Elevated blood pressure, antihypertensive medications and bone health in the population: revisiting old hypotheses and exploring future research directions

Blood pressure and bone metabolism appear to share commonalities in their physiologic regulation. Specific antihypertensive drug classes may also influence bone mineral density. However, current evidence from existing observational studies and randomised trials is insufficient to establish causal associations for blood pressure and use of blood pressure-lowering drugs with bone health outcomes, particularly with the risks of osteoporosis and fractures. The availability and access to relevant large-scale biomedical data sources as well as developments in study designs and analytical approaches provide opportunities to examine the nature of the association between blood pressure and bone health more reliably and in greater detail than has ever been possible. It is unlikely that a single source of data or study design can provide a definitive answer. However, with appropriate considerations of the strengths and limitations of the different data sources and analytical techniques, we should be able to advance our understanding of the role of raised blood pressure and its drug treatment on the risks of low bone mineral density and fractures. As elevated blood pressure is highly prevalent and blood pressure-lowering drugs are widely prescribed, even small effects of these exposures on bone health outcomes could be important at a population level.

Plasma lipids and risk of aortic valve stenosis: a Mendelian randomization study

Background

Aortic valve stenosis is commonly considered a degenerative disorder with no recommended preventive intervention, with only valve replacement surgery or catheter intervention as treatment options.

Purpose

We sought to assess the causal association between exposure to lipid levels and risk of aortic stenosis.

Methods

Causality of association was assessed using two-sample Mendelian Randomization (MR) framework through different statistical methods. MR approach uses instrumental variable analysis to mimic the randomization process that underpins causal inference in clinical trials. It takes advantage of the naturally-occurring random allocation of alleles inherited by offspring from their parents during the formation of the zygote. We retrieved summary estimations of 157 genetic variants that have been shown to be associated with plasma lipid levels in the Global Lipids Genetics Consortium that included 188,577 participants, mostly European ancestry, and genetic association with aortic stenosis as the main outcome from a total of 432,173 participants in the UK Biobank. Secondary negative control outcomes included aortic regurgitation and mitral regurgitation.

Results

The odds ratio (OR) for developing aortic stenosis per unit increase in lipid parameter was 1.52 (95% confidence interval [CI], 1.22 to 1.90; per 0.98 mmol/L) for low-density lipoprotein (LDL) cholesterol, 1.03 (95% CI, 0.80 to 1.31; per 0.41 mmol/L) for high-density lipoprotein (HDL) cholesterol, and 1.38 (95% CI 0.92 to 2.07; per 1 mmol/L) for triglycerides. There was no evidence of a causal association between any of the lipid parameters and aortic or mitral regurgitation.

Conclusion

Lifelong exposure to high LDL-cholesterol increases the risk of symptomatic aortic stenosis, suggesting that LDL-lowering treatment may be effective in its prevention.

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): British Heart Foundation

Cardiometabolic disease, comorbidities and risk of death: findings using data from large-scale electronic health records

Background

Myocardial infarction (MI), stroke and diabetes are separately associated with increased risk of mortality but it is uncertain if their combined effects are proportional, amplified or less than the expected risk of each disease individually. In addition, patients with these conditions tend to also have other long-term comorbidities. How the relationship between cardiometabolic disease and risk of death is modified by the presence of comorbidity is unclear.

Purpose

We investigated the separate and combined effects of MI, stroke and diabetes on all-cause mortality, and examined the impact of comorbidity on these associations.

Methods

We selected a patient cohort of 2,007,731 (51% women) aged ≥16 years at registration with their general practice, using large-scale UK primary care electronic health records that were linked to the national death registry. We identified patients with a recorded diagnosis of MI, stroke, diabetes or none before 2005 (baseline), and classified the patient cohort into mutually exclusive categories of their baseline disease status. For each group, we also extracted information on another major 53 long-term conditions prior to baseline. The cohort was followed until death, deregistration from the practice or censored at the end of study (31 Dec 2014). We used Cox regression, and tested for departure from additivity and multiplicativity to assess interaction.

Results

At baseline, the mean age of the cohort was 51 (SD=18) years and 7% (N=145,910) had a cardiometabolic disease. Over an average follow-up of 7 (SD=3) years, 270,036 died (mean age of death=79 years). After adjusting for baseline age and sex, the hazard ratio (HR) (95% confidence interval [CI]), relative to those without cardiometabolic disease, were as follows: diabetes=1.53 (1.51 to 1.55), MI=1.54 (1.51 to 1.56), stroke=1.87 (1.84 to 1.90), diabetes and MI=2.16 (2.09 to 2.23), MI and stroke=2.39 (2.28 to 2.49), diabetes and stroke=2.56 (2.47 to 2.65), and all three=3.17 (2.95 to 3.41). After adjusting for the 53 comorbidities, the HR (95% CI) were attenuated: diabetes=1.37 (1.35 to 1.39), MI=1.25 (1.23 to 1.27), stroke=1.49 (1.46 to 1.52), diabetes and MI=1.60 (1.55 to 1.65), MI and stroke=1.52 (1.45 to 1.59), diabetes and stroke=1.91 (1.84 to 1.98), and all three=1.77 (1.64 to 1.91). The results did not materially changed with adjustment for smoking and deprivation level. Test for interaction revealed some minor synergistic effects when cardiometabolic disease co-occurred but excess risks were lower than expected for two combined vs individual disease effects; no significant interaction was seen for all three vs individual disease effects.

Conclusion

MI, stroke and diabetes are associated with excess mortality, which was partly due to associated chronic conditions. We found no evidence that the co-occurrence of these three conditions contribute to a higher excess mortality than expected from each of them separately.

Funding Acknowledgement

Type of funding source: Public Institution(s). Main funding source(s): NIHR Oxford Biomedical Research Centre; Oxford Martin School, University of Oxford

An interpretable model for incident heart failure prediction with uncertainty estimation

Background

Forecasting incident heart failure is a critical demand for prevention. Recent research suggested the superior performance of deep learning models on the prediction tasks using electronic health records. However, even with a relatively accurate predictive performance, the major impediments to the wider use of deep learning models for clinical decision making are the difficulties of assigning a level of confidence to model predictions and the interpretability of predictions.

Purpose

We aimed to develop a deep learning framework for more accurate incident heart failure prediction, with provision of measures of uncertainty and interpretability.

Methods

We used a longitudinal linked electronic health records dataset, Clinical Practice Research Datalink, involving 788,880 patients, 8.3% of whom had an incident heart failure diagnosis. To embed the uncertainty estimation mechanism into the deep learning models, we developed a probabilistic framework based on a novel transformer deep learning model: deep Bayesian Gaussian processes (DBGP). We investigated the performance of incident heart failure prediction and uncertainty estimation for the model and validated it using an external held-out dataset. Diagnoses, medications, and age for each encounter were included as predictors. By comparing the uncertainty, we investigated the possibility of identifying the correct predictions from wrong ones to avoid potential misclassification. Using model distillation meant to mimic a well-trained complex model with simple models, we investigated the importance of associations between diagnoses, medications and heart failure with an interpretable linear regression component learned from DBGP.

Results

The DBGP achieved high precision with 0.941 as AUROC for external validation. More importantly, it showed the uncertainty information could distinguish the correct predictions from wrong ones, with significant difference (p-value with 500 samples) between distribution of uncertainties for negative predictions (3.21e-69 between true negative and false negative), and positive predictions (3.39e-22 between true positive and false positive). Utilising the distilled model, we can specify the contribution of each diagnosis and medication to heart failure prediction. For instance, Losartan/Fosinopril, Bisoprolol and Left bundle-branch block showed strong association to heart failure incidence with coefficient 0.11 (95% CI: 0.10, 0.12), 0.09 (0.08, 0.11) and 0.09 (0.07, 0.11) respectively; Peritoneal adhesions, Trochanteric bursitis and Galactorrhea showed strong disassociations with coefficient −0.07 (−0.09, −0.05), −0.07 (−0.09, −0.04) and −0.06 (−0.08, −0.04) individually.

Conclusions

Our novel probabilistic deep learning framework adds a measure of uncertainty the prediction and helps to mitigate misclassification. Model distillation provides an opportunity to interpret deep learning models and offers a data-driven perspective for risk factor analysis.

Funding Acknowledgement

Type of funding source: Public Institution(s). Main funding source(s): Oxford Martin School,University of Oxford; NIHR Oxford Biomedical Research Centre, University of Oxford

Blood pressure lowering treatment for prevention of cardiovascular events in patients with atrial fibrillation: an individual-participant data meta-analysis

Background

Randomised evidence showing that pharmacological blood pressure (BP) lowering can reduce cardiovascular risk of patients with atrial fibrillation (AF) is limited.

Purpose

This study aimed to compare the effect of BP-lowering treatment on fatal and non-fatal cardiovascular outcomes in patients with and without AF overall and by major drug classes.

Methods

We extracted individual participant data from all trials with over 1,000 person-years of follow-up that had randomly assigned patients to different classes of BP-lowering drugs, BP-lowering drugs vs placebo, or to more vs less intensive BP-lowering regimens. We investigated the effects of BP-lowering treatment on a composite endpoint of major cardiovascular events (stroke, ischaemic heart disease or heart failure) according to AF status at baseline using fixed-effect one-stage individual participant data meta-analyses based on Cox proportional hazards models stratified by trial.

Findings

Twenty-two trials were included with 188,570 patients, of whom 13,266 (7%) had AF at baseline. Patients with AF had lower BP at baseline than patients without AF (143/84 mmHg, SD 21/12mmHg) versus 155/88 mmHg, SD 21/13 mmHg, respectively). Meta-regression showed that relative risk reductions were proportional to trial-level intensity of BP lowering, both in patients with and without AF. The hazard ratio for major cardiovascular events was 0.91 in patients with AF (95% confidence interval [0.83–1.00]) and 0.91 without AF (95% confidence interval [0.88–0.93]) for each 5-mmHg reduction in systolic BP, with no difference between subgroups (p=0.91) (Figure 1). Similar patterns were observed for individual components of the composite primary outcome. In patients with AF, there was no evidence that treatment effects varied according to baseline systolic BP or use of specific drug classes.

Conclusion

This study demonstrated that BP-lowering treatment reduces the risk of major cardiovascular events in patients with AF to a similar extent to that of patients without AF, even when baseline BP is below recommended treatment thresholds. Owing to their higher absolute cardiovascular risk, treatment in patients with AF is likely to result in greater absolute risk reduction than in patients without AF. Guidelines should be updated to clearly recommend pharmacological BP lowering for prevention of cardiovascular events in patients with AF.

Figure 1. Forest plot

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): British Heart Foundation

BEHRT-HF: an interpretable transformer-based, deep learning model for prediction of incident heart failure

Background/Introduction

Predicting incident heart failure has been challenging. Deep learning models when applied to rich electronic health records (EHR) offer some theoretical advantages. However, empirical evidence for their superior performance is limited and they remain commonly uninterpretable, hampering their wider use in medical practice.

Purpose

We developed a deep learning framework for more accurate and yet interpretable prediction of incident heart failure.

Methods

We used longitudinally linked EHR from practices across England, involving 100,071 patients, 13% of whom had been diagnosed with incident heart failure during follow-up. We investigated the predictive performance of a novel transformer deep learning model, “Transformer for Heart Failure” (BEHRT-HF), and validated it using both an external held-out dataset and an internal five-fold cross-validation mechanism using area under receiver operating characteristic (AUROC) and area under the precision recall curve (AUPRC). Predictor groups included all outpatient and inpatient diagnoses within their temporal context, medications, age, and calendar year for each encounter. By treating diagnoses as anchors, we alternatively removed different modalities (ablation study) to understand the importance of individual modalities to the performance of incident heart failure prediction. Using perturbation-based techniques, we investigated the importance of associations between selected predictors and heart failure to improve model interpretability.

Results

BEHRT-HF achieved high accuracy with AUROC 0.932 and AUPRC 0.695 for external validation, and AUROC 0.933 (95% CI: 0.928, 0.938) and AUPRC 0.700 (95% CI: 0.682, 0.718) for internal validation. Compared to the state-of-the-art recurrent deep learning model, RETAIN-EX, BEHRT-HF outperformed it by 0.079 and 0.030 in terms of AUPRC and AUROC. Ablation study showed that medications were strong predictors, and calendar year was more important than age. Utilising perturbation, we identified and ranked the intensity of associations between diagnoses and heart failure. For instance, the method showed that established risk factors including myocardial infarction, atrial fibrillation and flutter, and hypertension all strongly associated with the heart failure prediction. Additionally, when population was stratified into different age groups, incident occurrence of a given disease had generally a higher contribution to heart failure prediction in younger ages than when diagnosed later in life.

Conclusions

Our state-of-the-art deep learning framework outperforms the predictive performance of existing models whilst enabling a data-driven way of exploring the relative contribution of a range of risk factors in the context of other temporal information.

Funding Acknowledgement

Type of funding source: Private grant(s) and/or Sponsorship. Main funding source(s): National Institute for Health Research, Oxford Martin School, Oxford Biomedical Research Centre

Genetic susceptibility, elevated blood pressure and risk of atrial fibrillation

Background

Whether elevated blood pressure (BP) is a modifiable risk factor for atrial fibrillation (AF) is not established.

Purpose

We tested (1) whether the association between BP and risk of AF is causal, (2) whether it varies according to individual's genetic susceptibility for AF, and (3) the extent to which specific BP-lowering drugs are expected to reduce this risk.

Methods

First, causality of association was assessed through two-sample Mendelian Randomization (MR), using data from two independent genome-wide association studies that included a total of one million European population. Second, UK Biobank individual participant data of 329,237 participants at baseline was used to study the effect of BP on AF according to genetic susceptibility of developing AF. Third, a possible treatment effect with BP-lowering drug classes on AF risk was predicted through genetic variants in druggable genes that code proteins related to the function of each drug class. Estimated drug effects were compared with effects on incident coronary heart disease, for which direct trial evidence exists.

Results

The two-sample MR analysis indicated that on average each 10-mm Hg increment in systolic BP increased the risk of AF (odds ratio [OR]: 1.23 [1.11 to 1.36]). This association was replicated in the UK biobank using individual participant data. However, in a further genetic risk-stratified analysis, there was evidence for a linear gradient in the relative effects of systolic BP on AF; while there was no conclusive evidence of an effect in those with low genetic risk, a strong effect was observed among those with high genetic susceptibility for AF (Figure). The indirect comparison of predicted treatment effects using genetic proxies for three main drug classes (angiotensin-converting enzyme inhibitors, beta-blockers and calcium channel blockers) suggested similar average effects for prevention of atrial fibrillation and coronary heart disease.

Conclusions

The association between elevated BP and higher risk of AF is likely to be causal, suggesting that BP-lowering treatment may be effective in AF prevention. However, average effects masked clinically important variations, with a more pronounced effect in individuals with high genetic susceptibility.

Figure 1

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): British Heart Foundation

Effect of blood pressure lowering treatment on the risk of atrial fibrillation: an individual-participant data meta-analysis

Background

Although observational studies have suggested an association between elevated blood pressure (BP) and increased risk of atrial fibrillation (AF), randomised evidence on the effects of pharmacological blood pressure lowering on the risk of new-onset AF remains limited.

Purpose

To investigate the effects of pharmacological BP lowering on the risk of AF overall and stratified by baseline risk of AF and by drug class.

Methods

We extracted individual participant data from trials with over 1,000 person-years of follow-up that had randomly assigned patients to different classes of BP-lowering drugs, BP-lowering drugs vs placebo, or to more vs less intensive BP-lowering regimens. We investigated the effects of BP lowering on the risk of new-onset AF using fixed-effect one-stage individual participant data meta-analyses based on Cox proportional hazards models stratified by trial.

Results

Twenty-one trials were included with a total of 194,041 patients, in whom 6,357 new-onset and 516 recurrent AF events were recorded. The hazard ratio for new-onset AF was 1.01, 95% CI [0.95–1.07] per each 5-mmHg reduction in systolic BP, and meta-regression suggested that treatment effects were similar irrespective of the intensity of systolic BP reduction. Patients were overall at low risk of AF at baseline (median 2.3%, IQR [1.2–3.4%] at 5 years), and there was no evidence of heterogeneity in treatment effects across thirds of risk and 10-mmHg strata of baseline systolic BP (Figure). There was also no clear evidence that treatment effects differed between drug classes when renin-angiotensin-aldosterone system inhibitors and calcium channel blockers were compared with placebo and/or standard treatment.

Conclusion

In a low-risk population, pharmacological BP lowering did not reduce the risk of new-onset AF. Further research is needed to understand whether the effects would be different in high-risk individuals, and to better clarify the existence of class-specific effects.

Figure 1. Forest plot

Funding Acknowledgement

Type of funding source: Foundation. Main funding source(s): British Heart Foundation

Association between comorbidities and blood pressure trajectories in patients with hypertension

Background

Our knowledge of how to better manage elevated blood pressure in presence of comorbidities is limited; in part due to exclusion or underrepresentation of multimorbid patients from major clinical trials.

Purpose

To investigate the burden and types of comorbidities in patients with hypertension, to assess how such comorbidities and other variables affect blood pressure levels over time.

Methods

The study was conducted using linked electronic health records from the UK Clinical Practice Research Datalink study from its inception on 1 January 1985 to 30 September 2015. Using linked electronic health records, we compared systolic blood pressure levels among 295,487 patients with diagnosed hypertension by type and numbers of major comorbidities from at least 5 years before to up to 10 years after hypertension diagnosis. We used a multiple landmark cohort design in order to investigate associations prospectively with time-updated information that takes advantage of the dynamic nature of electronic health records.

Results

Time-updated multivariable linear regression analyses showed that the presence of more comorbidities was independently associated with lower blood pressure during follow-up. This negative association was not specific to particular types of comorbidities; although associations were stronger in those with pre-existing cardiovascular disease. Tracking patients backwards to years prior to hypertension diagnosis revealed that the association between comorbidities and blood pressure were even more pronounced in years before hypertension diagnosis. Despite substantial declines in blood pressure in the first year after diagnosis, subsequent changes were modest, with no evidence of a more rapid decline in those with more or specific types of comorbidities.

Conclusions

Blood pressure levels at which patients were diagnosed with hypertension varied substantially and were lower when patients had more comorbidities. This early selection bias was a key determinant of long-term differences in blood pressure by comorbidity status and provides an additional explanation for the lower blood pressure in multimorbid patients.

Mean SBP (mmHg) over time

Funding Acknowledgement

Type of funding source: Public Institution(s). Main funding source(s): National Institute for Health Research Oxford Biomedical Research Centre, Rhodes Trust and Clarendon Fund

Stratified effects of blood pressure-lowering treatment on long-term blood pressure: an individual patient-level meta-analysis involving 50 randomised trials and 334,219 participants

Background

Meta-analyses of randomised controlled trials (RCT) have shown the efficacy of pharmacologic lowering of blood pressure (BP) in reducing cardiovascular disease (CVD) risk. While efficacy has been shown across important patient characteristics, meta-analysis based on aggregate data could not fully account for potential sources of variation due to individual-level characteristics. Moreover, it is unclear if any variation in treatment effects due to patient characteristics are reflected in differential effects of BP-lowering treatment on long-term BP according to these characteristics.

Purpose

We determined the effects of BP-lowering treatment on repeated measures of blood pressure, identified trial- and participant-level sources of heterogeneity, and examined consistency of these BP-lowering effects across different patient characteristics.

Methods

We conducted an individual patient-level data meta-analysis (N=50 trials) using one-stage approach. We classified trials according to trial design: drug comparison (N=28), placebo-controlled (N=21) and BP-lowering intensity (N=8) trials. We fitted mixed models with fixed treatment effects and fixed time effect, random intercepts at trial and participant level, and a random slope for time at participant level. We adjusted for age, sex and baseline BP (except when used as stratification factor). We used likelihood ratio test and Akaike information criterion to compare models.

Results

This meta-analysis included 334,219 (42% women) participants. At baseline, mean age=65 (SD=9) years, among whom 18% were current smokers, 47% had cardiovascular disease, 29% had diabetes, and 73% were previously on BP-lowering medication. Participants had an average of 8 BP measurements over 4 years of mean follow-up. For drug comparison trials, mean differences (95% confidence interval) in systolic BP (SBP) and diastolic BP (DBP) between comparison arms were 1.3 (1.2 to 1.3) mmHg and 0.5 (0.5 to 0.5) mmHg, respectively; for placebo-controlled trials, the SBP and DBP differences were 4.2 (4.0 to 4.3) mmHg and 1.9 (1.9 to 2.0) mmHg, respectively; and for BP-lowering intensity trials, the SBP and DBP differences were 8.2 (8.0 to 8.4) mmHg and 3.7 (3.6 to 3.9) mmHg, respectively. However, BP reduction differed by duration of follow-up, type of trial. In particular, for placebo-controlled and BP-intensity trials, heterogeneity in BP reductions according to patient characteristics such as baseline BP, age, sex, prior CVD, diabetes and non-randomised anti-hypertensive use were observed.

Conclusion

This study shows the role of pharmacologic agents in effectively reducing long-term BP across individuals with a wide range of characteristics. The magnitude of BP reduction varied by several patient characteristics. This might have implications for investigation and explanation of any differential effects of BP treatment on major clinical outcomes.

Funding Acknowledgement

Type of funding source: Public Institution(s). Main funding source(s): British Heart Foundation; NIHR Oxford Biomedial Research Centre

P1548Long-term past, current and usual systolic blood pressure and incident cardiovascular disease: risk prediction using large-scale, routinely recorded clinical data

Background

The impact of long-term exposure to elevated systolic blood pressure (SBP) on future cardiovascular disease (CVD) in “real-world” settings, and its relevance to risk prediction, are less investigated.

Purpose

To examine the risk of incident CVD in relation to long-term past, current, and usual SBP, and compare their predictive performance, using evidence from large-scale electronic health records (EHR).

Methods

Using data extracted from UK primary care linked EHR, we applied a landmark cohort study design, by including patients aged 40 (N≈64,000), 50 (N≈80,000) and 60 (N≈67,000) years (y) at study entry who had recorded SBP and with no prior CVD or previous antihypertensive or lipid-lowering prescriptions at baseline. We estimated past SBP (mean, time-weighted mean, and variability recorded up to 10 years prior to baseline) and usual SBP (correcting current values for past time-dependent SBP variability). We used Cox regression to estimate hazard ratio (HR), and applied Bayesian analysis within a machine learning framework in developing and validating models. To evaluate predictive performance of the models, we used discrimination (area under the curve [AUC]) and calibration metrics. The outcome was incident CVD (first hospitalisation for or death from coronary heart disease or stroke/transient ischaemic attack). Analyses were conducted separately for each age cohort.

Results

After a mean follow-up of 8 years, the numbers of patients who developed incident CVD were over 1000 (40y), 3000 (50y) and 5000 (60y). Higher past, current and usual SBP values were separately and independently associated with increased incident CVD risk. Per 20-mmHg rise in SBP, the HR (95% credible interval [CI]) for current SBP for ages 40, 50 and 60 years were 1.18 (1.08 to 1.26), 1.22 (1.18 to 1.30) and 1.22 (1.19 to 1.24); the corresponding HR were stronger in magnitude for past SBP (mean and time-weighted mean) and usual SBP (HR ranged from: 40y=1.31 to 1.41, 50y=1.39 to 1.45 and 60y=1.32 to 1.48). For each age cohort, the AUC (95% CI) for the model that included current SBP, sex, smoking, deprivation, diabetes and lipid profile in the validation sample were: 40y=0.739 (0.730 to 0.746), 50y=0.750 (0.716 to 0.810), and 60y=0.647 (0.642 to 0.658). Adding past SBP mean, time-weighted mean or variability to this model were associated with modest increases in the AUC and all models showed good calibration. Small improvements in the AUC were similarly observed when evaluating models separately for men and women within each age cohort.

Conclusion

Using multiple SBP recordings from patients' EHR showed stronger associations with incident CVD than a single SBP measurement, but their addition to multivariate risk prediction models had negligible effects on model performance.

Acknowledgement/Funding

Oxford Martin School and National Institute for Health Research Oxford Biomedical Research Centre

6129Temporal trends and patterns in cause-specific mortality and hospitalisations after incident heart failure: a longitudinal analysis of 86,000 individuals

Background

The past two decades have brought considerable improvements in heart failure care. Clinical trials have demonstrated effectiveness of several different treatments in reducing mortality and hospitalisations, and observational studies have shown that these treatments are increasingly being used in many countries. Little is known about whether these changes have been reflected in patient outcomes in routine clinical settings.

Methods

We used anonymised electronic health records that link information from primary care, secondary care, and the national death registry to investigate 86,000 individuals with newly diagnosed heart failure between 2002 and 2013 in the UK. We computed all-cause and cause-specific mortality rates and number of hospitalisations in the first year following diagnosis. We used Poisson regression models to calculate category-specific rate ratios and 95% confidence intervals, adjusting for patients' age, sex, region, socioeconomic status and 17 major comorbidities.

Findings

One year after initial heart failure diagnosis, all-cause mortality rates were high (32%) and did not change significantly over the period of study (adjusted rate ratio (RR) 2013 vs 2002: 0.94 [0.88, 1]). Overall rates masked diverging trends in cause-specific outcomes: a decline in cardiovascular mortality (RR: 0.74 [0.68, 0.81]) was offset by an increase in non-cardiovascular mortality (RR: 1.28 [1.17, 1.39]), largely due to infections and chronic respiratory conditions. Sub-group analyses further showed that overall mortality declined among patients under 80 years of age (RR 2013 vs 2002: 0.79 [0.71, 0.88]), although not in older age groups (RR 2013 vs 2002: 0.97 [0.9, 1.06]). After cardiovascular causes (43%), the major causes of death identified in 2013 were neoplasms (15%), respiratory conditions (12%), and infections (11%). Hospital admissions within a year of heart failure diagnosis were common (1.15 hospitalisations per patient-year at risk), changed little over time (RR: 0.96 [0.92, 0.99]), and were largely (60%) due to non-cardiovascular causes.

Interpretation

Despite increased use of life-saving interventions, overall mortality and hospitalisations following a new diagnosis of heart failure have changed little over the past decade. Improved prognosis among young and middle-aged patients marks an important achievement and attests of complex barriers to progress in elderly patients. The shift from cardiovascular to non-cardiovascular causes of death suggest that management of associated comorbidities might offer additional opportunities to improve patients' prognosis.

Acknowledgement/Funding

British Heart Foundation, National Institute for Health Research, UK Research and Innovation.

P5732Effects of blood pressure lowering drugs in heart failure: a systematic review and meta-analysis of randomised controlled trials

Background

Observational studies have reported a J-shaped relationship between blood pressure (BP) and all-cause and cardiovascular mortality in patients with heart failure (HF). Although decreasing BP significantly reduces the risk of fatal and non-fatal cardiovascular outcomes in the general population across a range of baseline BP categories, the extent to which those findings are applicable to HF patients and whether the relationship holds true when baseline BP is very low remain unclear. Therefore, it is yet to be established whether the observed J-shaped relationship between BP and clinical outcomes in patients with HF is causal and/or modified by antihypertensive treatment.

Purpose

We aimed to combine evidence from all HF trials that have investigated the effects of drugs with BP-lowering properties to assess (1) the extent to which such drugs reduce BP in HF, (2) the association between the net change in BP between treatment arms and cause-specific outcomes, and (3) whether treatment effects (including benefits and potential harms) vary according to baseline BP.

Methods

We conducted a systematic review and meta-analysis including randomised clinical trials of drugs with BP-lowering properties conducted in patients with chronic HF with at least 300 patient-years follow-up.

Results

We included a total of 37 trials (91,950 patients) and showed that treatment with drugs with BP-lowering properties significantly reduced SBP by 2.0 mmHg in all trials and by 2.4 mmHg in placebo-controlled trials (Figure 1). There was no evidence that BP reduction in placebo-controlled trials varied across strata of baseline BP, but there was suggestive evidence for differential effects by drug class, with renin-angiotensin-aldosterone system inhibitors reducing SBP by 3.2 mmHg (95% CI [−4.0, −2.4]), whilst BB appeared to have a neutral effect on BP. There was no evidence that the relative risk reduction afforded by treatment with BP-lowering drugs on all-cause mortality, cardiovascular mortality and HF hospitalisation was significantly different across categories of baseline BP. There was also no strong evidence for heterogeneity of treatment effect on adverse events leading to treatment discontinuation by baseline BP. Meta-regression did not show significant associations between the magnitude of BP reduction achieved in each trial and risk of those clinical outcomes.

Figure 1

Conclusions

Treatment with drugs with BP-lowering properties resulted in a small but significant decrease in SBP in patients with HF irrespective of baseline BP. There was no evidence that the effects of those drugs differed across the range of baseline SBP, thus supporting the efficacy and safety of those drugs in patients with low baseline BP. Data from published reports was insufficient to adequately investigate whether BP-dependent mechanisms contribute to the effect of BP-lowering drugs on clinical outcomes in patients with HF.

Acknowledgement/Funding

None

Comparison of regional fat measurements by dual-energy X-ray absorptiometry and conventional anthropometry and their association with markers of diabetes and cardiovascular disease risk

Background/Objectives:

Fat distribution is a strong and independent predictor of type 2 diabetes (T2D) and cardiovascular disease (CVD) and is usually determined using conventional anthropometry in epidemiological studies. Dual-energy X-ray absorptiometry (DXA) can measure total and regional adiposity more accurately. Nonetheless, whether DXA provides more precise estimates of cardiovascular risk in relation to total and regional adiposity is not known. We determined the strength of the associations between DXA- and conventional anthropometry determined fat distribution and T2D and CVD risk markers.

Subjects/Methods:

Waist (WC) and hip circumference (HC) and DXA was used to measure total and regional adiposity in 4950 (2119 men) participants aged 29–55 years from the Oxford Biobank without pre-existing T2D or CVD. Cross-sectional associations were compared between WC and HC vs. DXA-determined regional adiposity (all z-score normalised) with impaired fasting glucose, hypertriglyceridemia, hypertension and insulin resistance (IR).

Results:

Following adjustment for total adiposity, upper body adiposity measurements showed consistently increased risk of T2D and CVD risk markers except for abdominal subcutaneous fat in both sexes, and arm fat in men, which showed protective associations. Among upper adiposity depots, visceral fat mass showed stronger odds ratios (OR) ranging from 1.69 to 3.64 compared with WC 1.07–1.83. Among lower adiposity depots, HC showed modest protection for IR in both sexes (men: OR 0.80 (95% confidence interval 0.67, 0.96); women: 0.69 (0.56, 0.86)), whereas gynoid fat and in particular leg fat showed consistent and strong protective effects for all outcomes in both men and women. The differential effect of body fat distribution on CVD and T2D were more pronounced at higher levels of total adiposity.

Conclusions:

Compared with DXA, conventional anthropometry underestimates the associations of regional adiposity with T2D and CVD risk markers. After correcting for overall adiposity, greater subcutaneous fat mass in particular in the lower body is protective relative to greater android or visceral adipose tissue mass.

TaylorActive--Examining the effectiveness of web-based personally-tailored videos to increase physical activity: a randomised controlled trial protocol

BACKGROUND

Physical inactivity levels are unacceptably high and effective interventions that can increase physical activity in large populations at low cost are urgently needed. Web-based interventions that use computer-tailoring have shown to be effective, though people tend to 'skim' and 'scan' text on the Internet rather than thoroughly read it. The use of online videos is, however, popular and engaging. Therefore, the aim of this 3-group randomised controlled trial is to examine whether a web-based physical activity intervention that provides personally-tailored videos is more effective when compared with traditional personally-tailored text-based intervention and a control group.

METHODS/DESIGN

In total 510 Australians will be recruited through social media advertisements, e-mail and third party databases. Participants will be randomised to one of three groups: text-tailored, video-tailored, or control. All groups will gain access to the same web-based platform and a library containing brief physical activity articles. The text-tailored group will additionally have access to 8 sessions of personalised physical activity advice that is instantaneously generated based on responses to brief online surveys. The theory-based advice will be provided over a period of 3 months and address constructs such as self-efficacy, motivation, goal setting, intentions, social support, attitudes, barriers, outcome expectancies, relapse prevention and feedback on performance. Text-tailored participants will also be able to complete 7 action plans to help them plan what, when, where, who with, and how they will become more active. Participants in the video-tailored group will gain access to the same intervention content as those in the text-tailored group, however all sessions will be provided as personalised videos rather than text on a webpage. The control group will only gain access to the library with generic physical activity articles. The primary outcome is objectively measured physical activity. Secondary outcomes include website engagement and retention, quality of life, depression, anxiety, stress, sitting time, sleep and psychosocial correlates of physical activity. Outcomes will be measured at baseline, 3, and 9 months.

DISCUSSION

This study presents an ideal opportunity to study the effectiveness of an isolated feature within a web-based physical activity intervention and the knowledge generated from this study will help to increase intervention effectiveness.

TRIAL REGISTRATION

Australian New-Zealand Clinical Trial Registry: ACTRN12615000057583 . Registered 22 January 2015. CQUniversity Ethics Project Number: H14/07-163.

Serum sex hormone-binding globulin and testosterone in relation to cardiovascular disease risk factors in young men: a population-based study

OBJECTIVE

Reduced sex hormone-binding globulin (SHBG) concentration predicts insulin resistance and type 2 diabetes, but its association with cardiovascular disease (CVD) risk is unclear. We examined the association between SHBG and cardiovascular risk factors, independently of total testosterone (TT), in young men.

DESIGN

Observational, cross-sectional study.

SETTING

General community.

PARTICIPANTS

The study included 2716 men aged 31 years in the Northern Finland Birth Cohort in 1996 with clinical examination data and fasting blood samples.

OUTCOME VARIABLES

Blood pressure (BP), lipids and C-reactive protein (CRP) as biological CVD risk markers.

RESULTS

SHBG concentration was significantly and inversely related to systolic and diastolic BP, triglycerides and CRP, but positively to HDL cholesterol after adjusting for insulin, BMI, waist circumference, smoking, education and physical activity (all P<0.05). These linearly graded associations persisted with additional adjustment for TT. SHBG was significantly associated with total cholesterol only with adjustment for covariates and TT (P<0.05). The direction and magnitude of associations between TT and risk factors were variable, but further adjustment for insulin, adiposity and SHBG showed positive associations between TT and BP, total and LDL-cholesterol and triglycerides and an inverse association with CRP (all P<0.05), but its relation with HDL-cholesterol was no longer significant.

CONCLUSIONS

In this cohort of young adult men, higher SHBG concentration was associated with a more favourable CVD risk profile, independently of TT. SHBG concentration modified the associations of TT with CVD risk factors.

Farming environment and prevalence of atopy at age 31: prospective birth cohort study in Finland

BACKGROUND

Cross-sectional studies have shown an association between the farming environment and a decreased risk of atopic sensitization, mainly related to contact with farm animals in the childhood.

OBJECTIVE

Investigate the association of a farming environment, especially farm animal contact, during infancy, with atopic sensitization and allergic diseases at the age of 31.

METHODS

In a prospective birth cohort study, 5509 subjects born in northern Finland in 1966 were followed up at the age of 31. Prenatal exposure to the farming environment was documented before or at birth. At age 31, information on health status and childhood exposure to pets was collected by a questionnaire and skin prick tests were performed.

RESULTS

Being born to a family having farm animals decreased the risk of atopic sensitization [odds ratio (OR) 0.67; 95% confidence interval (CI) 0.56-0.80], atopic eczema ever (OR 0.77; 95% CI 0.66-0.91), doctor-diagnosed asthma ever (OR 0.74; 95% CI 0.55-1.00), allergic rhinitis at age 31 (OR 0.87; 95% CI 0.73-1.03) and allergic conjunctivitis (OR 0.86; 95% CI 0.72-1.02) at age 31. There was a suggestion that the reduced risk of allergic sensitization was particularly evident among the subjects whose mothers worked with farm animals during pregnancy, and that the reduced risk of the above diseases by farm animal exposure was largely explained by the reduced risk of atopy. Having cats and dogs in childhood revealed similar associations as farm animals with atopic sensitization.

CONCLUSION AND CLINICAL RELEVANCE

Contact with farm animals in early childhood reduces the risk of atopic sensitization, doctor-diagnosed asthma and allergic diseases at age 31.

Cardiorespiratory fitness and body mass index of 9-11-year-old English children: a serial cross-sectional study from 1998 to 2004

OBJECTIVE

To examine the changes over time in cardiorespiratory fitness and body mass index (BMI) of children.

DESIGN

Serial cross-sectional, population-based study.

SETTING

Primary schools in Liverpool, UK.

PARTICIPANTS

A total of 15,621 children (50% boys), representing 74% of eligible 9-11-year olds in the annual school cohorts between 1998/9 and 2003/4, who took part in a 20m multi-stage shuttle run test (20mMST).

MAIN OUTCOME MEASURES

Weight, height, BMI (kg/m(2)) and obesity using the International Obesity Taskforce definition.

RESULTS

Median (95% confidence interval) 20mMST score (number of runs) fell in boys from 48.9 (47.9-49.9) in 1998/9 to 38.1 (36.8-39.4) in 2003/4, and in girls from 35.8 (35.0-36.6) to 28.1 (27.2-29.1) over the same period. Fitness scores fell across all strata of BMI (P<0.001). Moreover, BMI increased over the same 6-year period even among children in fittest third of 20mMST.

CONCLUSION

In a series of uniform cross-sectional assessments of school-aged children, BMI increased whereas cardiorespiratory fitness levels decreased within a 6-year period. Even among lean children, fitness scores decreased. Public health measures to reduce obesity, such as increasing physical activity, may help raise fitness levels among all children - not just the overweight or obese.

Abdominal obesity and respiratory function in men and women in the EPIC-Norfolk Study, United Kingdom

Poor respiratory function and obesity are associated with all-cause and cardiovascular disease mortality. Obese persons may also have impaired lung function, but the mechanism is unclear. The authors investigated the relation between abdominal pattern of obesity and respiratory function in the European Prospective Investigation into Cancer and Nutrition-Norfolk (EPIC-Norfolk) cohort in Norfolk, United Kingdom. This analysis included 9,674 men and 11,876 women aged 45-79 years with no known preexisting serious illness who had complete anthropometric and respiratory function measures obtained at a health visit between 1993 and 1997. Waist:hip ratio was used to assess abdominal obesity, and forced expiratory volume in 1 second (FEV1) and forced vital capacity (FVC), obtained by spirometry, were used to assess respiratory function. Both FEV1 and FVC were linearly and inversely related across the entire range of waist:hip ratio in both men and women. This relation persisted after adjustment for age, body mass index, cigarette smoking, social class, physical activity index, prevalent bronchitis/emphysema, and prevalent asthma. The association remained significant among nonobese nonsmokers without preexisting respiratory disease. In the general adult population, abdominal fat deposition may play a role in the impairment of respiratory function among the abdominally obese.