Atherosclerotic Cardiovascular Disease Risk Prediction Research Articles

One-size-fits-all versus risk-category-based screening interval strategies for cardiovascular disease prevention in Chinese adults: a prospective cohort study

In non-high-risk individuals, risk-category-based atherosclerotic cardiovascular disease (ASCVD) screening strategies may be more cost-effective than one-size-fits-all approaches. However, current decisions are constrained by a lack of research evidence. We aimed to explore appropriate risk-category-based screening interval strategies for non-high-risk individuals in ASCVD primary prevention in the Chinese population. We used data from 28,624 participants in the China Kadoorie Biobank (CKB) who had completed at least two field surveys. The risk assessment tools were the 10-year ASCVD risk prediction models developed based on the CKB cohort. We constructed multistate Markov models to model disease progression and estimate transition probabilities between different risk categories. The total person-years spent unidentified in the high-risk state over a 10-year period were calculated for each screening interval protocol. We also estimated the number of ASCVD events prevented, quality-adjusted life years (QALYs) gained, and costs saved when compared to the 3-yearly screening protocol. When compared to the uniform 3-yearly protocol, most risk-category-based screening interval protocols would identify more high-risk individuals timely, thus preventing more ASCVD events and gaining QALYs. A few of them would reduce total health-care costs. The protocol, which used 6-year, 3-year, and 2-year screening intervals for low-risk, intermediate-low-risk, and intermediate-high risk individuals, was optimal, and would reduce the person-years spent unidentified in the high-risk category by 17.9% (95% CI: 13.1%-21.9%), thus preventing an estimated 113 thousand (95% CI: 83-138) hard ASCVD events for Chinese adults aged 30-79 over a 10-year period. When using a lower cost of statin therapy, more screening protocols would gain QALYs while saving costs. For the primary prevention of ASCVD, risk-category-based screening protocols outperformed the one-size-fits-all approach in the Chinese population. This work was supported by National Natural Science Foundation of China (82192904, 82388102, 82192900) and grants (2023YFC2509400) from the National Key R&D Program of China. The CKB baseline survey and the first re-survey were supported by a grant from the Kadoorie Charitable Foundation in Hong Kong. The long-term follow-up is supported by grants from the UK Wellcome Trust (212946/Z/18/Z, 202922/Z/16/Z, 104085/Z/14/Z, 088158/Z/09/Z), grants (2016YFC0900500) from the National Key R&D Program of China, National Natural Science Foundation of China (81390540, 91846303, 81941018), and Chinese Ministry of Science and Technology (2011BAI09B01).

The association between objectively-measured sedentary behavior patterns and predicted 10-year ASCVD risk

This study aims to investigated the association between sedentary behavior (SB) and predicted 10-year atherosclerotic cardiovascular disease (ASCVD) risk and determine whether the associations differ by how the behavior is accumulated, in US middle-aged and older adults. Cross-sectional data were derived from national health and nutrition examination survey (NHANES) 2003–2006. Seven-day wearing of accelerometer was used to assess SB pattern, exported as total SB, bouts of 1–9, 10–29, 30–59 and ≥ 60 min SB. Predicted 10-year ASCVD risk was calculated using validated pooled cohort equations. Linear regression was used to estimate adjusted coefficients. A total of 2327 participants were enrolled with mean age of 56.9 and mean predicted 10-year ASCVD risk of 10.7%. We observed significant associations of total SB and its longer accumulated patterns with higher 10-year predicted ASCVD risk, in a linear fashion and independent of a list of covariates. A 30 min increment per day of total SB, bouts in 10–29, bouts in 30–59 and bouts in ≥ 60 min were associated with 0.14, 0.14, 0.23 and 0.12% higher multivariable-adjusted 10-year predicted ASCVD risk. There are significant associations of total SB as well as its longer accumulated patterns with higher 10-year predicted ASCVD risk, independent of a list of covariates and in a linear fashion. The result indicates that reducing total sedentary time and interrupting long duration of prolonged SB, could be meaningful to for public guideline to lessen the personal and public health burden of cardiovascular health.

Predicted Cardiovascular Risk by the PREVENT Equations in US Adults With Stage 1 Hypertension.

The 2017 American College of Cardiology/American Heart Association blood pressure guideline recommends initiation of antihypertensive medication for adults with stage 1 hypertension (systolic blood pressure, 130-139 mm Hg, or diastolic blood pressure, 80-89 mm Hg) and 10-year atherosclerotic cardiovascular disease (ASCVD) risk ≥10% estimated by the pooled cohort equations (PCEs). In 2023, the American Heart Association published the predicting risk of cardiovascular disease events (PREVENT) equations to estimate ASCVD and total cardiovascular disease risk. We analyzed US National Health and Nutrition Examination Survey data from 2013 to 2020 for 1703 adults aged 30 to 79 years without self-reported cardiovascular disease with stage 1 hypertension. We estimated 10-year ASCVD risk by the PCEs and 10-year ASCVD and total cardiovascular disease risk by the base PREVENT equations. Analyses were weighted to represent noninstitutionalized US adults with stage 1 hypertension. Mean 10-year ASCVD risk was 5.4% (95% CI, 5.0%-5.9%) and 2.9% (95% CI, 2.7%-3.1%) using the PCEs and PREVENT equations, respectively. The proportion with 10-year ASCVD risk of 10% to <15% and ≥15% was 8.1% and 7.8% estimated by the PCEs, respectively, and 3.0% and 0.3% estimated by the PREVENT equations, respectively. No participants had a 10-year ASCVD risk ≥10% on the PREVENT equations and <10% on the PCEs, while 12.5% had a 10-year ASCVD risk ≥10% on the PCEs and <10% on the PREVENT equations. The mean 10-year total cardiovascular disease risk estimated by the PREVENT equations was lower than the mean 10-year ASCVD risk on the PCEs. Among US adults with stage 1 hypertension, the 10-year predicted ASCVD risk estimated by the PREVENT equations was approximately half the risk estimated by the PCEs.

Associations of 10-year atherosclerotic cardiovascular disease risk scores with cerebral small vessel disease: the PolyvasculaR Evaluation for Cognitive Impairment and vaScular Events (PRECISE) study.

10-year atherosclerotic cardiovascular disease (ASCVD) risk scores were useful for predicting large vessel disease, but the relationships between them and cerebral small vessel disease (CSVD) were unclear. Our study aimed to evaluate associations of 10-year ASCVD risk scores with CSVD and its magnetic resonance imaging (MRI) markers. Community-dwelling residents from the PolyvasculaR Evaluation for Cognitive Impairment and vaScular Events study were included in this cross-sectional study. At baseline, we collected data related to the Framingham Risk Score (FRS), pooled cohort equation (PCE), prediction for ASCVD risk in China (China-PAR) and Systematic COronary Risk Evaluation model 2 (SCORE2), and classified participants into low, moderate and high groups. Participants underwent brain MRI scans. We evaluated white matter hyperintensity (WMH), lacunes, cerebral microbleeds (CMBs) and enlarged perivascular spaces in basal ganglia (BG-EPVS) according to criteria of Wardlaw and Rothwell, and calculated total CSVD score and modified total CSVD score. A total of 3063 participants were included, and 53.5% of them were female. A higher FRS was associated with higher total CSVD score (moderate vs. low: cOR 1.89, 95% CI 1.53-2.34; high vs. low: cOR 3.23, 95%CI 2.62-3.97), and the PCE, China-PAR or SCORE2 score was positively related to total CSVD score (P < 0.05). Moreover, higher 10-year ASCVD scores were associated with higher odds of WMH (P < 0.05), lacunes (P < 0.05), CMBs (P < 0.05) and BG-EPVS (P < 0.05). The 10-year ASCVD scores were positively associated with CSVD and its MRI markers. These scores provided a method of risk stratification in the population with CSVD.

Establishment and validation of prediction model for atherosclerotic cardiovascular disease in patients with hyperuricemia.

To construct a risk prediction model for atherosclerotic cardiovascular disease (ASCVD) in patients with hyperuricemia. Data in this study were obtained from the National Health and Nutrition Examination Survey (NHANES) (2007-2010). Participants from Huashan Hospital were included as an external validation. Logistic regression analysis was used to explore the relevant factors of ASCVD in patients with hyperuricemia. The discriminability of the model was evaluated using the area under the curve (AUC) statistic of the receiver operating characteristic curve. Hosmer-Lemeshow test, correction curve and decision curve analysis (DCA) were used to evaluate the model. A total of 389 patients collected from the NHANES were included in the final analysis. Logistic regression analysis showed that age, creatinine (Cr), glucose (Glu), serum uric acid (SUA), and history of gout were predictive factors for ASCVD in hyperuricemia (HUA) patients. These predictive factors were used to construct a nomogram. And 157 patients from NHANES were in the internal validation group and 136 patients from Huashan Hospital were in the external validation group. The AUC values of the three groups were 0.943, 0.735, and 0.664. The p values of the Hosmer-Lemeshow test were .568, .600, and .763. The calibration curve showed consistency between the nomogram and the actual observed values. The DCA curve indicated that the model has good clinical practicality. This study constructed the ASCVD risk prediction model for HUA patients, which is beneficial for medical staff to detect high-risk populations of ASCVD in the early stage.

Health guidance intervention mitigates the escalation of predicted comprehensive cardiovascular disease risk: Interrupted time series design

IntroductionMetabolic syndrome increases the risk of cardiovascular events. Despite the Japanese healthcare system's efforts in early detection and intervention, particularly through Specific Health Checkup and Guidance programs, research on their effectiveness is limited. This study evaluated the impact of Specific Health Guidance on the predicted risk of atherosclerotic cardiovascular disease (ASCVD) in working-age Japanese individuals. MethodsEmploying an Interrupted Time Series (ITS) design, this study compared the trends in predicted ASCVD risk and each individual risk factor used for the prediction of ASCVD risk before and after intervention in individuals participating in the guidance. ResultsAnalyses based on the ITS design indicated that participation in Specific Health Guidance programs, specifically the intensive level program, mitigates the increase trend of the predicted ASCVD risk. On the other hand, the impact on the trends of individual cardiovascular risk factors was minimal. ConclusionsThe intensive level Specific Health Guidance appeared to reduce the increasing trend in ASCVD risk, emphasizing the importance of comprehensive risk assessment in evaluating health interventions. However, the results are limited owing to the specific demographics and short evaluation period. Further research is necessary to understand the long-term impacts and broader applicability.

Atherosclerotic cardiovascular disease risk among Ghanaians: A comparison of the risk assessment tools.

ObjectivesRisk stratification is a cornerstone for preventing atherosclerotic cardiovascular disease (ASCVD). Ghana has yet to develop a locally derived and validated ASCVD risk model. A critical first step towards this goal is assessing how the commonly available risk models perform in the Ghanaian population. This study compares the agreement and correlation between four ASCVD risk assessment models commonly used in Ghana. MethodsThe Ghana Heart Study collected data from four regions in Ghana (Ashanti, Greater Accra, Northern, and Central regions) and excluded people with a self-declared history of ASCVD. The 10-year fatal/non-fatal ASCVD risk of participants aged 40–74 was calculated using mobile-based apps for Pooled Cohort Equation (PCE), laboratory-based WHO/ISH CVD risk, laboratory-based Framingham risk (FRS), and Globorisk, categorizing them as low, intermediate, or high risk. The risk categories were compared using the Kappa statistic and Spearman correlation. ResultsA total of 615 participants were included in this analysis (median age 55 [Inter quartile range 46, 64]) years with 365 (59.3 %) females. The WHO/ISH risk score categorized 504 (82.0 %), 58 (9.4 %), and 53 (8.6 %) as low-, intermediate-, and high-risk, respectively. The PCE categorized 345 (56.1 %), 181 (29.4 %), and 89 (14.5 %) as low-, intermediate- and high-risk, respectively. The Globorisk categorized 236 (38.4 %), 273 (44.4 %), and 106 (17.2 %) as low-, intermediate-, and high-risk, respectively. Significant differences in the risk categorization by region of residence and age group were noted. There was substantial agreement between the PCE vs FRS (Kappa = 0.8, 95 % CI 0.7 – 0.8), PCE vs Globorisk (Kappa = 0.6; 95 % CI 0.6 – 0.7), and FRS vs Globorisk (Kappa = 0.6; 95 % CI 0.6 – 0.7). However, there was only fair agreement between the WHO vs Globorisk (Kappa = 0.3; 95 % CI 0.3–0.4) and moderate agreement between the WHO vs PCE and WHO vs FRS. ConclusionThere are significant differences in the ASCVD risk prediction tools in the Ghanaian population, posing a threat to primary prevention. Therefore, there is a need for locally derived and validated tools.

Comparative Analysis of Three Atherosclerotic Cardiovascular Disease Risk Prediction Models in Individuals Aged 75 and Older.

To evaluate the performance of the Framingham cardiovascular risk score (FRS)/pooled cohort equations (PCE)/China prediction for atherosclerotic cardiovascular disease (ASCVD) risk (China-PAR model) in a prospective cohort of Chinese older adults. We assessed 717 older adults aged 75-85 years without ASCVD at the baseline from the Sichuan province of China. The participants were followed annually from 2011 to 2021. We obtained the participants' information through the medical records of physical examination and evaluated their 10-year ASCVD risk using FRS, PCE, and China-PAR. We further evaluated the predictive abilities of three assessment models. During the 10-year follow-up, 206 participants developed ASCVD, with an incidence rate of 28.73%. The FRS and China-PAR moderately underestimated the risk of ASCVD (22.1% and 12.4%, respectively), but while PCE overestimated the risk (36.1%). FRS and China-PAR were found to underestimate the risk of ASCVD (26% and 63%, respectively) for men, while PCE overestimated the risk by 8%; For women, FRS and China-PAR were found to underestimate the risk of ASCVD (14% and 35%, respectively), while PCE overestimated the risk by 88%. The 10-year ASCVD risk was found to be overestimated by PCE. China-PAR had the most accurate predictions in women, while FRS was particularly well-calibrated in males. All three risk models have good discrimination, with FRS and PCE being well-calibrated in men and all three being well-calibrated in women. Therefore, accurate risk models are warranted to facilitate the prevention of ASCVD at the baseline among Chinese older adults.

Incorporating longitudinal history of risk factors into atherosclerotic cardiovascular disease risk prediction using deep learning

It is increasingly clear that longitudinal risk factor levels and trajectories are related to risk for atherosclerotic cardiovascular disease (ASCVD) above and beyond single measures. Currently used in clinical care, the Pooled Cohort Equations (PCE) are based on regression methods that predict ASCVD risk based on cross-sectional risk factor levels. Deep learning (DL) models have been developed to incorporate longitudinal data for risk prediction but its benefit for ASCVD risk prediction relative to the traditional Pooled Cohort Equations (PCE) remain unknown. Our study included 15,565 participants from four cardiovascular disease cohorts free of baseline ASCVD who were followed for adjudicated ASCVD. Ten-year ASCVD risk was calculated in the training set using our benchmark, the PCE, and a longitudinal DL model, Dynamic-DeepHit. Predictors included those incorporated in the PCE: sex, race, age, total cholesterol, high density lipid cholesterol, systolic and diastolic blood pressure, diabetes, hypertension treatment and smoking. The discrimination and calibration performance of the two models were evaluated in an overall hold-out testing dataset. Of the 15,565 participants in our dataset, 2170 (13.9%) developed ASCVD. The performance of the longitudinal DL model that incorporated 8 years of longitudinal risk factor data improved upon that of the PCE [AUROC: 0.815 (CI 0.782–0.844) vs 0.792 (CI 0.760–0.825)] and the net reclassification index was 0.385. The brier score for the DL model was 0.0514 compared with 0.0542 in the PCE. Incorporating longitudinal risk factors in ASCVD risk prediction using DL can improve model discrimination and calibration.

Performance of ASCVD Risk Prediction Models in Individuals With Inflammatory Bowel Disease: A UK Biobank Study.

Lay Summary This study suggests that the current atherosclerotic cardiovascular disease risk prediction models used in clinical practice performed better in the non–inflammatory bowel disease (IBD) cohort compared with IBD, highlighting the need for a more specific risk prediction model tailored to the IBD population.

The potential interaction between chemosensitivity and the development of cardiovascular disease in obstructive sleep apnea

ObjectivesChemosensitivity is an essential part of the pathophysiological mechanisms of obstructive sleep apnea (OSA). Not only does OSA have a certain relationship with the comorbidity of cardiovascular disease (CVD) but also chemosensitivity plays a crucial role in the development of CVD. This study aims to investigate the potential interaction between chemosensitivity and the development of CVD in OSA. MethodsA total of 169 participants with suspected OSA were included. Data were gathered on the parameters of polysomnography and baseline clinical features. Peripheral chemosensitivity was evaluated by employing the rebreathing test. The lifetime CVD risk was computed using the China-PAR (Prediction for atherosclerotic CVD Risk in China) risk equation. ResultsAfter controlling for covariates, participants with chemosensitivity levels in the second and fifth quantiles tended to hold an increased proportion of high lifetime CVD risk (OR 10.90, 95%CI [2.81–42.28]; OR 6.78, 95%CI [1.70–27.05], respectively). The diagnosis of OSA would significantly increase the 10-year and lifetime CVD risks in participants with low chemosensitivity, while no such differences were found in participants with high chemosensitivity. ConclusionHigher lifetime CVD risk was associated with participants who had greater peripheral chemosensitivity. In terms of the CVD outcomes, adult patients with a relatively low level of chemosensitivity may be primarily related to their diagnosis of OSA, whereas adult patients with a relatively high level of chemosensitivity may be more strongly associated with their elevated levels of chemosensitivity rather than OSA.

Atherosclerotic cardiovascular disease events among adults with high predicted risk without established risk factors

ObjectiveAge is the strongest contributor to 10-year predicted atherosclerotic cardiovascular disease (ASCVD) risk. Some older adults have a predicted ASCVD risk ≥7.5 %, without established risk factors. We sought to compare ASCVD incidence among adults with predicted ASCVD risk ≥7.5 %, with and without established ASCVD risk factors, to adults with predicted risk <7.5 %. MethodsWe analyzed data from REasons for Geographic and Racial Differences in Stroke study participants, 45–79 years old, without ASCVD or diabetes, not taking statins and with low-density lipoprotein cholesterol 70–189 mg/dL. Participants were categorized into 3 groups based on their 10-year predicted ASCVD risk and presence of established risk factors: <7.5 %, ≥7.5 % with established risk factors and ≥7.5 % without established risk factors. Established risk factors included smoking, systolic blood pressure ≥130 mmHg or antihypertensive medication use, total cholesterol ≥200 mg/dL, or high-density lipoprotein cholesterol <50 mg/dL for women (<40 mg/dL for men). Participants were followed for ASCVD events. ResultsAmong 11,115 participants, 911 incident ASCVD events occurred over a median of 11.1 years. ASCVD incidence rates were 3.6, 12.8, and 9.8 per 1,000 person-years for participants with predicted risk <7.5 %, predicted risk ≥7.5 % with established risk factors and predicted risk ≥7.5 % without established risk factors, respectively. Compared to adults with predicted risk <7.5 %, hazard ratios for incident ASCVD in participants with risk ≥7.5 % with and without established risk factors were 3.58 (95 %CI 3.03 – 4.21) and 2.72 (95 %CI 1.91–3.88), respectively. ConclusionsAdults with a 10-year predicted ASCVD risk ≥7.5 % but without established risk factors had a high ASCVD incidence.

Abstract 11599: High-Density Lipoprotein Cholesterol and Statin Use for Primary Prevention Among US Adults With Intermediate 10-year Predicted ASCVD Risk

Introduction: Although there is no strong evidence of a causal relationship between high-density lipoprotein cholesterol (HDL-C) with atherosclerotic cardiovascular disease (ASCVD), HDL-C is considered “good cholesterol” by many which may impact decision making around statin use. We determined if statin use for primary prevention differed by HDL-C levels among adults without ASCVD, diabetes mellitus (DM), or low-density lipoprotein cholesterol (LDL-C) &gt;190 mg/dL but an intermediate 10-year predicted ASCVD risk. Methods: National Health and Nutrition Examination Surveys (NHANES) participants who were non-pregnant, 40-75 years with a 10-year ASCVD risk 7.5-&lt;20% and fasting lipids, and without ASCVD, DM, LDL-C &gt; 190 mg/dL, or triglycerides ≥500 mg/dL from 2013-2020 were included. Multivariable Poisson regression with robust error variance estimated adjusted prevalence and prevalence ratios (PRs) and 95% confidence intervals (CIs) for statin use associated with HDL-C levels (low [&lt;40 mg/dL], normal [40-60 mg/dL], high [ &gt; 60 mg/dL]). Models were adjusted for age, race, ethnicity, sex, education, healthcare access within 12 months, routine location for health care, health insurance, LDL-C, and body mass index. Results: Among 999 included US adults (representing 23.7 million adults) with a 10-year predicted ASCVD risk 7.5-&lt;20%, the multivariable adjusted prevalence of statin use was 32.5%, 28.6%, and 24.4% among adults with low HDL-C, normal HDL-C, and high HDL-C, respectively. After adjustment, the PR for statin use among adults with low HDL-C was 1.14 (95% CI 1.01,1.18) and high HDL-C was 0.85 (95% 0.83,0.88) vs. adults with normal HDL-C (Figure 1). Discussion In the current analysis, adults with high HDL-C were less likely to be taking a statin, while adults with low HDL-C were more likely. HDL-C may influence the decision to use statin medication among adults with intermediate 10-year ASCVD risk despite evidence that HDL-C level is not causally related to ASCVD.

Genome-Wide Association Study of Cardiovascular Resilience Identifies Protective Variation in the CETP Gene.

Background The risk of atherosclerotic cardiovascular disease (ASCVD) increases sharply with age. Some older individuals, however, remain unaffected despite high predicted risk. These individuals may carry cardioprotective genetic variants that contribute to resilience. Our aim was to assess whether asymptomatic older individuals without prevalent ASCVD carry cardioprotective genetic variants that contribute to ASCVD resilience. Methods and Results We performed a genome-wide association study using a 10-year predicted ASCVD risk score as a quantitative trait, calculated only in asymptomatic older individuals aged ≥70 years without prevalent ASCVD. Our discovery genome-wide association study of N=12 031 ASCVD event-free individuals from the ASPREE (Aspirin in Reducing Events in the Elderly) trial identified 2 independent variants, rs9939224 (P<5×10-8) and rs56156922 (P<10-6), in the CETP (cholesteryl ester transfer protein) gene. The CETP gene is a regulator of plasma high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and lipoprotein(a) levels, and it is a therapeutic drug target. The associations were replicated in the UK Biobank (subpopulation of N=13 888 individuals aged ≥69 years without prevalent ASCVD). Carriers of the identified CETP variants (versus noncarriers) had higher plasma high-density lipoprotein cholesterol levels, lower plasma low-density lipoprotein cholesterol levels, and reduced risk of incident ASCVD events during follow-up. Expression quantitative trait loci analysis predicted the identified CETP variants reduce CETP gene expression across various tissues. Previously reported associations between genetic CETP inhibition and increased risk of age-related macular degeneration were not observed among the 3917 ASPREE trial participants with retinal imaging and genetic data available. Conclusions Common genetic variants in the CETP gene region are associated with cardiovascular resilience during aging. Registration URL: https://www.clinicaltrials.gov; Unique identifier: NCT01038583.

Utility of coronary artery calcium in refining 10-year ASCVD risk prediction using a Thai CV risk score.

Coronary artery calcium (CAC) scanning is a valuable additional tool for calculating the risk of cardiovascular (CV) events. We aimed to determine if a CAC score could improve performance of a Thai CV risk score in prediction of 10-year atherosclerotic cardiovascular disease (ASCVD) risk for asymptomatic patients with CV risk factors. This was a retrospective cohort study that enrolled asymptomatic patients with CV risk factors who underwent CAC scans between 2005 and 2013. The patients were classified as low-, intermediate-, or high-risk (<10%, 10%-<20%, and ≥20%, respectively) of having ASCVD within 10-years based on a Thai CV risk score. In each patient, CAC score was considered as a categorical variable (0, 1-99, and ≥100) and natural-log variable to assess the risk of developing CV events (CV death, non-fatal MI, or non-fatal stroke). The C statistic and the net reclassification improvement (NRI) index were applied to assess whether CAC improved ASCVD risk prediction. A total of 6,964 patients were analyzed (mean age: 59.0 ± 8.4 years; 63.3% women). The majority of patients were classified as low- or intermediate-risk (75.3% and 20.5%, respectively), whereas only 4.2% were classified as high-risk. Nearly half (49.7%) of patients had a CAC score of zero (no calcifications detected), while 32.0% had scores of 1-99, and 18.3% of ≥100. In the low- and intermediate-risk groups, patients with a CAC ≥100 experienced higher rates of CV events, with hazard ratios (95% CI) of 1.95 (1.35, 2.81) and 3.04 (2.26, 4.10), respectively. Incorporation of ln(CAC + 1) into their Thai CV risk scores improved the C statistic from 0.703 (0.68, 0.72) to 0.716 (0.69, 0.74), and resulted in an NRI index of 0.06 (0.02, 0.10). To enhance the performance of the Thai CV risk score, a revision of the CV risk model was performed, incorporating ln(CAC + 1), which further increased the C statistic to 0.771 (0.755, 0.788). The addition of CAC to traditional risk factors improved CV risk stratification and ASCVD prediction. Whether this adjustment leads to a reduction in CV events and is cost-effective will require further assessment.

Machine Learning Models for ASCVD Risk Prediction in an Asian Population - How to Validate the Model is Important.

Atherosclerotic cardiovascular disease (ASCVD) is prevalent worldwide including Taiwan, however widely accepted tools to assess the risk of ASCVD are lacking in Taiwan. Machine learning models are potentially useful for risk evaluation. In this study we used two cohorts to test the feasibility of machine learning with transfer learning for developing an ASCVD risk prediction model in Taiwan. Two multi-center observational registry cohorts, T-SPARCLE and T-PPARCLE were used in this study. The variables selected were based on European, U.S. and Asian guidelines. Both registries recorded the ASCVD outcomes of the patients. Ten-fold validation and temporal validation methods were used to evaluate the performance of the binary classification analysis [prediction of major adverse cardiovascular (CV) events in one year]. Time-to-event analyses were also performed. In the binary classification analysis, eXtreme Gradient Boosting (XGBoost) and random forest had the best performance, with areas under the receiver operating characteristic curve (AUC-ROC) of 0.72 (0.68-0.76) and 0.73 (0.69-0.77), respectively, although it was not significantly better than other models. Temporal validation was also performed, and the data showed significant differences in the distribution of various features and event rate. The AUC-ROC of XGBoost dropped to 0.66 (0.59-0.73), while that of random forest dropped to 0.69 (0.62-0.76) in the temporal validation method, and the performance also became numerically worse than that of the logistic regression model. In the time-to-event analysis, most models had a concordance index of around 0.70. Machine learning models with appropriate transfer learning may be a useful tool for the development of CV risk prediction models and may help improve patient care in the future.

Social disadvantage, coronary artery calcium, and their interplay in the prediction of atherosclerotic cardiovascular disease events

Background and aimsSocial determinants of health (SDOH) are key for the identification of populations at increased risk of atherosclerotic cardiovascular disease (ASCVD). However, whether at the individual level SDOH improve current ASCVD risk prediction paradigms beyond traditional risk factors and the coronary artery calcium (CAC) score, is unknown. We evaluated the interplay between CAC and SDOH in ASCVD risk prediction. MethodsMESA is a prospective study of US adults free of clinical ASCVD at baseline. We used an SDOH index inclusive of 14 determinants from 5 domains. The index ranged 0–1 and was divided into quartiles, with higher ones representing worse SDOH. Cox regression was used to evaluate the adjusted associations between CAC, SDOH, their interplay, and ASCVD events. The C-statistic was computed to assess improvement in risk discrimination for prediction of ASCVD events. ResultsWe included 6479 MESA participants (50% with CAC = 0, 24% CAC>100). ASCVD incidence increased with increasing CAC scores across SDOH quartiles. The lowest incidence was noted in those with CAC = 0 and favourable SDOH (2/1000 person-years) and highest in those with CAC>100 and most unfavourable SDOH (20.6/1000 person-years). While CAC was strongly associated with ASCVD across SDOH quartiles, SDOH was weakly associated with ASCVD across CAC strata. CAC improved the discriminatory ability of all prediction models beyond traditional risk factors, the improvement in C-statistic ranging +0.02 - +0.05. Improvements with SDOH were smaller, and were none on top of CAC. ConclusionsCAC improves ASCVD risk stratification across the spectrum of social vulnerability, while SDOH fail to improve risk prediction beyond traditional RFs and CAC.

Ambient PM2.5 and its components associated with 10-year atherosclerotic cardiovascular disease risk in Chinese adults

BackgroundExposure to particulate matter with aerodynamic diameters less than 2.5 µm (PM2.5) may increase the risk of 10-year atherosclerotic cardiovascular disease (ASCVD) risk. While PM2.5 is comprised of various components, the evidence on the correlation of its components with 10-year ASCVD risk and which component contributes most remains limited. MethodsData were derived from the baseline assessments of China Multi-Ethnic Cohort (CMEC). In total, 69,722 individuals aged 35–74 years were included into this study. The annual average concentration of PM2.5 and its components (black carbon, ammonium, nitrate, sulfate, organic matter, soil particles, and sea salt) were estimated by satellite remote sensing and chemical transport models. The ASCVD risk of individuals was calculated by the equations from the China-PAR Project (prediction for ASCVD risk in China). The relationship between single exposure to PM2.5 and its components and predicted 10-year ASCVD risk was assessed using the logistic regression model. The effect of joint exposure was estimated, and the most significant contributor was identified using the weighted quantile sum approach. ResultsTotally 69,722 participants were included, of which 95.8 % and 4.2 % had low and high 10-year ASCVD risk, respectively. Per standard deviation increases in the 3-year average concentration of PM2.5 mass (odds ratio [OR] 1.23, 95 % confidence interval [CI]: 1.12–1.35), black carbon (1.21, 1.11–1.33), ammonium (1.21, 1.10–1.32), nitrate (1.25, 1.14–1.38), organic matter (1.29, 1.18–1.42), sulfate (1.17, 1.07–1.28), and soil particles (1.15, 1.04–1.26) were related to high 10-year ASCVD risk. The overall effect (1.19, 1.11–1.28) of the PM2.5 components was positively associated with 10-year ASCVD risk, and organic matter had the most contribution to this relationship. Female participants were more significantly impacted by PM2.5, black carbon, ammonium, nitrate, organic matter, sulfate, and soil particles compared to others. ConclusionLong-term exposure to PM2.5 mass, black carbon, ammonium, nitrate, organic matter, sulfate, and soil particles were positively associated with high 10-year ASCVD risk, while sea salt exhibited a protective effect. Moreover, the organic matter might take primary responsibility for the relationship between PM2.5 and 10-year ASCVD risk. Females were more susceptible to the adverse effect.

Evaluation of Large-Scale Proteomics for Prediction of Cardiovascular Events

Whether protein risk scores derived from a single plasma sample could be useful for risk assessment for atherosclerotic cardiovascular disease (ASCVD), in conjunction with clinical risk factors and polygenic risk scores, is uncertain. To develop protein risk scores for ASCVD risk prediction and compare them to clinical risk factors and polygenic risk scores in primary and secondary event populations. The primary analysis was a retrospective study of primary events among 13 540 individuals in Iceland (aged 40-75 years) with proteomics data and no history of major ASCVD events at recruitment (study duration, August 23, 2000 until October 26, 2006; follow-up through 2018). We also analyzed a secondary event population from a randomized, double-blind lipid-lowering clinical trial (2013-2016), consisting of individuals with stable ASCVD receiving statin therapy and for whom proteomic data were available for 6791 individuals. Protein risk scores (based on 4963 plasma protein levels and developed in a training set in the primary event population); polygenic risk scores for coronary artery disease and stroke; and clinical risk factors that included age, sex, statin use, hypertension treatment, type 2 diabetes, body mass index, and smoking status at the time of plasma sampling. Outcomes were composites of myocardial infarction, stroke, and coronary heart disease death or cardiovascular death. Performance was evaluated using Cox survival models and measures of discrimination and reclassification that accounted for the competing risk of non-ASCVD death. In the primary event population test set (4018 individuals [59.0% women]; 465 events; median follow-up, 15.8 years), the protein risk score had a hazard ratio (HR) of 1.93 per SD (95% CI, 1.75 to 2.13). Addition of protein risk score and polygenic risk scores significantly increased the C index when added to a clinical risk factor model (C index change, 0.022 [95% CI, 0.007 to 0.038]). Addition of the protein risk score alone to a clinical risk factor model also led to a significantly increased C index (difference, 0.014 [95% CI, 0.002 to 0.028]). Among White individuals in the secondary event population (6307 participants; 432 events; median follow-up, 2.2 years), the protein risk score had an HR of 1.62 per SD (95% CI, 1.48 to 1.79) and significantly increased C index when added to a clinical risk factor model (C index change, 0.026 [95% CI, 0.011 to 0.042]). The protein risk score was significantly associated with major adverse cardiovascular events among individuals of African and Asian ancestries in the secondary event population. A protein risk score was significantly associated with ASCVD events in primary and secondary event populations. When added to clinical risk factors, the protein risk score and polygenic risk score both provided statistically significant but modest improvement in discrimination.

Lipoprotein(a) and the pooled cohort equations for ASCVD risk prediction: The Multi-Ethnic Study of Atherosclerosis

Background and aimsLipoprotein(a) [Lp(a)] is an independent risk factor for atherosclerotic cardiovascular disease (ASCVD) but is not included in the Pooled Cohort Equations (PCE). We aimed to assess how well the PCE predict 10-year event rates in individuals with elevated Lp(a), and whether the addition of Lp(a) improves risk prediction. MethodsWe compared observed versus PCE-predicted 10-year ASCVD event rates, stratified by Lp(a) level and ASCVD risk category using Poisson regression, and evaluated the association between Lp(a) > 50 mg/dL and ASCVD risk using Cox proportional hazards models in the Multi-Ethnic Study of Atherosclerosis (MESA). We evaluated the C-index and net reclassification improvement (NRI) with addition of Lp(a) to the PCE. ResultsThe study population included 6639 individuals (20%, n = 1325 with elevated Lp(a)). The PCE accurately predicted 10-year event rates for individuals with elevated Lp(a) with observed event rates falling within predicted limits. Elevated Lp(a) was associated with increased risk of CVD events overall (HR 1.27, 95% CI 1.00–1.60), particularly in low (HR 2.45, 95% CI 1.40–4.31), and high-risk (HR 1.41, 95% CI 1.02–1.96) individuals. Continuous NRI (95% CI) with the addition of Lp(a) to the PCE for CVD was 0.0963 (0.0158–0.1953) overall, and 0.2999 (0.0876, 0.5525) among low-risk individuals. ConclusionsThe PCE performs well for event rate prediction in individuals with elevated Lp(a). However, Lp(a) is associated with increased CVD risk, and the addition of Lp(a) to the PCE improves risk prediction, particularly among low-risk individuals. These results lend support for increasing use of Lp(a) testing for risk assessment.