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.

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.

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.

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.

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.

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.

Development, validation and comparison of multivariable risk scores for prediction of total stroke and stroke types in Chinese adults: a prospective study of 0.5 million adults.

BACKGROUND AND PURPOSE: Low-income and middle-income countries have the greatest stroke burden, yet remain understudied. This study compared the utility of Framingham versus novel risk scores for prediction of total stroke and stroke types in Chinese adults. METHODS: China Kadoorie Biobank (CKB) is a prospective study of 512 726 adults, aged 30-79 years, recruited from 10 areas in China in 2004-2008. By 1 January 2018, 43 234 incident first stroke cases (36 310 ischaemic stroke (IS); 8865 haemorrhagic stroke (HS)) were recorded in 503 842 participants with no history of stroke at baseline. We compared the predictive utility of the Framingham Stroke Risk Profile (FSRP) with novel CKB stroke risk scores and included recalibration, refitting, stratifying by study area and addition of other risk factors. Discrimination was assessed using area under the receiver operating characteristic curve (AUC) and calibration was assessed using Greenwood-Nam-D'Agostino χ2 statistics. RESULTS: Incidence of total stroke varied fivefold by area in China. The FSRP had good discrimination for total stroke (AUC (95% CI); men: 0.78 (0.77 to 0.79), women: 0.77 (95% CI 0.76 to 0.78)), but poor calibration (χ2; men: 1,825, women: 3,053), substantially underestimating absolute risks. Recalibration reduced χ2 by >80%, but did not improve discrimination. Refitting the FSRP did not materially improve discrimination, but further improved calibration. Stratification by area improved discrimination (AUC; men: 0.82 (0.82 to 0.83); women: 0.82 (0.82 to 0.83)), but not calibration. Adding other risk factors yielded modest, but statistically significant, improvements in the AUCs. The findings for IS and HS were similar to those for total stroke. CONCLUSIONS: The FSRP reliably differentiated Chinese adults with incident stroke, but substantially underestimated the absolute risks of stroke. Novel local risk prediction equations that took account of differences in stroke incidence within China enhanced risk prediction of total stroke and major stroke pathological types.

Utility of single versus sequential measurements of risk factors for prediction of stroke in Chinese adults.

Absolute risks of stroke are typically estimated using measurements of cardiovascular disease risk factors recorded at a single visit. However, the comparative utility of single versus sequential risk factor measurements for stroke prediction is unclear. Risk factors were recorded on three separate visits on 13,753 individuals in the prospective China Kadoorie Biobank. All participants were stroke-free at baseline (2004-2008), first resurvey (2008), and second resurvey (2013-2014), and were followed-up for incident cases of first stroke in the 3 years following the second resurvey. To reflect the models currently used in clinical practice, sex-specific Cox models were developed to estimate 3-year risks of stroke using single measurements recorded at second resurvey and were retrospectively applied to risk factor data from previous visits. Temporal trends in the Cox-generated risk estimates from 2004 to 2014 were analyzed using linear mixed effects models. To assess the value of more flexible machine learning approaches and the incorporation of longitudinal data, we developed gradient boosted tree (GBT) models for 3-year prediction of stroke using both single measurements and sequential measurements of risk factor inputs. Overall, Cox-generated estimates for 3-year stroke risk increased by 0.3% per annum in men and 0.2% per annum in women, but varied substantially between individuals. The risk estimates at second resurvey were highly correlated with the annual increase of risk for each individual (men: r = 0.91, women: r = 0.89), and performance of the longitudinal GBT models was comparable with both Cox and GBT models that considered measurements from only a single visit (AUCs: 0.779-0.811 in men, 0.724-0.756 in women). These results provide support for current clinical guidelines, which recommend using risk factor measurements recorded at a single visit for stroke prediction.

Stroke risk prediction using machine learning: a prospective cohort study of 0.5 million Chinese adults.

OBJECTIVE: To compare Cox models, machine learning (ML), and ensemble models combining both approaches, for prediction of stroke risk in a prospective study of Chinese adults. MATERIALS AND METHODS: We evaluated models for stroke risk at varying intervals of follow-up (<9 years, 0-3 years, 3-6 years, 6-9 years) in 503 842 adults without prior history of stroke recruited from 10 areas in China in 2004-2008. Inputs included sociodemographic factors, diet, medical history, physical activity, and physical measurements. We compared discrimination and calibration of Cox regression, logistic regression, support vector machines, random survival forests, gradient boosted trees (GBT), and multilayer perceptrons, benchmarking performance against the 2017 Framingham Stroke Risk Profile. We then developed an ensemble approach to identify individuals at high risk of stroke (>10% predicted 9-yr stroke risk) by selectively applying either a GBT or Cox model based on individual-level characteristics. RESULTS: For 9-yr stroke risk prediction, GBT provided the best discrimination (AUROC: 0.833 in men, 0.836 in women) and calibration, with consistent results in each interval of follow-up. The ensemble approach yielded incrementally higher accuracy (men: 76%, women: 80%), specificity (men: 76%, women: 81%), and positive predictive value (men: 26%, women: 24%) compared to any of the single-model approaches. DISCUSSION AND CONCLUSION: Among several approaches, an ensemble model combining both GBT and Cox models achieved the best performance for identifying individuals at high risk of stroke in a contemporary study of Chinese adults. The results highlight the potential value of expanding the use of ML in clinical practice.

The Human Leukocyte Antigen Locus and Rheumatic Heart Disease Susceptibility in South Asians and Europeans.

Rheumatic heart disease (RHD), an autoinflammatory heart disease, was recently declared a global health priority by the World Health Organization. Here we report a genome-wide association study (GWAS) of RHD susceptibility in 1,163 South Asians (672 cases; 491 controls) recruited in India and Fiji. We analysed directly obtained and imputed genotypes, and followed-up associated loci in 1,459 Europeans (150 cases; 1,309 controls) from the UK Biobank study. We identify a novel susceptibility signal in the class III region of the human leukocyte antigen (HLA) complex in the South Asian dataset that clearly replicates in the Europeans (rs201026476; combined odds ratio 1.81, 95% confidence intervals 1.51-2.18, P = 3.48×10-10). Importantly, this signal remains despite conditioning on the lead class I and class II variants (P = 0.00033). These findings suggest the class III region is a key determinant of RHD susceptibility offering important new insight into pathogenesis while partly explaining the inconsistency of earlier reports.

BMI and Cause-Specific Hospital Admissions and Costs: The UK Biobank Cohort Study.

OBJECTIVE: To estimate the effect of BMI on cause-specific hospital admissions and costs in men and women is not well understood, and this study's aim is to address this. METHODS: For 451,320 men and women aged 40 years or older recruited into the UK Biobank, followed up for 6 years on average, this study estimated annual rates and costs (at 2016 UK prices) of hospital admissions, overall and by diagnostic category (using International Classification of Diseases, Tenth Revision chapters), in relation to BMI. RESULTS: Among those who were never smokers, a 2-kg/m2 higher BMI (above 20 kg/m2 ) was associated with a 6.2% (99% CI: 5.3% to 7.2%) higher admission rate and an 8.6% (99% CI: 7.3% to 10.0%) higher annual cost in men and with a 5.7% (99% CI: 4.9% to 6.6%) higher admission rate and an 8.4% (99% CI: 7.5% to 9.3%) higher annual cost in women. Higher BMI was associated with higher admission rates and costs for many types of health conditions, particularly for endocrine, skin, and musculoskeletal disorders in both men and women and for circulatory diseases in men. CONCLUSIONS: BMI beyond healthy weight is strongly associated with higher annual rates of hospital admission and higher costs in both men and women across a wide range of health conditions.

Association of FADS1/2 Locus Variants and Polyunsaturated Fatty Acids With Aortic Stenosis.

Importance: Aortic stenosis (AS) has no approved medical treatment. Identifying etiological pathways for AS could identify pharmacological targets. Objective: To identify novel genetic loci and pathways associated with AS. Design, Setting, and Participants: This genome-wide association study used a case-control design to evaluate 44 703 participants (3469 cases of AS) of self-reported European ancestry from the Genetic Epidemiology Research on Adult Health and Aging (GERA) cohort (from January 1, 1996, to December 31, 2015). Replication was performed in 7 other cohorts totaling 256 926 participants (5926 cases of AS), with additional analyses performed in 6942 participants from the Cohorts for Heart and Aging Research in Genomic Epidemiology (CHARGE) Consortium. Follow-up biomarker analyses with aortic valve calcium (AVC) were also performed. Data were analyzed from May 1, 2017, to December 5, 2019. Exposures: Genetic variants (615 643 variants) and polyunsaturated fatty acids (ω-6 and ω-3) measured in blood samples. Main Outcomes and Measures: Aortic stenosis and aortic valve replacement defined by electronic health records, surgical records, or echocardiography and the presence of AVC measured by computed tomography. Results: The mean (SD) age of the 44 703 GERA participants was 69.7 (8.4) years, and 22 019 (49.3%) were men. The rs174547 variant at the FADS1/2 locus was associated with AS (odds ratio [OR] per C allele, 0.88; 95% CI, 0.83-0.93; P = 3.0 × 10-6), with genome-wide significance after meta-analysis with 7 replication cohorts totaling 312 118 individuals (9395 cases of AS) (OR, 0.91; 95% CI, 0.88-0.94; P = 2.5 × 10-8). A consistent association with AVC was also observed (OR, 0.91; 95% CI, 0.83-0.99; P = .03). A higher ratio of arachidonic acid to linoleic acid was associated with AVC (OR per SD of the natural logarithm, 1.19; 95% CI, 1.09-1.30; P = 6.6 × 10-5). In mendelian randomization, increased FADS1 liver expression and arachidonic acid were associated with AS (OR per unit of normalized expression, 1.31 [95% CI, 1.17-1.48; P = 7.4 × 10-6]; OR per 5-percentage point increase in arachidonic acid for AVC, 1.23 [95% CI, 1.01-1.49; P = .04]; OR per 5-percentage point increase in arachidonic acid for AS, 1.08 [95% CI, 1.04-1.13; P = 4.1 × 10-4]). Conclusions and Relevance: Variation at the FADS1/2 locus was associated with AS and AVC. Findings from biomarker measurements and mendelian randomization appear to link ω-6 fatty acid biosynthesis to AS, which may represent a therapeutic target.

The Associations Between Seven Different Types of Physical Activity and the Incidence of Fracture at Seven Sites in Healthy Postmenopausal UK Women.

There is a paucity of information on associations between specific types of physical activity and fracture risk at different sites in otherwise healthy postmenopausal women. Therefore, we examined risk of fracture at seven different sites associated with seven different types of physical activity in the population-based prospective UK Million Women Study. A total of 371,279 postmenopausal women (mean age 59.8 years), rating their health as good or excellent and reporting participation in walking, cycling, gardening, doing housework, yoga, dance, and sports club activities, were followed for site-specific incident fracture through record linkage to national databases on day-case and overnight hospital admissions. Cox regression yielded adjusted relative risks (RRs) and, because of the large number of statistical tests done, 99% confidence intervals (CIs) for fracture at seven different sites in relation to seven different physical activities. During an average follow-up of 12 years, numbers with a first site-specific fracture were as follows: humerus (2341), forearm (1238), wrist (7358), hip (4354), femur (not neck) (617), lower leg (1184), and ankle (3629). For upper limb fractures there was significant heterogeneity across the seven activity types (test for heterogeneity p = 0.004), with gardening more than 1 hour/week associated with a lower risk (RR = 0.91; 99% CI, 0.86 to 0.96; p < 0.0001), whereas cycling more than 1 hour/week was associated with an increased risk (RR = 1.11; 99% CI, 1.00 to 1.23; p = 0.008). For fractures of the lower limb (including hip) there was no significant heterogeneity by type of activity, with significant approximately 5% to 15% reductions in risk associated with most activities, except cycling. For hip fractures, there was no significant heterogeneity by type of activity, but with significant 15% to 20% reductions in risk associated with walking for 1 hour/day and participating in yoga and sporting activities. Physical activity is a modifiable risk factor for fracture, but the effects differ between different types of activities and different fracture sites. © 2019 The Authors. Journal of Bone and Mineral Research published by American Society for Bone and Mineral Research.

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.

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.

Hospital costs in relation to body-mass index in 1·1 million women in England: a prospective cohort study.

BACKGROUND: Excess weight is associated with poor health and increased health-care costs. However, a detailed understanding of the effects of excess weight on total hospital costs and costs for different health conditions is needed. METHODS: Women in England aged 50-64 years were recruited into the prospective Million Women Study cohort in 1996-2001 through 60 NHS breast cancer screening centres. Participants were followed up and annual hospital costs and admission rates were estimated for April 1, 2006, to March 31, 2011, in relation to body-mass index (BMI) at recruitment, overall and for categories of health conditions defined by the International Classification of Diseases 10th revision chapter of the primary diagnosis at admission. Associations of BMI with hospital costs were projected to the 2013 population of women aged 55-79 years in England. FINDINGS: 1 093 866 women who provided information on height and weight, had a BMI of at least 18·5 kg/m2, and had no previous cancer at recruitment, were followed up for an average of 4·9 years from April 1, 2006 (12·3 years from recruitment), during which time 1·84 million hospital admissions were recorded. Annual hospital costs were lowest for women with a BMI of 20·0 kg/m2 to less than 22·5 kg/m2 (£567 per woman per year, 99% CI 556-577). Every 2 kg/m2 increase in BMI above 20 kg/m2 was associated with a 7·4% (7·1-7·6) increase in annual hospital costs. Excess weight was associated with increased costs for all diagnostic categories, except respiratory conditions and fractures. £662 million (14·6%) of the estimated £4·5 billion of total annual hospital costs among all women aged 55-79 years in England was attributed to excess weight (BMI ≥25 kg/m2), of which £517 million (78%) arose from hospital admissions with procedures. £258 million (39%) of the costs attributed to excess weight were due to musculoskeletal admissions, mainly for knee replacement surgeries. INTERPRETATION: Excess body weight is associated with increased hospital costs for middle-aged and older women in England across a broad range of conditions, especially knee replacement surgery and diabetes. These results provide reliable up-to-date estimates of the health-care costs of excess weight and emphasise the need for investment to tackle this public health issue. FUNDING: Cancer Research UK; Medical Research Council; National Institute for Health Research.

Body mass index and healthcare costs: a systematic literature review of individual participant data studies.

Excess weight is associated with increased total healthcare costs, but it is less well known how the associations between excess weight and costs vary across different types of healthcare service. We reviewed studies using individual participant data to estimate associations between body mass index and healthcare costs, and summarized how annual healthcare costs for overweight (body mass index 25 to <30 kg/m2 ) and obese (≥30 kg/m2 ) individuals compared with those for healthy weight individuals (18.5 to <25 kg/m2 ). EMBASE and MEDLINE were searched from January 1990 to September 2016, and 75 studies were included in the review. Of these, 34 studies presented adequate information to contribute to a quantitative summary of results. Compared with individuals at healthy weight, the median increases in mean total annual healthcare costs were 12% for overweight and 36% for obese individuals. The percentage increases in costs were highest for medications (18% for overweight and 68% for obese), followed by inpatient care (12% and 34%) and ambulatory care (4% and 26%). Percentage increases in costs associated with obesity were higher for women than men. The substantial costs associated with excess weight in different healthcare settings emphasize the need for investment to tackle this major public health problem.