Dyslipidemia, inflammation, calcification, and adiposity in aortic stenosis: a genome-wide study.

AIMS: Although highly heritable, the genetic etiology of calcific aortic stenosis (AS) remains incompletely understood. The aim of this study was to discover novel genetic contributors to AS and to integrate functional, expression, and cross-phenotype data to identify mechanisms of AS. METHODS AND RESULTS: A genome-wide meta-analysis of 11.6 million variants in 10 cohorts involving 653 867 European ancestry participants (13 765 cases) was performed. Seventeen loci were associated with AS at P ≤ 5 × 10-8, of which 15 replicated in an independent cohort of 90 828 participants (7111 cases), including CELSR2-SORT1, NLRP6, and SMC2. A genetic risk score comprised of the index variants was associated with AS [odds ratio (OR) per standard deviation, 1.31; 95% confidence interval (CI), 1.26-1.35; P = 2.7 × 10-51] and aortic valve calcium (OR per standard deviation, 1.22; 95% CI, 1.08-1.37; P = 1.4 × 10-3), after adjustment for known risk factors. A phenome-wide association study indicated multiple associations with coronary artery disease, apolipoprotein B, and triglycerides. Mendelian randomization supported a causal role for apolipoprotein B-containing lipoprotein particles in AS (OR per g/L of apolipoprotein B, 3.85; 95% CI, 2.90-5.12; P = 2.1 × 10-20) and replicated previous findings of causality for lipoprotein(a) (OR per natural logarithm, 1.20; 95% CI, 1.17-1.23; P = 4.8 × 10-73) and body mass index (OR per kg/m2, 1.07; 95% CI, 1.05-1.9; P = 1.9 × 10-12). Colocalization analyses using the GTEx database identified a role for differential expression of the genes LPA, SORT1, ACTR2, NOTCH4, IL6R, and FADS. CONCLUSION: Dyslipidemia, inflammation, calcification, and adiposity play important roles in the etiology of AS, implicating novel treatments and prevention strategies.

Body size in early life and the risk of postmenopausal breast cancer.

BACKGROUND: Greater early life adiposity has been reported to reduce postmenopausal breast cancer risk but it is unclear whether this association varies by tumour characteristics. We aimed to assess associations of early life body size with postmenopausal breast cancer and its subtypes, allowing for body size at other ages. METHODS: A total of 342,079 postmenopausal UK women who reported their body size at age 10, clothes size at age 20, and body mass index (BMI) at baseline (around age 60) were followed by record linkage to national databases for cancers and deaths. Cox regression yielded adjusted relative risks (RRs) of breast cancer, overall and by tumour subtype, in relation to body size at different ages. RESULTS: During an average follow-up of 14 years, 15,506 breast cancers were diagnosed. After adjustment for 15 potential confounders, greater BMI at age 60 was associated with an increased risk of postmenopausal breast cancer (RR per 5 kg/m2=1.20, 95%CI 1.18-1.22) whereas greater adiposity in childhood and, to a lesser extent, early adulthood, was associated with a reduced risk (0.70, 0.66-0.74, and 0.92, 0.89-0.96, respectively). Additional adjustment for midlife BMI strengthened associations with BMI at both age 10 (0.63, 0.60-0.68) and at age 20 (0.78, 0.75-0.81). The association with midlife adiposity was confined to hormone sensitive subtypes but early life adiposity had a similar impact on the risk of all subtypes. CONCLUSION: Early life and midlife adiposity have opposite effects on postmenopausal breast cancer risk and the biological mechanisms underlying these associations are likely to differ.

Body mass index and use and costs of primary care services among women aged 55-79 years in England: a cohort and linked data study.

BACKGROUND: Excess weight is associated with poor health and increased healthcare costs. There are no reliable data describing the association between BMI and the use and costs of primary care services in the United Kingdom. METHODS: Among 69,440 participants in the Million Women Study with primary care records in the Clinical Practice Research Datalink between April 2006 (mean age 64 years) and March 2014, the annual rates and costs of their primary care consultations, prescription medications, and diagnostic and monitoring tests were estimated in relation to their self-reported body mass index (BMI) at recruitment in 1996-2001 (mean age 56 years). Associations of BMI with annual costs were projected to all women in England aged 55-79 years in 2013. RESULTS: Over an average follow-up of 6.0 years, annual rates and mean costs were lowest for women with a BMI of 20 to <22.5 kg/m2 for consultations (7.0 consultations, 99% CI 6.8-7.1; £288, £280-£295) and prescription medications (27.0 prescribed items, 26.0-27.9; £227, £216-£237). Above 20 kg/m2, a 2 kg/m2 higher BMI (a 5 kg change in weight for a woman of average height) was associated with 5.2% (4.8-5.6) and 9.9% (9.2-10.6) higher mean annual consultation and prescription medication costs, respectively. Annual rates and mean costs of diagnostic and monitoring tests were similar for women with different BMIs. Among all women aged 55-79 years in England, excess weight accounted for an estimated 11% (£229 million/£2.2 billion) of all consultation costs and 20% (£384 million/£1.9 billion) of all prescription medication costs, of which 27% were for diabetes drugs, 19% for circulatory system drugs, and 13% for analgesics. CONCLUSIONS: Excess body weight is associated with higher use and costs of primary care services among women in England. Reducing the prevalence of excess weight could improve the health of women and reduce pressures on primary care.

Adult cancer risk in women who were breastfed as infants: large UK prospective study.

There are known short-term benefits in breastfed infants versus bottle-fed infants in terms of lower risks of infection and obesity in infancy and childhood, but the long-term effect on the risk of adult cancers is unclear. In a cohort of 1 in 4 UK women born in 1935-1950 we report the incidence of adult cancers in relation to having been breastfed in infancy. In median year 2001 (interquartile range 2000-2003) 548,741 women without prior cancer reported whether they had been breastfed. There was 81% agreement between women's report of having been breastfed and information on breastfeeding recorded when they were 2 years old. Participants were followed by record-linkage to national cancer registration, hospital admission and death databases. Cox regression yielded adjusted relative risks (RRs) and 95% confidence intervals (CI) by having been breastfed or not for eight cancer sites with > 2000 incident cases and for related conditions, where appropriate. Of the eight cancers examined here one association was highly statistically significant: an increase in colorectal cancer incidence among women who had been breastfed versus not (RR 1.18, 95% CI 1.12-1.24, n = 8651). To investigate further the findings for colorectal cancer, we studied eight other gastro-intestinal conditions, and found increased risks in women who had been breastfed versus not for benign colorectal polyps (RR 1.09, 95% CI 1.05-1.13, n = 17,677) and for appendicitis (RR 1.19, 95% CI 1.07-1.31, n = 2108). The greater risks of adult colorectal cancer, colorectal polyps and appendicitis associated with having been breastfed in infancy suggest possible long-term effects of infant feeding practices on the gastrointestinal tract. Further studies are required to clarify this novel association.

Frequent physical activity may not reduce vascular disease risk as much as moderate activity: large prospective study of women in the United Kingdom.

BACKGROUND: Although physical activity has generally been associated with reduced risk of vascular disease, there is limited evidence about the effects of the frequency and duration of various activities on the incidence of particular types of vascular disease. METHODS AND RESULTS: In 1998, on average, 1.1 million women without prior vascular disease reported their frequency of physical activity and many other personal characteristics. Three years later, they were asked about hours spent walking, cycling, gardening, and housework each week. Women were followed by record linkage to National Health Service cause-specific hospital admissions and death records. Cox regression was used to calculate adjusted relative risks for first vascular events in relation to physical activity. During an average of 9 years follow-up, 49,113 women had a first coronary heart disease event, 17,822 had a first cerebrovascular event, and 14,550 had a first venous thromboembolic event. In comparison with inactive women, those reporting moderate activity had significantly lower risks of all 3 conditions (P<0.001 for each). However, women reporting strenuous physical activity daily had higher risks of coronary heart disease (P=0.002), cerebrovascular disease (P<0.001), and venous thromboembolic events (P<0.001) than those reporting doing such activity 2 to 3 times per week. Risks did not differ between hemorrhagic and ischemic stroke, or between venous thromboembolic events with or without pulmonary embolism. CONCLUSIONS: Moderate physical activity is associated with a lower risk of coronary heart disease, venous thromboembolic event, and cerebrovascular disease than inactivity. However, among active women, there is little to suggest progressive reductions in risk of vascular diseases with increasing frequency of activity.

Age at menarche and risks of coronary heart and other vascular diseases in a large UK cohort.

BACKGROUND: Early menarche has been associated with increased risk of coronary heart disease (CHD), but most studies were relatively small and could not assess risk across a wide range of menarcheal ages; few have examined associations with other vascular diseases. We examined CHD, cerebrovascular disease, and hypertensive disease risks by age at menarche in a large prospective study of UK women. METHODS AND RESULTS: In 1.2 million women (mean±SD age, 56±5 years) without previous heart disease, stroke, or cancer, menarcheal age was reported to be 13 years by 25%, ≤10 years by 4%, and ≥17 years by 1%. After 11.6 years of follow-up, 73 378 women had first hospitalization for or death from CHD, 25 426 from cerebrovascular disease, and 249 426 from hypertensive disease. Using Cox regression, we calculated relative risks for each vascular outcome by single year of menarcheal age. The relationship was U-shaped for CHD. Compared with women with menarche at 13 years, the adjusted relative risk for CHD for menarche at ≤10 years of age was 1.27 (95% confidence interval, 1.22-1.31; P<0.0001) and for menarche at ≥17 years of age was 1.23 (95% confidence interval, 1.16-1.30; P<0.0001). U-shaped relationships were also seen for cerebrovascular and hypertensive disease, although the magnitudes of these risks for early and late menarche were smaller than those for CHD. CONCLUSIONS: In this cohort, the relation of age at menarche to vascular disease risk was U shaped, with both early and late menarche being associated with increased risk. Associations were weaker for cerebrovascular and hypertensive disease than for CHD.

Body size in early life and risk of lymphoid malignancies and histological subtypes in adulthood.

Risk of adult lymphoid malignancy is associated with recent adiposity. Some have reported apparent associations with adiposity in childhood or early adulthood, but whether these associations are independent of recent adiposity is unknown. Birth weight, body size at age 10 years, clothes size at age 20 years, and recent body mass index (BMI) were recorded in 745,273 UK women, mean age 60.1 (SD 4.9) at baseline, without prior cancer. They were followed for 11 years, during which time 5,765 lymphoid malignancies occurred. Using Cox regression, a higher risk of lymphoid malignancy was strongly associated with higher recent BMI (RR=1.33, 95%CI 1.17-1.51, for BMI 35+ vs <22.5 kg/m(2)), and this association remained essentially unchanged after adjustment for birth weight and body size at 10. Higher lymphoid malignancy risk was also associated with large size at birth, at age 10, and at age 20 years, but after adjustment for recent BMI, the significance of the associations with large size at birth and at age 10 years was sufficiently reduced that residual confounding by adult BMI could not be excluded; a weak association with large size at 20 years remained (adjusted RR =1.17, 95%CI 1.10-1.24 for large size at age 20 vs. medium or small size). We found no strong evidence of histological specificity in any of these associations. In conclusion, our findings suggest a possible role of adiposity throughout adulthood in the risk of lymphoid malignancy, but the independent contribution of body size at birth and during childhood appears to be small.

The role of health-related behavioural factors in accounting for inequalities in coronary heart disease risk by education and area deprivation: prospective study of 1.2 million UK women.

BACKGROUND: Some recent research has suggested that health-related behaviours, such as smoking, might explain much of the socio-economic inequalities in coronary heart disease (CHD) risk. In a large prospective study of UK women, we investigated the associations between education and area deprivation and CHD risk and assessed the contributions of smoking, alcohol consumption, physical activity and body mass index (BMI) to these inequalities. METHODS: After excluding women with heart disease, stroke or cancer at recruitment, 1,202,983 women aged 56 years (SD 5 years) on average, were followed for first coronary event (hospital admission or death) and for CHD mortality. Relative risks of CHD were estimated by Cox regression, and the extent to which any association could be accounted for by smoking, alcohol, physical inactivity, and BMI was assessed by calculating the percentage reduction in the relevant likelihood-ratio (LR) statistic after adjustment for these factors, separately and together. RESULTS: A total of 71,897 women had a first CHD event (hospital admission or death) and 6032 died from CHD during 12 years follow-up. In analyses adjusted by age, birth cohort and region of residence only, lower levels of education and greater deprivation were associated with higher risks of CHD (P heterogeneity < 0.0001 for each); associations for education were found within every level of deprivation and for deprivation were found within every level of education. Smoking, alcohol consumption, physical inactivity and BMI accounted for most of the associations (adjustment for all four factors together reduced the LR statistics for education and for deprivation by 76 % and 71 %, respectively, for first CHD event; and by 87 % and 79 %, respectively, for CHD mortality). Of these four factors, adjustment for smoking resulted in the largest reduction in the LR statistic. Given the large reduction in the predictive values of education and deprivation after adjustment for only four health-related behavioural factors recorded just at recruitment, residual confounding might plausibly account for the remaining associations. CONCLUSIONS: Most of the association between CHD risk and education and area deprivation in UK women is accounted for by health-related behaviours, particularly by smoking and to a lesser extent by alcohol consumption, physical inactivity and BMI.

Validity over time of self-reported anthropometric variables during follow-up of a large cohort of UK women.

BACKGROUND: In prospective epidemiological studies, anthropometry is often self-reported and may be subject to reporting errors. Self-reported anthropometric data are reasonably accurate when compared with measurements made at the same time, but reporting errors and changes over time in anthropometric characteristics could potentially generate time-dependent biases in disease-exposure associations. METHODS: In a sample of about 4000 middle-aged UK women from a large prospective cohort study, we compared repeated self-reports of weight, height, derived body mass index, and waist and hip circumferences, obtained between 1999 and 2008, with clinical measurements taken in 2008. For self-reported and measured values of each variable, mean differences, correlation coefficients, and regression dilution ratios (which measure relative bias in estimates of linear association) were compared over time. RESULTS: For most variables, the differences between self-reported and measured values were small. On average, reported values tended to be lower than measured values (i.e. under-reported) for all variables except height; under-reporting was greatest for waist circumference. As expected, the greater the elapsed time between self-report and measurement, the larger the mean differences between them (each P < 0.001 for trend), and the weaker their correlations (each P < 0.004 for trend). Regression dilution ratios were in general close to 1.0 and did not vary greatly over time. CONCLUSION: Reporting errors in anthropometric variables may result in small biases to estimates of associations with disease outcomes. Weaker correlations between self-reported and measured values would result in some loss of study power over time. Overall, however, our results provide new evidence that self-reported anthropometric variables remain suitable for use in analyses of associations with disease outcomes in cohort studies over at least a decade of follow-up.

Source of dietary fibre and diverticular disease incidence: a prospective study of UK women.

BACKGROUND: Previous prospective studies have found the incidence of intestinal diverticular disease decreased with increasing intakes of dietary fibre, but associations by the fibre source are less well characterised. We assessed these associations in a large UK prospective study of middle-aged women. METHODS AND FINDINGS: During 6 (SD 1) years follow-up of 690 075 women without known diverticular disease who had not changed their diet in the last 5 years, 17 325 were admitted to hospital or died with diverticular disease. Dietary fibre intake was assessed using a validated 40-item food questionnaire and remeasured 1 year later in 4265 randomly-selected women. Mean total dietary fibre intake at baseline was 13.8 (SD 5.0) g/day, of which 42% came from cereals, 22% from fruits, 19% from vegetables (not potatoes) and 15% from potatoes. The relative risk (95% CI) for diverticular disease per 5 g/day fibre intake was 0.86 (0.84 to 0.88). There was significant heterogeneity by the four main sources of fibre (p<0.0001), with relative risks, adjusted for each of the other sources of dietary fibre of 0.84 (0.81 to 0.88) per 5 g/day for cereal, 0.81 (0.77 to 0.86) per 5 g/day for fruit, 1.03 (0.93 to 1.14) per 5 g/day for vegetable and 1.04 (1.02 to 1.07) per 1 g/day for potato fibre. CONCLUSIONS: A higher intake of dietary fibre is associated with a reduced risk of diverticular disease. The associations with diverticular disease appear to vary by fibre source, and the reasons for this variation are unclear.

Hospital admissions in relation to body mass index in UK women: a prospective cohort study.

BACKGROUND: Adiposity is associated with many adverse health outcomes but little direct evidence exists about its impact on the use of health care services. We aim to describe the relationship between body mass index (BMI) and rates of hospital admission in middle-aged UK women. METHODS: Among 1,251,619 Million Women Study participants, 50- to 64-years old at entry into the study, routine data on hospital admissions were used to estimate hospitalization rates according to BMI after standardization for age, region of recruitment, socioeconomic status, reproductive history, smoking status, hormonal therapy use and alcohol intake. Proportional hazards models were used to estimate adjusted relative risks of hospitalization separately for 25 common types of admission. RESULTS: During an average of 9.2 years follow-up, there were 2,834,016 incident hospital admissions. In women with BMIs (in kg/m2) of <22.5, 22.5 to <25, 25 to <30, 30 to <35 and 35+ standardized admission rates (and 95% confidence intervals (CIs)) per woman over a 10-year period were 2.4 (2.4 to 2.4), 2.4 (2.3 to 2.4), 2.6 (2.6 to 2.6), 3.0 (3.0 to 3.0) and 3.5 (3.4 to 3.5), respectively (P-value for heterogeneity <0.001). The relative increase in admission rates per 5 kg/m2 increase in BMI was 1.12 (1.12 to 1.13). This relationship did not vary materially by age. Corresponding average durations of stay (in days) per hospital visit within the same categories of BMI were: 3.1 (3.1 to 3.2), 2.8 (2.7 to 2.8), 2.9 (2.9 to 2.9), 3.2 (3.1 to 3.2) and 3.8 (3.7 to 3.8), respectively (P <0.001).Significant increases in the risk of admission with increasing BMI were observed for 19 of the 25 types of hospital admission considered. BMI was most strongly associated with admissions with diabetes, knee-replacement, gallbladder disease and venous thromboembolism, but marked associations were found with many other common categories of admission including cataracts, carpal tunnel syndrome and diverticulitis. CONCLUSIONS: Among women 50- to 84-years old in England, around one in eight hospital admissions are likely to be attributable to overweight or obesity, translating to around 420,000 extra hospital admissions and two million extra days spent in hospital, annually.

Marital status and ischemic heart disease incidence and mortality in women: a large prospective study.

BACKGROUND: Being married has been associated with a lower mortality from ischemic heart disease (IHD) in men, but there is less evidence of an association for women, and it is unclear whether the associations with being married are similar for incident and for fatal IHD. We examined the relation between marital status and IHD incidence and mortality in the Million Women Study. METHODS: A total of 734,626 women (mean age 60 years) without previous heart disease, stroke or cancer, were followed prospectively for hospital admissions and deaths. Adjusted relative risks (RRs) for IHD were calculated using Cox regression in women who were married or living with a partner versus women who were not. The role of 14 socio-economic, lifestyle and other potential confounding factors was investigated. RESULTS: 81% of women reported being married or living with a partner and they were less likely to live in deprived areas, to smoke or be physically inactive, but had a higher alcohol intake than women who were not married or living with a partner. During 8.8 years of follow-up, 30,747 women had a first IHD event (hospital admission or death) and 2,148 died from IHD. Women who were married or living with a partner had a similar risk of a first IHD event as women who were not (RR = 0.99, 95% confidence interval (CI) 0.96 to 1.02), but a significantly lower risk of IHD mortality (RR = 0.72, 95% CI 0.66 to 0.80, P <0.0001). This lower risk of IHD death was evident both in women with and without a prior IHD hospital admission (respectively: RR = 0.72, 95% CI 0.60 to 0.85, P <0.0001, n = 683; and 0.70, 95% CI 0.62 to 0.78, P <0.0001, n = 1,465). These findings did not vary appreciably between women of different socio-economic groups or by lifestyle and other factors. CONCLUSIONS: After adjustment for socioeconomic, lifestyle and other factors, women who were married or living with a partner had a similar risk of developing IHD but a substantially lower IHD mortality compared to women who were not married or living with a partner.

Incidence of gastrointestinal cancers by ethnic group in England, 2001-2007.

OBJECTIVE: To compare the incidence of six gastrointestinal cancers (colorectal, oesophageal, gastric, liver, gallbladder and pancreatic) among the six main 'non-White' ethnic groups in England (Indian, Pakistani, Bangladeshi, Black African, Black Caribbean and Chinese) to each other and to Whites. METHODS: We analysed all 378 511 gastrointestinal cancer registrations from 2001-2007 in England. Ethnicity was obtained by linkage to the Hospital Episodes Statistics database and we used mid-year population estimates from 2001-2007. Incidence rate ratios adjusted for age, sex and income were calculated, comparing the six ethnic groups (and combined 'South Asian' and 'Black' groups) to Whites and to each other. RESULTS: There were significant differences in the incidence of all six cancers between the ethnic groups (all p<0.001). In general, the 'non-White' groups had a lower incidence of colorectal, oesophageal and pancreatic cancer compared to Whites and a higher incidence of liver and gallbladder cancer. Gastric cancer incidence was lower in South Asians but higher in Blacks and Chinese. There was strong evidence of differences in risk between Indians, Pakistanis and Bangladeshis for cancer of the oesophagus, stomach, liver and gallbladder (all p<0.001) and between Black Africans and Black Caribbeans for liver and gallbladder cancer (both p<0.001). CONCLUSIONS: The risk of gastrointestinal cancers varies greatly by individual ethnic group, including within those groups that have traditionally been grouped together (South Asians and Blacks). Many of these differences are not readily explained by known risk factors and suggest that important, potentially modifiable causes of these cancers are still to be discovered.

Body mass index and incident coronary heart disease in women: a population-based prospective study.

BACKGROUND: A high body mass index (BMI) is associated with an increased risk of mortality from coronary heart disease (CHD); however, a low BMI may also be associated with an increased mortality risk. There is limited information on the relation of incident CHD risk across a wide range of BMI, particularly in women. We examined the relation between BMI and incident CHD overall and across different risk factors of the disease in the Million Women Study. METHODS: 1.2 million women (mean age=56 years) participants without heart disease, stroke, or cancer (except non-melanoma skin cancer) at baseline (1996 to 2001) were followed prospectively for 9 years on average. Adjusted relative risks and 20-year cumulative incidence from age 55 to 74 years were calculated for CHD using Cox regression. RESULTS: After excluding the first 4 years of follow-up, we found that 32,465 women had a first coronary event (hospitalization or death) during follow-up. The adjusted relative risk for incident CHD per 5 kg/m2 increase in BMI was 1.23 (95% confidence interval (CI) 1.22 to 1.25). The cumulative incidence of CHD from age 55 to 74 years increased progressively with BMI, from 1 in 11 (95% CI 1 in 10 to 12) for BMI of 20 kg/m2, to 1 in 6(95% CI 1 in 5 to 7) for BMI of 34 kg/m2. A 10 kg/m2 increase in BMI conferred a similar risk to a 5-year increment in chronological age. The 20 year cumulative incidence increased with BMI in smokers and non-smokers, alcohol drinkers and non-drinkers, physically active and inactive, and in the upper and lower socioeconomic classes. In contrast to incident disease, the relation between BMI and CHD mortality (n=2,431) was J-shaped. For the less than 20 kg/m2 and ≥35 kg/m2 BMI categories, the respective relative risks were 1.27 (95% CI 1.06 to 1.53) and 2.84 (95% CI 2.51 to 3.21) for CHD deaths, and 0.89 (95% CI 0.83 to 0.94) and 1.85 (95% CI 1.78 to 1.92) for incident CHD. CONCLUSIONS: CHD incidence in women increases progressively with BMI, an association consistently seen in different subgroups. The shape of the relation with BMI differs for incident and fatal disease.

Generating synthetic mixed-type longitudinal electronic health records for artificial intelligent applications.

The recent availability of electronic health records (EHRs) have provided enormous opportunities to develop artificial intelligence (AI) algorithms. However, patient privacy has become a major concern that limits data sharing across hospital settings and subsequently hinders the advances in AI. Synthetic data, which benefits from the development and proliferation of generative models, has served as a promising substitute for real patient EHR data. However, the current generative models are limited as they only generate single type of clinical data for a synthetic patient, i.e., either continuous-valued or discrete-valued. To mimic the nature of clinical decision-making which encompasses various data types/sources, in this study, we propose a generative adversarial network (GAN) entitled EHR-M-GAN that simultaneously synthesizes mixed-type timeseries EHR data. EHR-M-GAN is capable of capturing the multidimensional, heterogeneous, and correlated temporal dynamics in patient trajectories. We have validated EHR-M-GAN on three publicly-available intensive care unit databases with records from a total of 141,488 unique patients, and performed privacy risk evaluation of the proposed model. EHR-M-GAN has demonstrated its superiority over state-of-the-art benchmarks for synthesizing clinical timeseries with high fidelity, while addressing the limitations regarding data types and dimensionality in the current generative models. Notably, prediction models for outcomes of intensive care performed significantly better when training data was augmented with the addition of EHR-M-GAN-generated timeseries. EHR-M-GAN may have use in developing AI algorithms in resource-limited settings, lowering the barrier for data acquisition while preserving patient privacy.

Independent external validation of the QRISK3 cardiovascular disease risk prediction model using UK Biobank.

OBJECTIVE: To externally evaluate the performance of QRISK3 for predicting 10 year risk of cardiovascular disease (CVD) in the UK Biobank cohort. METHODS: We used data from the UK Biobank, a large-scale prospective cohort study of 403 370 participants aged 40-69 years recruited between 2006 and 2010 in the UK. We included participants with no previous history of CVD or statin treatment and defined the outcome to be the first occurrence of coronary heart disease, ischaemic stroke or transient ischaemic attack, derived from linked hospital inpatient records and death registrations. RESULTS: Our study population included 233 233 women and 170 137 men, with 9295 and 13 028 incident CVD events, respectively. Overall, QRISK3 had moderate discrimination for UK Biobank participants (Harrell's C-statistic 0.722 in women and 0.697 in men) and discrimination declined by age (<0.62 in all participants aged 65 years or older). QRISK3 systematically overpredicted CVD risk in UK Biobank, particularly in older participants, by as much as 20%. CONCLUSIONS: QRISK3 had moderate overall discrimination in UK Biobank, which was best in younger participants. The observed CVD risk for UK Biobank participants was lower than that predicted by QRISK3, particularly for older participants. It may be necessary to recalibrate QRISK3 or use an alternate model in studies that require accurate CVD risk prediction in UK Biobank.

Multimorbidity of cardiovascular disease subtypes in a prospective cohort of 1.2 million UK women.

OBJECTIVE: Cardiovascular multimorbidity (CVM) is the co-occurrence of multiple cardiovascular disease subtypes (CVDs) in one person. Because common patterns and incidence of CVM are not well-described, particularly in women, we conducted a descriptive study of CVM in the Million Women Study, a large population-based cohort of women. METHODS: UK women aged 50-64 years were followed up using hospital admissions and mortality records for an average of 19 years. CVM was defined as having ≥2 of 19 selected CVDs. The age-specific cumulative incidence of CVM between age 60 and 80 years was estimated. The numbers and proportions of individual, pairs and other combinations of CVDs that comprised incident CVM were calculated. For each individual CVD subtype, age-standardised proportions of the counts of other co-occurring CVDs were estimated. RESULTS: The age-specific likelihood of having CVM nearly doubled every 5 years between age 60 and 80 years. Among 1.2 million women without CVD at study baseline, 16% (n=196 651) had incident CVM by the end of follow-up. Around half of all women with CVM had a diagnosis of ischaemic heart disease (n=102 536) or atrial fibrillation (n=96 022), almost a third had heart failure (n=72 186) and a fifth had stroke (n=40 442). The pair of CVDs with the highest age-adjusted incidence was ischaemic heart disease and atrial fibrillation (18.95 per 10 000 person-years). Over 60% of individuals with any given CVD subtype also had other CVDs, after age standardisation. CONCLUSIONS: CVM is common. The majority of women with any specific CVD subtype eventually develop at least one other. Clinical and public health guidelines for CVD management should acknowledge this high likelihood of CVM.

Heterogeneity in the diagnosis and prognosis of ischemic stroke subtypes: 9-year follow-up of 22,000 cases in Chinese adults.

BACKGROUND: Reliable classification of ischemic stroke (IS) etiological subtypes is required in research and clinical practice, but the predictive properties of these subtypes in population studies with incomplete investigations are poorly understood. AIMS: To compare the prognosis of etiologically classified IS subtypes and use machine learning (ML) to classify incompletely investigated IS cases. METHODS: In a 9-year follow-up of a prospective study of 512,726 Chinese adults, 22,216 incident IS cases, confirmed by clinical adjudication of medical records, were assigned subtypes using a modified Causative Classification System for Ischemic Stroke (CCS) (large artery atherosclerosis (LAA), small artery occlusion (SAO), cardioaortic embolism (CE), or undetermined etiology) and classified by CCS as "evident," "probable," or "possible" IS cases. For incompletely investigated IS cases where CCS yielded an undetermined etiology, an ML model was developed to predict IS subtypes from baseline risk factors and screening for cardioaortic sources of embolism. The 5-year risks of subsequent stroke and all-cause mortality (measured using cumulative incidence functions and 1 minus Kaplan-Meier estimates, respectively) for the ML-predicted IS subtypes were compared with etiologically classified IS subtypes. RESULTS: Among 7443 IS subtypes with evident or probable etiology, 66% had SAO, 32% had LAA, and 2% had CE, but proportions of SAO-to-LAA cases varied by regions in China. CE had the highest rates of subsequent stroke and mortality (43.5% and 40.7%), followed by LAA (43.2% and 17.4%) and SAO (38.1% and 11.1%), respectively. ML provided classifications for cases with undetermined etiology and incomplete clinical data (24% of all IS cases; n = 5276), with area under the curves (AUC) of 0.99 (0.99-1.00) for CE, 0.67 (0.64-0.70) for LAA, and 0.70 (0.67-0.73) for SAO for unseen cases. ML-predicted IS subtypes yielded comparable subsequent stroke and all-cause mortality rates to the etiologically classified IS subtypes. CONCLUSION: This study highlighted substantial heterogeneity in prognosis of IS subtypes and utility of ML approaches for classification of IS cases with incomplete clinical investigations.

Vascular disease in women: comparison of diagnoses in hospital episode statistics and general practice records in England.

BACKGROUND: Electronic linkage to routine administrative datasets, such as the Hospital Episode Statistics (HES) in England, is increasingly used in medical research. Relatively little is known about the reliability of HES diagnostic information for epidemiological studies. In the United Kingdom (UK), general practitioners hold comprehensive records for individuals relating to their primary, secondary and tertiary care. For a random sample of participants in a large UK cohort, we compared vascular disease diagnoses in HES and general practice records to assess agreement between the two sources. METHODS: Million Women Study participants with a HES record of hospital admission with vascular disease (ischaemic heart disease [ICD-10 codes I20-I25], cerebrovascular disease [G45, I60-I69] or venous thromboembolism [I26, I80-I82]) between April 1st 1997 and March 31st 2005 were identified. In each broad diagnostic group and in women with no such HES diagnoses, a random sample of about a thousand women was selected for study. We asked each woman's general practitioner to provide information on her history of vascular disease and this information was compared with the HES diagnosis record. RESULTS: Over 90% of study forms sent to general practitioners were returned and 88% of these contained analysable data. For the vast majority of study participants for whom information was available, diagnostic information from general practice and HES records was consistent. Overall, for 93% of women with a HES diagnosis of vascular disease, general practice records agreed with the HES diagnosis; and for 97% of women with no HES diagnosis of vascular disease, the general practitioner had no record of a diagnosis of vascular disease. For severe vascular disease, including myocardial infarction (I21-22), stroke, both overall (I60-64) and by subtype, and pulmonary embolism (I26), HES records appeared to be both reliable and complete. CONCLUSION: Hospital admission data in England provide diagnostic information for vascular disease of sufficient reliability for epidemiological analyses.