Cardiovascular Risk Prediction Research Articles

Association of ABO blood group with risk of coronary stent thrombosis

BackgroundABO blood group's influence on cardiovascular risk, particularly in venous thromboembolism and coronary artery disease (CAD), is well-studied, with non-O blood groups associated with heightened CAD risk. However, its impact on stent thrombosis remains an unexplored area, prompting the question of whether ABO blood groups are also associated with risk of early stent thrombosis. ObjectivesThe primary objective of this study was to analyze the impact of ABO blood groups on the occurrence of early (≤30 days) stent thrombosis. MethodsThe study included 10,714 consecutive patients who underwent percutaneous coronary intervention (PCI) with implantation of drug-eluting stents (DES) at a tertiary care hospital. Among these, 78 patients (0.73 %) experienced early stent thrombosis. Propensity score matching was conducted using cardiovascular risk factors and predictors of stent thrombosis, including age, sex, diabetes, hypertension, smoking, hypercholesterinemia, and clinical presentation. ResultsThe presence of non-0 blood groups (blood groups A, B and AB; OR 1.48; 95 % CI 0.74–2.97; p = 0.27) and of A-antigen (blood groups A, AB; OR 0.93; 95 % CI 0.51–1.84; p = 0.89) was not associated early stent thrombosis, respectively. In contrast, patients with B-antigen (blood groups B, AB) were at higher risk of early stent thrombosis as compared to patients with blood group 0 (OR 2.48; 95 % CI 1.08–5.69; p = 0.019). ConclusionThe presence of blood group antigen B (blood groups B and AB) emerged as a significant factor associated with early stent thrombosis. Further investigations are warranted to elucidate the specific biological mechanisms through which ABO blood group antigens could influence stent thrombosis.

A Proteomics-Based Approach for Prediction of Different Cardiovascular Diseases and Dementia.

Many studies have explored whether individual plasma protein biomarkers improve cardiovascular disease risk prediction. We sought to investigate the use of a plasma proteomics-based approach in predicting different cardiovascular outcomes. Among 51 859 UK Biobank participants (mean age, 56.7 years; 45.5% male) without cardiovascular disease and with proteomics measurements, we examined the primary composite outcome of fatal and nonfatal coronary heart disease, stroke, or heart failure (major adverse cardiovascular events), as well as additional secondary cardiovascular outcomes. An exposome-wide association study was conducted using relative protein concentrations, adjusted for a range of classic, demographic, and lifestyle risk factors. A prediction model using only age, sex, and protein markers (protein model) was developed using a least absolute shrinkage and selection operator-regularized approach (derivation: 80% of cohort) and validated using split-sample testing (20% of cohort). Their performance was assessed by comparing calibration, net reclassification index, and c statistic with the PREVENT (Predicting Risk of CVD Events) risk score. Over a median 13.6 years of follow-up, 4857 participants experienced first major adverse cardiovascular events. After adjustment, the proteins most strongly associated with major adverse cardiovascular events included NT-proBNP (N-terminal pro B-type natriuretic peptide; hazard ratio [HR], 1.68 per SD increase), proADM (pro-adrenomedullin; HR, 1.60), GDF-15 (growth differentiation factor-15; HR, 1.47), WFDC2 (WAP four-disulfide core domain protein 2; HR, 1.46), and IGFBP4 (insulin-like growth factor-binding protein 4; HR, 1.41). In total, 222 separate proteins were predictors of all outcomes of interest in the protein model, and 86 were selected for the primary outcome specifically. In the validation cohort, compared with the PREVENT risk factor model, the protein model improved calibration, net reclassification (net reclassification index +0.09), and c statistic for major adverse cardiovascular events (+0.051). The protein model also improved the prediction of other outcomes, including ASCVD (c statistic +0.035), myocardial infarction (+0.023), stroke (+0.024), aortic stenosis (+0.015), heart failure (+0.060), abdominal aortic aneurysm (+0.024), and dementia (+0.068). Measurement of targeted protein biomarkers produced superior prediction of aggregated and disaggregated cardiovascular events. This study represents an important proof of concept for the application of targeted proteomics in predicting a range of cardiovascular outcomes.

Abstract 4138446: Ceramides Surpass LDL-Cholesterol in Predicting Cardiovascular Mortality in Diabetic and Non-Diabetic Patients

The power of cardiovascular risk prediction is biased by comorbidities (including T2DM), medical treatment, and age, which impact cholesterol levels. Consequently, when applying LDL-C, the true cardiovascular risk of such confounded patients is often underestimated. However, it is not clear whether this constraint for risk prediction may likewise apply to other lipid risk markers in particular ceramides. In this observational cohort study, we recorded cardiovascular mortality for 16 years in 1195 patients. Their median age was 67 years, all of them had a high preexisting cardiovascular risk, 51% were taking statins, and 30% had T2DM. We found a U-shaped association between cardiovascular mortality and LDL-C. We thus stratified the population according to the vertex into high (≥150mg/dL) and lower LDL-C (<150mg/dL) and also according to the T2DM status. A Cox regression analysis revealed that LDL-C failed to predict cardiovascular mortality in any subgroup. In contrast, no U-shaped but a linear association was found for ceramide-based markers. They were able to significantly predict cardiovascular mortality after stratification and even after multivariate adjustment. We thus propose that ceramide-based predictors rather than LDL-C should be applied for a more accurate cardiovascular risk prediction in diabetic and other high-risk patients.

Abstract 4121042: Novel genetic risk loci of arterial stiffness in Asian ancestry individuals

Background: Arterial stiffness serves as an independent predictor of cardiovascular disease risk. Pulse-wave velocity (PWV) is a non-invasive measurement of arterial stiffness. However, the genetic risk variants associated with arterial stiffness in the Asian ancestry population remain largely unknown. This lack of knowledge hinders the development of targeted preventive and treatment strategies. Hypothesis: Arterial stiffness is associated with genotyped SNP loci that have genetic relationships with metabolic syndrome phenotypes. Methods: For PWV, we performed genome-wide association study and gene-based genetic association study in 912 individuals from the Tongji-BGI research cardiovascular health (TORCH) cohort. We then used linkage disequilibrium score regression to estimate pairwise genetic correlations between arterial stiffness, blood pressure, lipids, and diabetes. Results: Eleven risk loci associated with arterial stiffness were discovered at genome-wide significance ( P < 5 × 10 −8 ), including three in intronic (rs978453410 mapping to PPP2R5E; rs146270927 mapping to ECI1; rs189517402 mapping to CNN2 ). UBE2U and ECI1 were identified as the PWV-associated genes by gene-based analysis. Genetic correlation analyses revealed inter-related of PWV with systolic blood pressure, mean arterial pressure, total cholesterol and diabetes. Conclusions: This research firstly identified eleven novel genetic variants associated with arterial stiffness in Asian ancestry individuals. Further research aimed at reducing arterial stiffness may benefit from the identification of SNPs and genes like ECI1, which is related to the oxidation of unsaturated fatty acids in the mitochondria.

Abstract 4137527: Evaluation of 10-Year Atherosclerotic Cardiovascular Risk Prediction Performance using the PREVENT versus Pooled Cohort Equations in a US Integrated Healthcare System

Background: The American Heart Association published the new PREVENT equations for estimating atherosclerotic cardiovascular disease (ASCVD) risk. Question: Do the PREVENT equations improve risk prediction for 10-year ASCVD compared with the pooled cohort equations (PCEs) in a large diverse population? Aim: To assess ASCVD risk prediction performance of the two PREVENT equations [PREVENT Base, PREVENT Base plus optional predictors including urine albumin-to-creatinine ratio, glycated hemoglobin, and social deprivation index (PREVENT Full)] compared with the PCEs. Methods: We included adults aged 40-75 years without a history of ASCVD or diabetes from Kaiser Permanente Southern California in 2009 and followed them through 2019. Outcome was incident ASCVD defined as myocardial infarction, fatal coronary heart disease, fatal and nonfatal stroke. We compared model discrimination (Harrell’s C), mean calibration (estimated as the ratio of predicted to observed event rates), and calibration curve among the overall population and stratified by sex and race/ethnicity. Results: Of the 559,111 adults (mean age 54, 11% Black, 32% Hispanic), 10,695 developed an ASCVD event during a median follow-up of 10 years. Harrell’s C among the overall population was 0.741 (95% CI 0.736-0.745) for PREVENT Base, 0.743 (0.738-0.748) for PREVENT Full, and 0.741 (0.736-0.746) for the PCEs ( Table ). Compared with the PCEs, both PREVENT equations improved Harrell’s C in men but not in women, and in Black adults but not in other racial/ethnic groups. Both PREVENT equations were well-calibrated (mean calibration ranged 0.83-1.37; calibration slope ranged 0.71-1.28), while the PCEs overestimated 10-year ASCVD risk (mean calibration ranged 1.80-2.20; calibration slope ranged 0.34-0.45) ( Table & Figure ). Conclusion: Compared with the PCEs, both PREVENT Base and Full equations improved calibration in predicting 10-year ASCVD risk, with minimal improvement in discrimination in men and Black adults only.

Abstract 4117834: Utility of Abdominal Aortic Calcification Assessment for the Prediction of Major Adverse Cardiovascular Events in Liver Transplant Patients

Background: Cardiovascular disease is a leading cause of postoperative morbidity and mortality following liver transplantation (LT). Presence of abdominal aortic calcification (AAC) has been linked to cardiovascular events in the general population. We investigated whether AAC on pre-transplant computed tomography (CT) predicts major adverse cardiovascular events (MACE) post-LT. Methods: Consecutive LT patients between 2010-2018 from the Victorian Liver Transplant Unit (Australia) were included. Extent of AAC was quantified in a blinded fashion, with high AAC defined as a calcium score of ≥500. MACE was defined as any recorded episodes of acute coronary syndrome (ACS), ventricular arrhythmia, decompensated heart failure, or stroke. The primary outcome was post-LT MACE at 30 days and the secondary outcome was all-cause mortality at maximal follow-up. Results: Of 461 patients undergoing LT, 350 had suitable CTs for analysis, of which 90 (25.7%) had also undergone CT coronary angiography. High AAC was identified in 98 patients (28.0%). This finding demonstrated a moderate correlation with high coronary artery calcium score (CACS) ≥400 (r=0.52, p<0.001). Fifty-eight MACE occurred in 42 patients (12.0%) within 30-days. High AAC was associated with increased risk of 30-day MACE (OR=2.68 (95%CI 1.39-5.17), p=0.004), and long-term mortality (HR=2.52 (95%CI 1.29-4.92), p=0.007) at a median follow-up of 4.0 years. In addition, high AAC predicted an increased risk of ACS (OR 3.80 (95%CI 1.18-12.27), p=0.025), heart failure (OR 3.07 (95%CI 1.21-7.80), p=0.020) and atrial fibrillation (OR 2.04 (95%CI 1.02-4.07), p=0.048), with a trend to increased prevalence of ventricular arrhythmia (OR 2.67 (95%CI 0.84-8.50), p=0.10). Following multivariate analysis, high AAC remained a strong independent predictor of MACE (OR=3.49, 95%CI 1.21-10.06, p=0.021). Addition of AAC to the revised cardiac risk index (RCRI) improved model fit for predicting MACE (net reclassification improvement of 52.7% (p<0.001)). Conclusions: This study demonstrates for the first time that high AAC on routine abdominal CT scans is associated with a 3-fold increased risk of 30-day MACE post-LT, and improves cardiovascular risk prediction compared to traditional risk scores. Quantification of AAC may offer a simple method of improving cardiovascular risk assessment and implementation of preventative strategies in liver transplant patients, using pre-existing scans without additional cost or patient risk.

Abstract 4121454: Machine-extractable Markers in Chest Radiograph to Predict Cardiovascular Risk in Screening Population

Introduction: Recent research has shown that AI is able to assess biological aging and cardiovascular disease (CVD) risk using chest radiographs. However, the lack of explainability of such deep learning algorithms hinders clinical utility and adoption. This motivates the current study which searches for and tests the use of machine extractable quantitative features in chest radiographs to predict CVD risk in population screening. Method: Chest radiograph measurements characterizing cardiomediastinal geometry, aortic calcification and tortuosity were handpicked for development of a segmentation-based feature extraction algorithm. The algorithm was applied on the PLCO lung screening dataset for analysis. The association between measurement-based imaging features, clinical characteristics (age, sex, BMI, smoking status, hypertension, diabetes, liver disease) with CVD mortality and 10-year major adverse cardiovascular events (MACE) were analysed by using proportional hazard regression, with feature selection done by LASSO. Result: Of 29,453 eligible subjects, 5693 subjects from a single study centre were used for fitting of all models. The median follow-up time was 19 years. A total of 32 imaging features were extracted and analysed. For both 10-year MACE and CVD mortality, model using imaging features, age, and sex performed similarly to model using conventional risk factors, and a deep learning chest radiograph CVD risk model. Two imaging features, mediastinal width at valve-level [HR 1.36 (1.23-1.50)] and maximal lateral displacement of descending aorta [HR 1.29 (1.18-1.42)] were found to be prognostic. To the best of our knowledge, these features have not been reported previously. Conclusion: Quantitative imaging features can predict CVD risk in chest radiograph similar to deep learning models while providing feature interpretability and explainability. Two novel imaging features prognostic of CVD risk were found and shown to be complementary to conventional risk factors.

Development and Validation of a Predictive Model for Maternal Cardiovascular Morbidity Events in Patients With Hypertensive Disorders of Pregnancy.

Hypertensive disorders of pregnancy (HDP) are a major contributor to maternal morbidity, mortality, and accelerated cardiovascular (CV) disease. Comorbid conditions are likely important predictors of CV risk in pregnant people. Currently, there is no way to predict which people with HDP are at risk of acute CV complications. We developed and validated a predictive model for all CV events and for heart failure, renal failure, and cerebrovascular events specifically after HDP. Models were created using the Premier Healthcare Database. The inclusion criteria for the model dataset were delivery with an HDP with discharge from October 1, 2015 to December 31, 2020. Machine learning methods were used to derive predictive models of CV events occurring during delivery hospitalization (Index Model) or during readmission (Readmission Model) using a training set (60%) to estimate model parameters, a validation set (20%) to tune model hyperparameters and select a final model, and a test set (20%) to evaluate final model performance. The total model cohort consisted of 553,658 deliveries with an HDP. A CV event occurred in 6501 (1.2%) of the delivery hospitalizations. Multilabel neural networks were selected for the Index Model and Readmission Model due to favorable performance compared to alternatives. This approach is designed for prediction of multiple events that share risk factors and may cooccur. The Index Model predicted all CV events with area under the receiver operating curve (AUROC) 0.878 and average precision (AP) 0.239 (cerebrovascular events: AUROC 0.941, heart failure: AUROC 0.898, and renal failure: AUROC 0.885). With a positivity threshold set to achieve ≥90% sensitivity, model specificity was 65.0%, 83.5%, 68.6%, and 65.6% for predicting all CV events, cerebrovascular events, heart failure, and renal failure, respectively. CV events within 1 year of delivery occurred in 3018 (0.6%) individuals. The Readmission Model predicted all CV events with AUROC 0.717 and AP 0.022 (renal failure: AUROC 0.748, heart failure: AUROC 0.734, and cerebrovascular events AUROC 0.698). Feature importance analysis indicated that the presence of chronic renal disease, cardiac disease, pulmonary hypertension, and preeclampsia with severe features had the greatest effect on the prediction of CV events. Among individuals with HDP, our multilabel neural network model predicted CV events at delivery admission with good classification and events within 1 year of delivery with fair classification.

Extending the horizon in cardiovascular risk prediction: A synergy of conventional predictors and spirometry data

Aim. To assess a correlation between conventional cardiovascular risk factors and unprovoked pulmonary volumes and flows.Material and methods. The subjects (n = 153, median age 43 [32; 51] years) were divided into two groups depending on age. Questionnaires, anthropometry, biochemical blood analysis, spirometry without bronchodilation test, calculation of the lower limit of normal (LLN – low limit of normal) were carried out. Using the Mann – Whitney U-test, the two groups were compared quantitatively. When analyzing four-field conjugacy tables, a comparison was performed using the Pearson criterion χ2. The direction and closeness of the correlation were estimated using Spearman’s rank correlation coefficient. A predictive model was developed using the linear regression method. The differences were considered statistically significant at p < 0.05.Results. In group I (median age 29 [26; 33] years), reliable connections were established between the volume of forced expiratory volume in one second and high-sensitivity C-reactive protein (hsCRP); modified Tiffno index (m-TI) and total cholesterol, lowdensity lipoproteins, glucose. When analyzing the LLN data, reliable associations between m-IT and glucose (p = 0.004), hsCRP (p = 0.011) were obtained. In persons over 40 years of age with a smoking index of more than 10 packs/year, a relationship was found between a decrease in LLN m-TI and LLN of forced expiratory flow between 25 and 75% of forced vital capacity and an increase in creatinine (p < 0.001) and lipoprotein (a) (p = 0.03), respectively. Conclusion. There are differences in the relationship between the function of external respiration and laboratory indicators of cardiovascular risk in people before and after 40 years of age, which makes it possible to individualize the prediction of chronic non-communicable diseases.

Optimal QT Correction Formula for Older Chinese: Guangzhou Biobank Cohort Study

Introduction: To identify the optimal QT correction formula for generating corrected QT (QTc) and cutoffs for prolonged QTc, and examine the associations with mortality and cardiovascular disease (CVD) in older Chinese. Methods: A prospective study included 24,611 Chinese aged 50+ years and without CVD at 2003–2008 from Guangzhou Biobank Cohort Study. QT interval was corrected by Bazett, Fridericia, Framingham and Hodges formulas. The slope and R2 of the QTc and heart rate regression were used to determine the optimal correction formula. The 95th percentile of QTc was used to defined prolonged QTc. Cox regression was used to examine associations of prolonged QTc with mortality and CVD. The net reclassification index was calculated to assess risk reclassification. Results: During an average follow-up of 15.3 years, 5,261 deaths and 5,539 CVD occurred. Optimal heart correction was observed for the Hodges formula, and Bazett formula performed the worst. Prolonged QTc corrected by Fridericia, Framingham and Hodges formulas had similar association strength with all-cause mortality, CVD mortality and incident CVD (especially coronary heart disease, myocardial infarction and ischemic stroke), with hazard ratios approximately being 1.25, 1.40, and 1.15, respectively. They also improved risk reclassification for all-cause mortality, CVD mortality and incident CVD by approximately 5%, 10%, and 6%, respectively. However, prolonged QTc corrected by Bazett formula was not associated with incident CVD and did not improve risk reclassification. Conclusions: Hodges formula outperformed other formulas for heart rate correction. Fridericia, Framingham, and Hodges formulas can be used for death and cardiovascular risk prediction.

Tracing the Lipid Fingerprints of Preeclampsia.

Preeclampsia (PE) is the most common pregnancy-related complication responsible for maternal mortality and morbidity. PE pathogenesis is characterized by placental dysfunction, impaired invasion of trophoblast, and defective spiral artery remodelling. Even after many years of research on PE, the etiology and pathophysiology of PE is still elusive. Our earlier studies have shown deregulated maternal and placental fatty acid and lipid metabolism to be associated with the pathogenesis of PE. Currently available lipidomics data have shown that glycerophospholipids, sphingolipid and cholesterol metabolism are mainly altered in preeclampsia. Including these five metabolites (SM C28:1, SM C30:1, LPC C19:0, LPE C20:0, propane-1,3-diol) with currently used protein biomarkers like sFlt-1/PlGF will improve PE prediction. Similarly, CE17:1 and CER(d20:1/24:1) alongwith sFlt-1/PlGF makes a better prediction of PE than sFlt-1/PlGF alone A comprehensive map of lipid profiles in early pregnancy may provide an improved understanding of disease pathogenesis and will be useful predictive biomarkers. In this article, we aimed to summarize the significance of lipid metabolism in the preeclampsia pathogenesis and altered lipidome signatures in preeclampsia. We also discuss the future scope of lipidomics in aiding early prediction of PE and future cardiovascular risk in both mother and child.

Use of anthropometric parameters, body composition and conicity index as predictors of cardiovascular risk in overweight and obese women

Introduction: Obesity is a growing global public health problem. Currently, more than 2 billion people in the world are overweight, and around 670 million of them have been diagnosed with obesity. Objective: To relate anthropometric parameters, body composition markers and the conicity index as predictors of cardiovascular risk in overweight or obese adult women. Methods: This was a cross-sectional study of 62 women. Body weight, height, waist and hip circumference were assessed. Based on these data, body mass index, waist-to-hip ratio, waist-to-height ratio and conicity index were obtained. Total and visceral fat percentages were also determined. Results: The patients were characterized by a mean age of 39.7 years (SD± 11.58), a mean BMI of 33.72 kg/m2 (SD± 4.74), WC of 100.27 cm (SD± 11.86); WC of 116.09 cm (SD± 9.53); WHR of 0.86 cm (SD± 0.07); WHtR of 0.64 cm (SD±0.07); %BF of 48% (SD± 639.11); %GV of 9% (SD± 2), and IC of 1.27 (SD± 0.09). There was a statistically significant difference between WHR and the parameters of %BF (P-value: 0.022), %GV (P-value: 0.001), and IC (P-value: 0.000). There was also a significant association between WHtR and GV percentages (P-value: 0.000) and IC (P-value: 0.001). With regard to IC, the analysis showed a significant difference for %GV (P-value: 0.002) and WC (P-value: 0.000). Conclusion: The use of anthropometric indicators such as waist-to-height ratio, waist-to-hip ratio and conicity index is an easy-to-apply and low-cost tools that can be used to assess and diagnose cardiovascular risk in overweight women.

Icosapent ethyl by baseline small dense low-density lipoprotein cholesterol: an analysis of REDUCE-IT

Abstract Background Small dense low-density lipoprotein cholesterol (sdLDL-C) is associated with cardiovascular (CV) risk. Purpose We assessed if baseline sdLDL-C modified the effect of icosapent ethyl among patients in REDUCE-IT. Methods REDUCE-IT randomized patients to icosapent ethyl vs placebo. In this post hoc analysis, patients ≥45 years with CV disease or ≥50 years with diabetes and CV risk factors were included. Patients were required to have triglycerides of 135-499 mg/dL and low-density lipoprotein cholesterol (LDL-C) of 41-100 mg/dL. Exclusion criteria included severe heart failure, acute severe liver disease, or hemoglobin A1c >10%. Patients were stratified by baseline calculated sdLDL-C. A threshold of 46 mg/dL was chosen because sdLDL-C greater than this cutoff has predicted CV risk despite well-controlled LDL-C. The primary outcome included a composite of CV death, nonfatal myocardial infarction (MI), nonfatal stroke, coronary revascularization, or unstable angina. The key secondary composite outcome included CV death, nonfatal MI, or nonfatal stroke. Outcomes were assessed with Cox proportional hazard models and interaction analyses were conducted to assess heterogeneity in treatment effect by baseline sdLDL-C. Results REDUCE-IT included 8179 patients, with 8157 (99.7%) patients having sdLDL-C data. Among these patients, 7698 (94.4%) had sdLDL-C <46 mg/dL and 459 (5.6%) had sdLDL-C ≥46 mg/dL. Median sdLDL-C was 34.2 mg/dL (interquartile range [IQR]: 30.3, 38.4 mg/dL) in the lower sdLDL-C group and 48.4 mg/dL (IQR: 47.1, 50.6 mg/dL) in the higher sdLDL-C group (Table). The placebo group had a greater event rate in the higher sdLDL-C group compared to the lower sdLDL-C group (72.0 vs 56.4 events per 1000 p-y), though event rates in the icosapent ethyl group were similar by sdLDL-C group (44.1 vs 43.3 events per 1000 p-y, respectively). Icosapent ethyl reduced the rate of the primary outcome compared with placebo (17.2% vs 22.0%; hazard ratio [HR]: 0.75 [95% CI: 0.68, 0.83]; P<0.0001). The number needed to treat [NNT] to avoid one primary endpoint event was 21. In the lower sdLDL-C group, icosapent ethyl decreased rates of the primary outcome (43.3 events per 1000 patient-years [p-y]) compared with the placebo group (56.4 events per 1000 p-y) (17.3% vs 21.7%; HR: 0.77 [95% CI: 0.69, 0.85]; NNT: 22). Similarly, in the higher sdLDL-C group, icosapent ethyl decreased the rate of the primary outcome (44.1 events per 1000 p-y) compared with the placebo group (72.0 events per 1000 p-y) (16.6% vs 26.4%; HR: 0.58 [95% CI: 0.38, 0.88]; NNT: 10). There was no significant interaction between treatment benefit and baseline sdLDL-C for the primary (Pinteraction = 0.22) and key secondary outcome (Pinteraction = 0.82). Findings were overall similar for the other secondary outcomes (Figure). Conclusion Icosapent ethyl reduced CV outcomes in patients with high CV risk and elevated triglycerides irrespective of low or high sdLDL-C.

The added value of deep learning to submaximal exercise electrocardiogram for the 10-year prediction of major adverse cardiovascular and cerebrovascular events

Abstract Background The added value of exercise ECG data to traditional risk factors in predicting cardiovascular events remains unclear. Deep learning of ECG signals has demonstrated state-of-the-art performance in detecting subtle abnormalities indicative of cardiovascular disease. We investigated whether deep learning analysis of exercise ECGs could improve the prediction of major cardiovascular and cerebrovascular events (MACCE) with respect to established risk models in a large population-based cohort. Purpose To assess the added value of submaximal exercise ECG measurements to established cardiovascular risk prediction models. Methods We obtained ECG recordings from 41,076 UK Biobank participants without cardiovascular disease (CVD) who underwent a submaximal exercise ECG test. We obtained ECG recordings from 41,076 UK Biobank participants without cardiovascular disease (CVD) who underwent a submaximal exercise ECG test. We created two deep neural network ECG risk scores: one derived from conventional ECG parameters, measured at rest, peak exercise and late-stage recovery, and another based on the complete ECG (Q-ECG). We assessed the added value of conventional ECG parameters and Q-ECG risk scores to the UK's current prediction algorithm (QRISK3). We estimate the association between ECG parameters and MACCE, using Cox Proportional hazard models, following adjustment for traditional risk factors. All models were internally validated via 5-fold cross-validation and 1000 bootstrap iterations. Predictive performance was evaluated using Harrel's C-index, Net Reclassification Index (NRI) and net benefit. Findings: Incident MACCE was reported in 4,082 (9.9%) individuals in the study population and 3,463(9.7%) individuals with valid ECG parameters over a median follow-up period of 12.5 years. We found combined conventional ECG and Q-ECG scores were independently associated with MACCE, following adjustment for multiple testing: adjusted hazard ratio [HZ] = 1.76 (95% CI:1.63-1.91); HZ = 1.14 (95% CI:1.10-1.18), respectively, per standard deviation increase. Both conventional ECG parameters and Q-ECG were predictive of MACCE, independent of clinical risk factors, C-index = 0.64 (95%CI 0.63-0.65); net benefit = 0.09(95% CI 0.07 - 0.11) and C-index = 0.56 (95% CI 0.55-0.57); net benefit = 0.07(95% CI 0.05 - 0.09). ECG measurement's modestly improved model discrimination over the bassline QRISK3 risk score when combined with QRISK3 risk factors for conventional markers and Q-ECG score, respectively; ΔC-index 0.03 (95% CI: 0.02 – 0.04) and ΔC-index 0.03 (95% CI: 0.02 – 0.03); However, we observed no significant improvements in classification at the current recommended threshold of 10%. Conclusion In individuals without a history of prior cardiovascular disease, ECG measures independently predict the risk of MACCE. When combined with QRISK3, neural-network-derived ECG risk scores marginally improve cardiovascular risk prediction over QRISK3.

Cardiovascular risk prediction using metabolomic biomarkers and polygenic risk scores: a cohort study and modelling analyses

Abstract Background/Introduction Metabolomic biomarker scores and polygenic risk scores (PRS) have shown promise for improving cardiovascular disease (CVD) prediction, but have not yet been evaluated in the context of current prediction models (SCORE2) and ESC recommendations for 10-year prediction of fatal and non-fatal CVD. Purpose To assess whether metabolomics biomarkers and PRSs can improve 10-year CVD risk prediction when added to SCORE2, and whether improvements are meaningful at ESC 2021 recommended risk thresholds for treatment consideration. Methods Metabolomics biomarker scores were constructed and compared to PRS and SCORE2 in 170,000 UK Biobank participants (5,096 incident CVD cases) aged 40–69 without previous CVD, diabetes, or lipid-lowering treatment. Improvement in risk discrimination when added to SCORE2 was assessed using Harrel’s C-index. Improvement in risk stratification following ESC guideline risk thresholds was assessed using categorical net reclassification. Population modelling was subsequently applied to estimate the impact on CVD prevention if applied at scale. Results Risk discrimination provided by SCORE2 (C-index: 0.718) was similarly improved by addition of metabolomic biomarker scores (ΔC-index: 0.011 [0.009–0.014]) and PRSs (ΔC-index 0.009; [0.007–0.012]). Addition of both metabolomic biomarker scores and PRSs to SCORE2 yielded the largest improvement risk discrimination, with ΔC-index 0.019 (0.016–0.022). Concomitant improvements in risk stratification were observed in categorical net reclassification index, with net case reclassification of 13.04% (11.67–14.41%). Modelling metabolic biomarker scores and PRSs for targeted risk-reclassification increased the number of CVD events prevented per 100,000 screened from 201 to 370 (ΔCVDprevented: 170 [158–182]) while essentially maintaining the number of statins prescribed per CVD event prevented. Conclusions Combining metabolomic biomarker scores and PRSs with SCORE2 moderately enhances prediction of first-onset CVD, and could have substantial population health benefit if applied at scale.

Predictors of cardiovascular disease risk: findings from the National Survey of Health and Development Cohort

Abstract Background Cardiovascular disease (CVD) remains a significant cause of mortality and morbidity within the UK. CVDs and all-cause mortality can be predicted with varying levels of certainty with different models. Purpose This study aimed to explore the incremental effect of risk factors and respective biomarkers on CVD events risk estimation models. Methods The National Survey for Health and Development (NSHD) birth cohort study, including 2547 women 2815 men, was used to model relationships between conventional and emerging risk factors and cardiometabolic and vascular outcomes, including myocardial infarction and stroke between 1999 and 2009. Logistic regression or XGBoost (eXtreme Gradient Boosting) models were used to predict the outcomes (myocardial infarction and strokes). Model fit was assessed by comparing predicted and known incident events in a two-by-two error matrix for binary outcomes. Results In a model including both sex and smoking alcohol intake (β = -0.029g; p < 0.05), body mass index (BMI) (β = 0.095 kg/m2; p < 0.05), glycated haemoglobin (HbA1c) (β = 0.033 mmol/mol; p < 0.01) and total cholesterol/high density lipoprotein (TC/HDL) ratio (β = 0.459 mg/dL; p < 0.0001) waist-to-hip ratio (β = 5.632cm; p = 0.05) and waist circumference (β = 0.031cm; p < 0.01) were significant predictors of myocardial infarction. A lower risk was observed among those with higher average alcohol intake, whereas higher risk being observed among those with higher BMI, HbA1c levels, total cholesterol/HDL ratio, waist-to-hip ratio and waist circumference. For stroke, exercising 5+ times (β = -1.47; p < 0.01), systolic blood pressure (SBP) (β = 0.040mmHg; p < 0.001), diastolic blood pressure (DBP) (β = 0.027mmHg; p < 0.001), pulse pressure (β = 0.033mmHg; p < 0.01), triglycerides (β = 0.388; p < 0.05) and total cholesterol/HDL ratio (β = 0.419; p < 0.01) were all significant predictors of stroke. We found that higher values of SBP, DBP, pulse pressure, triglycerides and total cholesterol/HDL ratio were associated with higher risk of stroke, whereas higher levels of physical activity were associated with a lower risk. Additionally, by combining BMI, HbA1c and total cholesterol/HDL ratio to the baseline model it provided the greatest F1 score of all models (0.123) and the greatest precision (6.7%), with a recall of 72.4%. With XGBoost we observed a reduced F1 score from 0.123 to 0.079, although adding information on alcohol intake and waist to hip ratio to this model, as previous analysis indicated great improvements in prediction, the F1 score obtained using the XGBoost increased from to 0.079 to 0.103. Conclusion This research shows that clinical and laboratory measurements play a potentially powerful role in predicting health outcomes when using evidence-based risk calculators. Using large cohorts allows for real-life long-term associations to be explored and quantified with a higher level of certainty for replicability across other populations.

Interpretable machine learning model for predicting long-term cardiovascular mortality risk in chronic kidney disease patients using SHAP explainers

Abstract Background Patients with chronic kidney disease (CKD) exhibited a pronounced burden of cardiac mortality. However, early prediction of cardiovascular mortality risk in CKD patients remains an unmet medical need. Purpose This study aimed to evaluate the performance of machine learning (ML) algorithms and identify the most influential factors in predicting long-term cardiac mortality among CKD patients. Methods Six ML models were developed, with the most effective one chosen for predicting and categorising patients into high-risk and low-risk groups based on the maximal Youden’s index. Differences in survival rates were assessed using the log-rank test on Kaplan-Meier curves. Cox regression analysis was employed to investigate the relationship between ML-predicted risk scores and mortality. Furthermore, the SHapley Additive exPlanations (SHAP) method was implemented to provide personalised explanations for model decisions. Results The auto-encoder (AE) model achieved the highest AUC of 0.94, with a sensitivity of 80.95% and specificity of 96.34%. Compared to individuals in low-risk group, those in high-risk group exhibited a significantly elevated risk of cardiovascular mortality (adjusted hazard ratio [HR]: 48.24; P <0.001). According to the SHAP model, age, C-reactive protein, blood urea nitrogen and hypertension were identified as the four most influential factors in the AE model. Conclusions Our study successfully developed an AE machine learning model for predicting 10-year cardiovascular mortality among CKD patients. The combination of ML model and SHAP aids doctors in understanding key features in the model, improving their insight into decision-making related to disease severity assessment, thus facilitating early preventive strategies in clinical practice.

The added value of device measured physical activity to the prediction of incident cardiovascular disease.

Abstract Background Conventional risk prediction models for Major Adverse Cardio/Cerebrovascular Events (MACCE) primarily rely on non-modifiable factors such as age, sex, and physiological measurements, including LDL/HDL cholesterol and blood pressure readings, that are prone to variations across medical practices and are often unavailable in many parts of the world. In this study, we investigate the potential of inexpensive wearable measurements of physical activity to enhance MACCE prediction when physiological measurements are unavailable. Purpose To assess the added value of physical activity measurements, collected via wrist-worn accelerometers, to established cardiovascular risk prediction models. Additionally, we seek to investigate whether physical activity measurements can substitute for common physiological measurements in established cardiovascular risk models. Methods We obtained accelerometer-measured physical activity over 7 days between 2013 and 2015 from 69,898 UK Biobank (UKB) participants without a prior history of MACCE, defined as death or hospitalisation due to (1) myocardial infarction, (2) heart failure, or (3) Stroke/transient ischaemic attack. We assessed the added value of daily step counts, sleep duration, and a deep learning-derived activity health score (Fig 1), calculated from complete 7-day accelerometer recordings, to established clinical risk models (SCORE2, QRISK3 and AHA) before and after excluding measurements of cholesterol and systolic blood pressure (SBP). The primary outcome was the first recorded MACCE within 6 years. Risk scores were developed in men and women using Cox proportional hazards models. We assessed predictive performance using Harrel’s C-index, net reclassification index (NRI), and net benefit at the recommended treatment threshold for each model. Results Of the 69,898 UKB participants in the study, 3,386 (4.8%) experienced a MACCE over a 6-year follow-up period. In place of HDL ratio and SBP, the addition of deep-learned activity health scores modestly improved performance in the best clinical baseline, QRISK3 for Female participants ΔC-index women:0.003 (95% CI 0.002-0.004); Δnet benefit 0.02(0.01-0.03); NRI 0.04, (0.02-0.07), and outperformed manually extracted steps and sleep measurements. We report no significant improvements for men. We find a greater increase in model performance when both HDL ratio and SBP, plus activity health scores, are included in the model. ΔC-index women:0.008, (0.007-0.009); men: 0.003 (0.002-0.004) (Fig 2). Conclusion Our findings indicate that in the absence of cholesterol and systolic blood pressure, the addition of device-measured physical activity modestly improves the performance of clinical risk scores among individuals without prior cardiovascular disease (CVD). These findings could further refine intervention strategies for targeted prevention of CVD, particularly in settings where physiological measurements may be unavailable.

Composition of non-HDL cholesterol and residual cardiovascular risk in secondary prevention

Abstract Background Non-high-density lipoprotein cholesterol (non-HDL-C) is associated with atherosclerotic cardiovascular disease (ASCVD) and is a recommended secondary prevention treatment target. However, the composition of non-HDL-C, that is, the relative contribution of remnant cholesterol (remnant-C) and low-density-lipoprotein cholesterol (LDL-C), is not considered for cardiovascular risk prediction. Purpose To examine ASCVD risk by composition of non-HDL-C. Methods Based on the Western Denmark Heart Registry, we established a cohort of patients with ischemic heart disease, including patients with acute coronary syndrome or stable angina pectoris, undergoing coronary angiography for evaluation of coronary artery disease between 2011 and 2020. A complete lipid panel within 1 year after coronary angiography on statin treatment was required. Patients with triglycerides >4.5 mmol/L were excluded. Exposures were remnant-C, LDL-C, non-HDL-C, and composition of non-HDL-C. The latter was defined as remnant-C of non-HDL-C, and LDL-C of non-HDL-C, in percentages. Cox regression analyses obtained adjusted hazard ratios (aHR) for myocardial infarction, ASCVD (myocardial infarction, ischemic stroke, and cardiovascular death), and all-cause death occurring from 1 year after coronary angiography to end of follow-up November 1, 2022. Results The secondary prevention cohort included 42,341 statin-treated patients (67.7% were men; median [IQR] age 66 [58-73] years). During median 4.1 years of follow-up, 3,846 patients developed ASCVD. Per 1 mmol/L increase in remnant-C, LDL-C, and non-HDL-C, the aHR for ASCVD was 1.41 (95% CI 1.28-1.55), 1.12 (95% CI 1.08-1.17), and 1.17 (95% CI 1.13-1.21), respectively (Figure 1). The relative contribution of remnant-C and LDL-C was median (IQR) 24.6% (18.5-33.1) and 75.4% (66.9-81.5) of non-HDL-C, respectively. When remnant-C constituted a larger percentage of non-HDL-C (and LDL-C a lower percentage), the aHR for ASCVD showed a linear increase in risk whereas the opposite trend was observed for LDL-C (Figure 2). Thus, per 10% increase in remnant-C of non-HDL-C, the aHR of ASCVD was 1.06 (95% CI 1.03-1.09) compared to 0.94 (95% CI 0.91-0.97) per 10% increase in LDL-C of non-HDL-C. Results were comparable for myocardial infarction and all-cause death, and by age, sex, and other clinically relevant subgroups. Conclusions The composition of non-HDL-C is strongly associated with ASCVD risk, with the risk escalating as remnant-C constitutes a larger percentage, and LDL-C a smaller percentage, of non-HDL-C. These findings challenge current guidelines and demonstrate the value of considering composition of non-HDL-C in clinical practice for more accurate risk assessment.Figure 1Figure 2

The role of coronary artery calcium score and european score2 in primary prevention of cardiovascular disease

Abstract Background There is limited evidence about the role of using coronary artery calcium (CAC) score in conjunction with the European SCORE2 (systematic coronary risk evaluation) for the identification of individuals at higher risk of cardiovascular disease. Purpose To evaluate the relationship between CAC score and the European SCORE2 for cardiovascular risk stratification in an asymptomatic population in primary prevention. Methods This is a preliminary analysis from a single-centre, prospective, randomised 1:1, controlled trial assessing the use of CAC score vs standard preventive strategies in asymptomatic individuals undergoing primary cardiovascular prevention. The 10-year risk of cardiovascular events was assessed with either SCORE2 (ages 45-69) or SCORE2-OP (ages 70-89). Subjects were classified as 1) low to moderate risk, 2) high risk, and 3) very high risk according to SCORE2/SCORE2-OP. CAC score was assessed by ECG gated low-dose computed tomography with no contrast administration. A CAC score of zero was considered to reclassify risk one level downward in high or very-high risk patients. CAC score ≥ 100 was used to reclassify risk one level upward in patients with low to moderate risk and a CAC score > 1000 to reclassify risk upward in subjects with high risk. Results A total of 1,252 participants: 627 in the CAC group and 625 in the control group, median age 55 yrs, 25th-75th 50-61 yrs, with availability of SCORE2 and SCORE2-OP were included in this analysis. SCORE2/SCORE 2-OP was similar in the CAC and control groups. Among the subjects in the CAC group, there was a highly significant difference in SCORE2/SCORE 2-OP between participants with CAC score = zero and those with CAC > zero (median 5.6, 25th-75th 3.6-8.55 vs 2.9, 1.8-4.6, p< 0.001) as well as between those with CAC ≥ 100 and those with CAC score < 100 [median 6.8, 25th-75th (10.3-4.5) vs 3.3 (2.0-5.3), p<0.001]. The prevalence of CAC score = zero was 75% in the low to moderate risk group, 47% in the high risk, and 20% in the very high risk, p<0.001. By contrast the prevalence of CAC score ≥ 100 was 7% in the low to moderate risk group, 24% in the high risk, and 45% in the very high risk, p< 0.001. The proportion of patients classified as low to moderate risk group was 72% with CAC score vs 59% without CAC score, high risk 23% with CAC score vs 35% without CAC score, very high risk 5% vs 6% respectively, p<0.001. Using CAC score risk reclassification occurred in 48% downward and 2% upward in originally classified high risk subjects, 7% upward in originally classified low to moderate risk, and 20% downward in originally classified very high risk subjects. Conclusions CAC score may lead to risk reclassification in a substantial proportion of individuals, in particular among those originally classified as high risk by the SCORE2/SCORE2-OP. The cardiovascular risk prediction accuracy of adding CAC score to SCORE2/SCORE2-OP awaits validation at long-term follow-up.