2.2. Questionnaire

With assistance from our trained and experienced researchers, basic information and dietary items were obtained from the recruited mothers during every sampling period on personal sociodemographic characteristics, physical status, dietary intakes, medical histories, and other necessities. Explicitly, maternal relevant data were recorded at the following time points: (a) necessary personal information acquired at postpartum periods among 0–5 days (colostrum stage), 7–15 days (transitional milk stage), and 26–30 days (mature milk stage) when breastmilk samples were collected; and (b) detailed dietary intakes recorded for 3 days before the collection of breastmilk samples. A 32‐diet‐item food frequency questionnaire (FFQ) with practicable validity and reproducibility was adopted for the diet investigation of local lactating women. The items of food products were divided into 12 groups according to previous relevant reports and local Tibetan diet habits in Shigatse (Cheng et al., ²⁰⁰⁹; Z. Wang et al., ²⁰¹⁰). The feedback of FFQ was consisted of the frequency and consumption quantity of a certain food or beverage, and its existing form in certain relevant recipe. The evaluation of food intakes by FFQ was aimed to assess the level of nutrition intakes of local lactating women with adherence to traditional Tibetan diet pattern. Macronutrient intakes of carbohydrate, protein, fat, and energy were calculated according to the 2017 China Food Composition Tables (INFS, C. C., ²⁰¹⁷), corresponding to the items of macronutrients in human milk analysis.

2.3. Breastmilk sample collection

After finishing 3 × 24 hr daily diet questionnaires, participants were asked to collect complete either‐side breastmilk by electric breast pump in the next forenoon between 09:00 a.m. and 11:00 a.m. in a sanitary and quiet environment. 60‐ml aliquot of the well‐shaken breastmilk was then divided equally into 10‐ml sterilized tubes. All the breastmilk samples were prestored at −20°C for no more than 7 days before being transported by ice box to −80°C refrigerators until analysis.

2.4. Analysis of human milk samples

3.2. Characteristics of Tibetan diet pattern among lactating women in Shigatse

The description of diet pattern among Tibetan lactating women in Shigatse is revealed in Table 3. Summarily, the selection of certain food types has basically possessed a shared commonness by Tibetan diet tradition, with some varied characteristics by individual maternal preference during the lactation period. Specifically speaking in food categories, Zanba (a traditional Tibetan food, which is cooked by roasting highland barley flour, usually mixed with buttered tea or red meat broth (Qu et al., ²⁰¹⁵)) was the dominating cereal type for staple food, consumed by over 98% participants through all three lactation stages. Other cereal choices included noodle, rice, and deep‐fried dough stick. Particularly, Zanba seemed to be an exclusive choice of cereal among lactating women in colostrum stage, while the profile of cereal category would exhibit more diversity in the following transitional milk and mature milk stages. Moreover, Zanba was also used as the first type of cereal introduced to local Tibetan infants in about 1 month after birth according to previous publications (Dang et al., ²⁰⁰⁵). Such priority to Zanba in local maternal and infant diet pattern should be owed to the long‐held traditional diet culture among native Tibetans at all ages, especially in the remote rural areas (Wang et al., ²⁰¹⁰). High‐fat drinks of broth (soup of beef, mutton, or bone) and Tibetan buttered tea were another traditional dietary category commonly included in the daily diet of the participants. Using a native way from the unique Tibetan tradition, the cooking of Zanba with broth or Tibetan buttered tea is a typical main course in their daily life. As revealed in this study, the intake frequency of the combination of Zanba and broth or Tibetan buttered tea was approximately at least once per day for about 80%–90% participants, many of whom even consumed more than once in their everyday diet. Other traditional Tibetan teas, such as light black tea (pure black tea) and sweet tea (Tibetan milk tea), were the most common beverages for Tibetan lactating women in this cohort study. Unexpectedly somewhat, highland barley liquor was involved in the diet of some participants during their lactating period, which was quite surprising for the reason that the alcohol ingested by mothers might transfer and bring adverse healthy implication to infants. However, some researchers have considered that low‐level alcohol consumption during breastfeeding might not presumably linked with the potential adverse outcomes if manageable breastfeeding strategies can be adopted (Wilson et al., ²⁰¹⁷). Red meat, generally including beef and mutton, were the mainly consumed meat among the participants, which were usually cooked with Zanba and buttered tea or other traditional Tibetan staple foods, while white meat, such as poultry meat and fish, was rarely involved in their daily diet. Such inclination to red meat rather than white meat might be due to the high‐altitude environment and traditional Tibetan diet culture. Moreover, eggs were not a commonly involved food category in the participants' recipe, possibly on the account of less adoption upon poultry farming than livestock farming in such areas (Shang et al., ²⁰¹⁴). It is unsurprisingly that pure milk was not taken as a drink for the lactating women, since milk was conventionally used as the ingredient for traditional Tibetan drinks and beverages, such as sweet tea. Vegetables and fruits were not a daily acquisition for every lactating mother, especially in the colostrum stage. Additionally, the ordinary vegetables were root and tuber vegetables, such as turnips, carrots, tomatoes, and potatoes, while leafy vegetables and fruits were taken less frequently, which were generally spinach, Chinese cabbage, pakchoi, and apple and banana according to our detailed information.

Consequently, the lactating women in the rural areas of Shigatse were customarily prone to a traditional Tibetan dietary pattern, invariably including Zanba, Tibetan buttered tea and sweet tea, red meat, fatty soup, and root and tuber vegetables with the insufficiency of leafy vegetables, fruits, and white meat, which was in accordance with the findings of previous relevant reports (Ge et al., ¹⁹⁹⁷; Wang et al., ²⁰¹⁰). Referring to the advice from Chinese dietary guidelines, lactating mothers are recommended maintaining an assorted, balanced and moderate diet profiles, which consist of cereals (whole grains and legumes included), leafy and colored vegetables, fruits, various meats (white and red meats and seafood included), dairy products, nuts, etc. (Wang et al., ²⁰¹⁶). Moreover, the diet patterns of lactating mothers in low‐altitude regions of China, especially urban areas, are more diversified and individualized than in Shigatse, yet with certain deficiencies in some nutrients (Ding et al., ²⁰²⁰; He et al., ²⁰¹⁵; Huang & Hu, ²⁰²⁰; Wang et al., ²⁰²⁰).

3.3. Contents of macronutrients in human milk from Shigatse

The macronutrient contents of fat, crude protein, lactose, total solids, true protein, and energy in human milk are shown in Table 4, which were measured by MIRIS HMA human milk analyzer. The calibration of HMA is based on established standard methods, which are universally used with the recommendations of ISO (International Organization for Standardization) and IDF (International Dairy Federation), such as the Röse–Gottlieb method for fat content determination, and the Kjeldahl method for crude protein and true protein (crude protein minus nonprotein nitrogen) (Giuffrida et al., ²⁰¹⁹). The contents of crude protein, lactose, and true protein in human milk were significantly different (p < .05) between the three lactation stages of colostrum, transitional milk, and mature milk. Similar to previous relevant reports, protein and true protein content of human milk decreased with the progress of lactation, especially between the stages of colostrum and transitional milk (Bauer & Gerss, ²⁰¹¹; Larnkjaer et al., ²⁰⁰⁶). Meanwhile, the contents of fat, total solids, and energy in the Shigatse human milk samples were insignificantly different (p > .05) during the lactation period. The level of lactose in human milk samples was nonmonotonously increasing/decreasing during lactation.

3.4. Composition of fatty acids in human milk from Shigatse

The fatty acid composition of human milk from Shigatse is displayed in Table 5. 36 fatty acids were detected in the human milk samples, including 17 saturated fatty acids (SFAs), 8 mono‐unsaturated fatty acids (MUFAs), and 11 polyunsaturated fatty acids (PUFAs). Significant differences (p < .05) were observed among the contents of varied fatty acid categories, such as individual fatty acids (single SFA, MFA, and PUFA) and certain fatty acid ratio (ARA: DHA). Palmitic acid (C16:0) was the most abundant SFA and second most abundant fatty acids in all three lactation stages. The content of oleic acid (C18:1 n−9c) was of the highest level among all MUFAs and even all fatty acids in human milk of any lactation stage. The content of LA (linoleic acid, C18:2 n−6c) was the most abundant PUFA in human milk and increased significantly (p < .05). Through the three lactation stages, various functional unsaturated fatty acids, including ALA (α‐linolenic acid, C18:3 n−3), EPA (eicosapentaenoic acid, C20:5 n−3), and LA (linoleic acid, C18:2 n−6c) were increased significantly in content (p < .05), while DHA (docosahexaenoic acid, C22:6 n−3), inversely, was decreased significantly (p < .05) in content simultaneously. Particularly, DHA has been argued to be sensitive to dietary level, which is likely to perform obvious variation trends in content when maternal diet profiles alter. In this study, DHA contents in the human milk samples from Shigatse were lower than the samples from areas with different diet patterns, especially among the coastal regions with sufficient seafood, such as South Korea, Hangzhou of China, Philippines, and Japan (Jiang et al., ²⁰¹⁶; Kim et al., ²⁰¹⁷; Yuhas et al., ²⁰⁰⁶). Among the SFA category, butyric acid (C4:0), caproic acid (C6:0), caprylic acid (C8:0), capric acid (C10:0), lauric acid (C12:0), arachidic acid (C20:0), and lignoceric acid (C22:0) were found to have significant differences in content among the three lactation stages, especially between colostrum and transitional milk. Moreover, for the sums of certain fatty acid categories, including the sum of n−3 fatty acids, n−6 fatty acids, SFA, MUFA, PUFA, and TFA, no significant difference was observed through varying lactation stages. The above analytical results were consistent with the relevant findings of previous publications (Jiang et al., ²⁰¹⁶; Thakkar et al., ²⁰¹⁹). Additionally, the fatty acid profile of human milk in Shigatse was comparatively deficient than in Chinese low‐altitude regions, especially the urban or coastal areas, which might be partly due to different dietary preferences. It has been reported that lactating mothers in Chinese low‐altitude coastal areas possessed higher intake of long‐chain PUFA, including DHA and EPA, than the mothers in the inland areas, which was assumed to be influenced by different dietary habits (Jiang et al., ²⁰¹⁶; Peng et al., ²⁰⁰⁹). Meanwhile, the intakes of certain FAs, such ARA, might be insignificantly impacted by the differences between diet patterns in China (Peng et al., ²⁰⁰⁹).

3.5. Composition of free amino acids in human milk from Shigatse

As described in Table 6, 18 FAAs were detected in the human milk samples from Shigatse. Glu was the most abundant FAA among all FAAs during the lactation period. Tau was of the highest level among all essential FAAs and second highest among all FAAs in human milk samples. Specifically, it is suggested that free Glu in human milk has a down‐regulating effect on the appetite and growth of infants, and it can be served as a satiety signal corresponding to the receptors in the oral cavity and gastrointestinal tract during feeding, which might be beneficial to the infant obesity inhibition, and long‐term fitness (Larnkjær et al., ²⁰¹⁶). The contents of individual FAAs found to have significant differences (p < .05) in varying degrees among the three lactation stages. In particular, the contents of His, Tau, Ser, Arg, Glu, and Thr were significantly different (p < .05) between colostrum stage and the other two lactation stages, while the content of Ala in colostrum was significantly different with mature milk, but indifferent (p > .05) with transitional milk. Nine essential FAAs, His, Tau, Thr, Lys, Met, Val, Ile, Leu, and Phe, were detected in human milk through all three lactation stages. The whole content of essential FAAs was decreased significantly (p < .05) during colostrum and transitional milk stages, and remained insignificantly variable (p > .05) in mature milk stage. The content of total free amino acids was similar with the variation trend of essential FAAs through the lactation period, which was consistent with the results from other researches (Garcia‐Rodenas et al., ²⁰¹⁶; Zhang et al., ²⁰¹³).

3.6. Classification of maternal diet profiles through different lactation stages during the first month postpartum

Principal component analysis (PCA) was adopted to display the classification of maternal diet profiles through three lactation stages. From Figure 1a, the distribution of maternal diet profile in colostrum stage was obviously more centralized than that in transitional milk stage and mature milk stage, and the distributions of maternal diet profiles in transitional milk stage and mature milk stage were quite scattered within either category and some overlapped between the both categories. It might imply a more particular maternal diet profile in colostrum stage and some resemblance in diet profiles between transitional milk stage and mature milk stage. The contribution of certain food products to the classification of maternal diet profiles is shown in Figure 1b. The inner ellipse and outer ellipse revealed 50% and 100% of the explained variances, respectively. The food products of Zanba, other cereals (rather than Zanba), other vegetables (rather than leafy vegetables), and egg performed more correlative significance in the PCA determination of the lactation stage differences upon maternal diet profiles.

Open in a separate window

FIGURE 1

PCA and correlation analysis of stage differences in the food profile upon traditional Tibetan maternal diet pattern. (a) PCA score plot of food profiles in different lactation stages. (b) Correlation loadings of food products for PCA classification

3.7. Correlation analysis of macronutrients between maternal diet profiles and human milk during different lactation stages

As shown in Table 7, the intake level of protein was lower than the recommended nutrient intake (RNI) of protein for Chinese lactating mothers; meanwhile, the intake level of carbohydrate was higher than the RNI of carbohydrate for Chinese lactating mothers. Besides, all the macronutrient intakes were fallen in the ranges of previous reports regarding the nutrient intakes of mothers from the relevant rural areas in western China (Cheng et al., ²⁰⁰⁹; Liu et al., ²⁰¹⁵; Wang et al., ²⁰¹⁰). Moreover, the intakes of energy and fat were through similar variation trends during the three lactation stages: initially held at significantly low (p < .05) levels in the colostrum stage, then increased significantly (p < .05) in the transitional milk stage and decreased significantly (p < .05) in the following mature milk stage. Furthermore, the intakes of protein were at low levels with insignificantly varying degrees (p > .05) through the lactation stages, while the intakes of carbohydrate increased significantly (p < .05) from colostrum stage to transitional milk stage and kept at a relatively steady level (p > .05) afterward. Correlated with the diet frequency of food products among participants (Table 6) and their living style, the above results of daily macronutrient intakes can be interpreted as follows: (a) dietary patterns—according to the Tibetan dietary habit, the main type of food products was the staple food, basically cereals, among which highland barley was at a dominating dietary level. As is known, cereals are abundant in carbohydrate and energy rather than in protein and fat (e.g., highland barley in average: 62 g/100 g in carbohydrate and 298 kcal/100 g in energy, but 10.2 g/100 g in protein and 1.2 g/100 g in fat; rice in average: 76 g/100 g in carbohydrate and 352 kcal/100 g in energy, but 12 g/100 g in protein and 0.8 g/100 g in fat) (INFS, C. C., ²⁰¹⁷), which would thus lead to a relatively more carbohydrate and less protein intake profiles in their daily diet; and (b) lactation stages. According to the findings of dietary frequency survey, the diet options were varying between either lactation stages. Explicitly, lactating women in general were prone to a monotonous diet pattern in colostrum stage, merely including Zanba, buttered tea, and red meat and bone soup, which could be reflected in the shortage of daily energy, protein, and fat intakes. The diet profiles favored relatively more diversified combination in transitional milk stage, with more occurrence of other cereals, eggs, leafy vegetables, and fruits, which significantly (p < .05) enhanced the intake values of energy, fat, and carbohydrate. In mature milk stage, the selection of food species was further more various and individualized in each food category, while such variability was confined to the substitution within the same food category, which did not present obvious differences across different food categories. Moreover, the intake values of energy and fat in mature milk stage were significantly lower than the intakes in transitional milk stage, which might be due to the alteration of maternal nutritional needs since the extended duration of lactation period.

Bivariate correlation analysis was processed upon a homogeneous comparison of macronutrients between daily maternal diet intakes and the corresponding human milk samples by SPSS. Interpretively, there are mainly two constituent parts of carbohydrates in human milk, namely lactose and oligosaccharides. Thereinto, lactose consists the principal proportion (generally over 90%) in human milk to ensure appropriate nutrition and development needs for infants, while oligosaccharides in human milk, primarily as prebiotics, exert different effects in the infant gut (Gridneva et al., ²⁰¹⁹). Therefore, it is admissible to correlate lactose in human milk with carbohydrates in maternal diets upon the aspect of macronutrients.

As displayed in Table 7, only the item of protein displayed varying significant degrees of correlation through the three lactation stages, whereas other macronutrients in human milk were uncorrelated with the relevant macronutrients in daily intakes. Such results were not quite parallel with some previous relevant reports (Aumeistere et al., ²⁰¹⁹; da Cunha et al., ²⁰⁰⁵). In this research, the intake of protein was negatively weak‐correlated between human milk and maternal diet in the colostrum stage, then positively medium‐correlated in the transitional milk stage, and finally uncorrelated in the mature milk stage. Meanwhile, protein intakes from maternal daily diet were lower than the values of Chinese RNIs through all the three lactation stages. However, the protein contents in human milk samples did not possess obvious inadequacy accordingly in the corresponding lactation stages (Table 3), which was consistent with some relevant publications (Shi et al., ²⁰¹¹; Zhu et al., ²⁰¹⁷). Such consequence might reveal that protein shortage in daily maternal diet was still a restrictive factor for the lactating women, especially in the colostrum and transitional milk stages. Benignantly, the nutritional metabolism involving multiple constituents in maternal digestion tract and mammary gland would be conducive to the maintenance of protein content in human milk. Meanwhile, the overcapacity contents of carbohydrate in maternal daily diets did not induce the excessive proportions of carbohydrate in their human milk, which indicates that the internal metabolism might be more influential than the external maternal dietary profiles to the ultimate composition of human milk.

All the participants and volunteers in this study are heartily acknowledged for sample and data provision. Senior study nurse Yuzhen Ciren is acknowledged for the sample collection and data survey all along the cohort study. Guowen Wang and Tie Lin are appreciated for the sample transport from Shigatse to Beijing.