KR101839910B1

KR101839910B1 - Method of predicting cardiovascular disease risk using cardiovascular disease risk factors

Info

Publication number: KR101839910B1
Application number: KR1020150074303A
Authority: KR
Inventors: 김영학
Original assignee: 재단법인 아산사회복지재단
Priority date: 2015-01-09
Filing date: 2015-05-27
Publication date: 2018-05-08
Also published as: KR20160086730A

Abstract

The present disclosure relates to a method for predicting the cardiovascular disease risk using cardiovascular disease risk factors including the age of the subject, diabetes mellitus, hypertension hypertension, current smoking, family history of CHD, blood pressure, cholesterol level, white blood cell count, creatinine level, Obtaining an index value for a glycated hemoglobin level, and atrial fibrillation factors; Assigning a risk prediction score corresponding to each measured value to each factor; And a step of matching the total score of the risk predictive scores of each factor with the probability of cardiovascular disease events within the predicted time period to estimate the risk of cardiovascular disease using the cardiovascular risk factors .

Description

FIELD OF THE INVENTION [0001] The present invention relates to a method for predicting the risk of cardiovascular disease using cardiovascular risk factors,

Disclosure relates generally to methods for predicting cardiovascular risk using cardiovascular risk factors, and more particularly to methods for predicting cardiovascular risk using cardiovascular risk factors with good applicability and high predictability of cardiovascular risk will be.

Herein, the background art relating to the present disclosure is provided, and these are not necessarily meant to be known arts.

Cardiovascular disease (CVD) has been growing rapidly in Asian countries over the last century, with about half of all CVD patients worldwide taking place in the region. Therefore, prevention of CVD in Asia, including Korea, is an important issue in world health.

CVD risk factors are very similar to those of coronary heart disease (CHD). As a cardiovascular risk prediction model, the standard Framingham risk model developed in the West uses risk factors such as age, gender, diabetes mellitus, smoking, blood pressure, and cholesterol. However, the standard Framingham risk model lacks sufficient reflection on various biomarker factors, and a model for predicting cardiovascular risk using these biomarkers for Koreans has not yet been developed.

This will be described later in the Specification for Implementation of the Invention.

SUMMARY OF THE INVENTION Herein, a general summary of the present disclosure is provided, which should not be construed as limiting the scope of the present disclosure. of its features).

According to one aspect of the present disclosure, in a predicting method of cardiovascular disease risk using cardiovascular disease risk factors, , Age of the subject, diabetes mellitus, hypertension, current smoking, family history of CHD, blood pressure, cholesterol level, Each index value is obtained for the white blood cell count (WBC count), the creatinine level, the glycated hemoglobin level, and the atrial fibrillation factors. ; Assigning a risk prediction score corresponding to each measured value to each factor; And a step of matching the total score obtained by summing the risk prediction scores of each factor with the probability of cardiovascular disease events within the predicted period to estimate the risk of cardiovascular disease using the cardiovascular risk factors / RTI >

According to another aspect of the present disclosure, in a predicting method of cardiovascular disease risk using cardiovascular disease factors, Obtaining a standard risk factor having a statistical significance in the onset of cardiovascular disease and a set of index values of a biomarker factor; And a step of extracting a probability of occurrence of a cardiovascular disease within a predicted period using a risk prediction model in which a set of measured values is matched to a probability of occurrence of a cardiovascular disease within a predicted period. Methods of predicting disease risk are provided.

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram for explaining a target group of a prediction model development,
Figure 2 is a drawing of a prognostic nomogram to illustrate an example of a method for predicting cardiovascular disease risk according to the present disclosure;
Table 1 in FIG. 3 shows the predictor parameters,
FIG. 4 is a table for explaining the best-fitting predictive model and the performance of the Framingham Risk Model in Table 2,
FIG. 5 is a diagram for explaining Table 3 showing the performance of the Best-fitting Predictive Model and the Framingham Risk Model in the verification group,
FIG. 6 is a view for explaining Table I showing the importance scores of variables for CVD events determined using the Random Forest Method; FIG.
Figures 7 to 10 are illustrations of Table II illustrating the performance of possible prediction models for CVD events in a train population,
11 and 12 are diagrams for explaining Table III showing the performance of the prediction models available in the verification group,
13 is a view for explaining Table IV showing reclassification of objects according to possible prediction models,
FIG. 14 is a view for explaining Table V showing the Net Reclassification Index and Integrated Discrimination Improvement compared with the Framingham Risk Model; FIG.
FIG. 15 is a diagram for explaining an example of an automatic calculator for the risk of developing cardiovascular disease in Koreans within 3 years and 5 years by a method of predicting cardiovascular disease risk according to the present disclosure; FIG.

The present disclosure will now be described in detail with reference to the accompanying drawings.

In a predicting method (hereinafter referred to as a method for predicting cardiovascular disease risk) using cardiovascular disease (CVD) risk factors according to the present disclosure, Age, diabetes mellitus, hypertension, current smoking, family history of CHD, white blood cell count (WBC count), creatinine level (creatinine the index value or the observed value for the blood pressure, the glycated hemoglobin level, the atrial fibrillation, the blood pressure, and the cholesterol level factors, . Then, a risk prediction score corresponding to each measured value is given to each factor. Then, the total score of the risk predictive scores of each factor is matched to the probability of cardiovascular disease events (CVD events) within the predicted period.

For example, in a method for predicting cardiovascular risk, standard risk factors with statistical significance in CVD events, and biomarker factors, are obtained. The probability of cardiovascular disease incidence is then extracted within the prediction period using a risk prediction model that matches the set of measures to the probability of CVD events within the prediction period.

For example, standard risk factors consisting of age, diabetes, hypertension, smoking, family history of coronary artery disease, blood pressure, and cholesterol levels, leukocyte levels, creatinine levels, glycated hemoglobin levels, and atrial fibrillation We developed a best-fitting prediction model of cardiovascular disease risk using biomarker factors and used this optimal prediction model for a certain period of time 5 years), the probability of a CVD event is extracted from a person. Therefore, cardiovascular risk management can be performed according to these predictions. Using statistical means, we made predictive models with many factors, verified the factors that are important for the prediction of cardiovascular disease, compared the various prediction models, and selected optimal prediction models with excellent performance. These significant factors included biomarker factors as well as standard risk factors, as described above. Hereinafter, the method of predicting the cardiovascular disease risk, that is, the characteristics of the optimal prediction model, will be described through the description of the development and verification of prediction models.

Methods and Results

FIG. 1 is a diagram for explaining the target group of the prediction model development. In the group used for developing the optimal prediction model, 57,393 persons were diagnosed as Korean until the age of 30-80 without symptoms, and those who had no history of cardiovascular disease . Subjects were randomly divided into train cohort (n = 45,914) and validation cohort (n = 11,479). Possible 31 risk factors were assessed as CVD events predictors and predictive models were developed using statistically significant risk factors. In this disclosure, CVD events have been defined as a composite cardiac event of cardiovascular death, myocardial infarction, and stroke. To develop an optimal prediction model, C-index (95% confidence interval; 95% confidence interval) and Akaike Information Criterion (AIC) were used in the train population. The C-index and AIC are the means for evaluating the relative quality of the statistical model for a given set of data. The C-index and AIC compare each model to other models to assess the quality of each model. Thus, the C-index and AIC provide a means of model selection. The ratio of predicted values to observed values for CVD events in the validation group was compared by C-index and Nam D'Agostino X ² statistics. During a median follow-up period of 3.1, interquartile range, 1.9-4.3, 474 of the 458 subjects experienced CVD events.

FIG. 2 is a diagram for explaining an example of a prognostic nomogram for explaining an example of a method for predicting cardiovascular disease risk according to the present disclosure. In a train population, 11 risk factors, namely, age , C-index = 0.757 [0.726-0.788], and AIC = 7,207 using diabetes, hypertension, smoking, family history of coronary artery disease, leukocyte levels, creatinine level, glycated hemoglobin level, atrial fibrillation, blood pressure, and cholesterol level A prediction model was created. This optimal prediction model was tested in the validation group and worked well in terms of discrimination and calibration abilities (C-index = 0.760 [0.693-0.828], Nam and D'Agostino X ² statistic = 0.001 for 3 years; C-index = 0.782 [0.719-0.846], Nam and D'Agostino X ² statistic = 1.037 for 5 years).

1 and 2 will be further described below.

Data Sources

1, a total of 91,636 Koreans, age 30-80, were enrolled from January 2007 to June 2011 for those who underwent general health screening at Asan Medical Center, Seoul, Korea. Of these, 65,739 (72%) agreed to participate. If there is a CVD history (codes I00-99 in the International Classification of Diseases, 10th Revision) in the Health Insurance Review & Assessment Service (HIRA) database before the index day, or if the data is not available in the HIRA database . HIRA is a semi-governmental organization that systematically reviews medical systems to minimize unnecessary health care.

Korea has a National Health Insurance (NHI) system, and all healthcare providers are obligated to join this system. All NHI claims data are reviewed by HIRA. For the development of the predictive model according to the present disclosure, the HIRA database was adopted until December 2011. Patients with prior history of an index (eg, angina, MI, stroke, structural heart disease, percutaneous coronary intervention (PCI), previous cardiac procedure, or open heart surgery) were also excluded. Accordingly, as shown in Fig. 1, finally, 57,393 members were registered. The basic demographic data was obtained from a database maintained at the Asan Medical Center's Health Screening and Promotion Center. A systematic questionnaire prior to a general health examination revealed a history of previous medical history of angina (MI, stroke, structural heart disease, coronary revascularization, previous cardiac procedure, open heart surgery, hypertension, diabetes mellitus, hyperlipidemia, family history of CHD, smoking status, physical activity, education status, etc.) were collected.

Family history of CHD was defined as a CHD that occurred in a first-degree relative of any age. Physical activity is classified according to the frequency of exercise per week (eg> 5 days, 3-5 days, <3 days, and none), depending on the duration (eg> 1 hour, 20 minutes to 1 hour, and <20 minutes. In addition, the education status was classified into four levels (eg,> college or university graduate, college or university graduate, high school graduate, and ≤middle school graduate). In general health examinations, body values (eg body weight, body mass index, waist circumference, blood pressure) were measured and an electrocardiogram was performed. In addition, a complete blood count (erythrocyte sedimentation rate, fasting plasma glucose, glycated hemoglobin, uric acid, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, triglyceride, blood urea nitrogen, serum creatinine, ). The aspirin or statin medication history was obtained from the HIRA database.

CVD events are defined as a combination of cardiovascular death, myocardial infarction (MI), and stroke. By December 31, 2011, the deaths were confirmed by matching information to the death record. To this end, the death certificate of the National Statistical Office was confirmed in accordance with the resident registration number. The cause of death of the National Statistical Office is coded according to ICD-10. These recorded causes of death were used to develop predictive models. In the development of prediction models, deaths were classified as cardiovascular if the ICD-10 codes I00-99 and R96 were recorded, and the remaining ICD-10 codes were classified as non-cardiovascular. Myocardial infarction (MI) and stroke were identified using the hospitalization database of HIRA (ICD-10 codes I21-22 and I60-69).

Statistical Analysis

Subjects were randomly divided into two groups (train cohort, validation cohort). Predictive models were developed in the train cohort (n = 45,914) and tested in the validation cohort (n = 11,479). When multiple events occur, only the first occurrence is considered as CVD events. Statistical analysis was performed using the time of the first episode. A Cox proportional hazards model was used to construct a predictive model to assess the CVD event rate.

Table 1 in FIG. 3 shows the predictor parameters. To select prognostic factors, the random survival forest method was used and all covariates were included. Then, importance scores were calculated. Clinically or statistically insignificant variables were removed one by one in the building process of the predictive model.

Several candidate prediction models for CVD events prediction were then developed and compared. To avoid multicollinearity between the selected variables, the variation inflation factor was calculated. If this number is ≥ 5, then multi-collinearity is considered to exist. In this disclosure, two index variables representing cholesterol and blood pressure (two index variables representing patients' cholesterols and blood pressures) were produced. There is a close association between low-density cholesterol, high-density cholesterol, and total cholesterol, systolic pressures, and diastolic pressures as well. Since it has been anticipated to be involved, the parameters for cholesterol and blood pressure were created using statistical methods as follows.

blood pressure index = 0.5 * ([systolic pressure-115] / 15) +0.5 * ([diastolic pressure-75] / 100);

cholesterol index = 0.19 * ([low-density lipoprotein cholesterol-120] / 30) - 0.25 * ([high-density lipoprotein cholesterol-55] / 14) +0.16 * ([total cholesterol-190] / 34).

The proportional hazards assumption was examined using a scaled Schoenfeld residual test. The overall goodness-of-fit was tested by the Cramer-von-Mises test. The Cramer-von-Mises test can be used to examine the linearity assumption of predictors in a multivariable model. In addition, overfitting has been evaluated by calculating the slope shrinkage estimate, although over-fitting is not a problem in large target groups as in the present disclosure. That is, if the slope shrinkage was exceeded (0.85), the prediction model was considered overfitting.

In addition, the discriminant power of the predictive model and the discrimination and calibration abilities were measured by calculating the C-index (concordance index) and the Hosmer-Lemeshow type statistic (Nam and D'Agostino X ² statistic). The Best-fitting Predictive Model is based on the minimal loss of predictive ability and has clinical simplicity compared to the maximal-fitting model using all variables based on a score table. It was badly invented. After fitting several candidate prediction models, AIC values were also determined. To compare the performance of the predictive models for Framingham risk models, the predictive models were created using published equations, a net reclassification index, and integrated discrimination improvement were evaluated. For the development of a user-friendly nomogram, the rms package developed by Harrell in software (version 2.15.1) was used. All data analysis was performed using R software (version 2.15.1; R Foundation for Statistical Computing, Vienna, Austria). All reported p-values were two-sided and were considered statistically significant if the p-value was less than 0.05 (<0.05).

Results

Population Characteristics and Outcomes

Table 1, presented in Figure 3, shows the baseline characteristics of the target population. The mean age of the subjects was 42.1 ± 8.1 years, 33,727 (58.8%) were male, 5,653 (9.8%) were diabetic patients, hypertension patients, hyperlipidemia patients and smokers were 14,179 (24.7%) and 27,202 %) And 12,909 (24.0%), respectively. And 126 (0.2%) had hsCRP ≥ 2 mg / dL.

Of the 458 subjects, 474 underwent 35 cardiovascular deaths (79 MIs, and 360 strokes) during a median follow-up period of 3.1, interquartile range, 1.9-4.3.

FIG. 4 is a table for explaining Table 2 showing the performance of the Best-fitting Predictive Model and the Framingham Risk Model in a train group, FIG. 5 is a graph showing the best-fitting Predictive Model Table 3 illustrates the performance of the Framingham Risk Model.

4 and 5 will be further described below.

Cardiovascular Disease Predictors and Risk Model Development (Cardiovascular Disease Predictors and Risk Model Development)

FIG. 6 is a view for explaining Table I showing the importance score of the variables for the CVD events determined using the Random Forest Method. FIG. 31 potential variables were evaluated in the train cohort for inclusion in the predictive model. In the univariate analysis, age was the single best predictor of cardiovascular disease (HR = 1.086, 95% CI = 1.075-1.097, p <0.001). The variable importance scores were chosen as possible CVD event determinants for the random forest method (see Table 6), where the relative importance scores> 0.1. Age had the highest significance score.

Figures 7-10 are illustrations of Table II illustrating the performance of possible models for CVD events in a train population. Possible prediction models for CVD events were gradually developed using the C-index (95% CI) of the variables and their parameters, and the AIC values were calculated. The best-fitting predictive model was predictive model 11. Predictive model 11 was a predictive model made using 11 variables: age, diabetes, hypertension, smoking, family history of coronary artery disease, leukocyte count, creatine level, glycated hemoglobin level, atrial fibrillation, blood pressure, cholesterol level Table 2). For CVD events in predictive model 11, the C-index was 0.757 (0.726-0.788) and the AIC was 7,207.

Risk Model Validation

11 and 12 are diagrams for explaining Table III showing the performance of the prediction models available in the verification group. The table summarizes the performance of the prediction models in the validation cohort. The optimal predictive model (Model 11) was tested in the validation group, and the optimal predictive model showed good performance in terms of discrimination and scale scaling for CVD events. In other words, C-index = 0.760 [0.693-0.828 ], and ^{Nam and D'Agostino X 2 statistic =} 0.001 for 3 years, and the C-index = 0.782 [0.719-0.846] , and Nam and D'Agostino X ² statistic = 1.037 for 5 years (see Table 3 in FIG. 5). The Nam and D'Agostino X ² statistic for the optimal prediction model was smaller than those for the other models, indicating a better match between the risk estimate and the observed risk.

Risk Reclassification and Comparison with the Framingham Risk Model (Risk Reclassification and Comparison with the Framingham Risk Model)

FIG. 13 is a view for explaining Table IV showing the reclassification of objects according to the possible prediction models. Table shows a list of several persons classified into different risk groups by different prediction models. Depending on the forecasting models, the risk varies by 19 percent.

The Framingham risk model shows comparable discriminating and calibrating abilities in the validation cohort in Table 2 in Figure 4, train cohort in Figure 3, and Table 3 in Figure 5 in terms of discrimination and scale scaling.

FIG. 14 is a diagram for explaining the Table V showing the Net Reclassification Index and the Integrated Discrimination Improvement compared with the Framingham Risk Model, which is evaluated by a net reclassification index and integrated discrimination improvement. , The optimal prediction model is no less than the Framingham risk model and has a comparable predictive ability.

Clinical Application

According to this disclosure, a prognostic nomogram was developed that assessed the risk of CVD events (see FIG. 2). There is a point line on the upper side of the nomogram and below it there are the indexes of age, diabetes, hypertension, smoking, family history of coronary artery disease, leukocyte level, creatine level, glycated hemoglobin level, atrial fibrillation, blood pressure, cholesterol level (index). Each indices of these eleven factors, or a line up from each observation, assign the corresponding score in the score line to the risk predictive score of that factor. The total score of 11 risk prediction scores is shown on the total dotted line below the nomogram. Below the total dashed line,% probability of CVD event corresponding to the total score is displayed within 3 years, with probability within 5 years.

In the method of predicting the risk of cardiovascular disease according to the present example, in the process of assigning the risk prediction score to each factor, a nomogram matching the measured value range of each factor is used at least in part of the score line having the lowest and highest points In the process of matching the total score to the probability, the nomogram is used which matches the total score of the probability of occurrence of cardiovascular disease and the risk prediction score of each factor within the prediction period.

In this example, the start point of the range of the measured values of each factor in the nomogram is matched to correspond to the lowest point of the score line, and the range of the measured value of at least one of the ranges of the measured values of the factors, And this length may reflect the weight or importance of each factor. In this example, the length of the range of the measurement value in the nomogram is length in the order of age> cholesterol level = blood pressure> leukocyte count> glycosylated hemoglobin level> creatine level> atrial fibrillation> smoking> diabetes> family history. In nomogram, score line length = length of age measurement range = total length of line. Values for creatine levels, leukocyte counts, and HbA1c values are displayed in the nomograms.

In this example, the probability of developing a cardiovascular disease within 1 year is 1% -20% for the first dotted line segment (eg, total score 124-204) in the nomogram, and the second dotted dotted line segment (total score 100-214) The probability of developing cardiovascular disease within 5 years is 1% -50%.

FIG. 15 is a diagram for explaining an example of an automatic calculator for the risk of developing cardiovascular disease in Koreans within three years and five years by the method for predicting cardiovascular disease risk according to the present disclosure, The prognostic scores may be automatically calculated by the calculator, as shown in FIG.

Conclusion (DISCUSSION)

The predictive method (or predictive model) of cardiovascular risk according to this disclosure was developed to track 57,393 asymptomatic Koreans and track 3.1 (interquartile range, 1.9-4.3) years. The major findings in this disclosure are as follows: (1) the incidence of cardiovascular events is low in asymptomatic Koreans, (2) age among all risk factors is strongly associated with cardiovascular disease inventions, and (3) (4) standard risk factors, and predictions of cardiovascular risk according to the present disclosure based on biomarker factors. The model (the above-mentioned optimal predictive model = predictive model 11) showed excellent performance in predicting the incidence of cardiovascular disease in Koreans, and in particular the possibility of being applicable to East Asian people who share similar risk factors for cardiovascular disease .

The strength of this disclosure is that it was developed using reliable, well-controlled and reliable data from government or quasi-government organizations (the National Statistical Office and HIRA databases in Korea).

Important factors in the predictive model of cardiovascular risk according to this disclosure are age, diabetes, hypertension, smoking, family history of coronary artery disease, leukocyte levels, creatine levels, glycated hemoglobin levels, atrial fibrillation, blood pressure, cholesterol levels. Among these, age was the strongest predictor. This is not surprising because as age increases, it significantly affects the heart and cardiovascular system and significantly increases the risk of cardiovascular disease such as atherosclerosis, hypertension, stroke, MI, and atrial fibrillation. In contrast, other risk factors contribute much less to the risk of cardiovascular disease prediction.

In this example, to improve the predictive power of predictive models of cardiovascular risk, blood pressure and cholesterol were expressed as indices rather than as categories. By doing so, the potential multicollinearity is prevented by the correlation between blood pressure and cholesterol. In the final model, these cholesterol and blood pressure indices were independently associated with CVD events.

Various embodiments of the present disclosure will be described below.

(1) Cardiovascular Disease In the predicting method of cardiovascular disease risk using risk factors, the age, diabetes mellitus, hypertension hypertension, current smoking, family history of CHD, blood pressure, cholesterol level, white blood cell count (WBC count), creatinine level Obtaining an index value for each of the atrial fibrillation factors, the glycated hemoglobin level, and the atrial fibrillation factors; Assigning a risk prediction score corresponding to each measured value to each factor; And matching the total score of risk predictors of each factor to the probability of cardiovascular disease events within the predicted period.

(2) In a step of assigning a risk prediction score to each factor, a nomogram matching a measured value range of each factor is used for at least a part of a score line having a lowest point and a highest point, Wherein said normogram is used to match the sum of the probability of onset of cardiovascular disease and the risk predictive score of each factor within the predicted period, to predict cardiovascular risk using cardiovascular risk factors.

(3) standard risk factors consisting of age, diabetes, hypertension, smoking, family history of coronary artery disease, blood pressure, and cholesterol level, A method for predicting the risk of cardiovascular disease using a cardiovascular risk agent, characterized by obtaining each measurement of biomarker factors consisting of leukocyte count, creatine level, glycosylated hemoglobin level, and atrial fibrillation.

(4) The starting point of the range of the measured values of each factor in the nomogram is matched to correspond to the lowest point of the score line, and the range of the measured value of at least one of the ranges of the measured values of the factors is A method for predicting cardiovascular risk using cardiovascular risk factors characterized by another.

(5) The length of the range of the measurement value in the nomogram is the cardiovascular risk factor characterized by having a length in the order of age> cholesterol level = blood pressure> leukocyte count> glycosylated hemoglobin level> creatine level> atrial fibrillation> smoking> diabetes> Methods for the Risk Assessment of Cardiovascular Disease.

(6) cardiovascular disease probability is matched within 3 years for the first section of the total score in the Nomogram, and the probability of cardiovascular disease occurrence within 5 years is matched for the second section of the total score in the Nomogram Methods for predicting cardiovascular risk using risk factors.

(7) a method for predicting the risk of cardiovascular disease using a cardiovascular risk factor characterized in that, in the nomogram, the score line length = the length of the age measurement range = the total score line length.

(8) In the predicting method of cardiovascular disease risk using cardiovascular disease factors, standard risk with statistical significance in cardiovascular disease events Obtaining a set of measured values of a standard risk factor and a biomarker factor; And a step of extracting a probability of occurrence of a cardiovascular disease within a predicted period using a risk prediction model in which a set of measured values is matched to a probability of occurrence of a cardiovascular disease within a predicted period. Methods for predicting disease risk.

(9) Standard risk factors and biomarker factors are age, diabetes mellitus, hypertension, current smoking, family history of CHD, blood pressure, a cardiovascular risk factor characterized by a blood pressure, a cholesterol level, a white blood cell count, a creatinine level, a glycated hemoglobin level, and atrial fibrillation Methods for predicting cardiovascular risk.

(10) extracting the probability includes: providing a standard risk factor, and a risk prediction score corresponding to each measurement value obtained for the biomarker factor, to each factor; And a step of matching the total score of the risk predictive scores of each factor with the probability of occurrence of cardiovascular disease within the predicted period.

(11) In the process of assigning the risk prediction score to each factor, a nomogram matching the range of the measured values of the respective factors is used in at least a part of the score line having the lowest point and the highest point, Wherein said normogram is used to match the sum of the probability of onset of cardiovascular disease and the risk predictive score of each factor within the predicted period, to predict cardiovascular risk using cardiovascular risk factors.

According to the method for predicting cardiovascular risk according to the present disclosure, a predictive model of cardiovascular risk that can assess the risk of a subject's CVD events is used to provide a probability of CVD events within a predicted period.

Claims

Cardiovascular Disease In the predicting method of cardiovascular disease risk using risk factors,
A standard consisting of the age of the subject, diabetes mellitus, hypertension, current smoking, family history of CHD, blood pressure and cholesterol level There are standard risk factors and biomarker factors consisting of white blood cell count, creatinine level, glycated hemoglobin level, and atrial fibrillation. obtaining an index value for each factor;
Assigning a risk prediction score corresponding to each measured value to each factor, the risk prediction score reflecting the weight or importance of each factor to each factor; And
And comparing the total score of the risk predictors of each factor to the probability of cardiovascular disease events within the predicted period.

The method according to claim 1,
In the step of assigning the risk prediction score to each factor,
A nomogram is used in which a range of measured values of each factor is matched to at least a part of the score line having the lowest point and the highest point,
In the step of matching the probabilities,
Wherein said normogram is used to match the sum of the probability of onset of cardiovascular disease and the risk predictive score of each factor within the predicted period, to predict cardiovascular risk using cardiovascular risk factors.

delete

The method of claim 2,
The start point of the range of the measured values of each factor in the nomogram is matched to correspond to the lowest point of the score line,
A method for predicting the risk of cardiovascular disease using a cardiovascular risk factor characterized in that the range of the measured value of at least one of the ranges of the measured values of the parameters is different from the range of the measured values of the other factors.

The method of claim 2,
The length of the measurement range in the nomogram is
Method for predicting cardiovascular risk using cardiovascular risk factors characterized by age> cholesterol level = blood pressure> leukocyte count> glycated hemoglobin level> creatine level> atrial fibrillation>smoking>diabetes> family history.

The method of claim 2,
The probability of developing cardiovascular disease within 3 years is matched against the first segment of total score in the nomogram,
A method for predicting the risk of cardiovascular disease using a cardiovascular risk factor characterized in that the probability of developing cardiovascular disease within 5 years is matched to the second section of the total score in the nomogram.

The method of claim 2,
In the nomogram,
Score line length = length of age measurement value range = total score line length.

delete