Traditional Clinical Risk Factors Research Articles

Left atrial function for predicting atrial fibrillation: a machine learning approach

Abstract Background Common risk scores for atrial fibrillation (AF) are based on clinical risk factors, and little is known about the contribution of left atrial (LA) function parameters to risk assessment. Purpose To determine whether a machine learning-based model that includes LA function parameters outperforms a commonly used risk scores, the CHARGE-AF, in predicting AF. Methods Participants in the Atherosclerosis Risk in Communities (ARIC) Study (a community-based cohort study in the USA) who attended visit 5 (2011-2013) and without a history of AF were included in this analysis. Incident AF was ascertained from ECG during study visits, hospitalization discharge codes, and death certificates. Random survival forest methodology was used to build a prediction model for AF; candidate variables included traditional clinical risk factors and measures of LA structure and function. Dataset was randomly split into a training (75%) and testing (25%) cohort. For each variable, importance was measured based on the reduction in Harrell’s c-statistic if the variable was randomly permuted prior to obtaining predicted values. Variables significantly associated with AF were included in the final model. Model calibration was tested using the Greenwood-Nam-D'Agostino (GND) test. Discrimination ability of our model was assessed in the testing cohort using Harrell’s c-statistic and compared with that of the CHARGE-AF model. Results 4,909 participants (59% female, 22% Black, mean age 75 years) were included. During a median follow-up of 7.0 years (1st-3rd quartile: 5.4-7.8), 650 participants developed AF. Variables significantly associated with incident AF were race, triglycerides, systolic blood pressure, coronary heart disease, age, LA maximal volume index, NT-proBNP, and LA reservoir strain (Figure 1). LA reservoir strain was the variable most strongly associated with AF. The final model showed good discrimination (c-statistic (95% CI): 0.734 (0.698-0.768)). The GND test did not detect statistically significant miscalibration (GND p=0.08). Compared with our model, CHARGE-AF had lower discrimination ability (c-statistic (95% CI): 0.635 (0.594-0.677)) and the difference was statistically significant (difference in c-statistic (95% CI): 0.099 (0.043-0.152)). Conclusion A machine learning-based model with few variables including LA function outperforms a commonly used risk model in predicting AF. Incorporation of LA functional assessment may help inform clinicians regarding individual risk of AF.

Plasma Lipidomic Profiles Improve upon Traditional Risk Factors for the Prediction of Arterial Stiffness Among Patients with Type 2 Diabetes Mellitum: A Randomized, Placebo-Controlled Trial

Background: Exercise or vitamin D intervention can reduce the risk of arterial stiffness; however, the underlying mechanisms of lipid metabolism remain unexplored. To examine the effects of a 12-week moderate and vigorous exercise program (65–80% maximal heart rate, 60 min/time, 2~3 times/week) with or without vitamin D supplementation (1000 IU/day) on the reduction in arterial stiffness and further explore whether the effects of interventions could be associated with the basal lipidome among patients with Type 2 diabetes mellitum (T2DM). Method: 61 patients with T2DM were randomly assigned to the following groups: control (CON, n = 15), exercise (EX, n = 14), vitamin D (VD, n = 16), and exercise + vitamin D (EX + VD, n = 16). Arterial stiffness risk factors (ankle–brachial index (ABI); brachial–ankle pulse wave velocity (baPWV), systolic blood pressure (SBP), and diastolic blood pressure (DBP)) were evaluated before and after the intervention. The plasma lipidome was determined using ultra-performance liquid chromatography coupled with tandem mass spectrometry. Machine learning was applied to establish prediction models for the responsiveness to arterial stiffness. Result: Vitamin D supplementation could inhibit the decrease in the ankle–brachial index (mean ± SD: EX + VD and VD, −0.001 ± 0.058; EX + CON, −0.047 ± −0.089; p = 0.03). We observed high inter-individual variability in the arterial stiffness risk factors in response to the interventions. We also found that optimally selecting the lipid predictors at baseline, such as SM d44:6, LPE 18:2, and Hex2Cer 29:0, could enhance the predictive power by 100% for arm SBP changes in the exercise group. Basal levels of Cer (33:1) and GM3 (44:4) could enhance the predictive power by 100% for changes in baPWV in the vitamin D group. Conclusions: A 12-week vitamin D supplementation was beneficial in preventing arterial stiffness. Compared with traditional clinical risk factors, specific lipids at baseline could significantly improve the ability to predict intervention-induced changes in the reduction of arterial stiffness.

The Non-Targeted Lipidomic-Based Classifier Reveals Two Candidate Biomarkers for Ischemic Stroke in Hypertensive Individuals.

Traditional clinical risk factors are insufficient to estimate the residual risk of large-vessel ischemic stroke. Non-targeted lipidomic techniques provide an opportunity to evaluate these risks. Plasma samples were collected from 113 hypertensive individuals, including 55 individuals at high risk of ischemic stroke and 58 matched individuals, in a prospective nested case-control cohort. To identify dysregulated lipid metabolites, we conducted multivariate and univariate analyses. A classifier based on a cross-validated procedure was employed to select the optimal combination of lipid species and their ratios. We identified 23 dysregulated lipid species in patients with and without ischemic stroke, including 16 (69.6%) up-regulated and 7 (30.4%) down-regulated lipid species. Through internal cross-validation, the optimal combination of two lipid features (phosphatidylcholine 34:2 and triglyceride 18:1/18:1/22:1 / phosphatidylcholine 34:2, referred to as ischemic stroke-related 2 lipid features - IS2LP) was selected, leading to a more precise prediction probability for ischemic stroke within 3.9 years. In the comparison of different risk factors, the traditional risk score, the IS2LP risk score, and the combination of the traditional risk score with IS2LP yield AUC values of 0.613(95% CI:0.509-0.717), 0.833(95% CI:0.755-0.911), and 0.843(95% CI:0.777-0.916), respectively. The combination of the traditional risk score and IS2LP exhibited significantly improved discriminative performance, with an integrated discrimination improvement (IDI) of 0.31 (p<0.001) and a continuous net reclassification improvement (NRI) of 1.06 (p < 0.001) compared to the traditional risk score. We identified new lipidomic biomarkers associated with the futural event of large-vessel ischemic stroke. These lipid species could serve as potential blood biomarkers for assessing the residual risk of ischemic stroke in hypertensive individuals.

Cardiovascular disease risk: it is complicated, but race and ethnicity are key, a Bayesian network analysis.

Cardiovascular diseases are the leading cause of morbidity and mortality in the United States. Despite the complexity of cardiovascular disease etiology, we do not fully comprehend the interactions between non-modifiable factors (e.g., age, sex, and race) and modifiable risk factors (e.g., health behaviors and occupational exposures). We examined proximal and distal drivers of cardiovascular disease and elucidated the interactions between modifiable and non-modifiable risk factors. We used a machine learning approach on four cohorts (2005-2012) of the National Health and Nutrition Examination Survey data to examine the effects of risk factors on cardiovascular risk quantified by the Framingham Risk Score (FRS) and the Pooled Cohort Equations (PCE). We estimated a network of risk factors, computed their strength centrality, closeness, and betweenness centrality, and computed a Bayesian network embodied in a directed acyclic graph. In addition to traditional factors such as body mass index and physical activity, race and ethnicity and exposure to heavy metals are the most adjacent drivers of PCE. In addition to the factors directly affecting PCE, sleep complaints had an immediate adverse effect on FRS. Exposure to heavy metals is the link between race and ethnicity and FRS. Heavy metal exposures and race/ethnicity have similar proximal effects on cardiovascular disease risk as traditional clinical and lifestyle risk factors, such as physical activity and body mass. Our findings support the inclusion of diverse racial and ethnic groups in all cardiovascular research and the consideration of the social environment in clinical decision-making.

Abstract PO4-07-03: Machine learning breast cancer risk prediction using sequential past mammograms - a pilot study

Abstract Background and Aim Some studies have reported the use of mammogram AI deep learning algorithms to accurately predict the risk of future breast cancer development in women. In these studies, random and unrelated mammograms were independently used for training of the AI model. We propose a novel deep learning system using sequential past mammograms instead of random studies and aimed to determine if it can further improve risk prediction in a pilot study. Methods Pre-training of the backbone of a Swin-transformer machine learning model was performed by applying region-of-interest lesional masks of benign and malignant on a training set of 1644 mammograms derived from the open-access Curated Breast Imaging Subset of Digital Database for Screening Mammography dataset. To further develop the model, deep learning training and validation were performed on the Cohort of Screen-age Women (CSAW) open-access screening mammogram set which was derived from the long term data of 873 women who developed breast cancer for the first time and 7850 non-cancer women. We then compared the diagnostic accuracy of a model using sequential past mammograms for training against the standard, non-sequential model. For the sequential model, imaging segmentation of the potential cancer site on mammography was performed on all past, pre-cancer mammograms obtained within 60 to 2555 days of cancer diagnosis in the pre-processing stage for breast cancer cases. AI training was then carried out by combining the pre-cancer mammograms belonging to the same woman in a temporal sequence using a transformer encoder. The learning rate started from 1e-6 and decayed 10 times every 10 epochs. For the non-sequential model, we removed the temporal encoder and used random, unrelated mammogram studies for training. Validation was performed on a dataset consisting of 75 pre-cancer mammograms from breast cancer cases and 69 mammograms from non-cancer cases. Results Sensitivity, specificity and overall accuracy on the validation set were 77.3% (58/75), 68.1% (47/69) and (105/144) 72.9% respectively for the sequential model compared to 69.3% (52/75), 69.6% (48/69) and 69.4% (100/144) respectively for the non-sequential model. Conclusion The study showed that deep learning breast cancer risk prediction can be further improved by using sequential past mammograms instead of random mammograms for AI model training. This can also potentially enhance other AI risk prediction models that employ combined mammogram and traditional breast cancer clinical risk factors. A larger validation study will be useful. Contact information: email Lester Leong at lester.leong.c.h@singhealth.com.sg ROC curve for breast cancer risk prediction using AI deep learning training with sequential past mammograms. Citation Format: Lester Leong, Mingjie Xu, Weimin Huang, Erli Zhang, Engracia Loh, Sze Yiun Teo, Geok Hoon Lim, Veronique Kiak Mien Tan, Yirong Sim, Fredrik Strand, Ryan Tan. Machine learning breast cancer risk prediction using sequential past mammograms - a pilot study [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO4-07-03.

Educational attainment, severity and short-term prognosis of intracerebral haemorrhage

BackgroundEducational attainment is a critical social determinant of health that impacts the risk and severity of incident ischaemic stroke, but less is known of its impact on intracerebral haemorrhage (ICH)....

A predictive model for patent ductus arteriosus seven days postpartum in preterm infants: an ultrasound-based assessment of ductus arteriosus intimal thickness within 24 h after birth.

To develop a predictive model for patent ductus arteriosus (PDA) in preterm infants at seven days postpartum. The model employs ultrasound measurements of the ductus arteriosus (DA) intimal thickness (IT) obtained within 24 h after birth. One hundred and five preterm infants with gestational ages ranging from 27.0 to 36.7 weeks admitted within 24 h following birth were prospectively enrolled. Echocardiographic assessments were performed to measure DA IT within 24 h after birth, and DA status was evaluated through echocardiography on the seventh day postpartum. Potential predictors were considered, including traditional clinical risk factors, M-mode ultrasound parameters, lumen diameter of the DA (LD), and DA flow metrics. A final prediction model was formulated through bidirectional stepwise regression analysis and subsequently subjected to internal validation. The model's discriminative ability, calibration, and clinical applicability were also assessed. The final predictive model included birth weight, application of mechanical ventilation, left ventricular end-diastolic diameter (LVEDd), LD, and the logarithm of IT (logIT). The receiver operating characteristic (ROC) curve for the model, predicated on logIT, exhibited excellent discriminative power with an area under the curve (AUC) of 0.985 (95% CI: 0.966-1.000), sensitivity of 1.000, and specificity of 0.909. Moreover, the model demonstrated robust calibration and goodness-of-fit (χ2 value = 0.560, p > 0.05), as well as strong reproducibility (accuracy: 0.935, Kappa: 0.773), as evidenced by 10-fold cross-validation. A decision curve analysis confirmed the model's broad clinical utility. Our study successfully establishes a predictive model for PDA in preterm infants at seven days postpartum, leveraging the measurement of DA IT. This model enables identifying, within the first 24 h of life, infants who are likely to benefit from timely DA closure, thereby informing treatment decisions.

A Second Trimester Prediction Algorithm for Early-Onset Hypertensive Disorders of Pregnancy Occurrence and Severity Based on Soluble fms-like Tyrosine Kinase 1 (sFlt-1)/Placental Growth Factor (PlGF) Ratio and Uterine Doppler Ultrasound in Women at Risk.

Hypertensive disorders of pregnancy (HDPs) represent a significant source of severe maternal and fetal morbidity. Screening strategies relying on traditional medical history and clinical risk factors have traditionally shown relatively modest performance, mainly in the prediction of preeclampsia, displaying a sensitivity of 37% for the early-onset form and 29% for the late-onset form. The development of more accurate predictive and diagnostic models of preeclampsia in the early stages of pregnancy represents a matter of high priority. The aim of the present paper is to create an effective second trimester prediction algorithm of early-onset HDP occurrence and severity, by combining the following two biochemical markers: a soluble fms-like tyrosine kinase 1 (sFlt-1)/placental growth factor (PlGF) ratio and uterine artery Doppler ultrasound parameters, namely the pulsatility index (PI) and the resistivity index (RI), in a population of high-risk pregnant women, initially assessed through traditional risk factors. A prospective single-center observational longitudinal study was conducted, in which 100 women with singleton pregnancy and traditional clinical and medical history risk factors for preeclampsia were enrolled at 24 weeks of gestation. Shortly after study enrollment, all women had their sFlt-1 and PlGF levels and mean uterine artery PI and RI determined. All pregnancies were followed up until delivery. Receiver operating characteristic (ROC) analysis established algorithms based on cutoffs for the prediction of the later development of preeclampsia: PI 1.25 (96.15% sensitivity, 86.49% specificity), RI 0.62 (84.6% sensitivity, 89.2% specificity) and sFlt-1/PlGF ratio 59.55 (100% sensitivity, 89.2% specificity). The sFlt-1/PlGF ratio was the best predictor for preeclampsia, as it displayed the highest area under the curve (AUC) of 0.973. The prediction algorithm for the severe form of preeclampsia, complicated by fetal growth restriction leading to preterm birth, antepartum fetal demise or acute fetal distress with a cerebro-placental ratio of <one consisted of the following cutoffs: PI 1.44 (93.75% sensitivity, 95.24% specificity), RI 0.69 (87.5% sensitivity, 100% specificity) and sFlt-1/PlGF ratio 102.74 (93.75% sensitivity, 95.2% specificity). These algorithms may significantly enhance the prediction accuracy of preeclampsia compared to traditional risk factors. The combination of the sFlt-1/PlGF ratio with mean uterine PI and RI in particular displayed an improved performance in the prediction of severe preeclampsia with the above-mentioned complications, compared to the biochemical markers or uterine Doppler parameters used alone. Therefore, HDP screening strategies should increasingly focus on implementing such algorithms for women who are initially regarded as high risk based on traditional risk factors, in order to properly diagnose HDP and properly limit or manage the later maternal and fetal complications.

Predicting heart failure outcomes by integrating breath-by-breath measurements from cardiopulmonary exercise testing and clinical data through a deep learning survival neural network.

Mathematical models previously developed to predict outcomes in patients with heart failure (HF) generally have limited performance and have yet to integrate complex data derived from cardiopulmonary exercise testing (CPET), including breath-by-breath data. We aimed to develop and validate a time-to-event prediction model using a deep learning framework using the DeepSurv algorithm to predict outcomes of HF. Inception cohort of 2490 adult patients with high-risk cardiac conditions or HF underwent CPET with breath-by-breath measurements. Potential predictive features included known clinical indicators, standard summary statistics from CPETs, and mathematical features extracted from the breath-by-breath time series of 13 measurements. The primary outcome was a composite of death, heart transplant, or mechanical circulatory support treated as a time-to-event outcomes. Predictive features ranked as most important included many of the features engineered from the breath-by-breath data in addition to traditional clinical risk factors. The prediction model showed excellent performance in predicting the composite outcome with an area under the curve of 0.93 in the training and 0.87 in the validation data sets. Both the predicted vs. actual freedom from the composite outcome and the calibration of the prediction model were excellent. Model performance remained stable in multiple subgroups of patients. Using a combined deep learning and survival algorithm, integrating breath-by-breath data from CPETs resulted in improved predictive accuracy for long-term (up to 10 years) outcomes in HF. DeepSurv opens the door for future prediction models that are both highly performing and can more fully use the large and complex quantity of data generated during the care of patients with HF.

Pooled prevalence of lymphopenia in all-cause hospitalisations and association with infection: a systematic review and meta-analysis

BackgroundLymphopenia is defined as a decrease below normal value (often 1.0 x 109 cells/L) of blood circulating lymphocyte count. In the general population, lymphopenia is associated with an increased risk of hospitalisation secondary to infection, independent of traditional clinical risk factors. In hospital, lymphopenia is associated with increased risk of healthcare-associated infection and mortality. By summarising lymphopenia’s prevalence and impact on clinical outcomes, we can identify an at-risk population and inform future studies of immune dysfunction following severe illness.MethodsPeer-reviewed search strategy was performed on three databases. Primary objective was to summarise the pooled prevalence of lymphopenia. Primary outcome was infection including pre-existing lymphopenia as a risk factor for admission with infection and as an in-hospital risk factor for healthcare-associated infection. Secondary outcomes were length of stay and mortality. Mortality data extracted included in-hospital, 28/30-day (‘early’), and 90-day/1-year (‘late’) mortality. Meta-analysis was carried out using random-effects models for each outcome measure. Heterogeneity was assessed using I2 statistic. Joanna Briggs Institute checklist for cohort studies was used to assess risk of bias. The protocol was published on PROSPERO.ResultsFifteen observational studies were included. The pooled prevalence of lymphopenia in all-cause hospitalisations was 38% (CI 0.34-0.42, I2= 97%, p< 0.01). Lymphopenia was not associated with an infection diagnosis at hospital admission and healthcare associated infection (RR 1.03; 95% CI 0.26-3.99, p=0.97, I2 = 55% and RR 1.31; 95% CI 0.78-2.20, p=0.31, I2=97%, respectively), but was associated with septic shock (RR 2.72; 95% CI 1.02-7.21, p=0.04, I2 =98%). Lymphopenia was associated with higher in-hospital mortality and higher ‘early’ mortality rates (RR 2.44; 95% CI 1.71-3.47, p < 0.00001, I2 = 89% and RR 2.05; 95% CI 1.64-2.56, p < 0.00001, I2 = 29%, respectively). Lymphopenia was associated with higher ‘late’ mortality (RR 1.59; 1.33-1.90, p < 0.00001, I2 = 0%).ConclusionsThis meta-analysis demonstrates the high prevalence of lymphopenia across all-cause hospitalisations and associated increased risk of septic shock, early and late mortality. Lymphopenia is a readily available marker that may identify immune dysfunctional patients. Greater understanding of immune trajectories following survival may provide insights into longer-term poor clinical outcomes.

Abstract 14503: Chemokine Ligand 18 (CCL18) is a Long-Term Predictor of All-Cause Mortality, but Did Not Predict CVD Events

Introduction: Chemokines mediate leukocyte recruitment and activation. Chemokine ligand 18 (CCL18) is mainly expressed by a broad range of cells such as monocytes/macrophages and dendritic cells. It is highly expressed in chronic inflammatory diseases, and locally in atherosclerotic plaques, particularly at sites of reduced stability. Circulating CCL18 levels are transiently elevated in patients admitted with unstable angina pectoris, but data on myocardial infarction (MI) are scarce. Aim: To assess the utility of CCL18 as a prognostic marker of CVD in patients hospitalized for chest pain of suspected coronary origin. Methods: The population consisted of 871 consecutive patients (ClinicalTrials.gov Identifier: NCT00521976). Stepwise Cox regression models, applying continuous log e -transformed values, were fitted for each of the biomarkers with all-cause mortality and cardiac death within 24-months and all-cause mortality within median 7 years (y) as the dependent variables, adding an analysis of 1) all-cause mortality or MI, and 2) all-cause mortality or MI or stroke at 7 y follow up (FU). The hazard ratio (HR) (95% CI) was assessed in a univariate and multivariable model, adjusting for traditional clinical cardiovascular risk factors, adding BNP, hsCRP, and TnT. Results: Plasma samples from 849 patients were available. By 24 months FU, 138 (15.8%) patients had died; 86 were cardiac deaths. The univariate analysis showed a positive, significant association between CCL18 and total death [HR 1.55 (95% 1.30-1.83), p=0.000], and for cardiac death [HR 1.32 (95% 1.06-1.64), p=0.013]. Applying multivariable analysis, the associations were no longer significant. At 7 y FU, a total of 332 (38.1%) patients had died. After multivariable adjustment, CLL18 was independently associated with all-cause mortality [HR 1.14 (95% CI, 1.01-1.29), p=0.030], but not for MI (n=203) or stroke (n=55). The positive association between CCL18 and all-cause mortality was lost when combined with CVD events; all-cause mortality or MI (n=404, p=0.75), and all-cause mortality or MI or stroke (n=422, p=0.79). Conclusions: CCL18 predicts long-term all-cause mortality, but had no independent prognostic bearing on short-term cardiac death and CVD events during long-term FU.

Abstract 14288: α1-Antichymotrypsin Complex (SERPINA3) is an Independent Predictor of Long-Term All-Cause Mortality, but Did Not Predict CVD Events

Introduction: SERPINA3 is an acute phase protein triggered by inflammation. It inhibits neutrophil cathepsin which cleaves protease-activated receptor 4 (PAR-4) via a thrombin-independent pathway, and mast cell chymase which mediates the production of angiotensin-II via an ACE-independent pathway. SERPINA3 is upregulated after an acute myocardial infarction (MI), but data on its long-term prognostic value in MI patients are scarce. Aim: To assess the utility of SERPINA3 as a prognostic marker in patients hospitalized for chest pain of suspected coronary origin. Methods: A total of 871 consecutive patients were included (ClinicalTrials.gov Identifier: NCT00521976). Stepwise Cox regression models, applying continuous log e -transformed values, were fitted for each of the biomarkers with all-cause mortality and cardiac death within median 24-months (mo) and all-cause mortality within median 7 years (y) as the dependent variables, adding an analysis of 1) all-cause mortality or MI, and 2) all-cause mortality or MI or stroke at 7 y follow up (FU). The hazard ratio (HR) (95% CI) was assessed in a univariate and multivariable model, adjusting for traditional clinical cardiovascular risk factors, adding BNP, hsCRP, TnT. Results: Plasma samples from 847 patients were available. At 24 mo FU, 138 (15.8%) patients had died, of which 86 were cardiac deaths. The univariate analysis showed a positive, significant association between SERPINA3 and total death [HR 1.42 (95% 1.19-1.68), p =0.000], but was not significant for cardiac death. Associations after multivariable adjustment were statistically non-significant. At 7 y FU, 332 (38.1%) patients had died. After adjustment, SERPINA3 was independently associated with all-cause mortality [HR 1.14 (95% CI, 1.02-1.28), p=0.022], but not for MI (n=203), and stroke (n=55), respectively, and remained essentially unaffected for all-cause mortality or MI combined (n=404) [HR 1.12 (95% CI, 1.01-1.24), p=0.031], slightly weakened when stroke also was added (n=422) [HR 1.10 (95% CI, 1.00-1.22), p=0.060]. Conclusions: SERPINA3 predicts long-term all-cause mortality, but had no prognostic bearing on short-term cardiac death. Introducing non-fatal CVD events during long-term FU did not improve its prognostic value.

Abstract 13405: Cardiac Computed Tomography Variables for Prediction of New-Onset Heart Failure: Multi-Ethnic Study of Atherosclerosis

Introduction: Heart failure (HF) prediction models utilize clinical variables. The purpose of this study was to determine if addition of cardiac computed tomography (CT) variables could improve discrimination of HF prediction scores. Methods: Participants of the Multi-Ethnic Study of Atherosclerosis (MESA), a cohort between 45 to 84 years old and free of clinical cardiovascular disease at baseline, with complete clinical data and cardiac CT were included to evaluate new-onset HF during the follow-up period. Cardiac CT variables analyzed included left ventricular size index (LVSi) and calcification of the coronary arteries (CAC), aortic valve (AVC), mitral valve (MVC), and thoracic aorta (TAC). We evaluated if CT variables improved the Pooled Cohort equations to Prevent HF (PCP-HF) score (an internally and externally validated model of 33,010 patients predicting 10-year risk of new-onset HF). Results: Among 6,667 MESA participants (52.7% female), 426 events of new-onset HF occurred during the follow-up period. After the Cox model was adjusted for cardiac CT and PCP-HF score variables, CAC (HR 1.10; 95% CI 1.05-1.15; p&lt;0.001), AVC (HR 1.08; 95% CI 1.03-1.13; p=0.002), MVC (HR 1.06; 95% CI 1.01-1.12; p=0.015), and LVSi (HR 1.14; 95% CI 1.10-1.18; p&lt;0.001) were all significantly associated with new-onset HF. After using time-dependent ROC and stratifying by sex, AUC discrimination of new-onset HF with the PCP-HF score was improved by the inclusion of cardiac CT variables. In females, the AUC for 10-year risk of new-onset HF improved from 0.80 to 0.86 with the addition of cardiac CT variables. In males, the AUC for 10-year risk of new-onset HF improved from 0.77 to 0.84 with the addition of cardiac CT variables. Conclusions: In addition to traditional clinical risk factors, cardiac CT variables provide anatomic risk factors that improve the prediction of new-onset HF. Coronary and valvular calcification may signal poorly controlled clinical risk factors that are often associated with HF. LVSi may reflect left ventricular remodeling associated with HF. More studies are needed to further elucidate the relationship between cardiac CT anatomic risk factors and HF.

Development and validation of a simple model to predict functionally significant coronary artery disease in Chinese populations: A two-center retrospective study

ObjectivesThis study sought to derive and validate a simple model combining traditional clinical risk factors with biomarkers and imaging indicators easily obtained from routine preoperative examinations to predict functionally significant coronary artery disease (CAD) in Chinese populations. MethodsWe developed five models from a derivation cohort of 320 patients retrospective collected. In the derivation cohort, we assessed each model discrimination using the area under the receiver operating characteristic curve (AUC), reclassification using the integrated discrimination improvement (IDI) and net reclassification improvement (NRI), calibration using the Hosmer-Lemeshow test, and clinical benefit using decision curve analysis (DCA) to derive the optimal model. The optimal model was internally validated by bootstrapping, and external validation was performed in another cohort including 96 patients. ResultsThe optimal model including 5 predictors (age, sex, hyperlipidemia, hs-cTnI and LVEF) achieved an AUC of 0.807 with positive NRI and IDI in the derivation cohort. Moreover, the Hosmer-Lemeshow test showed a good fit, and the DCA demonstrated good clinical net benefit. The C-statistic calculated by bootstrapping internal validation was 0.798, and the calibration curve showed adequate calibration (Brier score = 0.179). In the external validation cohort, the optimal model performance was acceptable (AUC = 0.704; Brier score = 0.20). Finally, a nomogram based on this model was constructed to facilitate its use in clinical practice. ConclusionsA simple model combined clinical risk factors with hs-cTnI and LVEF improving the prediction of functionally significant CAD in Chinese populations. This attractive model may be a choice for clinicians to risk stratification for CAD.

Metabolomic Profiling of Cholesterol Efflux Capacity in a Multiethnic Population: Insights From MESA.

Impaired cholesterol efflux capacity (CEC) is a novel lipid metabolism trait associated with atherosclerotic cardiovascular disease. Mechanisms underlying CEC variation are unknown. We evaluated associations of circulating metabolites with CEC to advance understanding of metabolic pathways involved in cholesterol efflux regulation. Participants enrolled in the MESA (Multi-Ethnic Study of Atherosclerosis) who underwent nuclear magnetic resonance metabolome profiling and CEC measurement (N=3543) at baseline were included. Metabolite associations with CEC were evaluated using standard linear regression analyses. Repeated ElasticNet and multilayer perceptron regression were used to assess metabolite profile predictive performance for CEC. Features important for CEC prediction were identified using Shapley Additive Explanations values. Greater CEC was significantly associated with metabolite clusters composed of the largest-sized particle subclasses of VLDL (very-low-density lipoprotein) and HDL (high-density lipoprotein), as well as their constituent apo A1, apo A2, phospholipid, and cholesterol components (β=0.072-0.081; P<0.001). Metabolite profiles had poor accuracy for predicting in vitro CEC in linear and nonlinear analyses (R2<0.02; Spearman ρ<0.18). The most important feature for CEC prediction was race, with Black participants having significantly lower CEC compared with other races. We identified independent associations among CEC, the largest-sized particle subclasses of VLDL and HDL, and their constituent apolipoproteins and lipids. A large proportion of variation in CEC remained unexplained by metabolites and traditional clinical risk factors, supporting further investigation into genomic, proteomic, and phospholipidomic determinants of CEC.

Development of a biomarker panel for assessing cardiovascular risk in diabetic patients with chronic limb-threatening ischemia (CLTI): a prospective study

BackgroundLower-extremity endovascular revascularization (LER) is often required for diabetic patients with chronic limb threatening ischemia (CLTI). During the post-revascularization period patients may unpredictably experience major adverse cardiac events (MACE) and major adverse limb events (MALE). Several families of cytokines are involved in the inflammatory process that underlies the progression of atherosclerosis. According to current evidence, we have identified a panel of possible biomarkers related with the risk of developing MACE and MALE after LER. The aim was to study the relationship between a panel of biomarkers - Interleukin-1 (IL-1) and 6 (IL-6), C-Reactive Protein (CRP), Tumor Necrosis Factor-α (TNF-α), High-Mobility Group Box-1 (HMGB-1), Osteoprotegerin (OPG), Sortilin and Omentin-1- at baseline, with cardiovascular outcomes (MACE and MALE) after LER in diabetic patients with CLTI.MethodsIn this prospective non-randomized study, 264 diabetic patients with CLTI undergoing endovascular revascularization were enrolled. Serum levels of each biomarker were collected before revascularization and outcomes’ incidence was evaluated after 1, 3, 6 and 12 months.ResultsDuring the follow-up period, 42 cases of MACE and 81 cases of MALE occurred. There was a linear association for each biomarker at baseline and incident MACE and MALE, except Omentin-1 levels that were inversely related to the presence of MACE or MALE. After adjusting for traditional cardiovascular risk factors, the association between each biomarker baseline level and outcomes remained significant in multivariable analysis. Receiver operating characteristics (ROC) models were constructed using traditional clinical and laboratory risk factors and the inclusion of biomarkers significantly improved the prediction of incident events.ConclusionsElevated IL-1, IL-6, CRP, TNF-α, HMGB-1, OPG and Sortilin levels and low Omentin-1 levels at baseline correlate with worse vascular outcomes in diabetic patients with CLTI undergoing LER. Assessment of the inflammatory state with this panel of biomarkers may support physicians to identify a subset of patients more susceptible to the procedure failure and to develop cardiovascular adverse events after LER.

LEVELS OF THE ADIPOKINE CHEMERIN HELP TO PREDICT PREECLAMPSIA AND ITS COMPLICATIONS

Objective: The adipokine chemerin is increased in preeclampsia (PE). Here, we evaluated its potential as a biomarker to predict PE or fetal complications. Design and method: We measured chemerin in 523 women with suspected or confirmed PE, as a secondary analysis of a prospective cohort study that evaluated the use of the sFlt-1 (soluble Fms-like tyrosine kinase-1)/PlGF (placental growth factor) ratio. The concordance (c)-statistic was used to evaluate the predictive value of chemerin or the sFlt-1/PlGF ratio on top of the traditional clinical risk factors (TCF) model involving gestational age, parity, proteinuria, and mean arterial pressure (MAP). Results: At the time of biomarker measurement, women had a median gestational age of 35.6 weeks (range 31.7-38.0). Serum chemerin was elevated (P&lt;0.001) in women with confirmed PE (n = 152) or gestational hypertension (n = 95) vs. women without hypertensive disease of pregnancy (n = 276; 252±129 and 236±113 vs. 198±83 ng/mL). Moreover, serum chemerin levels were higher (P&lt;0.001) in pregnancies with fetal growth restriction or fetal death (n = 78) vs. those without fetal complications (n = 361; 289±172 vs. 210±95 ng/mL). To predict PE, the TCF model yielded a C-index of 0.90, while adding chemerin or the sFlt-1/PlGF ratio to the model raised this to 0.95, with no further increase occurring when adding both at the same time. To predict fetal complications, the C-index was 0.76 for the TCF model only. This increased to 0.78 and 0.80 when adding chemerin and the ratio, respectively, and to 0.82 when adding both. Conclusions: Chemerin levels are elevated in women with PE and gestational hypertension, and chemerin levels might be used to predict PE and fetal complications.

Predictive value of single-nucleotide polymorphism signature for nephrolithiasis recurrence: a 5-year prospective study.

Genetic variations are linked to kidney stone formation. However, the association of single nucleotide polymorphism (SNPs) and stone recurrence has not been well studied. This study aims to identify genetic variants associated with kidney stone recurrences and to construct a predictive nomogram model using SNPs and clinical features to predict the recurrence risk of kidney stones. We genotyped 49 SNPs in 1001 patients who received surgical stone removal between Jan 1 and Dec 31 of 2012. All patients were confirmed stone-free by CT scan and then received follow-up at least 5 years. SNP associations with stone recurrence were analyzed by Cox proportion hazard model. A predictive nomogram model using SNPs and clinical features to predict the recurrence risk of kidney stones was developed by use of LASSO Cox regression. The recurrence rate at 3, 5, 7 years were 46.8%, 71.2%, and 78.4%, respectively. 5 SNPs were identified that had association with kidney stone recurrence risk. We used computer-generated random numbers to assign 500 of these patients to the training cohort and 501 patients to the validation cohort. A nomogram that combined the 14-SNPs-based classifier with the clinical risk factors was constructed. The areas under the curve (AUCs) at 3, 5 and 7 years of this nomogram was 0.645, 0.723, and 0.75 in training cohort, and was 0.631, 0.708, and 0.727 in validation cohort, respectively. Results show that the nomogram presented a higher predictive accuracy than those of the SNP classifier or clinical factors alone. SNPs are significantly associated with kidney stone recurrence and should add prognostic value to the traditional clinical risk factors used to assess the kidney stone recurrence. A nomogram using clinical and genetic variables to predict kidney stone recurrence has revealed its potential in the future as an assessment tool during the follow-up of kidney stone patients.

Cardiovascular Disease Risk Assessment Using Traditional Risk Factors and Polygenic Risk Scores in the Million Veteran Program

Primary prevention of atherosclerotic cardiovascular disease (ASCVD) relies on risk stratification. Genome-wide polygenic risk scores (PRSs) are proposed to improve ASCVD risk estimation. To determine whether genome-wide PRSs for coronary artery disease (CAD) and acute ischemic stroke improve ASCVD risk estimation with traditional clinical risk factors in an ancestrally diverse midlife population. This was a prognostic analysis of incident events in a retrospectively defined longitudinal cohort conducted from January 1, 2011, to December 31, 2018. Included in the study were adults free of ASCVD and statin naive at baseline from the Million Veteran Program (MVP), a mega biobank with genetic, survey, and electronic health record data from a large US health care system. Data were analyzed from March 15, 2021, to January 5, 2023. PRSs for CAD and ischemic stroke derived from cohorts of largely European descent and risk factors, including age, sex, systolic blood pressure, total cholesterol, high-density lipoprotein (HDL) cholesterol, smoking, and diabetes status. Incident nonfatal myocardial infarction (MI), ischemic stroke, ASCVD death, and composite ASCVD events. A total of 79 151 participants (mean [SD] age, 57.8 [13.7] years; 68 503 male [86.5%]) were included in the study. The cohort included participants from the following harmonized genetic ancestry and race and ethnicity categories: 18 505 non-Hispanic Black (23.4%), 6785 Hispanic (8.6%), and 53 861 non-Hispanic White (68.0%) with a median (5th-95th percentile) follow-up of 4.3 (0.7-6.9) years. From 2011 to 2018, 3186 MIs (4.0%), 1933 ischemic strokes (2.4%), 867 ASCVD deaths (1.1%), and 5485 composite ASCVD events (6.9%) were observed. CAD PRS was associated with incident MI in non-Hispanic Black (hazard ratio [HR], 1.10; 95% CI, 1.02-1.19), Hispanic (HR, 1.26; 95% CI, 1.09-1.46), and non-Hispanic White (HR, 1.23; 95% CI, 1.18-1.29) participants. Stroke PRS was associated with incident stroke in non-Hispanic White participants (HR, 1.15; 95% CI, 1.08-1.21). A combined CAD plus stroke PRS was associated with ASCVD deaths among non-Hispanic Black (HR, 1.19; 95% CI, 1.03-1.17) and non-Hispanic (HR, 1.11; 95% CI, 1.03-1.21) participants. The combined PRS was also associated with composite ASCVD across all ancestry groups but greater among non-Hispanic White (HR, 1.20; 95% CI, 1.16-1.24) than non-Hispanic Black (HR, 1.11; 95% CI, 1.05-1.17) and Hispanic (HR, 1.12; 95% CI, 1.00-1.25) participants. Net reclassification improvement from adding PRS to a traditional risk model was modest for the intermediate risk group for composite CVD among men (5-year risk >3.75%, 0.38%; 95% CI, 0.07%-0.68%), among women, (6.79%; 95% CI, 3.01%-10.58%), for age older than 55 years (0.25%; 95% CI, 0.03%-0.47%), and for ages 40 to 55 years (1.61%; 95% CI, -0.07% to 3.30%). Study results suggest that PRSs derived predominantly in European samples were statistically significantly associated with ASCVD in the multiancestry midlife and older-age MVP cohort. Overall, modest improvement in discrimination metrics were observed with addition of PRSs to traditional risk factors with greater magnitude in women and younger age groups.

Polygenic Risk Score, Lifestyles, and Type 2 Diabetes Risk: A Prospective Chinese Cohort Study

The aim of this study was to generate a polygenic risk score (PRS) for type 2 diabetes (T2D) and test whether it could be used in identifying high-risk individuals for lifestyle intervention in a Chinese cohort. We genotyped 80 genetic variants among 5024 participants without non-communicable diseases at baseline in the Wuxi Non-Communicable Diseases cohort (Wuxi NCDs cohort). During the follow-up period of 14 years, 440 cases of T2D were newly diagnosed. Using Cox regression, we found that the PRS of 46 SNPs identified by the East Asians was relevant to the future T2D. Participants with a high PRS (top quintile) had a two-fold higher risk of T2D than the bottom quintile (hazard ratio: 2.06, 95% confidence interval: 1.42–2.97). Lifestyle factors were considered, including cigarette smoking, alcohol consumption, physical exercise, diet, body mass index (BMI), and waist circumference (WC). Among high-PRS individuals, the 10-year incidence of T2D slumped from 6.77% to 3.28% for participants having ideal lifestyles (4–6 healthy lifestyle factors) compared with poor lifestyles (0-2 healthy lifestyle factors). When integrating the high PRS, the 10-year T2D risk of low-clinical-risk individuals exceeded that of high-clinical-risk individuals with a low PRS (3.34% vs. 2.91%). These findings suggest that the PRS of 46 SNPs could be used in identifying high-risk individuals and improve the risk stratification defined by traditional clinical risk factors for T2D. Healthy lifestyles can reduce the risk of a high PRS, which indicates the potential utility in early screening and precise prevention.