LDL(low-density lipoprotein) cholesterol estimation formula and estimation method thereof

Abstract

The present invention relates to a precise estimation formula for estimating low-density lipoprotein cholesterol and a method for deriving the same. According to a specific formula, it is possible to estimate LDL cholesterol levels with high accuracy, thereby providing highly accurate LDL levels without direct measurement of LDL cholesterol, and consequently, enabling the prediction of cardiovascular disease.

Description

Technology Field

[0001] The present invention relates to a precise estimation formula for estimating low-density lipoprotein cholesterol and an estimation method using the same, and provides a novel method for calculating LDL cholesterol levels, thereby enabling the provision of accurate LDL cholesterol levels. Background Technology

[0003] Low-density lipoprotein cholesterol (LDL cholesterol) is one of the lipoproteins that transport cholesterol in the blood; it is known to accumulate on the walls of blood vessels, causing health problems such as heart attacks and strokes. High LDL cholesterol levels can lead to issues such as hypercholesterolemia, dyslipidemia, atherosclerosis, and angina. Accordingly, LDL cholesterol levels are measured to assess an individual's risk, and generally, a concentration of less than 100 mg / dL is considered optimal. Total cholesterol, HDL cholesterol, triglycerides, and LDL cholesterol are measured to evaluate bodily abnormalities, specifically the risk of heart disease.

[0004] Although LDL cholesterol can be measured directly, the method is complex and requires significant cost and time. Therefore, a method of estimating concentration from other factors is used, and a formula generally known as the Friedwald equation is used to evaluate LDL cholesterol. The Friedwald formula is an estimation formula derived from the concentrations of total cholesterol, HDL cholesterol, and triglycerides, and is calculated according to the following equation.

[0005] Friedwald Formula: Low-density cholesterol (LDL) = Total cholesterol - HDL cholesterol - Triglycerides / 5

[0006] However, this formula is valid only when triglycerides are less than 400 mg / dL and has a problem in that it applies a value divided by 5 uniformly to everyone. Since various factors such as smoking, age, hypertension, family history of heart disease, and diabetes can play a role in the cholesterol assessment for each individual, modifications to the method for evaluating LDL cholesterol are required accordingly. Due to the problems with this formula, modified formulas based on various studies are being developed to evaluate LDL cholesterol; however, despite the use of these formulas, there is still a problem where there is a slight difference between the estimated values ​​and the direct measurement results.

[0007] In this regard, the following formulas have been disclosed by Martin, Hattori, Chen, Puavilai, and Choi, respectively, as modified formulas.

[0008]

[0009] However, even when using the modified formula, there is a problem in that accuracy varies significantly depending on country, ethnicity, genetics, environment, race, etc., and it fails to provide accurate values ​​based on the concentrations of total cholesterol, HDL cholesterol, and triglycerides.

[0010] Accordingly, the inventors discovered a new formula for estimating LDL cholesterol and confirmed that its error is low, thereby completing the present invention. Prior art literature

[0012] Yuichi Hatoori et al., Development of approximate formula for LDL-chol, LDL-apo B and LDL-chol / LDL-apo B as indices of hyperapobetalipoproteinemia and small dense LDL, Atherosclerosis 138 (1998), 289-299.Yunqin Chen et al., A modified formula for calculating low-density lipoprotein cholesterol values, Lipids in Health and Disease 2010, 9:52Rihwa Choi et al., Validation of multiple equations for estimating low-density lipoprotein cholesterol levels in Korean adults, Lipids in Health and Disease (2021) 20:111. The problem to be solved

[0013] The present invention provides a novel method for calculating LDL cholesterol levels to measure LDL cholesterol levels with greater accuracy and, accordingly, to provide more accurate information to the subject. means of solving the problem

[0015] To achieve the above objective, the present invention provides a method for predicting cardiovascular disease in a subject, comprising the following steps.

[0016] (a) A step of measuring high-density lipoprotein cholesterol (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC; Total-cholesterol) in blood separated from a subject;

[0017] (b) a step of measuring low-density lipoprotein cholesterol according to Formula 1 using three components: high-density lipoprotein cholesterol (HDLC), triglycerides (TG), and total cholesterol (TC);

[0018] Formula 1: LDLC = 0.94 × Total Cholesterol (TC) - 0.94 × High-Density Lipoprotein Cholesterol (HDLC) - 0.12 × Triglycerides (TG)

[0019] (c) A step of determining the level of low-density lipoprotein cholesterol according to the judgment criteria in accordance with step (b) above.

[0020] In one embodiment of the present invention, the cardiovascular disease is selected from the group consisting of hypercholesterolemia, hyperlipidemia, arteriosclerosis and atherosclerosis.

[0021] In one aspect of the present invention, the cardiovascular disease is hyperlipidemia.

[0022] In one aspect of the present invention, in step (a), high-density lipoprotein cholesterol, triglycerides, and total cholesterol are measured by an enzymatic method.

[0023] In one embodiment of the present invention, in step (c), the low-density lipoprotein cholesterol judgment criteria are (1) adequate: < 100 mg / dL; (2) normal: 100~129 mg / dL; (3) borderline: 130~159 mg / dL; (4) high: 160~189 mg / dL; and (5) very high: ≥ 190 mg / dL.

[0024] In one embodiment of the present invention, Formula 1 of step (b) is applied to Mongoloids.

[0025] In one embodiment of the present invention, the measurement method of step (b) has a misclassification rate of 20% or less.

[0026] In one aspect of the present invention, the measurement method of step (b) is the coefficient of determination (R 2 ; Coefficient of Determination) is 0.9 or higher.

[0027] In one embodiment of the present invention, the measurement method of step (b) has a Root Mean Square Error (RMSE) of 8.0 or less.

[0029] In addition, the present invention provides a method for providing the level of low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol of a subject according to Formula 1 below, using three components of high-density lipoprotein cholesterol (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC; Total-cholesterol) of the subject.

[0030] Formula 1: LDLC = 0.94 × Total Cholesterol (TC) - 0.94 × High-density lipoprotein cholesterol (HDLC) - 0.12 × Triglycerides (TG).

[0032] In addition, the present invention

[0033] (a) A step of selecting public big data to derive an estimation formula for low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol;

[0034] (b) a step of selecting groups of low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol, high-density lipoprotein (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC; Total-cholesterol) from the big data of step (a) above;

[0035] (c) A step of deriving and comparing low-density lipoprotein cholesterol estimation formulas using machine learning;

[0036] (d) a step of verifying the validity of the value of low-density lipoprotein cholesterol estimated in step (c); a method for deriving a low-density lipoprotein cholesterol estimation formula is provided.

[0037] In one aspect of the present invention, step (a) is to select the Korea National Health and Nutrition Examination Survey (KNHANES) data as public big data.

[0038] In one aspect of the present invention, step (c) derives an estimation formula through the least squares method.

[0039] In one aspect of the present invention, step (c) compares the estimation equation through hyperparameters via grid search of machine learning.

[0040] In one aspect of the present invention, step (d) is the Root Mean Square Error (RMSE) and the coefficient of determination (R 2 It is to verify using one or more methods selected from a group composed of ; Coefficient of Determination).

[0041] In addition, in one embodiment of the present invention, step (d) is to verify based on NCEP-ATP III 6 categories.

[0042] In addition, the present invention provides a computer-readable recording medium on which a program for implementing the above method is recorded. Effects of the invention

[0044] The present invention has the advantage of being able to determine LDL levels with high accuracy without direct measurement of LDL cholesterol by discovering a new formula for estimating LDL cholesterol and revealing that the error is significantly low. Brief explanation of the drawing

[0046] Figure 1 is a figure showing a design for deriving an estimation formula according to the present invention. Figures 2a to 2c show the misclassification of LDL cholesterol values ​​according to NCEP ATP III criteria in a test data set. Figures 3a through 3n illustrate the correlation between actual LDL cholesterol values ​​and estimated LDL cholesterol values ​​in a cross-validation dataset. ((A) Friedewald, (B) DeLong, (C) Rao, (D) Hattori, (E) Anandaraja, (F) Puaviai, (G) Vujovic, (H) Chen and Zhang, (I) de Cordova, (J) Martin, (K) Sampson, (L) Choi, and (M) Formula of the present invention and (N) Model of the present invention) Figures 4a through 4n show confusion matrices for estimated LDL cholesterol values. ((A) Friedewald, (B) DeLong, (C) Rao, (D) Hattori, (E) Anandaraja, (F) Puaviai, (G) Vujovic, (H) Chen and Zhang, (I) de Cordova, (J) Martin, (K) Sampson, (L) Choi, and (M) Formula of the present invention and (N) Model of the present invention) Specific details for implementing the invention

[0047] Terms and words used in this specification and claims are not limited to their ordinary or dictionary meanings, and must be interpreted in a meaning and concept consistent with the technical spirit of the invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention.

[0048] Throughout the specification, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, terms such as "…part," "…unit," "module," and "device" as used in the specification refer to a unit that processes at least one function or operation, which may be implemented as a combination of hardware and / or software.

[0049] Throughout the specification, the term “and / or” should be understood to include all combinations that can be presented from one or more related items. For example, the meaning of “Item 1, Item 2 and / or Item 3” is not only Item 1, Item 2 or Item 3, but also any combination of items that can be presented from two or more of Items 1, Item 2 or Item 3.

[0050] Throughout the specification, identification codes (e.g., a, b, c, ...) for each step are used for convenience of description and do not limit the order of the steps; the steps may occur in a different order than specified unless a specific order is clearly indicated in the context. That is, the steps may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

[0052] This relates to a method for predicting cardiovascular disease in subjects, comprising the following steps.

[0053] (a) A step of measuring high-density lipoprotein cholesterol (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC; Total-cholesterol) in blood separated from a subject;

[0054] (b) a step of measuring low-density lipoprotein cholesterol according to Formula 1 using three components: high-density lipoprotein cholesterol (HDLC), triglycerides (TG), and total cholesterol (TC);

[0055] Formula 1: LDLC = 0.94 × Total Cholesterol (TC) - 0.94 × High-Density Lipoprotein Cholesterol (HDLC) - 0.12 × Triglycerides (TG)

[0056] (c) A step of determining the level of low-density lipoprotein cholesterol according to the judgment criteria in accordance with step (b) above.

[0058] In one embodiment of the present invention, the cardiovascular disease is selected from the group consisting of hypercholesterolemia, hyperlipidemia, arteriosclerosis and atherosclerosis.

[0059] In one aspect of the present invention, the cardiovascular disease is hyperlipidemia.

[0060] In one embodiment of the present invention, the high-density lipoprotein cholesterol, triglycerides, and total cholesterol of step (a) are measured by an enzymatic method.

[0061] In one embodiment of the present invention, in step (c), the low-density lipoprotein cholesterol judgment criteria are (1) adequate: < 100 mg / dL; (2) normal: 100~129 mg / dL; (3) borderline: 130~159 mg / dL; (4) high: 160~189 mg / dL; and (5) very high: ≥ 190 mg / dL.

[0062] In one embodiment of the present invention, the judgment criteria in step (c) further include criteria for judging high cholesterol levels, total cholesterol levels, and triglycerides.

[0063] The criteria for judging the above high cholesterol levels are (1) adequate: < 200 mg / dL; (2) borderline: 200-239 mg / dL; (3) high: ≥ 200 mg / dL.

[0064] The criteria for the above total cholesterol levels are (1) low: < 40 mg / dL; (2) appropriate: 40-59 mg / dL; (3) high: ≥ 60 mg / dL.

[0065] The criteria for the above triglyceride levels are (1) adequate: < 150 mg / dL; (2) borderline: 150-199 mg / dL; (3) high: 200-499 mg / dL; (4) very high: ≥ 500 mg / dL.

[0066] In addition, in one embodiment of the present invention, the judgment criteria in step (c) further include the presence or absence of smoking and the presence or absence of diabetes. In addition, in one embodiment of the present invention, the judgment criteria in step (c) may further include age (male ≥ 45 years, female ≥ 55 years), family history, and hypertension (systolic blood pressure 140 mmHg or higher or diastolic blood pressure 90 mmHg or higher).

[0068] In one embodiment of the present invention, the high-density lipoprotein cholesterol, triglycerides, and total cholesterol are measured by an enzymatic method.

[0069] In one aspect of the present invention, the calculation method is applied to Mongoloids. In a specific aspect of the present invention, the calculation method is applied to Koreans.

[0070] In one embodiment of the present invention, the calculation method has a misclassification rate of 20% or less. In a specific embodiment of the present invention, the calculation method has a misclassification rate of 19.2% or less.

[0071] In one aspect of the present invention, the calculation method is a coefficient of determination (R 2 ; Coefficient of Determination) is 0.9 or higher.

[0072] In one embodiment of the present invention, the calculation method has a Root Mean Square Error (RMSE) of 8.0 or less.

[0073] The calculation method according to the present invention does not show a significant difference in accuracy even if the triglyceride level increases to 200 mg / dL or higher, and can solve the problem of underestimation that existing formulas have.

[0075] In addition, the present invention relates to a method for providing the low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol level of a subject according to Formula 1, comprising three components of the subject's high-density lipoprotein cholesterol (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC; Total-cholesterol).

[0076] In addition, the present invention

[0077] (a) A step of selecting public big data to derive an estimation formula for low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol;

[0078] (b) a step of selecting groups of low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol, high-density lipoprotein (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC; Total-cholesterol) from the big data of step (a) above;

[0079] (c) A step of deriving and comparing low-density lipoprotein cholesterol estimation formulas using machine learning;

[0080] (d) a step of verifying the validity of the value of low-density lipoprotein cholesterol estimated in step (c); the present invention relates to a method for deriving a low-density lipoprotein cholesterol estimation formula comprising: (d) a step of verifying the validity of the value of low-density lipoprotein cholesterol estimated in step (c).

[0081] In one aspect of the present invention, step (a) involves selecting data from the Korea National Health and Nutrition Examination Survey (KNHANES) as public big data. Although public big data in Korea's healthcare sector includes data from the National Health Insurance Corporation (NHIS), the Health Insurance Review & Assessment Service (HIRA), and the Korea National Health and Nutrition Examination Survey, there are many limitations to conducting analysis using data from the National Health Insurance Corporation and the Health Insurance Review & Assessment Service. Specifically, due to high data usage costs, limited usage time, and the inconvenience of having to visit the National Health Insurance Corporation in person, the present invention utilizes data from the Korea National Health and Nutrition Examination Survey.

[0082] In the present invention, the cholesterol results from the National Health and Nutrition Examination Survey are conducted starting from the 3rd year of the 4th period (2009), and are used to derive a conversion formula to align the measurements from clinical testing institutions with the true values ​​based on the evaluation results. Since the data is based on a sample design targeting the entire population of Korea, it can be interpreted as an expanded result for the Korean people. Furthermore, since cholesterol results have been measured in the National Health and Nutrition Examination Survey data since 2009, the data from 2009 onwards can be selected as big data, and all big data collected from 2009 to the present can be utilized.

[0083] In addition, in the present invention, a new estimation formula can be derived based on measured values ​​rather than estimated values ​​calculated through the Friedewald formula for low-density cholesterol levels in National Health and Nutrition Examination Survey data.

[0084] In one embodiment of the present invention, step (c) derives an estimation formula through the least squares method. In the present invention, by setting low-density lipoprotein cholesterol, high-density lipoprotein cholesterol, triglycerides, and total cholesterol as parameters, a linear equation can be derived through the least squares method in step (c). Additionally, an optimal β value can be derived according to the least squares method.

[0085] In one aspect of the present invention, step (c) involves comparing estimation formulas using hyperparameters through grid search of machine learning. In the present invention, optimal hyperparameters can be derived by utilizing machine learning techniques to compare results between estimation formulas. Through this, an optimal estimation formula for low-density cholesterol levels can be derived.

[0086] In one aspect of the present invention, step (d) is the Root Mean Square Error (RMSE), coefficient of determination (R 2 The method is to verify using one or more methods selected from a group consisting of (Coefficient of Determination). Specifically, the accuracy of the estimation formula of the present invention can be evaluated by comparing it with an existing formula using the verification method according to the present invention, and the verification method can be evaluated using one or more, two or more, or three or more methods, and accordingly, the estimation formula with the highest accuracy can be verified.

[0087] In addition, in one embodiment of the present invention, step (d) is verified based on NCEP-ATP III 6 categories. Specifically, the method for calculating the numerical value of low-density lipoprotein cholesterol according to the present invention shows a misclassification rate of 18.9% (Weighted Kappa=0.919(0.003)) with respect to the NCEP-ATP III criteria.

[0088] In the present invention, the verification method may utilize two or more methods in combination for verification. For clinical application, the estimation formula may be estimated using the values ​​of observed variables, and

[0089] In addition, the present invention relates to a computer-readable recording medium on which a program for implementing the above method is recorded.

[0091] The present invention will be explained in detail below through examples and experimental examples.

[0092] However, the following examples and experimental examples are merely illustrative of the present invention, and the content of the present invention is not limited to the following examples and experimental examples.

[0094] <Example> Measurement of Low-Density Lipoprotein (LDL) Cholesterol

[0095] To estimate LDL cholesterol, raw data from the Korea National Health and Nutrition Examination Survey (2009–2019) containing measurements of triglycerides and other substances were utilized. Subjects were selected based on the availability of direct measurements for triglycerides (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL), and high-density lipoprotein cholesterol (HDL), which are used in the Friedewald formula; a total of 18,837 subjects were included in the study. For database cross-validation, the data were randomly split in a 7:3 ratio, and 5-fold cross-validation was used to derive and compare the results.

[0096] In addition, the optimal β value was derived by minimizing the error from the straight line or curve obtained from the actual data using the least squares method, and the optimal hyperparameters were found and constructed using XGBoost grid search as a prediction model utilizing machine learning techniques, and the following equation was derived by comparing the results between the estimation equations.

[0097] [Equation 1]

[0098] LDL Cholesterol = 0.94 × Total Cholesterol (TC) - 0.94 × High-Density Lipoprotein Cholesterol (HDL) - 0.12 × Triglycerides (TG)

[0100] Using the formula derived from this, the error between the actual LDL cholesterol value and the estimated LDL cholesterol value was evaluated. The error was measured using the Root Mean Square Error (RMSE) and the Coefficient of Determination (R²). 2 Comparisons were made using ). In addition, to verify the extent to which the estimation formula can be utilized in clinical practice, the degree of misclassification based on the LDL values ​​estimated according to the NCEP-ATP III 6 categories and the actual LDL values ​​was examined, and the degree of agreement was confirmed using the Weighted Kappa coefficient.

[0101] In addition, LDL values ​​were estimated using the existing estimation formula and the estimation formula of the present invention, and the results for the five datasets divided according to 5-fold cross-validation were R 2 , and were compared using RMSE. The following equation was used for the existing estimation formula.

[0102] - Friedewald style:

[0103] LDL Cholesterol = Total Cholesterol - HDL Cholesterol - Triglycerides / 5

[0104] - Martin style:

[0105] LDL Cholesterol = Total Cholesterol - HDL Cholesterol - Triglycerides / Adjustment Factor

[0106] - Chio style:

[0107] LDL Cholesterol = Total Cholesterol - 0.87 × HDL Cholesterol - 0.13 × Triglycerides

[0108] - Sampson formula:

[0109] LDL Cholesterol = Total Cholesterol / 0.948 - HDL Cholesterol / 0.971 - (Triglycerides / 8.56 + Triglycerides × Non-HDL Cholesterol / 2140 - Triglycerides 2 / 16100) - 9.44

[0110] In addition, a total of 12 equations were compared using [DeLong equation / Rao equation / Hattori equation / Anandaraja equation / Puaviai equation / Vujovic equation / Chen and Zhang equation / de Cordova equation].

[0112] Comparison results

[0113] The baseline characteristics and lipid profiles of the subjects are shown in Table 1 below. The total number of subjects was 18,837, consisting of a training dataset (13,185 subjects, 70.00%) and a test dataset (5,652 subjects, 30.00%), with 52.38% male and 47.62% female overall.

[0114]

[0116] In addition, the RMSE values ​​were compared using data sets divided according to 5-fold cross-validation, and the results are shown in Table 2.

[0117]

[0119] By comparing the results according to each formula, when the triglyceride (TG) level is less than 400 mg / dL, R when using the estimation formula according to the present invention when compared with the above formula of Friedewald, Choi, and Sampson 2All were the highest at 0.94, and the RMSE in the training dataset was (cv1 : 7.97; cv2 : 8.04; cv3 : 8.13; cv4 : 8.14; cv5 : 8.04), and the RMSE in the test dataset was 7.96, which was confirmed to have the lowest error when compared to the actual LDL cholesterol values ​​measured.

[0120] In addition, the misclassification rate was 20.7% for the Sampson formula (Weighted Kappa=0.912(0.003)), 43.0% for the Choi formula (Weighted Kappa=0.831(0.004)), 19.4% for the Martin formula (Weighted Kappa=0.917(0.003)), and 18.9% for the formula according to the present invention (Weighted Kappa=0.917(0.003)), with the misclassification rate being the lowest and the Weighted Kappa being the highest. It was confirmed that the results obtained by the formula according to the present invention significantly reduced the tendency of the Friedewald estimation formula to underestimate LDL cholesterol values, and were lower than the results of XGBoost (misclassification rate=19.0%(Weighted Kappa=0.917(0.003)).

[0122] Therefore, it was confirmed that the estimation formula according to the present invention can be usefully used for estimating low-density lipoprotein cholesterol, as it exhibits high accuracy in actual LDL cholesterol values ​​compared to existing formulas when estimating low-density lipoprotein cholesterol and also shows a low misclassification rate. In particular, it resolved the problem of existing formulas, which is that inaccuracy tends to increase as triglycerides (TG) increase above 200 mg / dL, accuracy is low when exceeding 400 mg / dL, and is relatively inaccurate when the resulting LDL cholesterol value is 70 mg / dL or lower.

[0123] In addition, based on statistical trends, Koreans show higher triglyceride (TG) levels and lower total cholesterol (TC), low-density lipoprotein (LDL), and high-density lipoprotein (HDL) levels compared to Westerners; therefore, it was confirmed that the results obtained according to the above formula can demonstrate higher accuracy for Koreans.

Claims

Claim 1 (a) a step of measuring high-density lipoprotein cholesterol (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC) of blood separated from a subject; (b) a step of deriving a low-density lipoprotein cholesterol value according to Equation 1 using three elements of high-density lipoprotein cholesterol (HDLC), triglycerides (TG), and total cholesterol (TC): Equation 1: LDLC = 0.94 × Total Cholesterol (TC) - 0.94 × High-density Lipoprotein Cholesterol (HDLC) - 0.12 × Triglycerides (TG); and (c) a step of making a judgment according to a judgment criterion using the low-density lipoprotein cholesterol value derived according to step (b), wherein the method of measuring low-density lipoprotein cholesterol according to Equation 1 in step (b) is characterized by having a misclassification rate of 20% or less. Claim 2 A method for predicting cardiovascular disease according to claim 1, wherein the cardiovascular disease is selected from the group consisting of hypercholesterolemia, hyperlipidemia, arteriosclerosis, and atherosclerotic arteriosclerosis. Claim 3 A method for predicting cardiovascular disease according to claim 1, wherein the cardiovascular disease is hyperlipidemia. Claim 4 A method for predicting cardiovascular disease according to claim 1, wherein in step (a), high-density lipoprotein cholesterol, triglycerides, and total cholesterol are measured by an enzymatic method. Claim 5 A method for predicting cardiovascular disease according to claim 1, wherein in step (c), the low-density lipoprotein cholesterol judgment criteria are (1) adequate: < 100 mg / dL; (2) normal: 100~129 mg / dL; (3) borderline: 130~159 mg / dL; (4) high: 160~189 mg / dL; and (5) very high: ≥ 190 mg / dL. Claim 6 A method for predicting cardiovascular disease according to claim 1, wherein Equation 1 of step (b) is applied to Mongoloids. Claim 7 delete Claim 8 In claim 1, the low-density lipoprotein cholesterol measurement method according to Equation 1 of step (b) above has a coefficient of determination (R 2 A cardiovascular disease prediction method with a Coefficient of Determination of 0.9 or higher. Claim 9 A cardiovascular disease prediction method according to claim 1, wherein the low-density lipoprotein cholesterol measurement method according to Equation 1 of step (b) has a Root Mean Square Error (RMSE) of 8.0 or less. Claim 10 A method for providing a subject's low-density lipoprotein (LDLC) cholesterol level according to the following Formula 1 using three components of the subject's high-density lipoprotein cholesterol (HDLC; HDL-cholesterol), triglycerides (TG; Triglycerides), and total cholesterol (TC), characterized in that the misclassification rate of the low-density lipoprotein cholesterol level according to Formula 1 is 20% or less: Formula 1: LDLC = 0.94 × Total Cholesterol (TC) - 0.94 × High-density Lipoprotein Cholesterol (HDLC) - 0.12 × Triglycerides (TG). Claim 11 (a) a step of selecting public big data to derive an estimation formula for low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol; (b) a step of selecting groups of low-density lipoprotein (LDLC; LDL-cholesterol) cholesterol, high-density lipoprotein (HDLC; HDL-cholesterol), triglycerides (TG; Triglyceride), and total cholesterol (TC; Total-cholesterol) from the big data of step (a); (c) a step of deriving and comparing an estimation formula for low-density lipoprotein cholesterol using machine learning; (d) a step of verifying the validity of the values ​​of low-density lipoprotein cholesterol estimated in step (c); wherein the misclassification rate of the estimation formula is 20% or less. Claim 12 In paragraph 11, the above step (a) is a method for deriving a low-density lipoprotein cholesterol estimation formula, wherein the Korea National Health and Nutrition Examination Survey (KNHANES) data is selected as public big data. Claim 13 A method for deriving a low-density lipoprotein cholesterol estimation formula, wherein step (c) above derives the estimation formula through the least squares method. Claim 14 A method for deriving a low-density lipoprotein cholesterol estimation formula, wherein step (c) is comparing the estimation formula through hyperparameters via grid search of machine learning. Claim 15 In paragraph 11, the above step (d) is the Root Mean Square Error (RMSE) and the coefficient of determination (R 2 A method for deriving a low-density lipoprotein cholesterol estimation formula, which is verified by one or more methods selected from a group consisting of ; Coefficient of Determination). Claim 16 A method for deriving a low-density lipoprotein cholesterol estimation formula, wherein step (d) is verified based on NCEP-ATP III 6 categories in claim 11. Claim 17 A computer-readable recording medium having a program recorded thereon for implementing the method of any one of paragraphs 11 through 16.

Images