Ethical approval
This study complies with all ethical regulations relevant to research involving human participants and was carried out in accordance with the criteria set by the Helsinki declaration. The UKBB study was approved by the National Research Ethics Committee (June 17, 2011 (Reference res11 / NW / 0382) and was extended on May 10, 2016 (Reference res16 / NW / 0274)). The collection, storage and analyzes of bioscales, genetic data and data derived from electronic health files within the framework of the PMBB study are approved as part of the IRB protocol of the University of Pennsylvania # 813913. UKBB and PMBB cohorts have given written informed consent to use their samples and data for medical research purposes. The use of deactivation data for these biobanks for this specific research has been covered by existing approvals. All data has been managed in accordance with relevant data protection and confidentiality regulations. No additional ethical approval was necessary for this specific analysis of existing approved data sets. This study respected the requirements for strengthening the declaration of declaration of observation studies in epidemiology (Strobe).
Study population
The UKBB is a major prospective observation cohort study which recruited more than 500,000 adults from 22 Centers in the United Kingdom between 2006 and 2010. Participants aged 40 to 69 were registered and were followed for subsequent health events. The UKBB has collected in -depth reference data, in particular demography, life factors and physical measures, as well as biological samples for genotyping and analysis of biomarkers. For this study, we included women of European origin who had at least one living birth and available genetic data. The complete protocol of the UKBB study is accessible to the public for reference33.
The PMBB study (i.e. a large-scale academic medical biobank) has not recruited participants in ambulatory environments non-selectively. These participants gave access to their electronic health file data and generated genomic and biomarkers data34. All international diagnostic codes of International Disease Classification (CIM) -9 and ICD -10, clinical imagery and laboratory measures until July 2020 have been extracted from electronic health files. The work flows underlying this study are illustrated in Fig. 1 And 2.
Hypertensive pregnancy disorders and comorbidities
In the UKBB study, the participants provided details on their reproductive history, including parity, during a basic survey. The HDP was defined as gestational hypertension, preeclampsia, lightning or superimposed preeclampsia. This identification was based on self-evaluations during registration or corresponding ICD codes obtained from primary care or hospital files (additional table 14). Similarly, in the PMBB study, the HDP was defined using relevant ICD codes.
To assess the risk of HDP according to HDP-PRS, age at first pregnancy and the presence of a disease giving a high risk of HDP before pregnancy have been selected as covariable, in accordance with clinical directives for high factors Risk for HDP35,,36. According to these guidelines, high HDP risk diseases include hypertension, meal diabetes and dyslipidemia. The presence of high-risk HDP disease before pregnancy was determined by self-assessment or diagnosis with ICD codes relevant to each disease (additional table (table 15) which occurred before the first living birth.
Cardiovascular results
To analyze the ASCVD Incident and its association with HDP-PRS, participants with congenital heart disease were excluded to eliminate the possible association between congenital heart disease and CV results (Additional methods contains relevant diagnostic codes). The prevalent metabolic comorbidities, in particular hypertension, diabetes mellitus and dyslipidemia, were used as adjusted covariable and were determined either by self-declaration to registration, or by ICD codes, as described in Additional methods.
ASCVD incident has been defined as a diagnosis of coronary coronary disease, myocardial infarction, an ischemic stroke, a disease of the peripheral artery or an aortic aneurysm after the inscription of participants without preexisting CVD . In addition, myocardial infarction has been defined algorithmic using UKBB data. For each new ASCVD, participants with pre -existing disease during registration were excluded from the analysis. For example, participants with a preexisting coronary disease during registration were excluded from the analysis of the new coronary disease, which assured that recurring coronary disease was not wrongly counted as a disease Coronarian new appearance.
Variable
During the registration process in the UKBB study, participants provided information on their socio-demographic characteristics, their health / medical history and their lifestyle / environment factors by a self-administration questionnaire and interviews basic in person and basic interviews in person.
According to AHA, four factors mainly define lifestyle behavior; These include current smoking status, obesity, physical activity and eating habits37,,38. Smoking status has been classified as a current or non-smoking smoker. Obesity was defined as a BMI ≥ 30 kg / m2 According to the International Classification of the World Health Organization. Regarding physical activity, participants were classified as having a healthy lifestyle if they reported more than five days a week of moderate or vigorous activity. Eating habits were defined following recommendations on CV health food priorities, which classified common food components such as fruits, vegetables, whole grains, fish, dairy products, refined cereals, meats transformed and unprocessed meats. Eating habits were considered healthy if participants adhere to at least half of CV health food recommendations, as evaluated using a food frequency questionnaire39. Collectively, lifestyle behavior has been classified into three groups – infarnglaws (0–1 healthy lifestyle factor)40Intermediary (2 healthy lifestyle factors) and favorable (≥ 3 healthy lifestyle factors). More detailed descriptions and definitions of the variables considered in lifestyle behavior can be found in Additional methods.
The metabolic state of health has been identified according to the presence of the five components of the dishes according to the criteria of the TDI consensus report41. The metabolic state of health has been classified as three groups – ideals (factor of 0 to 1 food), intermediaries (factors 2 to 3 food) and poor (factors ≥ 4 dishes). Detailed descriptions and definitions of the variables considered in the dishes can be found in the additional table 16 And Additional methods.
In the PMBB cohort, smoking status and obesity (BMI ≥ 30) have been used in restrictively as variable for replication analysis.
Control and imputation of the quality of genotype data
The genotyping and quality control (QC) and imputation procedures followed standard practices and were carried out using a pair of cohort genotyping platform. More details are provided in Additional methods.
British Biobank
The UKBB samples (version 3; March 2018) have been genotyped for more than 800,000 SNP using the Affytrix UK Bileve Axiom table or the Axiom Affytrix UKBB Table. After QC and imputation, 164,500 European participants (white British) were deemed eligible for the validation of genetic analyzes.
Penn Medicine Biobank
PMBB data consisted of 43,623 samples that were genotypical using a GSA genotyping network. After exclusion, 982 women parou of the participants in European (non-Hispanic) ancestry and 1019 women participating parouls with African-American ascendance (non-Hispanic black) were considered eligible for replication analysis.
Polygenic risk score
The HDP-PRS has been generated using summary statistics from a large-scale GWAS HDP (13,071 cases and 177,808 controls) of the Finngen Consortium (R8V4 data gel)42. The PRS has been calculated using the PRS-CS Polygenic Predictional Prediction method43. The individual PRS were determined by applying version 1.90 of Plink using the command – score and were calculated from beta coefficients such as the weighted sum of risk alleles44. The details of the PRS analysis are described in Additional methods.
Statistical analysis
The demographic and clinical characteristics are presented as the average ± SD or the number (percentage). Continuous variables have been compared using the student t Test, unidirectional anova or mann-whiteney U Test, as the case may be. The categorical variables were compared using the Test of the Square Chi or the exact Fisher test, if applicable.
To assess the risk of HDP according to HDP-PRS, we used a multiviated logistical regression model to assess the association between HDP-PRS and HDP. We have calculated the IC or and 95% after adjustment for age to first living birth, BMI, smoking status, the first ten ancestry PCs and the type of genotyping network in the model of multivariate logistical regression.
In primary analysis, the association between HDP-PRS and the CV results of the new appearance was examined using the multivariate regression analysis of Cox. Adjustments have been made for history of HDP, from age to first living birth, BMI, smoking status, first ten ancestry PCs and the genotypes network to calculate HR and CI 95%. The gold and HR of the PRS for HDP were used as quantitative variables reported by SD, and the categorical variables have been defined as follows: Low (low (<20%), intermédiaire (20–80%), élevé (80–99%), et très élevé (> 99%). Subsequently, we carried out joint association analyzes to study the interaction between genetic risk, lifestyle and the status of dishes. In addition, we have carried out sensitivity analyzes based on the ASCVD subtypes.
For replication analysis, the impact of maintaining a favorable lifestyle in different genetic risk groups has been studied using the Square Chi test and COX regression analysis in a PMBB cohort independent. Among the women of 2001 Paroue of the PMBB cohort, the cases with ASCVD new appearance were few, but the conditional regression analysis of the COX was possible for incident hypertension, considering smoking status and obesity (BMI ≥ 30) as lifestyle variables.
All statistical tests have been bilateral and the statistical service has been set to P<0.05. All statistical analyzes have been carried out using statistical software R (version 4.1.0; R Foundation for Statistical Computing, Vienne, Austria) and Plink version 1.9044. Details of statistical analyzes are described in Additional methods.
Summary of reports
Further information on research design is available in the Summary of portfolio reports of nature linked to this article.