PUFA, antioxidants and α-:γ-tocopherol ratio. Meanwhile, mature milk mainly consisted of lutein, whereas lycopene and β-carotene predominated among carotenoids in colostrum. Polarity-associated lipoprotein transfer mechanisms may be involved in the transformation of milk content and composition. In spite of the distinct absolute concentrations among different studies, the balances of nutrients and their variation tendency during lactation were rather comparative. The ratios of nutrients might indicate more significance than the measurements of individual nutrients.
Acknowledgements
This research was supported by Abbott Nutrition Research & Development Center, Shanghai, China. M. C., B. W., Y. Z. and J. Z. conceived and designed the study protocols. B. W., L. Z. and L. S. contributed to the subject recruitment. R. G., K. W. and Z. L. conducted the sample collection. R. G. and K. W. conducted the sample determinations. K. W., Y. M. and J. Z. wrote the manuscript. M. C. and B. W. were primarily responsible for the final contents. All the authors read and approved the final manuscript.
None of the authors has any conflicts of interest to declare.
