Study design and study population
This cross-sectional study was conducted in Xiangtan, Hunan Province, located in South Central China. All children aged 4–7 years in a single primary school were invited to participate. In total, 435 children were recruited; 10 were excluded due to incomplete dietary assessment. Experimental strategy for linking dietary patterns and lead exposure in children was shown in Fig. 1.
Dietary assessment and food grouping
Dietary intake information was collected using a validated food frequency questionnaire (FFQ) to evaluate children’s dietary intake(Chinese Center for Disease Control and Prevention 2015). Caregivers were asked to recall each child’s intake (including estimated portion size and frequency of each food item) over the previous 12 months. The frequency was recorded in terms of times per day, week, month, or year; portion sizes were expressed in grams or milliliters. The mean daily intake of each food item was calculated using the estimated portion size and frequency. Total of 55 food items were further categorized into 19 food groups based on similarities in nutrient profiles or processing methods. These foods included rice, wheat flour, coarse cereals, soybeans and their products, meat, poultry, eggs, fish, shrimp, crab and shellfish, leafy vegetables, leafless vegetables, tubers, fresh beans, fungi and algae, fruits, milk and its products, nuts, beverages, and snacks. The FFQ was administered by a well-trained dietitian.
Principal component analysis was used to identify dietary patterns, as in our previous study (Huang, et al. 2019). The main dietary patterns were identified based on eigenvalue scree plots, factor interpretability, and variance explained. Factor loadings represent correlation coefficients between food items and dietary patterns; factor loadings > |0.3| were regarded as primary contributors. Factor scores (i.e., summed intakes of each food group weighted according to factor loading) were calculated for each pattern and each individual. Factor scores were categorized into four quartiles, such that Q1 was weakly related to the dietary pattern and Q4 was strongly related to the dietary pattern.
Lead intake
Lead intake of each child was estimated based on the food intake in this study and food lead concentration data from previous study (Jin, et al. 2014). The lead concentration (mean lead in each food category) was based on lead measurements in 2077 food samples from 23 food categories during 2007–2010(Supplement 1).
Blood sampling and analysis
Peripheral venous blood samples (3mL/child) were drawn from the cubital vein using a vacuum tube with an anticoagulant. BLLs were measured using graphite furnace atomic absorption spectrometry (ContrAA 700; Analytik Jena GmbH, Jena, Germany). Recorded values were the means of triplicate sample analyses. The limit of detection for Pb was 4.0µg/L. The value below the limit of detection was calculated as half of the detection limit.
Assessment of growth and development
Weight (kg) and height (cm) were measured with an electronic height and weight measurement instrument. Body Mass Index (BMI) was calculated by weight (kg) dividing square of height (m).An appropriate cuff size and a mercury sphygmomanometer were used to measure BP. Systolic blood pressure (SBP) and diastolic blood pressure (DBP), expressed in mmHg, were recorded from two consecutive BP measurements. Cognitive performance was assessed using the Clinical Memory Scale Test, developed by the Institute of Psychology of the Chinese Academy of Science(Xu,SL., et al. 1996). This test was mainly used to evaluate short-term memory performance, including directed memory, associative memory, free image recall, meaningless image recognition, and associative image memory. First, each section was scored according to the results of a test for each student. Then, the score was changed to a scale and the total value was determined. Last, the total scale score was converted into a memory quotient (MQ) according to age group, which reflected comprehensive memory performance.
Other related variables
Information regarding demographic characteristics was collected during the FFQ interview. These variables included age, sex, caregiver group (parents or grandparents/others), caregiver occupation (public institution, non-public institution, or unemployed), caregiver education (college and above, senior, or junior and below), and annual family economic level (≥ 50,000 yuan, 20,000–50,000 yuan, or < 20,000 yuan). Age was divided into two groups (4–5 years or 6–7 years).
Ethics approval and consent to participate
All participants provided written informed consent. The study protocol was approved by the Ethics Committee of the Hunan Provincial Center for Disease Control and Prevention.
Statistical analysis
Continuous variables were expressed as mean and standard deviation (SD) for normal distribution or medians and inter-quartile ranges (IQR) for Skewed distribution. Categorical variables were expressed as numbers and percentages. BLLs were categorized into two groups according 50th percentile (P50). The demographic characteristics were compared according to BLLs, using analysis of Variance (ANOVA) or chi-square test. Correlation of factor scores with lead intake was assessed using Pearson correlation analysis. Logistic regression analysis was used to investigate the correlations of BLLs with dietary patterns. Next, log-Pb was calculated in children in group of > P50. The associations between log-Pb and dietary pattern were tested in linear regression analysis. P < 0.05 was considered to indicate statistical significance. Statistical analyses were performed using SPSS software (version 13.0; SPSS, Inc., Chicago, IL, USA).