Prediction of low-density lipoprotein cholesterol levels using machine learning methods.

OBJECTIVE: Low-density lipoprotein cholesterol (LDL-C) has been commonly calculated by equations, but their performance has not been entirely satisfactory. This study aimed to develop a more accurate LDL-C prediction model using machine learning methods. METHODS: The study involved predicting directly measured LDL-C, using individual characteristics, lipid profiles, and other laboratory results as predictors. The models applied to predict LDL-C values were multiple regression, penalized regression, random forest, and XGBoost. Additionally, a novel 2-step prediction model was developed and introduced. The machine learning methods were evaluated against the Friedewald, Martin, and Sampson equations. RESULTS: The Friedewald, Martin, and Sampson equations had root mean squared error (RMSE) values of 12.112, 8.084, and 8.492, respectively, whereas the 2-step prediction model showed the highest accuracy, with an RMSE of 7.015. The LDL-C levels were also classified as a categorical variable according to the diagnostic criteria of the dyslipidemia treatment guideline, and concordance rates were calculated between the predictive values obtained from each method and the directly measured ones. The 2-step prediction model had the highest concordance rate (85.1%). CONCLUSION: The machine learning method can calculate LDL-C more accurately than existing equations. The proposed 2-step prediction model, in particular, outperformed the other machine learning methods.

Indirect calculation of LDL using thirteen equations in Pakistani population.

BACKGROUND: Owing to the atherogenic properties, low density lipoprotein cholesterol (LDL-C) is the primary target for treatment and diagnosis of cardiovascular diseases (CVDs), hence accurate measurement of LDL-C is critical. Despite the availability of direct measurement assays for LDL-C, it is routinely calculated by Friedewald equation in clinical settings in Pakistan mostly due to financial constraints. However, the validity of this equation is impacted by several factors, therefore several other equations have been developed for the calculation of LDL-C. MATERIALS AND METHODS: LDL-C of 39,385 individuals measured directly by homogenous assays (dLDL) was compared with LDL-C calculated by thirteen equations (cLDL-C). Stratifications based on different lipids i.e., triglycerides (TG), total cholesterol (TC), high-density lipoprotein (HDL) were made to check the validity of these equations across all ranges of lipid profile. The correlation and median difference between dLDL and cLDL-C was statistically analyzed. RESULTS: Overall Teerakanchana equation displayed a strong positive correlation (ρ = 0.967) and least median difference (-8.81) with dLDL, followed by Martin equation (ρ = 0.967). For higher TG ranges (>500 mg/dL), Teerakanchana equation had the least median difference (1.31) and a strong correlation (ρ = 0.800). CONCLUSION: Our data suggest that Teerakanchana equation may be employed as an alternative to Friedewald equation for Pakistani population.

Evaluation of Martin's equation for LDL-C estimation in type 2 diabetes mellitus Egyptian patients.

INTRODUCTION: Type 2 Diabetes Mellitus has characteristic dyslipidemia. Low-density lipoprotein cholesterol (LDL-C) measurement plays a role in cardiovascular risk assessment and management. Friedewald equation (FE) has several limitations. This study aims to evaluate the effectiveness of Martin equation (ME) in Egyptian patients, especially those with type 2 diabetes. METHODS: A cross-sectional study was conducted on 454 diabetic and non-diabetic patients who were referred to the internal medicine outpatient clinic. Lipid profile was assessed by Cobas 8000 Modular Analyzer. RESULTS: The LDL-C was estimated by both FE and ME. In diabetic patients, LDL-C estimated by FE was underestimated with a bias of -3.9 ±â€¯5.3 mg/dL (p = .04). But LDL-C estimated by ME was not significantly different compared to directly measured LDL-C. FE underestimate LDL-C with a bias of -4.6 ±â€¯6.4 mg/dL (p = .042) in uncontrolled diabetic patients. A non-significant difference in both uncontrolled patients and controlled ones was detected by ME. FE had lower sensitivity and specificity (80% and 88.9 respectively) compared to the ME (95.9% sensitivity, and 95.6% specificity). ME was not influenced by triglyceride levels (p = .34). CONCLUSION: The ME improves concordance of calculated LDL-C with a direct LDL-C assay in Egyptian diabetic patients.

Martin's Equation as the Most Suitable Method for Estimation of Low-Density Lipoprotein Cholesterol Levels in Korean Adults.

BACKGROUND: Friedewald equation is the most widely used method for estimating low-density lipoprotein cholesterol (LDL-C) level. However, due to potential over- or underestimation, many studies have used a modified equation. This study aimed to compare estimates by 4 different equations to directly measured LDL-C concentrations in order to propose the most appropriate method for LDL-C estimation in the Korean population. METHODS: We studied data of 4,350 subjects that included total cholesterol, high-density lipoprotein cholesterol (HDL-C), triglyceride (TG), and LDL-C concentrations that had been measured at one university hospital in Seoul. We investigated 4 equations: LDL-C by Friedewald's original equation (LDL-CF) and its 3 modifications. Pearson correlation analysis was performed to compare these estimates to the direct measurement. RESULTS: Pearson correlation analysis revealed a good correlation among all 4 estimated LDL-C values and the directly measured LDL-C value. The Pearson coefficients were 0.951 for LDL-CF, 0.917 for LDL-C by Hatta equation (LDL-CH), 0.968 for LDL-C by Puavilai equation (LDL-CP), and 0.983 for LDL-C by Martin equation (LDL-CM). Martin equation (LDL-CM) resulted in the best approximation (mean difference from the direct measurement, 5.5 mg/dL; mean percentage difference from the direct measurement, 5.1%) and the best agreement with the direct measurement (86.1%). LDL-CP resulted in the second-best approximation (mean difference, 7.0 mg/dL; mean percentage difference, 6.2%; concordance, 82.5%). LDL-CM was found to be less influenced by TG and HDL-C levels than by LDL-CF. CONCLUSION: Estimates by Martin equation had the best agreement with direct LDL-C concentrations and both Martin and Puavilai equations were superior to Friedewald equation for estimating LDL-C concentrations in Korean adults.

Martin's Equation as the Most Suitable Method for Estimation of Low-Density Lipoprotein Cholesterol Levels in Korean Adults / 가정의학회지

BACKGROUND: Friedewald equation is the most widely used method for estimating low-density lipoprotein cholesterol (LDL-C) level. However, due to potential over- or underestimation, many studies have used a modified equation. This study aimed to compare estimates by 4 different equations to directly measured LDL-C concentrations in order to propose the most appropriate method for LDL-C estimation in the Korean population. METHODS: We studied data of 4,350 subjects that included total cholesterol, high-density lipoprotein cholesterol (HDL-C), triglyceride (TG), and LDL-C concentrations that had been measured at one university hospital in Seoul. We investigated 4 equations: LDL-C by Friedewald's original equation (LDL-C(F)) and its 3 modifications. Pearson correlation analysis was performed to compare these estimates to the direct measurement. RESULTS: Pearson correlation analysis revealed a good correlation among all 4 estimated LDL-C values and the directly measured LDL-C value. The Pearson coefficients were 0.951 for LDL-C(F), 0.917 for LDL-C by Hatta equation (LDL-C(H)), 0.968 for LDL-C by Puavilai equation (LDL-C(P)), and 0.983 for LDL-C by Martin equation (LDL-C(M)). Martin equation (LDL-C(M)) resulted in the best approximation (mean difference from the direct measurement, 5.5 mg/dL; mean percentage difference from the direct measurement, 5.1%) and the best agreement with the direct measurement (86.1%). LDL-C(P) resulted in the second-best approximation (mean difference, 7.0 mg/dL; mean percentage difference, 6.2%; concordance, 82.5%). LDL-C(M) was found to be less influenced by TG and HDL-C levels than by LDL-C(F). CONCLUSION: Estimates by Martin equation had the best agreement with direct LDL-C concentrations and both Martin and Puavilai equations were superior to Friedewald equation for estimating LDL-C concentrations in Korean adults.