Data science in large cohort studies / 中华流行病学杂志

Large cohort study gained its popularity in biomedical research and demonstrated its application in exploring disease etiology and pathogenesis, improving the prognosis of disease, as well as reducing the burden of diseases. Data science is an interdisciplinary field that uses scientific methods from computer science and statistics to extract insights or knowledge from data in a specific domain. The results from the combination of the two would provide new evidence for developing the strategies and measures on disease prevention and control. This review included a brief introduction of data science, descriptions on characteristics of large cohort data according to the development of the study design, and application of data science at each stage of a large cohort study, as well as prospected the application of data science in the future large cohort studies.

Interpretation for the group standards in data management for large population-based cohorts / 中华流行病学杂志

Precision medicine became the key strategy in development priority of science and technology in China. The large population-based cohorts become valuable resources in preventing and treating major diseases in the population, which can contribute scientific evidence for personalized treatment and precise prevention. The fundamental question of the achievements above, therefore, is how to construct a large population-based cohort in a standardized way. The Chinese Preventive Medicine Association co-ordinated experienced researchers from Peking University and other well-known institutes to write up two group standards Technical specification of data processing for large population-based cohort study (T/CPMA 001-2018) and Technical specification of data security for large population-based cohort study (T/CPMA 002-2018), on data management. The standards are drafted with principles of emphasizing their scientific, normative, feasible, and generalizable nature. In these two standards, the key principles are proposed, and technical specifications are recommended in data standardization, cleansing, quality control, data integration, data privacy protection, and database security and stability management in large cohort studies. The standards aim to guide the large population-based cohorts that have been or intended to be established in China, including national cohorts, regional population cohorts, and special population cohorts, hence, to improve domestic scientific research level and the international influence, and to support decision-making and practice of disease prevention and control.

Study on genetic structure differences and adjustment strategies in different areas of China / 中华流行病学杂志

Objective: To describe the genetic structure of populations in different areas of China, and explore the effects of different strategies to control the confounding factors of the genetic structure in cohort studies. Methods: By using the genome-wide association study (GWAS) on data of 4 500 samples from 10 areas of the China Kadoorie Biobank (CKB), we performed principal components analysis to extract the first and second principal components of the samples for the component two-dimensional diagram generation, and then compared them with the source of sample area to analyze the characteristics of genetic structure of the samples from different areas of China. Based on the CKB cohort data, a simulation data set with cluster sample characteristics such as genetic structure differences and extensive kinship was generated; and the effects of different analysis strategies including traditional analysis scheme and mixed linear model on the inflation factor (λ) were evaluated. Results: There were significant genetic structure differences in different areas of China. Distribution of the principal components of the population genetic structure was basically consistent with the geographical distribution of the project area. The first principal component corresponds to the latitude of different areas, and the second principal component corresponds to the longitude of different areas. The generated simulation data showed high false positive rate (λ=1.16), even if the principal components of the genetic structure was adjusted or the area specific subgroup analysis was performed, λ could not be effectively controlled (λ>1.05); while, by using a mixed linear model adjusting for the kinship matrix, λ was effectively controlled regardless of whether the genetic structure principal component was further adjusted (λ=0.99). Conclusions: There were large differences in genetic structure among populations in different areas of China. In molecular epidemiology studies, bias caused by population genetic structure needs to be carefully treated. For large cohort data with complex genetic structure and extensive kinship, it is necessary to use a mixed linear model for association analysis.

Relationship between educational level and long-term changes of body weight and waist circumference in adults in China / 中华流行病学杂志

Objective: To evaluate the association of educational level with anthropometric measurements at different adult stages and their long-term changes in adults who participated in the second re-survey of China Kadoorie Biobank (CKB). Methods: The present study excluded participants who were aged >65 years, with incomplete or extreme measurement values, or with major chronic diseases at baseline survey or re-survey. The weight at age 25 years was self-reported. Body height, body weight and waist circumference at baseline survey (2004-2008) and re-survey (2013-2014) were analyzed. Results: The present study included 3 427 men and 6 320 women. Both body weight and waist circumference (WC) increased with age. From age 25 years to baseline survey (mean age 45.2±6.5), the mean weight change per 5-year was (1.70±2.63) kg for men and (1.27±2.10) kg for women. From baseline survey to re-survey (53.2±6.5), the mean changes per 5-year for body weight were (1.12±2.61) kg for men and (0.90±2.54) kg for women; and that for WC was (3.20±3.79) cm for men and (3.83±3.85) cm for women. Among women, low educational level was consistently associated with higher body mass index (BMI) and WC at age 25 years, baseline survey and re-survey. Among men, low educational level was associated with higher BMI at age 25 years. At baseline survey and re-survey, the educational level in men was not statistically associated with BMI; but men who completed junior or senior high school showed slight higher WC and increase of WC from baseline survey to re-survey than other male participants. Conclusions: Body weight and WC increased with age for both men and women. The associations of educational level with BMI and WC were different between men and women.

Prevalence of 'healthy lifestyle' in Chinese adults / 中华流行病学杂志

Objective: To examine the prevalence of 'healthy lifestyle' from data extracted from the China Kadoorie Biobank (CKB) of 0.5 million adults from ten areas across China. Methods: After excluding participants with self-reported histories of coronary heart disease, stroke or cancer, a total of 487 198 participants at baseline (2004-2008) and 22 604 participants at second survey (2013- 2014), were included for analysis. 'Healthy lifestyle' was defined as haing the following characteristics: a) never smoking or having stopped smoking for reasons other than illness; b) alcohol drinking <25 g/day (men)/<15 g/day (women); c) diet rich in vegetables, fruits, legumes and fish, but low in red meat; d) upper quarter of the physical activity level; e) body mass index of 18.5-23.9 kg/m(2) and waist circumstance <85 cm (men)/80 cm (women). We calculated the healthy lifestyle scores (HLS) by counting the number of all the healthy lifestyle factors, with a range from 0 to 6. Results: At baseline, prevalence rates of the above five healthy lifestyles (except physical activity) were 70.6%, 92.6%, 8.7%, 52.6% and 59.0%, respectively, with the mean HLS being 3.1±1.2. Most participants (81.4%) had2-4 healthy components, while only 0.7% (0.2% in men and 1.0% in women) of all the participants had all six healthy lifestyles. Participants who were women, at younger age, with more schooling and rural residents, were more likely to adhere to the healthy lifestyle. After ten years, the mean HLS showed a slight decrease. Conclusion: The prevalence of optimal lifestyles in Chinese adults appeared extremely low. Levels of 'healthy lifestyle' varied greatly among those populations with different socio-demographic characteristics across the ten areas in China.

Family history and risk of coronary heart disease / 中华流行病学杂志

Objective: To evaluate the association of family history with risk of major coronary events (MCE) and ischemic heart disease (IHD). Methods: After excluding participants with heart disease, stroke or cancer at baseline survey, a total of 485 784 participants from the China Kadoorie Biobank, who had no missing data on critical variables, were included in the analysis. Cox regression analysis was used to estimate the hazard ratios (HR) and 95% CI. Subgroup analyses were performed according to the baseline characteristics. Results: During a median of 7.2 years of follow-up, we documented 3 934 incident cases of MCE and 24 537 cases of IHD. In multivariable-adjusted models, family history was significantly associated with risk of MCE and IHD. The adjusted HRs (95%CI) were 1.41 (1.19-1.65) and 1.25 (1.18-1.33), respectively. History of disease among siblings was more strongly associated with early-onset MCE than parental history (HR=2.97, 95%CI: 1.80-4.88). Moreover, the association of family history with MCE and IHD was stronger in persons who were overweight or obesive, and the association between family history and MEC was stronger in smokers. Conclusion: This large-scale, prospective study indicated that family history was an independent risk factor for MCE and IHD in China. The intervention targeting major known lifestyle risk factors and the management of chronic diseases should be strengthened for Chinese population, especially for the individuals with family history were at high risk.

Levels of calcaneus bone mineral density in adults from 10 regions of China / 中华流行病学杂志

Objective: To describe the regional and population-related differences in calcaneus bone mineral density (BMD) across ten regions of China. Methods: Based on the results: from the second Re-survey of China Kadoorie Biobank project, in which 5% of the surviving participants were interviewed during 2013-2014 and 24 677 participants aged 38-87 years were included in the study. We excluded those people with missing data for BMD and important variables. Calcaneus BMD was measured using the quantitative ultrasound bone densitometer. We analyzed four indexes, including broadband ultrasound attenuation (BUA), speed of sound (SOS), stiffness index (SI), and T score. Results The average calcaneus BMDs of the present population were: BUA (109.7±12.6) dB/MHz, SOS (1 554.7±45.6) m/s, SI (88.3±18.8), T score (-0.74±1.28). Urban residents showed higher calcaneus BMD, so as in men. The calcaneus BMD decreased by age, with a larger decline seen in women. Current smokers and postmenopausal women presented lower calcaneus BMD, while in those who frequently drank milk or yogurt or being physically more active, had higher calcaneus BMD. Conclusion: Calcaneus BMD varied greatly among people from the ten regions of CKB study and among participants having different demographic characteristics, lifestyle behaviors or health conditions.

Association between obesity and DNA methylation among the 7-16 year-old twins / 中华流行病学杂志

Objective: On whole-genome scale, we tried to explore the correlation between obesity-related traits and DNA methylation sites, based on discordant monozygotic twin pairs. Methods: A total of 90 pairs of 6-17 year-old twins were recruited in Chaoyang district, Yanqing district and Fangshan district in Beijing in 2016. Information on twins was gathered through a self-designed questionnaire and results: from physical examination, including height, weight and waist circumference of the subjects under study. DNA methylation detection was chosen on the Illumina Human Methylation EPIC BeadChip. R 3.3.1 language was used to read the DNA methylation signal under quality control on samples and probes. Ebayes function of empirical Bayes paired moderated t-test was used to identify the differential methylated CpG sites (DMCs). VarFit function of empirical Bayes paired moderated Levene test was used to identify the differentially variables CpG sits (DVCs) in obese and normal groups. Results According to the obesity discordance criteria, we collected 23 pairs of twins (age range 7 to 16 years), including 12 male pairs. A total of 817 471 qualified CpG loci were included in the genome-wide correlation analysis. According to the significance level of FDR set as <0.05, no positive sites would meet this standard. When DMC CpG site cg05684382, with the smallest P value (1.26E-06) as on chromosome 12, the DVC CpG site cg26188191 with the smallest P value (6.44E-06) appeared in CMIP gene on chromosome 16. Conclusions: In this study, we analyzed the genome-wide DNA methylation and its correlation with obesity traits. After multiple testing corrections, no positive sites were found to have associated with obesity. However, results from the correlation analysis demonstrated sites cg05684382 (chr: 12) and cg26188191 (chr: 16) might have played a role in the development of obesity. This study provides a methodologic reference for the studies on discordance twins related problems.

Effect of family environment in childhood and adolescence on mental health in adulthood / 中华流行病学杂志

Objective: To explore the relationship of family environment in childhood and adolescence and mental health in adulthood. Methods: A total of 791 subjects aged ≥25 years were selected through the Chinese National Twin Registry (CNTR). The short-form of Family Environment Scale-Chinese Version (FES-CV) was used to evaluate the family environment during childhood and adolescence in three dimensions: relationship, system maintenance and personal growth. The mental health status in adulthood was assessed with the Chinese version of 6-item Kessler Psychological Distress Scale (K6). The generalized linear mixed model was used to examine their relationship. Results: About 4.6% of the subjects reported general or worse mental health status. Compared with the subjects with good mental status, statistical difference was observed only in parenting way among twins (living together or not). After adjusting the potential confounders, such as age, sex, zygosity, education and lifestyle (smoking, drinking and physical activity), good family relationship and system maintenance had a positive effect on mental health, with the OR (95%CI) of 0.66 (0.51-0.87) and 0.70 (0.50-0.98) respectively. Furthermore, parenting way did not modify the effect of family environment on mental health status in adulthood (interaction: P>0.05). In each scale, scores of cohesion and organization were positively correlated with mental health, while the score of conflict was negatively correlated with the mental health. Conclusion: Good family relationship and system maintenance in childhood and adolescence had a positive impact on mental health in adulthood.

Association between central obesity and risk for heart disease in adults in China: a prospective study / 中华流行病学杂志

Objective: To understand the association between central obesity, assessed by waist circumference, and the risks for ischemic heart disease (IHD), major coronary event (MCE), and IHD death. Methods: After excluding participants with heart disease, stroke, cancer, COPD, and diabetes at baseline survey, we included a total of 428 595 participants from the China Kadoorie Biobank for the analysis. The baseline survey was conducted from June 2004 to July 2008. We used Cox proportional hazards model to estimate the hazard ratio (HR) and 95%CI. Results: During an average 9.1 years of follow-up (3 803 637 person-years), we documented 26 900 incident cases of IHD, 4 320 cases of MCE, and 2 787 of deaths from IHD. After adjustment for possible confounders and BMI, central obesity was found to be associated with increased risks for IHD, MCE, and IHD death. The adjusted HRs (for the participants who had waist circumference of 85.0-89.9 cm in men and 80.0-84.9 cm in women were 1.13 (95%CI: 1.09-1.17) for IHD, 1.15 (95%CI: 1.05-1.26) for MCE and 1.11 (95%CI: 0.98-1.24) for IHD death. The respective HRs for those central obese participants (men ≥90.0 cm, women ≥85.0 cm) were 1.29 (95%CI: 1.24-1.34), 1.30 (95%CI: 1.17-1.44) and 1.32 (95%CI: 1.16-1.51). Further stratification analysis according to BMI showed that the risks for incident IHD, MCE, and IHD death increased along with the increase of waist circumference even in the participants with normal weight. Conclusion: This large-scale prospective study revealed that central obesity was an independent risk factor for IHD in adults in China, and the risk of IHD would increase with the increase of waist circumference.

Fujian Tulou Family Cohort Study: study design and characteristics of participants and pedigrees in baseline investigation / 中华流行病学杂志

Objective: To describe the study design, the characteristics of participants as well as the pedigrees included in the baseline survey of Fujian Tulou Family Cohort Study. Methods: Fujian Tulou Family Cohort Study was a prospective open cohort study with a biological sample bank. A baseline survey was conducted in Tulou areas of Nanjing county in Fujian province from 2015 to 2018, including questionnaire survey, physical and biochemical indicators examinations, and blood sample collection in adults aged ≥18 years. In addition, family relationship of the participants was also recorded. The pedigree information of the juveniles under 18 years old were also collected. Results: The baseline survey included 2 727 individuals in two clans, of whom 2 373 (87.0%) were adults, and 2 126 participants completed questionnaires, physical examinations and biochemical tests. The average age of the 2 126 participants was (57.9±13.3) years, with 39.4% being males. The current smoking rates in male and female participants were 41.2% and 2.1%, respectively. The corresponding rates of current alcohol consumption were 19.0% and 2.6%. For common chronic diseases, the prevalence rates were 51.3% for hypertension, 9.7% for diabetes and 26.7% for hyperlipemia according to the self-reported disease diagnoses, health examination results and biochemical examination results in class Ⅱ or Ⅲ hospitals. Based on the family relationship information and genealogical data, 710 pedigrees were finally identified, consisting of 5 087 family members. The numbers of five, four, three, and two generations pedigrees were 3, 88, 238 and 381, respectively. The pairs of the first to the fifth degree relatives were 12 039, 2 662, 1 511, 202 and 31, respectively. Conclusion: The establishment of Fujian Tulou Family Cohort provides valuable resources for exploring the genetic risk factors, environmental risk factors and gene-environment interactions contributing to the risk of common chronic diseases.

Epidemiological characteristics of household fuel use in 10 areas of China / 中华流行病学杂志

Objective: To describe the characteristics of cooking and heating fuel use in participants from the China Kadoorie Biobank (CKB) study. Methods: The CKB study recruited 512 891 adults from 10 areas in China during 2004-2008. Information on cooking fuel and heating fuel was collected using a questionnaire in baseline survey. The proportions of various fuels used in different areas, in different populations, and at different time points were calculated and compared. Results: Overall, 52.1% participants used solid fuel for cooking or heating. Rural areas had higher prevalence of solid fuel use than urban areas. The percentage of participants using solid fuel for cooking was 36.1% (coal 20.1%, wood/charcoal 16.0%); The percentage of participants using solid fuel for heating was 36.7% (coal 22.7%, wood/charcoal 14.0%). The prevalence of solid fuel use and the fuel type mainly used varied widely across 10 areas. The proportion of clean fuel use was lower in less-educated and lower-income people. Household coal and wood/charcoal use showed a declining trend, which was more remarkable in urban areas. Conclusion: There are still a large number of rural residents and people with low income relying on solid fuel in China, which is a serious public health concern.

Associations between family history of major chronic diseases and healthy lifestyles in Chinese adults / 中华流行病学杂志

Objective: To examine the associations between family history of major chronic diseases and healthy lifestyle in adults in China. Methods: Data were from the baseline and second survey of China Kadoorie Biobank study, which were conducted during 2004-2008 and 2013-2014, respectively. After excluding participants with self-reported histories of coronary heart disease, stroke, cancer or diabetes, a total of 461 213 adults from baseline survey and 20 583 adults from second survey were included in the current study. Participants who reported a family history of acute myocardial infarction, stroke, cancer, or diabetes of any first-degree relative (i.e., biological father, mother, or siblings) were defined as having a family history of major chronic diseases. Healthy lifestyles were defined as current nonsmoking, non-excessive alcohol drinking, eating vegetables and fruits daily, upper quarter of physical activity level, body mass index (BMI) of (18.5-23.9) kg/m(2), and waist- to-hip ratio (WHR) <0.90 (man)/<0.85 (women). Results: At baseline survey, 36.5% of the participants had family history of major chronic diseases. Proportions of the above six healthy lifestyles were 70.5%, 93.0%, 18.0%, 25.0%, 53.4%, and 43.5%, respectively. Compared with participants without family history, the proportions of current nonsmoking, non-excessive drinking, normal BMI, and normal WHR were lower in participants with family history of major chronic diseases, while the proportions of eating vegetables and fruits daily, and being physically active, were higher. In general, the absolute differences in these proportions between participants with and without a family history were only slight. Similar results were observed when other family history status (the type or number of disease, the category or number of affected family members) were analysed. The association between family history of major chronic diseases and healthy lifestyles was consistently observed in the second survey 10 years later. Conclusion: In Chinese population, adults with family history of major chronic diseases did not adopt healthier lifestyles.