Women of reproductive age are especially at risk of developing or worsening obesity, caused by excessive weight gain during pregnancy(Reference Mannan, Doi and Mamun1). Postpartum weight retention (PPWR) is a short- and long-term risk factor for overweight and obesity in women(Reference Picciano2,Reference Kirkegaard, Stovring and Rasmussen3) . Additionally, PPWR incurs an increased risk for complications in subsequent pregnancies(Reference Poston, Caleyachetty and Cnattingius4,Reference Bogaerts, Van den Bergh and Ameye5) . In China, the prevalence of high PPWR at 2 years postpartum was 41·5 % in 2013 (high PPWR was defined as ≥5 kg)(Reference Wang, Yang and Pang6). PPWR may be particularly harmful, as its damaging effects cross generations. Recent studies reported that PPWR may be associated with the incidence of diabetes, heart disease and hypertension(Reference Huang, Brown and Curran7,Reference Rooney, Schauberger and Mathiason8) . Moreover, it contributes to an increased risk of obesity in the offspring, causing an intergenerational cycle of obesity(Reference Williams, Mackenzie and Gahagan9).
Learning about potential determinants of PPWR is critical for the development of effective interventions to optimise the trajectory of weight gain, particularly among those mothers at high risk. At first, there are many previous researches on the relationships between prepregnancy BMI on either PPWR or offspring weight, which indicated conflicting results(Reference Goldstein, Abell and Ranasinha10,Reference Martin, Hure and Macdonald-Wicks13) . Then, many studies have found that the main reason for PPWR was gestational weight gain (GWG)(Reference Jaakkola, Hakala and Isolauri14,Reference Subhan, Shulman and Yuan18) . Some studies have indicated that dietary intake is associated with PPWR. A previous cross-sectional survey among lactating women in south central China suggested that pattern with a high intake of fresh vegetables (non-leafy), soya milk, probiotics and algae, and fresh legumes was negatively associated with PPWR(Reference Huang, Li and Hu19). Higher energy, protein, carbohydrate and fat intake in diet were significantly associated with higher 6 months PPWR(Reference Fadzil, Shamsuddin and Wan Puteh20). Also, lactation is a critical period for nutritional needs. A representative traditional Chinese postpartum diet (also called a confinement diet or a diet for puerperium rest and recovery in China, which consists of too much animal foods such as eggs and various soups of chicken or trotters) has an excessive amount of protein and energy which potentially increases the risk for postpartum weight gain over the short-term, obesity and related health problems in the long-term(Reference Chan, Nelson and Leung21,Reference Boghossian, Yeung and Lipsky23) . Nutritional requirements increase not only to support infant growth and development but also to promote maternal postpartum recovery(Reference Barrera, Valenzuela and Chamorro24,Reference Vonnahme, Lemley and Caton25) . Intake of nutrients during this critical period may have an important effect on maternal weight(Reference Jaakkola, Hakala and Isolauri14).
As suggested from some previous survey, there is marked association of Se nutritional status with human body weight. Some studies founded that blood/serum Se showed positive relationships with obesity or BMI in children, adolescent and adults from all over the world, including China(Reference Fan, Zhang and Bu26,Reference Cavedon, Manso and Negro30) . Among them, an intervention study showed that Se supplementation can significantly reduce the body fat mass. A cross-sectional study supported an inverse association between fingernail Se levels and the risk of obesity in Chinese children(Reference Xu, Chen and Zhou31). Another cross-sectional study also indicated a trend of low Se biomarkers in the overweight/obese group, although the differences were not statistically significant (P > 0·05)(Reference Larvie, Doherty and Donati32). In addition to functions of immune, endocrine, cardiovascular, reproductive, nervous systems and anti-cancer activity, it is well known that Se plays a significant and dimness role in modulating insulin signalling, and consequently carbohydrate and lipid metabolism(Reference Avery and Homann33–Reference Steinbrenner40). During recent decades, some epidemiological surveys around the world also showed that indicators of diabetes or the metabolic syndrome were positively associated with Se content in plasma or nails, as well as selenoprotein P in plasma(Reference Ogawa-Wong, Berry and Seale41,Reference Oo, Misu and Saito46) . However, to our known, no survey for this relationship for lactating women was reported.
In consideration of the potential roles of selenoproteins in thyroid function which are critical for body growth and energy production, a cohort study evaluated the association of low Se status with hypothyroidism during pregnancy and the association of maternal low thyroid function with infant birth size, which indicated that low Se status during pregnancy may associate with low thyroid function that was related to low birth weights(Reference Kohrle, Jakob and Contempre47–Reference Guo, Zhou and Xu50). A prospective observational study in 2014 that involved 126 pregnant women between 28 and 32 weeks gestation revealed an association between lower maternal Se levels and delivery of small-for-gestational age infants suggesting Se deficiency as a possible risk factor for intra-uterine growth retardation(Reference Kumpulainen, Salmenperä and Siimes51).
The purpose of this study was to investigate the association between daily Se intake with PPWR among Chinese lactating women, and their infants’ physical development.
Methods
Study population
This cross-sectional study recruited 450 lactating mothers from three geographical locations with different Se levels in soil, a Se-deficient area (Liangshan, Sichuan province), a Se-sufficient city (Xicheng District, Beijing) and a Se-toxic region (Enshi, Hubei province), during the first week after delivery in 2014. All participants had delivered a normal single infant at full term and intended to breastfeed their child. As their own cultures and collection of precious breast milk, some subjects did not go back to local hospital again postpartum, especially from these minority areas (Enshi and Liangshan). The eligibility criteria included subjects whose infants were breastfed fully at the end of the third month postpartum, healthy according to self-evaluation and no smoking. Exclusion criteria included women with the metabolic syndrome or chronic diseases, such as diabetes mellitus, hypertension or goitre during pregnancy and lactation. We excluded individuals who have extreme energy intakes: <2093 or >14 650 kJ/d), as identified by the Goldberg equation modified by Black(Reference Black52). Thereby, 264 healthy lactating women were included in this study, thirty-seven from Liangshan, 128 from Beijing and ninety-nine from Enshi, which was acceptable for the requirement of >80 % power, regarding Cohen’s d criteria (power = 0·80, 95 % CI, α = 0·05, and dz = 0·50), based on the Se levels in serum and human milk. All 264 pairs of healthy mothers and their infants were divided into three groups according to daily Se intake. The flow chart of participants is shown in Fig. 1. Written informed consent was obtained from all subjects prior to study participation. This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures were approved by the Ethics Committee of National Institute for Nutrition and Health, Chinese Center for Disease Control and Prevention. This study design and examination procedures have been described in detail elsewhere(Reference Han, Liu and Lu53).
Flow chart of participants.
Demographic information collection
Information on socio-demographic factors including each mother’s age, infant’s sex, delivery mode (vaginal delivery, forceps delivery or caesarean delivery), parity, exercising and sleeping time were collected by questionnaires.
Dietary assessment
Dietary information was collected using a 24-h food record method on three consecutive days at the 42nd day postpartum. Participants were trained to record meals and snacks consumed by themselves during a 24-h period, for three consecutive days. The weight and volume of each food or beverage were estimated with different sizes of household measuring tools such as bowls, cups and spoons; besides, pictures of each food in the raw and cooked state with reference objects to assure the data were as reliable as possible.
Since this investigation in these participants was conducted in 2014, the daily dietary intakes of energy and three macronutrients were calculated based on the new version of the China Food Composition Tables(Reference Yang54).
Anthropometric measurements
Prepregnancy weight, weight before delivery and current body weight were recorded in kg. The prepregnancy weight, weight before delivery, newborn weights and lengths were obtained from the mothers’ Maternal-Infant Health Handbooks. The prepregnancy weight was self-reported by each subject. At the 42nd day postpartum, the current maternal weight (to the nearest 100 g) was measured with an electronic scale (TANITA HD370) and height (to the nearest 0·5 cm) was measured by a trained data collector using a measuring tape. The weight (to the nearest 10 g, Seca 335 in Poitiers) and length (to the nearest 0·1 cm, YSC-2 measuring machine) of their infants were also measured. All results were recorded as a mean value, which was calculated from two separate measurements. In the case when a third measurement was taken, a median was used in place of the mean when the difference between the two former measurements was more than 10 %(Reference Han, Liu and Lu53). GWG was calculated as the weight before delivery minus the prepregnancy weight. PPWR was calculated as the current body weight minus the prepregnancy weight. According to Institute of Medicine guidelines, prepregnancy BMI was categorised as underweight, normal, overweight or obese and GWG was categorised as insufficient, optimal or excess weight gain (Appendix 1)(55).
Infant physical evaluation
We used the WHO Anthro software to calculate the Z-scores of weight-for-length (WLZ), length-for-age (LAZ), weight-for-age (WAZ) according to infant’s sex, birth date and check-up date. Based on the WHO growth standards, stunting was defined as LAZ < −2; underweight was defined as WAZ < −2; wasting was defined as weight lighter than the corresponding weight of WLZ of −2 for particular length and sex; overweight means as weight heavier than the corresponding weight of WLZ of 2 for specific length and sex and defined WAZ > 2; obesity was defined as weight heavier than the corresponding weight of WLZ of 3 for particular length and sex(56). Outliers in outcomes based on the WHO standards (LAZ < −6, LAZ > 6, WAZ < −6, WAZ > 5, WLZ < −5 and WLZ > 5) were also dropped (n 4)(56).
Sample collection
At the 42nd day postpartum interview, a 10 ml sample of breast milk and a 5 ml blood sample (collected in a heparin sodium tube for plasma) were collected from all the subjects. All the above steps were performed by local hospital nurses. To minimise discomfort and maximise participation, mothers could collect breast milk at any time and no time constraints were given for when to express it with respect to their infant’s feeding. Blood (3 ml) was centrifuged at 10 000 g at 4 °C for 10 min to separate the plasma. The samples of milk and plasma were stored at −80°C until chemical analysis was performed(Reference Han, Liu and Lu53).
Laboratory analysis
Samples of plasma and breast milk collected at the 42nd day postpartum interview were used for the determination of the Se contents, the procedure of which has been described in details previously(Reference Han, Liu and Lu53). To be specific, a previously heated (25°C) and shaken plasma sample or breast milk sample was digested by a CEM MARS Xpress microwave system (CEM). The cooled, digested samples were diluted to 10 ml with ultrapure water and analysed for total Se content by inductively coupled plasma MS (Agilent 8800). Three independent replicates were conducted, and the respective blanks were considered in the final results. Accuracy of the Se analysis was assessed during each batch of analysis using a standard reference material (SRM 1549, National Institute of Standards and Technology).
The method for the calculation of dietary Se daily intake has also been previously described in details(Reference Han, Liu and Lu53). In brief, as lacking the sufficient data for Se content in each region, the dietary Se intakes were estimated from the plasma Se concentrations by employing the following formula: log (daily Se intake (μg/d)) = 1·623 log (plasma Se concentration (mg/l)) + 3·433(Reference Yang, Yin and Zhou57).
Based on the recommended dietary Se intake for lactating Chinese women(58), 264 early lactating Chinese women were classified into three groups: the insufficient Se group (<78 μg Se/d), the optimal Se group (78–400 μg Se/d) and the excessive Se group (>400 μg Se/d).
Statistical analysis
Epidata software 3.1 was applied to build a database. Means and standard deviations were used to express the values of normally distributed data and medians and 25th and 75th percentiles to express skewed data. The comparison between group differences after ANOVA was evaluated by least significant difference, and when indicated for non-parametric analysis, we used the Mann–Whitney U or Kruskal–Wallis test, and χ 2 test for categorical variables. The group changes of infants’ Z-scores between at birth and at the 42nd day were evaluated by the paired-samples t test.
The directed acyclic graph (Fig. 2) used to select the variables was created using DAGitty version 3.0(Reference Textor, van der Zander and Gilthorpe59). Multivariable linear regression was conducted to examine the relationships between dietary intake of Se and PPWR, adjusting the following covariates: age, parity, GWG, prepregnancy BMI, dietary intakes of protein fat, carbohydrate and total energy and time of physical activity and sleep, based on Fig. 2(a). Multinomial logistic regression was applied to estimate the associations of change of mothers’ weight with their infants’ physical Z-scores at birth (adjusting for age, parity and household income) and maternal Se nutritional status (including daily Se intake and Se content in human milk) with their infants’ Z-scores at 42nd day (adjusting for age, parity, Z-scores at birth, daily dietary intake of total energy and three macronutrients and household income), based on Fig. 2(b). The dependent variables were the categories of WLZ, LAZ and WAZ, according to WHO growth standards(Reference Roman, Jitaru and Barbante49).
Directed acyclic graph representing the causal assumptions used for covariate selection ((a) the multivariable linear regression analysis; (b) the multinomial logistic regression analysis). 
, Exposure; 
, outcome; 
, ancestor of exposure; 
, ancestor of outcome; 
, ancestor of exposure and outcome; 
, adjusted variable; 
, unobserved (latent); 
, other variables; 
, causal path. PPWR, postpartum weight retention; GWG, gestational weight gain.
All of the analyses were performed using SPSS version 25.0 (IBM Corp.). G-Power (version 3.1.9) was used for power analysis. Statistical significance was accepted at a two-sided P value of <0·05. The variance inflation factor was used for identifying multicollinearity. No multicollinearity was observed in our multivariate analysis (variance inflation factor < 2·0).
Results
Table 1 shows the demographic characteristics of the lactating women grouped by three different levels of daily Se intake. The lactating women had a mean age of 27 years, and most of these women delivered first time (72·73 %).
Characteristics of socio-demographics by participants’ daily dietary selenium intake
(Mean values and standard deviations; numbers and percentages)
a,b Mean values within a row with unlike superscript letters were significantly different (P < 0·05).
* Subjects in this group have insufficient Se daily intake (<78 μg Se/d).
† Subjects in this group have optimal Se daily intake (78–400 μg Se/d).
‡ Subjects in this group have excessive Se daily intake (>400 μg Se/d).
§ Values expressed as means and standard deviations for continuous variables (compared by ANOVA and least significant difference).
|| Values expressed as numbers and percentages for categorical data (compared by χ 2 test).
The weight characteristics of mothers are presented in Table 2. No significant difference of prepregnancy BMI was observed (P = 0·706). GWG in the insufficient Se group and the optimal Se group (15·59 (sd 5·43) and 17·08 (sd 6·72) kg, respectively) were significantly higher than those in the excessive Se group (12·87 (sd 6·63) kg; P < 0·01), and the same in PPWR (8·17 (sd 6·17) and 7·86 (sd 6·25) kg v. 4·49 (sd 6·02) kg; P < 0·01).
Weight and BMI of participants by daily selenium intake
(Mean values and standard deviations; numbers and percentages)
GWG, gestational weight gain; PPWR, postpartum weight retention.
a,b Mean values within a row with unlike superscript letters were significantly different (P < 0·05).
* Subjects in this group have insufficient Se daily intake (<78 μg Se/d).
† Subjects in this group have optimal Se daily intake (78–400 μg Se/d).
‡ Subjects in this group have excessive Se daily intake (> 400 μg Se/d).
§ Values expressed as means and standard deviations for continuous variables (compared by ANOVA and least significant difference).
|| Values expressed as numbers and percentages for categorical data (compared by χ 2 test).
¶ At 42nd day postpartum.
Dividing all subjects into four subgroups by the quartiles of the PPWR, significant differences for GWG, prepregnancy BMI and dietary intakes of energy, protein and Se were investigated among women with different quartiles of PPWR (P = 0·000, 0·000, 0·004, 0·005 and 0·001, respectively, shown in Table 3). On the basis of the directed acyclic graph (Fig. 2(a)) and our univariate analysis (Table 3), multivariable linear regression analyses were conducted to examine the relationships between dietary daily Se intake and PPWR, adjusting for other potential confounders, including GWG, prepregnancy BMI, dietary intakes of protein, age, parity, dietary intakes of fat, carbohydrate and total energy and time of physical activity and sleep. The association of dietary daily intake Se with PPWR are listed in Table 4. The linear regression model after adjusting indicated that PPWR significantly decreased with dietary intake of Se (P < 0·05). The distribution curves of Z-scores at two different time points (at birth and 42nd day) for infants of all participants are shown in Appendix 2. The infants were also divided into the same three groups, followed by their mothers. Table 5 presents the WAZ, LAZ and WLZ of infants at different time points (at birth and 42nd day). At birth, most of all have optimal WLZ (89·39 %), LAZ (98·10 %) and WAZ (96·59 %). Compared with infants of lactating women in excessive Se group, infants in other two groups had higher LAZ and WAZ (P < 0·05). No significant difference of WLZ was observed in infants among three groups. At 42nd day, infants of mothers in the optimal Se group have the lowest WLZ and highest LAZ (P < 0·05). For WAZ, there was no significant difference observed in infants among three groups. Overall, the WAZ of infants increased significantly (P = 0·002) during the first 42-d period, especially in excessive Se group (P < 0·01) which had a significantly increased WLZ and WAZ (P = 0·000) as well as decreased LAZ (P = 0·000). In optimal Se intake group, WLZ of infants markedly decreased (P = 0·002), while LAZ significantly increased (P = 0·000). No significant changes of Z-scores were found in group of insufficient Se intake (P > 0·05). Consistent with infants at birth, most of all infant at 42nd day have adequate WLZ, LAZ and WAZ (77·27 %, 89·02 % and 96·97 %, respectively).
Distribution of age, gestational weight gain (GWG), prepregnancy BMI, time of physical activity and sleep, dietary intakes of total energy, three macronutrients and selenium among early lactation women with postpartum weight retention (PPWR) quartiles
(Mean values and standard deviations; median values and percentiles)
a,b,c,d Values within a row with unlike superscript letters were significantly different (P < 0·05).
* Values expressed as means and standard deviations for continuous variables (compared by ANOVA and least significant difference).
† Values expressed as medians and 25th and 75th percentiles for skewed data (compared by Kruskal–Wallis and Mann–Whitney U test).
‡ Physical activity is any physical activity.
Association of postpartum weight retention with dietary selenium intake*†
(Coefficient values and 95 % confidence intervals)
* Covariates are following: gestational weight gain (kg), prepregnancy BMI, dietary intakes of protein fat (g/d), carbohydrate (g/d) and total energy (kJ/d), age (years), parity, and time of physical activity (h/d) and sleep (h/d).
† Results were observed by multivariable linear regression analyses.
Z-scores of weight-for-length (WLZ), length-for-age (LAZ) and weight-for-age (WAZ) of infants at birth and 42nd day by daily selenium intake of their mothers
(Mean values and standard deviations; numbers and percentages)
a,b,c Mean values within a row with unlike superscript letters were significantly different (P < 0·05).
*P < 0.05, **P < 0.01 (differences between Z-scores of infants at birth and at 42nd day).
† Values expressed as means and standard deviations for continuous variables (compared by ANOVA and least significant difference).
‡ Values expressed as numbers and percentages for categorical data.
Our results indicated that each 1 kg of increase in GWG of mother was related to OR of 0·902 (95 % CI 0·878, 0·925) in underweight and 0·915 (95 % CI 0·844, 0·990) in wasting of her child, while each unit of increase in prepregnancy BMI was associated with OR of 0·887 (95 % CI 0·803, 0·970) in underweight and 0·798 (95 % CI 0·666, 0·955) in wasting of her infant (P < 0·05, Table 6).
Associations of infants’ abnormal Z-scores categories at birth with maternal prepregnancy BMI and gestational weight gain (GWG)*
(Odds ratios and 95 % confidence intervals)
WLZ, Z-scores of weight-for-length; LAZ, Z-scores of length-for-age; WAZ, Z-scores of weight-for-age.
* Covariates are following: age (years), parity and household income (Yuan (CNY)/person per year). The effects of change of mothers’ weight on infant’s Z-scores were estimated by multinomial logistic regression.
Multinomial logistic regression was conducted to estimate the associations of maternal Se nutritional status with their infant’s Z-scores (normal level Z-scores as a reference group) at 42nd day. Table 7 lists the OR of growth abnormally associated with dietary intake of Se and Se content in milk of their mothers. Just only WLZ was associated with dietary Se intake and Se level in breast milk (P < 0·05).
Associations of infants’ abnormal Z-scores categories at 42nd day with maternal daily selenium intake (μg/d) and selenium content in breast milk (μg/l)*
(Odds ratios and 95 % confidence intervals)
WLZ, Z-scores of weight-for-length; LAZ, Z-scores of length-for-age; WAZ, Z-scores of weight-for-age.
* Covariates are following: age (years), parity, Z-scores at birth, daily dietary intake of total energy (kJ/d) and three macronutrients (g/d) and household income (Yuan (CNY)/person per year). The effects of maternal Se nutritional status on infant’s Z-scores were estimated by multinomial logistic regression.
Discussion
Nowadays, more and more evidence shows that PPWR can develop into overweight and obesity, which is harmful to women health for a long time, and PPWR is also associated with adverse neonatal outcomes(Reference Mannan, Doi and Mamun1,Reference Williams, Mackenzie and Gahagan9) . Unexpectedly, in this study, Se levels in lactating mothers were also found to be negatively related to PPWR (P < 0·05). Reasonable? Many studies reported the relationship between serum/plasma Se and BMI or the prevalence of obesity in many countries, the results of which were contradictory. For example, in a study involving 6440 men and 6849 women from the USA, serum Se levels were inversely associated with increased BMI with −4 (95 % CI −5·5, −1·6) ng/ml and difference between the highest and the lowest quartiles was statistically significant(Reference Zhong, Lin and Nong60). A detailed examination of 573 school-age children from Madrid also demonstrated that both serum Se and Se intake of overweight children (BMI > P85) were 14 % and 27 %, respectively, lower than normal-weight children(Reference Ortega, Rodríguez-Rodríguez and Aparicio61). Conversely, a cross-sectional study of 245 adolescent girls from rural Vietnam suggested that BMI < 17 kg/m2 (OR 2·65, 95 % CI 1·25, 5·61) was found to be a risk factor for low serum Se levels (<70 ng/ml), but this was likely a reflection of poor nutritional intake(Reference Van Nhien, Yabutani and Khan27). A case–control study on Chinese adults suggested that, obesity, rather than insulin resistance, is central to the increase in selenoprotein P (SELENOP) level in serum(Reference Chen, Liu and Wilkinson28). However, several previous studies indicated that BMI was not associated with circulating Se concentrations(Reference Ghayour-Mobarhan, Taylor and New62,Reference Suadicani, Hein and Gyntelberg64) . The underlying mechanism remains to be clarified.
In vitro, several studies have demonstrated that Se had an essential role in adipogenesis(Reference Park, Kim and Nam65,Reference Suh and Lee73) . In vivo, both low- and over-expression of adipocyte selenoproteins may result in adipose tissue dysfunction contributing to adipocyte hypertrophy or dystrophy, IR and adipose tissue inflammation(Reference Tinkov, Ajsuvakova and Filippini74). The proposed role of Se in adipose tissue physiology and obesity pathogenesis is shown in Appendix 3 (Reference Donovan and Copeland75,Reference Pitts and Hoffmann80) . Adequate Se supply (including its transport with SELENOP), as well as normal selenoprotein expression, is essential for regulation of adipogenesis and physiological development of adipose tissue, further influencing its physiological functioning, including energy storage, endocrine and immune functions(Reference Tang, Li and Zhao81,Reference Zhao, Li and Tang82) .
Adenosine monophosphate-activated protein kinase is an energy status sensor that controls cellular energy homoeostasis and activates energy production processes by stimulating catabolic pathways and inactivating processes involved in ATP consumption(Reference Hardie83). GPX1 and/or SELENOP inhibited phosphorylation (activation) of key mediators, such as Akt and adenosine monophosphate-activated protein kinase, in energy metabolism in liver and/or skeletal muscle(Reference Steinbrenner and Sies37).
Moreover, insulin has a key role in the control of carbohydrate and lipid homoeostasis, inducing storage of metabolic fuels after food intake, and some selenoproteins (GPX and SELENOP) have recently found to take a part in IR, by reducing the reactive oxygen species needed in the insulin signalling process, and deactivating the energy status sensor adenosine monophosphate-activated protein kinase in liver(Reference Steinbrenner and Sies37,Reference Carreras, Ojeda, Nogales and Vinood84) . The current concept on potential mechanisms underlying insulin-antagonistic actions of Se, GPX1 and SELENOP is shown in Appendix 4 (Reference Steinbrenner, Speckmann and Pinto85).
That is to say, the Se nutritional status of lactating Chinese women was found as one possible cofactor of their PPWR. However, whether this impact of Se is short-term adaptation or long-term impact and its potential mechanism needs further studies.
In this study, GWG and prepregnancy BMI of mother were the indicators of their infants’ physical Z-scores at birth (Table 6). As far as we know, PPWR of these mothers came often from weight gain during pregnancy and weight gain during lactation. GWG of mother includes fetal tissue (related to the birth weight of her newborn) and placental tissues as well as weight gain. Similarly, weight gain during lactation includes weight of breast milk and weight gain during this period. PPWR is mainly from excessive weight gains during pregnancy and lactation, and it might have effects on the birth weight and weight gain during development of her offspring. Some previous studies also indicated that maternal blood Se concentration is regarded as a factor of small-for-gestational age newborns(Reference Guo, Zhou and Xu50,Reference Lewandowska, Sajdak and Lubiński86,Reference Tindell and Tipple87) .
Also in our study, in addition to GWG and prepregnancy BMI, maternal Se intake and Se content in milk of mother also were the indicator of her infant’s physical Z-scores at 42nd day postpartum (shown in Table 7): the optimal maternal Se intake and the rational Se content in milk have positive associations with growth and development of offspring (P < 0·05).
Se is related to the normal development of the body. First of all, it may directly participate in the normal formation and development of bones. Se deficiency can cause the prevalence of Kashin-Beck disease, but the underlying mechanism is unknown(Reference Wang, Yu and Liu88,Reference Xie, Liao and Yue89) . Of course, the interaction between Se and iodine promotes the development of the body, and its indirect mechanism is much clear: iodothyronine deiodinase, three selenoenzymes, can catalyse the conversion of thyroxine (T4) to active triiodothyronine (T3) as one of the critical hormones in normal growth and development of infant(Reference St Germain and Galton90). Thus, Se nutrition status could conceivably affect thyroid function in infants and further affect the growth of these infants, which was supported by some previous animal experiments(Reference Meinhold, Campos-Barros and Walzog91,Reference Hefnawy, Youssef and Aguilera92) .
Above all, some health policy should be formulated to prevent excessive maternal PPWR and abnormal Z-scores of infants. Initially, it is important to improve the quality of currently implemented general prenatal and postpartum care programmes, at the public health centres, especially those concerning nutritional guidelines. Care policy should include the following: (1) systematic weight surveillance throughout the entire gestational period; (2) specific nutritional counselling, such as healthy pregnancy diets appropriate for controlling GWG and postpartum diets considering Se dietary intake especially.
Several strengths of the current study should be noted; to the best of our knowledge, this is one of the first surveys to observe the relationship between dietary intakes of Se and PPWR in early lactating Chinese women as well as their offspring’s physiological development. Several potential confounding variables were controlled in our analysis. Our research has several limitations. Causal inference cannot be drawn from this study because it is a cross-sectional study during the early short-term postpartum. Thus, whether the impact of Se obtained from our results is short-term adaptation or long-term impact and its potential mechanism needs further studies. Besides, we used WHO standards to evaluate the nutritional status of infants, which may be not appropriate for Chinese children, and some studies found the WHO growth standard was not suitable for any region and age(Reference Rossiter, Colapinto and Khan93,Reference Roelants, Hauspie and Hoppenbrouwers95) . Additionally, there is no specific formula for lactating Chinese women to calculate the dietary Se intake daily from the concentration of plasma Se and the little difference in bioavailablity of Se between non-lactating adult women and lactating adult women may lead to overestimate the daily dietary Se intake in these mothers.
Conclusions
In conclusion, daily Se intake of lactating women was related negatively to PPWR. If maternal pregnancy BMI or GWG was too low, the risk of offspring’s developmental delay would be increased. In addition, the physical Z-scores at 42nd day of infants were associated with maternal dietary Se intakes and Se content in milk. Higher levels of daily Se intake and Se in milk of mothers could reduce the incidence of infant malnutrition. These results revealed that lactating women with a sufficient intake of proteins and total energy but lack of Se still had potential associations with excessive maternal PPWR and abnormal physical development of their offspring, and the causal relations should be confirmed by further longitudinal study.
Acknowledgements
The authors are grateful to the subjects participating in this study and to the doctors and nurses facilitating both the recruitment of participants and the interviews.
The authors’ contributions were as follows: F. H., Q. W. and Z. W. H. designed research; F. H., L. P. L, Y. J. C., J. Z. and S. J. W. conducted research; F. H., X. H. P., L. C. S., Q. W., Y. Q. L. and S. Z. analysed data; F. H. and Z. W. H. wrote the paper; Q. W. and Z. W. H. had primary responsibility for final content. All authors read and approved the final manuscript.
This work was supported by the Chinese Nutrition Society Nutrition Scientific Research Funds – Yili Nutrition and Health Research Fund (grant number 2013-013), the National Natural Science Foundation of China under the grant numbers 81973048, 81741032 and 81372989, Chinese Center for Disease Control and Prevention National Institute for Nutrition and Health, Nutrition Standards System Construction Project – Reference Selenium Intake for Infants, and the National Key Research and Development Program of China (grant number 2020YFC2006302).
The authors have no financial or personal conflicts of interest to declare.
Supplementary material
For supplementary material referred to in this article, please visit https://doi.org/10.1017/S0007114521000015
