Development and validation of modified risk prediction models for cardiovascular disease and its subtypes: The Hisayama Study

Home blood pressure stability score is associated with better cardiovascular prognosis: data from the nationwide prospective J-HOP study

A home blood pressure (BP)-centered strategy is emerging as the optimal approach to achieve adequate BP control in individuals with hypertension, but a simple cardiovascular risk score based on home BP level and variability is lacking. This study used prospective data from the Japan Morning Surge-Home Blood Pressure (J-HOP) extended study to develop a simple home BP stability score for the prediction of cardiovascular risk. The J-HOP extended study included 4070 participants (mean age 64.9 years) who measured home BP three times in the morning and evening for 14 days at baseline. During the mean 6.3-year follow-up, there were 260 cardiovascular events. A home BP stability score was calculated based on the average of morning and evening systolic BP (SBP; MEave), and three home BP variability metrics: average real variability (average absolute difference between successive measurements); average peak (average of the highest three SBP values for each individual), and time in therapeutic range (proportion of time spent with MEave home SBP 100-135 mmHg). There was a curvilinear association between the home BP stability score and the risk of cardiovascular events. Compared with individuals in the optimal home SBP stability score group (9-10 points), those in the very high-risk group (0 points) had significantly higher cardiovascular event risk during follow-up (adjusted hazard ratio 3.97, 95% confidence interval 2.22-7.09; p < 0.001), independent of age, sex, medication, cardiovascular risk factors, and office BP. These data show the potential for a simple home BP-based score to predict cardiovascular event risk in people with hypertension.

Home Blood Pressure Variability Risk Prediction Score for Cardiovascular Disease Using Data From the J-HOP Study

BACKGROUND

Home blood pressure (BP) is more closely associated with cardiovascular event risk than office BP, but cardiovascular risk prediction based on home BP variability is lacking. This study developed a simple cardiovascular event prediction score, including home BP variability data, from the J-HOP study (Japan Morning Surge-Home Blood Pressure).

METHODS

The J-HOP study extended follow-up from December 2017 to May 2018 generated the study data set (4231 patients). Cardiovascular events included fatal/nonfatal stroke (n=94), coronary heart disease (n=124), heart failure (n=42), and aortic dissection (n=8). Cox proportional hazards models were used to predict overall cardiovascular risk. Potential covariates included age, sex, body mass index, smoking, history of diabetes, statin use, history of cardiovascular disease, total cholesterol:high-density lipoprotein cholesterol ratio, office systolic BP (SBP), mean of morning-evening average (MEave), home SBP, and average real variability of MEave home SBP. A risk score and models were constructed, and model performance was assessed.

RESULTS

Model performance was best when average real variability of MEave SBP was included (C statistic, 0.760). The risk score assigns points for age (5-year bands), sex, cardiovascular disease history, high-density lipoprotein cholesterol, mean MEave home SBP, and average real variability of MEave home SBP. Estimated 10-year cardiovascular risk ranged from ≤0.6% (score ≤0) to >32% (score ≥26). Calibration 2 statistics values for the model (2.66) and risk score (5.29) indicated excellent goodness of fit.

CONCLUSIONS

This simple cardiovascular disease prediction algorithm, including day-by-day home BP variability, could be used as part of a home BP-centered approach to hypertension management in clinical practice.

Prediction of the 4-Year Incidence Risk of Ischemic Stroke in Healthy Japanese Adults: The Fukushima Health Database

AIM

Estimating the risk of developing ischemic stroke (IS) may assist health professionals in motivating individuals to modify their risk behavior.

METHODS

A predictive model was derived from 178,186 participants from Fukushima Health Database, aged 40-74 years, who attended the health checkup in 2014 and completed at least one annual health checkup by 2018 (Cohort I). Cox proportional hazard regression model was used to build a 4-year prediction model, thus the risk scores were based on the regression coefficients. External validation for the risk scores was conducted in another cohort of 46,099 participants following between 2015 and 2019 (Cohort II).

RESULTS

The 4-year cumulated incidence rate of IS was 179.80/100,000 person-years in Cohort I. The predictive model included age, sex, blood pressure, hypertension treatment, diabetes, low- and high-density lipoprotein cholesterol, smoking, walking pace, and body weight change of 3 kg within one year. Risk scores were interpreted based on the Cohort I predictive model function. The Harrell's C-statistics of the discrimination ability of the risk score model (95% confidence interval) was 0.744 (0.729-0.759) in Cohort I and 0.770 (0.743-0.797) in Cohort II. The overall agreement of the risk score probability of IS incidence for the observed/expected case ratio and 95% CI was 0.98 (0.92-1.05) in Cohort I and 1.08 (0.95-1.22) in Cohort II.

CONCLUSIONS

The 4-year risk prediction model revealed a good performance for IS incidence, and risk scores could be used to estimate individual incidence risk of IS. Updated models with additional confirmed risk variables may be needed.

Cost Effectiveness of the First-in-Class ARNI (Sacubitril/Valsartan) for the Treatment of Essential Hypertension in a Chinese Setting

OBJECTIVE

The aim of this study was to model the potential long-term disease progression and pharmacoeconomic value of sacubitril/valsartan for the treatment of essential hypertension from a Chinese healthcare system perspective.

METHODS

A Markov cohort model with five health states was constructed to simulate the incidence of acute cardiovascular events and cost per quality-adjusted life-year (QALY) gained with sacubitril/valsartan compared with allisartan isoproxil and valsartan over a lifetime horizon with an annual cycle. Multivariable risk regression models derived from China-PAR data accompanied by hazard ratios were used to transform the dual mechanism of sacubitril/valsartan to lower blood pressure and left ventricular mass index into long-term fatal and non-fatal cardiovascular risks. Efficacy data were calculated using a network meta-analysis integrated by the results of clinical trials. Healthcare costs were determined from a real-world study and published literature, supplemented by expert opinion. Utilities were derived from literature. Both costs and health outcomes were discounted at 5.0% annually, and prices corresponded to 2021. Model validation, deterministic and probabilistic sensitivity analyses were conducted to test the robustness of results.

RESULTS

For simulated patients with hypertension, sacubitril/valsartan reduced the rates of myocardial infarction by 6.67% and 6.39%, stroke by 9.38% and 8.98%, and heart failure hospitalization by 9.92% and 9.62% relative to allisartan isoproxil and valsartan, respectively. It was also associated with gains in life expectancy among hypertensive individuals of 0.362-0.382 years. Eventually, lifetime costs per patient were CN¥59,272 (US$9187) for sacubitril/valsartan, CN¥54,783 (US$8492) for allisartan isoproxil, and CN¥56,714 (US$8791) for valsartan; total QALYs were 11.38, 11.24, and 11.25, respectively. The incremental cost-effectiveness ratio was CN¥31,805/QALY (US$4930/QALY) compared with allisartan isoproxil, and CN¥19,247/QALY (US$2983/QALY) compared with valsartan, both of which are below the one time per-capita GDP of CN¥80,976/QALY (US$12,551/QALY) in China. Similar results were obtained in various extensive sensitivity analysis scenarios.

CONCLUSIONS

This was the first study to evaluate the cost effectiveness of sacubitril/valsartan in the treatment of hypertension. Sacubitril/valsartan compares favorably with allisartan isoproxil and valsartan in the Chinese setting, which is mainly due to its higher efficacy resulting in fewer cardiovascular events and ultimately less related mortality over time. The results could inform deliberations regarding reimbursement and access to this treatment in China and may provide reference for facilitating more reasonable and efficient allocation of limited resources in such low- and middle-income countries.

Heart diseases, hypertension and effects of antihypertensive medications: Is hypertension a true risk factor of heart diseases?

Background

Heart diseases (HD) are the leading cause of deaths in the world. Many studies have been done on the relationships among hypertension, HD and antihypertensive medications. Most of the studies find that hypertension is a significant risk factor of HD, but there are some studies in which hypertension is not a risk factor. As antihypertensive medications are routinely prescribed to prevent HD, it is necessary to evaluate the effects of these and other risk factors of HD.

Data and methods

The relationship between hypertension and HD was analyzed using 6,773,464 medical checkups obtained from the JMDC Claims Database obtained from January 2005 to September 2019. Factors potentially affecting HD, including blood pressures (BP) and usage of antihypertensive medications, were evaluated using 2,861,769 observations. To avoid the causality problem, probit models were used to analyze the probability of an individual who had no history of HD at year t developing HD by year t + 1.

Results

A positive relation between systolic blood pressure (SBP) and HD was found in the equation without any other covariates. However, the significant relation between HD and BP disappeared when the models contained various other factors as covariates. When a 10-year age or longer interval was used in the model, a positive relation between the two variables was found, suggesting that SBP works as a proxy variable. Taking antihypertensive medications greatly increases the probability of developing HD in the next year. Higher levels of cholesterols decrease the probability of developing HD.

Conclusion

Unlike many previous studies, no significant relationship between HD and hypertension was found in the models containing multiple covariates. The accepted relation might actually be spurious, and it is important to select covariates carefully. Taking antihypertensive medications appears to increase the probability of developing HD in the next year, suggesting the need for further research and greater caution in the use of antihypertensive medications.

A simplified prediction model for end-stage kidney disease in patients with diabetes

This study aimed to develop a simplified model for predicting end-stage kidney disease (ESKD) in patients with diabetes. The cohort included 2549 individuals who were followed up at Kyushu University Hospital (Japan) between January 1, 2008 and December 31, 2018. The outcome was a composite of ESKD, defined as an eGFR < 15 mL min⁻¹ [1.73 m]⁻², dialysis, or renal transplantation. The mean follow-up was 5.6 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pm$$\end{document}± 3.7 years, and ESKD occurred in 176 (6.2%) individuals. Both a machine learning random forest model and a Cox proportional hazard model selected eGFR, proteinuria, hemoglobin A1c, serum albumin levels, and serum bilirubin levels in a descending order as the most important predictors among 20 baseline variables. A model using eGFR, proteinuria and hemoglobin A1c showed a relatively good performance in discrimination (C-statistic: 0.842) and calibration (Nam and D’Agostino \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi$$\end{document}χ² statistic: 22.4). Adding serum albumin and bilirubin levels to the model further improved it, and a model using 5 variables showed the best performance in the predictive ability (C-statistic: 0.895, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\chi$$\end{document}χ² statistic: 7.7). The accuracy of this model was validated in an external cohort (n = 5153). This novel simplified prediction model may be clinically useful for predicting ESKD in patients with diabetes.

Development and Validation of a Risk Prediction Model for Atherosclerotic Cardiovascular Disease in Japanese Adults: The Hisayama Study

AIM

To develop and validate a new risk prediction model for predicting the 10-year risk of atherosclerotic cardiovascular disease (ASCVD) in Japanese adults.

METHODS

A total of 2,454 participants aged 40-84 years without a history of cardiovascular disease (CVD) were prospectively followed up for 24 years. An incident ASCVD event was defined as the first occurrence of coronary heart disease or atherothrombotic brain infarction. A Cox proportional hazards regression model was used to construct the prediction model. In addition, a simplified scoring system was translated from the developed prediction model. The model performance was evaluated using Harrell's C statistics, a calibration plot with the Greenwood-Nam-D'Agostino test, and a bootstrap validation procedure.

RESULTS

During a median of a 24-year follow-up, 270 participants experienced the first ASCVD event. The predictors of the ASCVD events in the multivariable Cox model included age, sex, systolic blood pressure, diabetes, serum high-density lipoprotein cholesterol, serum low-density lipoprotein cholesterol, proteinuria, smoking habits, and regular exercise. The developed models exhibited good discrimination with negligible evidence of overfitting (Harrell's C statistics: 0.786 for the multivariable model and 0.789 for the simplified score) and good calibrations (the Greenwood-Nam-D'Agostino test: P=0.29 for the multivariable model, 0.52 for the simplified score).

CONCLUSION

We constructed a risk prediction model for the development of ASCVD in Japanese adults. This prediction model exhibits great potential as a tool for predicting the risk of ASCVD in clinical practice by enabling the identification of specific risk factors for ASCVD in individual patients.

Changes in Body Weight and Concurrent Changes in Cardiovascular Risk Profiles in Community Residents in Japan: the Hisayama Study

AIM

We investigated the influence of weight change on concurrent changes in predicted cardiovascular disease (CVD) risk and individual CVD risk factors over time.

METHODS

A total of 2,140 community-dwellers aged 40-74 years participated in both 2002 and 2007 health examinations. Obesity was defined as body mass index ≥ 25 kg/m2. Weight trajectories were classified as: "stable obese" (obese at both examinations), "obese to nonobese" (obese in 2002 but nonobese in 2007), "nonobese to obese" (nonobese in 2002 but obese in 2007), or "stable nonobese" (nonobese at both examinations). We compared changes in the model-predicted risk for CVD and individual CVD risk factors across weight-change categories.

RESULTS

The predicted risk for CVD increased during 5 years in all groups; the increment in the predicted risk for CVD was smallest in the obese to nonobese participants and steepest in the nonobese to obese subjects. Compared with the stable obese participants, the obese to nonobese participants had greater favorable changes in waist circumferences, blood pressure, fasting plasma glucose, serum high-density lipoprotein cholesterol, serum triglycerides, and liver enzymes. For all these parameters, opposite trends were observed when comparing the nonobese to obese participants with the stable nonobese group.

CONCLUSIONS

We demonstrated the favorable association of losing weight in obese people and avoiding excessive weight gain in nonobese people with global risk of future CVD and individual CVD risk factors in a real-world setting. The findings could improve behavioral lifestyle interventions that provide information on the health consequences of weight change at health checkups.

Estimating Risk of Cardiovascular Disease Among Long-Term Colorectal Cancer Survivors: A Nationwide Cohort Study

Background

Concerns about a growing number of colorectal cancer survivors have emerged regarding cardiovascular disease (CVD) risks. However, there is not yet a predictive tool that can estimate CVD risk and support the management of healthcare as well as disease prevention in terms of CVD risk among long-term colorectal cancer survivors.

Aim

To develop predictive tools to estimate individualized overall and each subtype of CVD risk using a nationwide cohort in South Korea.

Methods and Results

A total of 4,709 newly diagnosed patients with colorectal cancer who survived at least 5 years in the National Health Insurance System were analyzed. Cox proportional hazard regression was used for the identification of independent risk factors for the derivation of predictive nomograms, which were validated in an independent cohort (n = 3,957). Age, fasting serum glucose, γ-glutamyl transpeptidase, Charlson comorbidity index, household income, body mass index, history of chemotherapy, cigarette smoking, and alcohol consumption were identified as independent risk factors for either overall CVD or each subtype of CVD subtype. Based on the identified independent risk factors, six independent nomograms for each CVD category were developed. Validation by an independent cohort demonstrated a good calibration with a median C-index of 0.687. According to the nomogram-derived median score, relative risks of 2.643, 1.821, 4.656, 2.629, 4.248, and 5.994 were found for overall CVD, ischemic heart disease, myocardial infarction, total stroke, ischemic stroke, and hemorrhage stroke in the validation cohort.

Conclusions

The predictive tools were developed with satisfactory accuracy. The derived nomograms may support the estimation of overall and individual CVD risk for long-term colorectal cancer survivors.

Risk Prediction Model for Incident Atrial Fibrillation in a General Japanese Population - The Hisayama Study

BACKGROUND

The risk prediction of incident atrial fibrillation (AF) is useful to prevent AF and its complications. The aim of this study is to develop a new risk prediction model for incident AF using the prospective longitudinal data from a general Japanese population.

METHODS AND RESULTS

A total of 2,442 community-dwelling AF-free residents aged ≥40 years were followed up from 1988 to 2012 (46,422 person-years). The development of AF was confirmed by a standard 12-lead electrocardiogram at repeated health examinations and by medical records at clinics or hospitals. The risk prediction model for incident AF was developed using a Cox proportional hazards model. During the follow up, 230 AF events were confirmed. Age, sex, systolic blood pressure, waist circumference, estimated glomerular filtration rate, abnormal cardiac murmur, high R-wave amplitude, and arrhythmia other than AF were selected for inclusion in the model. This model showed good discrimination (Harrell's c statistics: 0.785) and calibration (Greenwood-Nam-D'Agostino test: P=0.87) for AF risk at 10 years.

CONCLUSIONS

The new risk prediction model showed good performance on the individual risk assessment of the future onset of AF in a general Japanese population. As this model included commonly used clinical parameters, it may be useful for determining the requirements for the careful evaluation of AF, such as frequent electrocardiogram examinations in clinical settings, and subsequent reductions in the risk of AF-related complications.

An integrated review on machine learning approaches for heart disease prediction: Direction towards future research gaps

Objectives

To make a clear literature review on state-of-the-art heart disease prediction models.

Methods

It reviews 61 research papers and states the significant analysis. Initially, the analysis addresses the contributions of each literature works and observes the simulation environment. Here, different types of machine learning algorithms deployed in each contribution. In addition, the utilized dataset for existing heart disease prediction models was observed.

Results

The performance measures computed in entire papers like prediction accuracy, prediction error, specificity, sensitivity, f-measure, etc., are learned. Further, the best performance is also checked to confirm the effectiveness of entire contributions.

Conclusions

The comprehensive research challenges and the gap are portrayed based on the development of intelligent methods concerning the unresolved challenges in heart disease prediction using data mining techniques.

Development of a Cardiovascular Disease Risk Prediction Model Using the Suita Study, a Population-Based Prospective Cohort Study in Japan

AIM

To construct a risk prediction model for cardiovascular disease (CVD) based on the Suita study, an urban Japanese cohort study, and compare its accuracy against the Framingham CVD risk score (FRS) model.

METHODS

After excluding participants with missing data or those who lost to follow-up, this study consisted of 3,080 men and 3,470 women participants aged 30-79 years without CVD at baseline in 1989-1999. The main outcome of this study was incidence of CVD, defined as the incidence of stroke or coronary heart disease. Multivariable Cox proportional hazards models with stepwise selection were used to develop the prediction model. To assess model performance, concordance statistics (C-statistics) and their 95% confidence intervals (CIs) were calculated using a bootstrap procedure. A calibration test was also conducted.

RESULTS

During a median follow-up period of 16.9 years, 351 men and 241 women developed CVD. We formulated risk models with and without electrocardiogram (ECG) data that included age, sex, systolic blood pressure, diastolic blood pressure, high-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, diabetes mellitus, smoking, and urinary protein as risk factors. The C-statistics of the Suita CVD risk models with ECG data (0.782; 95% CI, 0.766-0.799) and without ECG data (0.781; 95% CI, 0.765-0.797) were significantly higher than that of the FRS model (0.768; 95% CI, 0.750-0.785).

CONCLUSIONS

The Suita CVD risk model is feasible to use and improves predictability of the incidence of CVD relative to the FRS model in Japan.

In-hospital morality associated with acute myocardial infarction was inversely related with the number of coronary risk factors in patients from a Japanese nation-wide real-world database

BACKGROUND

Hypertension, diabetes, dyslipidemia and smoking are established coronary risk factors for coronary heart disease in the general population. However, in Japanese patients with acute myocardial infarction (AMI), the impact of the number of coronary risk factors on in-hospital morality remains unclear.

METHODS

The Japan Acute Myocardial Infarction Registry (JAMIR) is a nationwide real-world database. We examined the association between the number of coronary risk factors and in-hospital mortality.

RESULTS

Data were obtained from total of 20462 AMI patients (mean age, 68.8 ± 13.3 years old; 15281 men, 5181 women). The prevalence of hypertension increased with advancing age, while the prevalence of smoking decreased with advancing age. The prevalence of diabetes and dyslipidemia were highest in middle age. A majority (76.9%) of the patients with AMI had at least 1 of these coronary risk factors. Overall, the number of coronary risk factor was relatively less in older subjects and women under 50 years old. Crude in-hospital mortality rates were 10.7%, 10.5%, 7.2%, 5.0% and 4.5% with 0, 1, 2, 3 and 4 risk factors, respectively. After adjusting for age and sex, there was an inverse association between the number of coronary risk factors and the in-hospital mortality (adjusted odds ratio [1.68; 95% confidence interval, 1.20-2.35] among individuals with 0 vs. 4 risk factors).

CONCLUSION

In the present study of Japanese patients with AMI, who received modern medical treatment, in-hospital mortality was inversely related to the number of coronary risk factors. To investigate the underlying reasons for these findings, further studies are needed.

Weight optimized neural network for heart disease prediction using hybrid lion plus particle swarm algorithm

Heart disease remains one of the significantcauses ofmortality and morbidity amongst the world's population. Predicting heart disease is considered as one of the vital issues in clinical data analysis. Since the number of data is rising gradually, it is muchcomplicatedforanalyzing and processing, and especially, it becomes difficult to maintain the e-healthcare data. Moreover, the prediction model under machine learning seems to be anessentialfacet in this research area. In this scenario, this paper aims to propose a new heart disease prediction model with the inclusion of specificprocesses like Feature Extraction, Record, Attribute minimization, and Classification. Initially, both statistical and higher-order statistical features are extracted under feature extraction. Subsequently, the record and attribute minimization carried out, where Component Analysis PCA plays its major role in solving the "curse of dimensionality."Finally, the prediction process takes place by the Neural Network (NN) model that intake the dimensionally reduced features. Moreover, the major intention of this paper deals with the accurate prediction. Hence, it is planned to influence the utility of meta-heuristic algorithms for the weight optimization of NN. This paper introduces a new hybrid algorithm termed Particle Swarm Optimization (PSO) merged LA update (PM-LU) algorithm that solves the above-mentioned optimization crisis, which hybrids the concept of Lion Algorithm (LA) and PSO algorithm. Finally, the efficiency of proposed work is compared over other conventional approaches and its superiority is proven with respect to certain performance measures. From the analysis, the presented PM-LU-NN scheme with regards to accuracy is 3.85%, 12.5%, 12.5%, 3.85%, and 7.41% better than LM-NN, WOA-NN, FF-NN, PSO-NN and LA-NN algorithms.

Impact of hypertension stratified by diabetes on the lifetime risk of cardiovascular disease mortality in Japan: a pooled analysis of data from the Evidence for Cardiovascular Prevention from Observational Cohorts in Japan study

Lifetime risk is an informative estimate to motivate people to change lifestyle behaviors, especially from a younger age, in public health education. The impact of the combination of hypertension and diabetes on the lifetime risk of cardiovascular mortality has not been investigated in Asian populations. A pooled analysis of individual data from nine cohorts was performed. A total of 57,339 Japanese men and women were eligible for the analysis. We used the modified Kaplan-Meier approach and estimated the remaining lifetime risk of cardiovascular mortality starting from the index age of 35 years. Participants were classified into four categories defined by hypertension and diabetes. The lifetime risk was increased in the order of those without either risk, those without hypertension but with diabetes, those with hypertension but without diabetes, and those with both risks. The lifetime risk of cardiovascular mortality at the 35-year index age was as follows: 7.8% in men and 6.2% in women for those without either hypertension or diabetes, 13.2% in men and 9.5% in women for those without hypertension but with diabetes, 17.2% in men and 11.7% in women for those with hypertension but without diabetes, and 19.4% in men and 15% in women for those with both risks. These findings reinforce the need for a life-course perspective in the management of hypertension and diabetes from a younger age.