Adjusting DBI-2016 to dietary balance index for Chinese maternal women and assessing the association between maternal dietary quality and postpartum weight retention: A longitudinal study

Diet is believed to play a major role in maternal recovery, postpartum weight retention (PPWR) is one of the challenges for Chinese women. However, the association between puerperal women’s diet and PPWR remained unclear and complicated in China. The study assessed the dietary quality of puerperal women using adjusted Chinese Dietary Balance Index-16 (DBI-16) and explored its associations with PPWR. Participants were enrolled in the Mother-Infant Cohort Study of China. Dietary intake and demographic characteristics were obtained by a semi-quantitative food frequency questionnaire and a self-designed questionnaire at 0–3 months postpartum. PPWR was calculated by the weight difference at 0-3months and 6-8months postpartum minus pre-pregnancy weight. Dietary quality was assessed using adjusted DBI-16. 316 puerperal women were enrolled. According to adjusted DBI-16, 84.8% of participants had an insufficient dietary intake (vegetables 84.8%, fruits 91.8%, dairy 87.3%, soybean 61.4% and aquatic foods 79.4%, respectively), 67.1% had an excessive intake (cereals 60%, meat 57.3% and eggs 64.9%, respectively), 98.4% had an imbalanced diet consumption. PPWR at 0–3, and 6–8 months were 6.0 (±5.1) kg and 5.2 (±7.7) kg, and the percentage of PPWR (≥5kg) were 63.0% and 52.8% respectively. Multivariable linear regression showed the intake of fish and shrimp at 0–3 months postpartum was negatively associated with PPWR at 6–8 months (β = -0.114, SE = 0.279, p < 0.05). The diet quality of Chinese puerperal women was unreasonable and imbalanced. Fish intake tended to be a favorable factor for postpartum weight loss.

Introduction to December 2015, in seven cities (Beijing, Taiyuan of Shanxi, Jinan of Shandong, Changsha of Hunan, Shiyan of Hubei, Chongqing, and Chengdu of Sichuan) from northern, eastern, central and southwestern districts around China. The inclusion criteria of participant women in this study were as follows: being at 0-3 months postpartum, being apparently healthy without severe acute or chronic diseases, adopting breastfeeding. And the non-lactating women, whose infant did not receive breast milk during the previous day of investigation, were excluded. Based on this, the clinicians explained the study protocol to 540 mothers while their first-time postpartum visiting at local maternal and child care clinics. Finally, informed written consent was obtained from 316 participants voluntarily (the response rate was 58.5%). Even though there was not formula calculating, the sample size was sufficient to draw conclusion referring to literatures [19,[23][24][25][26][27].
The mother-infant cohort was followed up to 2 years postpartum until December 2018. During follow-up period, the dietary and feeding practices, physical and mental development of child, anthropometric indices of postpartum women and related factors were investigated. The women's data at 0-3, 6-8 months postpartum were analyzed, including diet and weight data. The study was approved by the Peking University Institutional Review Board (Beijing, China, Approval Number IRB00001052-15020), and conducted according to the Declaration of Helsinki. The privacy of participant mothers and the confidentiality of their personal information would be protected.

Measurements
Dietary intake. Maternal dietary assessment was carried out at 0-3 months by face-toface interview with a 21-item food semi-quantitative Food Frequency Questionnaire (FFQ) adapted from the Chinese National Nutrition and Health Surveillance (2010-2013) and previously validated FFQ [28]. The original FFQ had 17 food groups, including 9 groups of fluids intake (different soups and drinks), that were thought important for maintaining lactation in Chinese traditional dietary culture. This study focused on the maternal health outcome (PPWR), so the original food groups were not fully suitable. On the other hand, the "Chinese Dietary Guideline" was revised in 2016, in which some foods were recommended specially for lactating women. Based on the above consideration, the original 17 food groups were modified to 21 groups in this study, namely rice, flour, coarse food grains, starchy roots, dark vegetables, light vegetables, fruit, livestock meat, poultry meat, eggs, aquatic products, marine fish, milk, legumes products, and soups, according to their nutritional significance, revised dietary guideline and puerperal women's eating habits.
Women were asked to recall the quantity and frequency of food they consumed within one month postpartum (not the past one month) at the baseline (0-3m postpartum) or the 2nd (6-8m) follow-up survey. Atlas of the Illustrative Food Pictures for Use in Dietary Intake Recall were referred to help the participants to estimate the food portion size. The quantity (g) was reported according to frequencies of consumption (never, day, week, and month), and daily food intakes were calculated by multiplying the frequency of consumption for portion size (g).
Adjusting DBI-2016 to dietary balance index for puerperal women. General DBI-2016 was developed for normal adults according to the Chinese dietary guidelines and Food Guide Pagoda (2016), which should be adjusted to meet the needs of puerperal women in the study. Revised DBI comprises five components, including (1) cereals; (2) vegetable and fruit; (3) bean and dairy; (4) animal food; (5) food variety. For each component and indicator, a score of '0' indicates an excellent dietary intake, the negative score demonstrates that insufficient food intake and the positive score demonstrates that excessive food intake.
Negative score (range -12~0) is used to assess inadequate intake levels of vegetables and fruits, bean and dairy, which the guidelines recommend that puerperal women should consume in a "sufficient" or "plenty" amount. The score of the food variety component also ranges from -12 to 0. The "food variety" is assessed by intake of 12 identified food subgroups including: 1) rice and products; 2) wheat and products; 3) corn, coarse grains and products, Starchy roots and products; 4) dark-colored vegetables; 5) light-colored vegetables; 6) fruits; 7) soybean and products; 8) milk and dairy products; 9) livestock meat and products; 10) poultry and game; 11) eggs; 12) fish and shellfish. If the intake of the food subgroup reached or exceeded the lowest recommended intake scored 0, else, a score of -1 is assigned. The lowest recommended intake is 5g for soybean and products, and the other 11 food subgroup is 25g.
Moreover, both positive and negative scores were used to evaluate the intake levels of cereals ((range -12~+12) and animal food (range -12~-8), which the guidelines recommend consuming in an "appropriate" amount.
In considering the difference in nutrient requirements in energy consumption, the scoring of food components is based on seven energy intake levels from 1800kcal/d to 3000kcal/d. Scoring details of the revised DBI can be found in Table 1. Based on the scores for each DBI component, three indicators of diet quality are calculated. Higher Bound Score (HBS) is the sum of all the positive scores, which indicates excessive food intake. Lower Bound Score (LBS) is the sum of all negative scores, which indicates inadequate food intake; Diet Quality Distance (DQD) is computed by the absolute values of both positive and negative scores, which indicates imbalanced food intake. The lower and upper bound of HBS, LBS as well as DQD were: 0 to 20, 0 to 60, 0 to 60, respectively.
Each indicator is further divided into five levels according to scores: 1) Score "0" represents "excellent" dietary intake; 2) Less than 20% of total score represents "good" dietary intake; 3) 20%~40% of total score represents "mild poor" dietary intake; 4) 40%~60% of total score represents "moderate poor" dietary intake; 5) Higher than 60% of total score represents "severe poor" dietary intake. Postpartum weight retention. Maternal self-reported and hospital-documented gestational weight gain (GWG) at 0.5~2.5 years after childbearing had been observed to be highly correlated [29]. Thus, in our study, body weight was reported by puerperal women themselves. The postpartum weight retention (PPWR) was calculated by subtracting the pre-pregnancy body weight from postpartum body weight at 0-3, 6-8 months postpartum. And gestational weight gain equaled to the weight before delivery minus the pre-pregnancy weight.
Related factors. We used a self-designed questionnaire to collect the sociodemographic characteristics (age, family affluent level, education level), gestational weight gain, delivery mode, feeding patterns. Education level included two levels: primary/secondary (high school or below), tertiary (bachelor degree or above). Feeding patterns were classified as follow: 1) Exclusive breastfeeding (breast milk was exclusively given without a very small amount of water or/and juice); 2) Predominant breastfeeding (breast milk was predominantly given with a very small amount of water or/and juice); 3) Mixed feeding (a mixture of breast milk and formula milk was given with or without some water or/an juice) [30]. Body mass index (BMI) was computed as weight in kilograms divided by the square of the height in meters (kg/m 2 ). The BMI cut-off points for Chinese adults were used to determine underweight (BMI < 18.5 kg/m 2 ), normal weight (BMI 18.5-23.9 kg/m 2 ) and overweight (BMI 24.0-27.9 kg/m 2 )and obesity(BMI � 28.0 kg/m 2 ) [31]. Household monthly income per capita reported in CHY (China Yuan, 1CNY = 0.14 USD) and classified into 3 categories: low (less than 3000 CNY), middle (3001-6000 CNY) and high (above 6000 CNY). C2-Vegetables and fruits -12-0 Vegetables 320-399g = -1; 360-449g = -1; 400-499 g = -1; 480-599g = -1; Score decreased by 1 with intake amount decreased by 80g; 0g = -6.

Statistical analysis
Data were input to EpiData3.1 software. All analyses were performed with SPSS software version 22.0 (SPSS Inc., Chicago, IL, USA). Numeric variables that normally distributed were given as mean ± standard deviation (SD), Numeric variables that non-normally distributed were given as median. Categorical variables were given as a percentage (%). Kruskal-Wallis test and Mann-Whitney U test were used to compare the distribution of HBS, LBS, DQD among socio-demographics and related factors. Multiple linear regression analysis was utilized to evaluate the association between DBI scores for each component (cereals, vegetables, fruits, dairy, soybean, meat and poultry products, fish and shrimp, eggs, as well as food variety) and postpartum weight retention at 6-8 months postpartum. And the covariates included demographic characteristics (age, education level, and family income), delivery mode, infant feeding pattern, pre-pregnancy BMI, and gestational weight gain, some of which were valued as below: 1)Education level: high school or below = 0, bachelor degree or above = 1; 2)Family income (CNY): less than 3000 = 0, 3001~6000 = 1, more than 6000 = 2; 3)Delivery mode: vaginal = 0, cesarean = 1; 4)Infant feeding pattern: exclusive breastfeeding = 0, predominant breastfeeding = 1, mixed feeding = 2; 5) Number of childbirth: more than one child = 0, one child = 1. Multi-collinearity among covariates was estimated though tolerance, which being greater than 0.1 indicates that there is no multicollinearity among covariates.
The level of significance was p values<0.05 (two sides significance level).

Characteristics of the study puerperal women
Characteristics of 316 puerperal women are presented in Table 2. The age range was 20-45 years and 83.5% of the participants were less than 35 years old. Most of the puerperal women (60.9%) had a bachelor degree or above as well as 71.4% had a middle or above household income (>3000 CNY). The majority (65.8%) was the first-time mother and 57.3% of the puerperal women adopted exclusive or predominant breastfeeding. The sample covered a wide range of pre-pregnancy BMI, but most of the participants (68.4%) had a normal weight, as well as 14.6% were overweight or obesity. Average gestational weight gain was 14.6 ±5.1kg. Postpartum weight retention at 0-3, and 6-8 months were 6.0 (±5.1) kg and 5.2 (±7.7) kg respectively, and percentage of PPWR (�5kg) was 63.0% and 52.8%.

Distribution of overall diet quality of puerperal women by adjusted DBI
The respectively. The results showed that more than two-thirds of puerperal women had an excessive dietary intake, more than four-fifths of them have an inadequate dietary intake, and most of them have an imbalanced diet. According to sociodemographic characteristics, there were no statistically differences for HBS, LBS and DQD (p>0.05), as shown in Table 4

PLOS ONE
Dietary quality of Chinese maternal women and weight retention

Dietary index for assessment of puerperal women
To our knowledge, this study was the first assessment of overall dietary quality using diet index method for Chinese puerperal women, and other investigations usually analyzed individual nutrients and food groups separately. The dietary quality depends on the existing dietary patterns, which were assessed generally by two approaches: a priori method based on prior nutrition knowledge translated into dietary guidelines, and posteriori method, in which dietary patterns are extracted once the dietary intake are investigated [9]. In the a priori method, the diet quality indices are constructed to quantify the healthiness of the dietary pattern based on existing scientific guideline [9]. We have finished assessing dietary patterns of Chinese puerperal women in the same cohort (manuscript under submission), and identified four dietary patterns through a posterior principal component analysis, including "plant food" pattern (characterized by high intake of rice and vegetables), "diverse" pattern (characterized by starchy roots, fruit, livestock meat and aquatic products, relatively healthy), "traditional northern" pattern (poultry meat, eggs and soup), and "marine-flour" pattern (flour, coarse food grains and marine foods). In this study the dietary quality of Chinese puerperal women was assessed using a priori approach, i.e. Chinese diet balance index, which has been proved to be an effective index to evaluate overall dietary quality by food adequacy and moderation, developed based on Chinese Dietary Guideline and the Chinese Food Pagoda 2016.
The general DBI-2016 was developed for normal adults and in the study which must be revised to meet the need of puerperal women. One important revision is to delete some DBI components like alcoholic beverages, which consumption was seldom in Chinese puerperal women. In our whole cohort population, the drinking percentage during puerperium was only 1.3% (n = 7), which has been deleted from the study subjects. Similarly, a study conducted by Fernandez-Barres et al. created a new tool, "the relative MD score", excluding alcohol from the scoring system [32]. We also adjusted other items, including animal foods and dairy products, which are rich in good protein and vitamin A, recommended a higher intake during the lactating period. Population with different physical activity levels require different energy, our study focused on puerperal women whose physical activity is usually defined as low level, and energy requirement is 2300kcal/d [33]. But in the general DBI, the energy levels include 2200Kcal and 2400Kcal without 2300Kcal. Considering the Chinese puerperal women are prescribed animal foods during "doing the month" period, which may exceed the energy recommendation [34], our study estimates 2400Kcal/d as energy level.
Considering the availability of foods and cultural dietary preferences, it is necessary to develop recognized diet indices based on national dietary recommendations, to evaluate the dietary quality of puerperal women for in-country comparison and trend analysis. However, such indices should be used carefully for cross-country comparisons because their generalizability might be limited, especially regarding the foods or food groups included in the index.

Dietary quality of puerperal women
The results showed the postpartum diet quality was unreasonable and imbalanced, 84.8% of puerperal women had an insufficient dietary intake (especially fruits, vegetables, dairy, beans and aquatic foods), 67.1% had an excessive intake (especially meat and eggs), and 98.4% had an imbalanced diet consumption, that was similar to a posterior analysis. The overall dietary pattern in a prior approach was like "traditional northern" pattern in a posterior approach. These findings are also consistent with other studies in Chinese population. Li et al reported that the puerperal lactating women in Fuzhou had an excessive intake of poultry and eggs, and inadequate intake of vegetables, fruits, milk, and bean products generally [35]. Mao et al found that postpartum women in the central district of China had abundant consumption of cereals and animal foods; however, their intake of dairy products, fruits, and vegetables failed to meet established recommendations [36]. The Chinese National Nutrition and Health Surveillance (2010)(2011)(2012)(2013) showed that during the first month postpartum the percentage of fruits and dairy consumption at least once a day were 67.4% and 38.3%, the median intake of vegetables, fruits and dairy products were 150.0, 40.0 and 0.0 g/d, respectively, much less than recommendation of Chinese Dietary Guideline [37]. Similar results were shown in other countries. A study reported that Spanish lactating women did not meet the recommendations for cereals, salad vegetables, fresh fruits and dairy [38]. Another showed that in North America only 13% and 21% of lactating women had adequate fruit and vegetables respectively, and 9% had a good quality diet [23]. Yu et al. applied dietary modelling method to develop individualized optimal diets, which meet the nutrient requirements for lactating women in urban China, they found that the large difference between observed and optimized diets suggests that the nutrient needs of lactating mothers in China may only be met after substantial dietary changes [39]. Vega-López S et al. using Health Eating Index 2010 assessed dietary quality of Mexican-American postpartum women, all participants in the study had HEI-2010 scores suggesting intake of diets with low quality or in need for improvement (55 ± 14 relative to a maximum score of 100) [40]. Above results indicated the dietary quality of puerperal women was imbalanced in different countries, despite the targeted foods were a little different.

Dietary quality and PPWR
Gestational weight gain and postpartum weight retention could increase maternal noncommunicable diseases risk and subsequent pregnancy, and the dietary quality was the main influencing factors. Our results showed the percentage of PPWR (�5kg) were 63.0% at 0-3 months and 52.8% at 6-8 months respectively. Considering dietary patterns present a more holistic view of the overall diet quality, which are currently on the frontier of the association examination between diet and chronic diseases risk [24].
Many other studies demonstrated the relations of energy intake with weight [19,24], but few studies confirmed the association between postnatal diet and PPWR [25]. Our study did not find an association between overall HBS, LBS, or DQD and PPWR, but revealed a negative correlation between the fish intake score and PPWR at 6-8 months, which indicated that fish intake tended to be a favourable factor for postpartum weight loss. López-Olmedo et al also found that postpartum weight changes were not related to the postpartum diet [25]. The American Infant Feeding Practices Study II showed that despite the apparent difference in the trajectory of weight change by diet quality as assessed by aMED or AHEI-2010 in the early part of the postpartum period, neither total mean weight retention nor the probability of substantial weight retention differed through the end of follow-up, instead, the total energy intake was a significant contributor of substantial weight retention [19]. Donne et al. conducted a study found that small oily fish consumption played a protective role in preterm birth risk, neonatal weight, length and head circumference [26]. Our study found the more fish and shrimp consumption, the lower PPWR at 6-8 postpartum. Agree with above, adequate intake of fish and shrimp is beneficial for pregnant women, postpartum women as well as newborns. Different from our results, a study, included 305 lactating women in south-central China, showed that a diet characterized by high intake of fresh vegetables, soy milk, and bacteria and algae was negatively associated with postpartum weight retention [27]. To date, studies on fish intake and postpartum weight retention are rare and need to be further explored.
Other study found that the postpartum women in higher socioeconomic status (SES) retained less weight [41], the higher SES might mean the higher health literacy of puerperal women and family members, more availability of healthy food, and much more professional care. But our study found that family economic status was not associated with PPWR, the reason perhaps is that the respondents in this study have a higher level of education, with 60.9% having a bachelor's degree or above. Further studies are needed to provide more evidence for the conclusion.
Our study has certain limitations. We focused on food intake, but not nutrient intake. Despite these limitations, developing DBI-L can not only quantitatively evaluate the dietary quality of lactating mothers, but also have great significance for further exploring the diet of lactating mothers and postpartum weight retention.

Conclusion
Overall, the dietary quality of puerperal women is an important factor influencing short-term and long-term maternal health and offspring development, it is important to support and engage women throughout this life-stage. To our knowledge, it is the first study to assess the overall dietary quality of Chinese puerperal women using diet index, and the results showed the puerperal women had an imbalanced diet, and fish intake tended to be a favourable factor for postpartum weight loss. Also, many things remained regretted and unclear, the validity and reliability of adjusted DBI should be evaluated, the association of maternal dietary quality with breast milk secretion and infants' development need to be explored, nutritional intervention should be designed, and so on. Furthermore, special attention should be given to these gaps.