The differences of lipid profiles between only children and children with siblings: A national survey in China

With the increasing number of the one-child family, it is important to investigate whether the only-child status is associated with dyslipidemia. Among a national sample of 65,347 Chinese children aged 6–17 years, 16,100 lipid profiles were available. Children’s height, weight, total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C) were measured. In comparison to children with siblings, only children (OC) were more likely to be boys and live in urban areas. OC had less physical activity, less fried food intake, but more meat and dairy intakes. OC had significantly higher levels of TC (3.97 ± 0.78 vs. 3.89 ± 0.77) and LDL-C (2.12 ± 0.65 vs. 2.06 ± 0.64) in the overall group, and also in the subgroups of rural boys and girls. The prevalence of hyper-TC (5.48% vs. 4.43%) and hyper-LDL-C (3.97% vs. 2.96%) were significantly higher in OC than their counterparts. Furthermore, we found higher odds of hyper-LDL-C [1.43 (1.12, 1.83)] in OC after adjustments. In the subgroup analysis, only-child status was associated with increased risk of hyper-TC [1.86 (1.06, 3.26)] and hyper-LDL-C [2.65 (1.14, 6.16)] among rural boys, and hyper-LDL-C among rural girls [2.20 (1.14, 4.22)]. In conclusion, higher levels of TC and LDL-C were found in OC especially for rural children. Being an only-child was associated with increased risk of hyper-LDL-C.

Scientific RepoRts | (2019) 9:1441 | https://doi.org/10.1038/s41598-018-37695-0 by social and environmental factors. Competing with and sharing with siblings have considered to be account for personality and behavior differences by birth order 20 . The doting parents and grandparents tended to feed OC more food, which made them more likely to be over-nourished 21 . Meanwhile, the OC had less opportunities for playing games and certain physical activities with siblings 22 . In modern China, the one-child policy was recognized as a leading contributor to the rising childhood obesity epidemic 17,23,24 . The increasing prevalence of overweight/obesity could result in a much larger population of children with CVD risks, such as abnormal lipid levels 25 . As lipid levels continue to track from childhood into adulthood, improving lipid profiles in children is of great significance for the prevention of adult CVD 26,27 . However, no studies has been conducted to directly investigate the association between only-child status and lipid profiles. A better understanding of lipid profiles in OC, is conducive to an in-depth analysis of the relationship between only-child status and cardiovascular risk factors. Therefore, using data from a national survey on 16,100 Chinese children, this study aimed to evaluate the differences of lipid profiles between OC and CWS. We also tested whether this association varied by sex and living area.

Results
The demographic and anthropometric characteristics of the participants are shown in Table 1. A total of 16,100 children aged 6-17 years (mean age: 11.08 years; boys: 50.97%) were included in this study. Compared with CWS, OC were more likely to be boys (53.76% vs. 44.10%, P < 0.001) and live in urban areas (65.44% vs. 40.97%, P < 0.001). Parents of OC had higher educational levels than their counterparts (P < 0.001). OC also had higher height, weight and BMI z-score (all P < 0.001). The characteristics of participants by sex and living area are presented in Supplemental Table S1.

Discussion
Using national data of 16,100 children, we found that OC had higher levels of TC and LDL-C than CWS, especially in rural China. In addition, being an only child was associated with increased risk of hyper-LDL-C especially among rural children.  Table 2. Lipid profiles of the only children and children with siblings by sex and living area. TC, total cholesterol; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol. Continuous variables are displayed as mean ± standard deviation. The P value was compared with only children and children with siblings. In the subgroups, P* value was adjusted for age, parental educational levels, monthly family incomes, BMI z-score, MVPA time, screen time and food intakes and a random effect for provinces; in the total group, *P value was additionally adjusted for sex and living areas.
Scientific RepoRts | (2019) 9:1441 | https://doi.org/10.1038/s41598-018-37695-0 The impact of family characteristic (e.g. birth order) on lipid profiles has been examined in previous studies 5,8,12 . It was hypothesised that first-borns may be at greater risk of CVD than their later born siblings 28 . A cohort study found that the first-borns showed significantly higher levels of TC and LDL-C in early adulthood 8 . In our study, we found that OC showed less favourable lipid profiles than their counterparts. To some extent, there may be some similar underlying mechanisms of higher metabolic risks in both OC and first-borns. While two studies found that there were no impact of birth order on children's lipid profiles, and one of them found no differences in stature between first-borns with or without siblings 5,12 . These two studies, however, had a very small sample size (312 and 85, respectively), and they didn't take into account the influence of behavioral and environmental factors.
Two tentative explanations could be offered to explain that OC showed elevated lipid levels. First, OC were all first-borns. There was evidence that the in-utero growth of first-born babies might be restrained and they had lower birth weight and accelerated early catch-up growth 28,29 , both of which were independent risk factors for metabolic and cardiovascular diseases 19,[30][31][32][33] . Second, their higher metabolic and cardiovascular risk may be partly resulted from the behavioral or environmental changes. It was suggested that the focus of attention and resources from OC's families might lead to over-feeding 23 . In our study, OC tended to have higher intakes of meat and dairy products. On the other hand, we found that OC had less PA time than CWS, although no significant difference was found in screen time. Siblings played an important role in interactions, activities and cooperative play, leading to more PA time in CWS 21 . The combination of higher food intake and reduction in PA might result in energy imbalance in OC and therefore lead to higher risk of obesity and other metabolic diseases 23,25 .
We also found higher prevalence of hyper-TC in OC, however, the difference became insignificant after adjustment for parental educational levels and family incomes. This suggested that the association between the  only-child status and dyslipidemia may be partly dependent on these factors. It was suggested that triglycerides level was inversely associated with education and salary in some developed areas 34,35 . However, other studies found that individuals of lower socioeconomic status (SES) had lower levels of TC and LDL-C in developing areas 36,37 . In our study, only-child families were characterized as higher in SES and OC had higher levels of TC and LDL-C, which were consistent with findings in developing areas, indicating greater CVD risks among OC. This phenomenon may be partly explained by that higher income levels was associated with higher consumption of high-fat, high-energy food in modern China 24 .
The differences of lipid profiles between OC and CWS were less pronounced in urban China, which might be relate to the similar parental attention and family resources between OC and CWS in urban areas. China's economy has developed rapidly during the past decades, especially in urban areas. Urban children had better health and nutritional status than rural children in China 38 . Under the low fertility regime, fewer children in each household might be provided similar attention and resources from their parents.
Another important result of this study was that OC in rural China had a more than double risk of having hyper-LDL-C compared to rural CWS. In fact, those rural parents of OC showed higher socio-economic and educational levels than the rural parents of CWS (Supplemental Table S1). In terms of health-related behaviors, rural OC had less MVPA time but higher intakes of meat and dairy products compared to rural CWS. These phenomena may provide some ideas to explain our finding, however, the underlying mechanism remains to be determined. Considering the abovementioned urban-rural differences, it is important to pay additional attention to blood lipid profiles in OC who live in rural areas.
The fertility rates have been declining throughout the world during the recent decades, especially in Europe and many Asian countries 39,40 . Globally, the total fertility rates declined from 4.984 children per woman in 1960 to 2.451 in 2015 41 . As a result, there has been a great increase in the number of one-child families. The health of the only/first-born children will be a worldwide public health challenge. China's one-child policy has resulted in a large number of one-child family, which provides us a context to study the related characteristics and brings the world a warning about the higher risks of metabolic and cardiovascular diseases in OC.
This study had several strengths. First, to our knowledge, this was the first national study to explore the association between only-child status and dyslipidemia. Second, this study considered both sex and urban-rural disparities. Although the one-child rule was enforced for both urban and rural residents in China, this rule was virtually  Table 3. Odds ratios of only-child status for dyslipidemia according to generalized linear mixed models. OR, odds ratio; CI, confidence interval; TC, total cholesterol; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; OC, only children; CWS, children with siblings. Dyslipidemia refers to at least one of the following: hyper-TC, hyper-TG, hyper-LDL-C, and hypo-HDL-C. In the subgroups, model 1 was adjusted for age, parental educational levels, monthly family incomes and a random effect for provinces; model 2 was adjusted for variables in model 1, MVPA time, screen time, and food intakes; model 3 was adjusted for variables in model 2 and BMI z-score. In the total group, model 1 was additionally adjusted for sex and living areas.
Scientific RepoRts | (2019) 9:1441 | https://doi.org/10.1038/s41598-018-37695-0 unenforceable in rural areas 1 . Furthermore, there has been a huge urban-rural disparity in many ways, including social economic status 42 , utilization 43 , dietary pattern and nutrition status 38 . In this study, the differences of TC and LDL-C levels between OC and CWS were particularly apparent in rural China. The only-child status was associated with increased risk of hyper-LDL-C among rural children. These results suggested that the only-child status may be a risk factor of dyslipidemia in rural children. Thus, it is better to take urban-rural disparities into account in the intervention programs for the childhood dyslipidemia. Limitations in this study should be highlighted. First, the causality of the relationships observed cannot be inferred because of the cross-sectional design. Second, there were some factors that we did not take into account, such as food preparation practices, health status of parents, parent-child interaction, etc. Third, self-report bias in lifestyle and socio-economic factors could not be excluded. Fourth, the brief self-reported dietary assessment was subjected to recall bias with not very high ICC. Fifth, there were significant differences in age between participants who were included in the analysis and those who were not. Nevertheless, we confirmed that the association between only-child status and lipid profiles was similar in different age group.

Conclusions
Higher levels of TC and LDL-C were found in OC especially for those who lived in rural areas. Being an only-child was associated with increased risk of hyper-LDL-C especially among rural children. Early prevention of childhood dyslipidemia in OC especially in Chinese rural areas is urgent needed. Furthermore, the urban-rural disparity of lipid profiles between OC and CWS should be well considered in the preventive guidelines and public health policies.

Methods
Design and Study Subjects. This study was based on the baseline data of a multi-centered school-based obesity intervention program (NCT02343588) 44 . The baseline survey was conducted between September 2013 and November 2013. Using a multistage cluster sampling design, 94 schools were selected from 7 provinces/ regions, which included Liaoning (northeast), Tianjin (north), Ningxia (northwest), Shanghai (east), Chongqing (west), Hunan (central), and Guangdong (south). Invitation letters, information sheets and a presentation containing study details were sent to the principals of selected schools. With the principal's permission, all students in the selected schools were invited to participate in the survey and a total of 65,347 children aged 6 to 17 years were recruited. Furthermore, a subsample was selected for blood collection. In brief, half of the schools in each province/region were randomly chosen, and half of the classes in each grade were selected from these schools. In each selected class, all the students were invited to take blood sampling. Children who had missing data on age, gender, only-child status, or living areas (total n = 707) were excluded. Finally, 16,100 lipid profiles were available in this study. Comparing to the overall sample, the subsample (n = 16, 100) was similar in sex distribution (boys: 50.97% vs. 51.60%, P > 0.05), but was slightly older (11.08 ± 3.25 years vs. 10.82 ± 3.30 years, P < 0.05).
The study was approved by the Ethical Committee of the Peking University and was performed in accordance with principles in the Declaration of Helsinki. Written informed consents were obtained from all participating students and their parents.
Anthropometric Measurements. Children's height and weight were measured by qualified technicians, with the child in light clothing without shoes. Body weight was measured to the nearest 0.1 kg, and height was measured to the nearest mm. Each subject's height and weight were measured twice and we calculated the average values. Body mass index (BMI) was expressed by the body weight (in kilograms) divided by height (in meters) squared and then transformed to a BMI z-score based on data from the sex-and age-specific World Health Organization Growth reference (5-19 years) 45 . Questionnaire Assessment. The self-reported questionnaires were designed to collect information on socio-demographic factors and other behavioral factors. Children and families' basic demographic information including age, sex, only child or not, living in urban or rural areas, provinces, parental educational levels, and monthly family incomes were filled out by parents. Children's health-related behaviors including physical activities (PA), screen time (television time and computer time), and food intakes were completed by parents and children together. Participants were asked about daily consumptions of vegetables, fruits, sugar-sweetened beverages (SSBs) and meat products, and weekly frequencies of having dairy products, high-energy food, fried food, and monthly frequency of having western fast food. Participants were asked about the frequency and time (hours and minutes) they spent daily on screen time, vigorous-intensity physical activities (VPA), and moderate-intensity physical activities (MPA) over the past 7 days, respectively. VPA and MPA were defined according to the International Physical Activity Questionnaire (IPAQ) 46 . Moderate to vigorous physical activities (MVPA) time was calculated as the sum of VPA and MPA time. In addition, a quartile method was used to class and code the intake of each food item for each child (1: bottom 25%, 2: 25%-50%, 3: 50%-75%, and 4: top 25%). The total food intake was calculated as the sum of the total scores and was quartered into four groups. The reliability and validity of the questionnaires of sedentary behaviors, PA, and dietary intakes were assessed in a sample of 298 primary school students in Guangzhou, China. Generally, the reliability coefficient of 0.40-0.75 and correlation coefficient of 0.20-0.60 are considered acceptable 47,48 . Our results indicate that the questionnaires have acceptable reliability and validity (e.g. the average Intra-class Correlation Coefficient (ICC) of PA was 0.460, and the average Spearman's Correlation Coefficient (SCC) of PA was 0.407, both P < 0.05).
Blood Lipid Profiles. After at least 12 hours of overnight fasting, venous blood specimens (5 ml) were collected into ethylenediaminetetraacetic acid (EDTA) vacuum tubes, for the measurement of serum lipid levels. The blood specimens were separated by centrifugation at 3000r for 15 minutes, aliquoted and stored at −80 °C Scientific RepoRts | (2019) 9:1441 | https://doi.org/10.1038/s41598-018-37695-0 until testing. Serum lipid levels, including total cholesterol (TC), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) were assessed at a biomedical analysis company, which was accredited by Peking University. TC and TG were measured by enzymatic methods; LDL-C and HDL-C were measured by clearance method. Serum lipids were categorized by using the Chinese criteria for dyslipidemia in children and adolescents 49 : high TC, ≥200 mg/dL (5.20 mmol/L); high TG, ≥150 mg/dL (1.76 mmol/L); high LDL-C, ≥130 mg/dL (3.38 mmol/L); and low HDL-C, ≤40 mg/dL (1.04 mmol/L). Statistical Analysis. Statistical analysis was performed using SPSS 22.0. Descriptive statistics were used to characterize the population. Continuous variables were presented as mean values and standard errors or deviations, while categorical variables were presented as proportions. Statistically significant differences of continuous variables with non-normal distribution (age, height, weight, BMI, BMI z-score, TC, TG, LDL-C, and HDL-C) and categorical variables between OC and CWS were determined using Mann-Whitney test and chi-square test, respectively. Linear mixed models were used in the comparison of lipid profiles between OC and CWS; and generalized linear mixed models were used to assess the association between the only-child status and dyslipidemia. All models were adjusted for age, sex, living areas, parental educational levels, monthly family incomes, BMI z-score, MVPA time, screen time, food intakes, and a random effect for provinces. In addition, we conducted stratified analysis by child's sex and urban/rural residence to further explore the related differences by these factors. A two-sided P value < 0.05 was considered significant.

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.