2.2. Study design
Dietary assessment. The semi-quantitative Food Frequency Questionnaire (FFQ) report asks all mothers to record their usual dietary intake in the past month. The three components of a food list, how often you eat a certain food, and the amount of food you eat at a time make up the FFQ. A total of 156 food items are on the list, mainly divided into 13 categories: staple foods, vegetables, fruits, aquatic products, livestock and poultry, poultry, eggs, milk and its products, soybeans and their products, nuts, edible oils, processed foods, seasonings, and beverages. This FFQ was validated with 3-day dietary records from lactating mothers. This information is collected by trained graduate students themselves. The original data of food amount in FFQ were input into EpiData software for verification. Then the daily intake of food group, energy and nutrition was calculated. The database was based on the Chinese Food Composition Table (6th edition) (Yang, 2018) and combined with the values obtained from the USDA National Nutrient Database for Standard Reference (Burke et al., 2016).
Sample collection. Breast milk (10-15 mL) was collected from each mother between 07:30 a.m. and 09:00 a.m. Breast milk samples are collected in a polypropylene tube and no preference for one breast. All sampling procedures are carried out to ensure that the light is avoided. The samples were quickly stored in a hospital refrigerator at 4 ℃, and then transported to a laboratory refrigerator in a lightproof container with ice, stored at −20 ℃ for one week, and then stored at −70 ℃ until analysis.
Quality control. Experts were invited to review and revise the questionnaire on baseline characteristics to ensure its scientificity. All researchers received rigorous training and evaluation before conducting the study. After the experiment was completed, baseline information and dietary data were examined with the participants. In order to improve the accuracy of dietary data in FFQ, participants estimated the size of the food component with the help of a food map with sufficient visual reference. Breast milk samples were collected in strict accordance with standard sampling techniques. The instruments and containers used for sampling were disinfected and used once. TAG concentration in breast milk was detected by the same instrument and the same batch of reagents. The main data were input into EpiData software twice by two trained researchers to verify its accuracy. In statistical analysis, the appropriate statistical methods are selected strictly according to the statistical requirements and data types.