Corrigendum to Neural network model outperforms conventional equations in LDL cholesterol estimation: A comparative study of 188,887 Chinese individuals with focus on hypertriglyceridemia [Atherosclerosis Plus, (62), December 2025, Pages 38-46

BackgroundAccurate estimation of low-density lipoprotein cholesterol (LDL-C) is crucial for atherosclerotic cardiovascular disease (ASCVD) risk management. Extensive international validation studies have demonstrated that traditional formulas (Friedewald, Martin/Hopkins, Sampson) often yield significant errors under conditions of extreme hypertriglyceridemia. This study aimed to assess the performance of these conventional formulas in Chinese populations and develop a novel neural network–based LDL-C estimation model [LDL-C(NN)].MethodsIn this retrospective study, we analyzed 188,887 lipid profiles—including total cholesterol, triglycerides, high-density lipoprotein cholesterol, and directly measured LDL-C—from Peking University Shenzhen Hospital using Mindray (outpatients, n = 83,731) and Beckman (inpatients, n = 105,156) systems. The test results from the two detection systems are non-overlapping. We used stratified random sampling based on TG levels to select 30,000 profiles from each of the two systems as the training dataset (60,000 profiles in total). Within this training dataset, 70 % of profiles were used for parameter learning, 15 % were used for early-stopping validation, and 15 % were used for post-training testing. The remaining profiles constituted the independent test set for the final performance evaluation (Mindray: n = 53,731; Beckman: n = 75,156). We then compared the performance of LDL-C(NN) with the Friedewald, Martin/Hopkins, and Sampson formulas using correlation coefficient (r), root mean square error (RMSE), Concordance Correlation Coefficient (CCC) and clinical risk stratification accuracy.ResultsCompared with directly measured LDL-C, LDL-C(NN) demonstrated higher correlation and lower RMSE than other traditional LDL-C equations in the Mindray system (r = 0.9778, RMSE = 0.1762 mmol/L; vs Friedewald quation: r = 0.8894, RMSE = 0.4783 mmol/L; vs Martin/Hopkins quation: r = 0.9658, RMSE = 0.2463 mmol/L; vs Sampson quation: r = 0.9548, RMSE = 0.2934 mmol/L, particularly patients with high triglycerides (TG levels, 9.03–13.56 mmol/L, neural network Model: CCC = 0.8750, vs Friedewald quation: CCC = 0.3320; vs Martin/Hopkins quation: CCC = 0.7278; vs Sampson quation: CCC = 0.4176). Beckman database shows the same performance. The clinical classification accuracy for LDL-C(NN) reached 87.5 % (Mindray) and 83.4 % (Beckman), surpassing that of other traditional LDL-C equations (66.6–78.7 %).ConclusionsBy overcoming the linear assumptions of conventional equations, the neural network–based model significantly improves LDL-C estimation in hypertriglyceridemia (especially≥9.03 mmol/L) and complex lipid profiles, thereby expanding the applicability of traditional formulas, while demonstrating robust performance across multiple analytical systems.

ATP III recognizes that detection of cholesterol disorders and other coronary heart disease (CHD) risk factors occurs primarily through clinical case finding. Risk factors can be detected and evaluated as part of a person's work-up for any medical problem. Alternatively, public screening programs can identify risk factors, provided that affected individuals are appropriately referred for physician attention. The identification of cholesterol disorders in the setting of a medical examination has the advantage that other cardiovascular risk factors—including prior CHD, PVD, stroke, age, gender, family history, cigarette smoking, high blood pressure, diabetes mellitus, obesity, physical inactivity—co-morbidities, and other factors can be assessed and considered prior to treatment. At the time of physician evaluation, the person's overall risk status is assessed. Thus, detection and evaluation of cholesterol and lipoprotein problems should proceed in parallel with risk assessment for CHD. The approach to both is described below. The guiding principle of ATP III is that the intensity of LDL-lowering therapy should be adjusted to the individual's absolute risk for CHD. In applying this principle, ATP III maintains that both short-term (≤10-year) and long-term (> 10-year) risk must be taken into consideration. Thus, treatment guidelines are designed to incorporate risk reduction for both short-term and long-term risk (composite risk). ATP III identifies three categories of risk for CHD that modify goals and modalities of LDL-lowering therapy: established CHD and CHD risk equivalents, multiple (2+) risk factors, and 0-1 risk factor (Table III.1-1). View this table: Table III.1-1. Categories of Risk for Coronary Heart Disease (CHD) ### a. Identification of persons with CHD and CHD risk equivalents Coronary heart disease . Persons with CHD are at very high risk for future CHD events (10-year risk >20 percent). Several clinical patterns constitute a diagnosis of CHD; these include history of acute myocardial infarction, evidence of silent myocardial infarction or myocardial ischemia, history of unstable angina and stable angina pectoris, and history …

Most of the laboratories make use of the Friedewald formula to assess low-density lipoprotein cholesterol (LDL-C). The accuracy of this approach, however, crucially depends on triglyceride concentrations. Since hypertriglyceridaemia is a characteristic trait of the lipid profile in chronic kidney disease (CKD), the present study examines the accuracy of the Friedewald formula in this population. It aims to derive and validate a more accurate equation for CKD. Cross-sectional study on two cohorts of subjects (overall n = 3.514) with estimated glomerular filtration rate (eGFR) <60 mL/min comparing directly measured LDL-C (LDL-Cmeas) as assessed by an enzymatic assay (Roche, Switzerland) to concentrations estimated by the Friedewald (LDL-CF) and the Martin's formula (LDL-CM). Accuracy was analysed by Bland-Altman and linear regression analyses. In the first cohort, a novel formula was derived to assess LDL-C in CKD. The formula was validated in Cohort 2. Cohort 1 comprised 1738 subjects, and Cohort 2 comprised 1776 subjects. The mean eGFR was 29.4 ± 14.4 mL/min. In Cohort 1, LDL-CF was highly correlated with LDL-Cmeas (R2 = 0.92) but significantly underestimated LDLmeas by 11 mg/dL. LDL-C = cholesterol - HDL - triglycerides/7.98 was derived as the optimal equation for the calculation of LDL-C in Cohort 1 and was successfully validated in Cohort 2 (bias of 1.6 mg/dL). The novel formula had a higher accuracy than both the Friedewald (bias -12.2 mg/dL) and the Martin's formula (bias -4.8 mg/dL). The Friedewald formula yields lower LDL-C concentrations in CKD than direct enzymatic measurements, which may lead to undersupply of this cardiovascular high-risk population in a treat-to-target approach.

To evaluate the accuracy of LDL cholesterol calculated with Friedewald's equation in the assessment of cardiovascular risk in NIDDM patients. The calculation of LDL cholesterol according to Friedewald's formula was compared with the measurement of LDL cholesterol separated by ultracentrifugation in 151 NIDDM patients with fairly good metabolic control (HbA1c < or =10%) and in 405 nondiabetic subjects. Measured and calculated LDL cholesterol was found to be well correlated in both diabetic (r = 0.95) and nondiabetic (r = 0.97) subjects. Compared with measured LDL cholesterol, the calculated LDL cholesterol differed by > or =10% in 34% of samples from diabetic patients and in 26% of samples from nondiabetic subjects (chi(2) = 3.885, P < 0.05). The percentage of error increased when the serum triglyceride (TG) level was > or =200 mg/dl (2.26 mmol/l) and when the ratio of VLDL cholesterol to TG was <0.20 or >0.29 in both groups of subjects. Although the percentage of error from calculated LDL cholesterol was greater in diabetic than in nondiabetic subjects because of the greater prevalence of hypertriglyceridemia in the former group, the misclassification of coronary heart disease risk, according to the cutoff points of the National Cholesterol Education Program (NCEP), was similar in the two groups (25% in diabetic and 22% in nondiabetic subjects). In both groups of patients, the misclassification of coronary heart disease risk was higher when calculation of LDL cholesterol produced values near the cutoff points. Although accuracy in the estimation of LDL cholesterol is less than ideal, Friedewald's equation seems to be of value in the correct assignment of coronary heart disease risk classes in the great majority of diabetic as well as nondiabetic subjects. Caution must be exercised for subjects in whom calculated LDL cholesterol is close to the cutoff points of the NCEP guidelines.

Background Calculated LDL cholesterol (LDL-C) is a standard component of the lipid panel and an important risk factor in managing atherosclerotic cardiovascular disease (ASCVD). Due to the assumptions in the derivation of the Friedwald formula, fasting (≥8 hours) and triglyceride levels ≤400mg/dL have long been recommended. More recent data have shown that in most cases fasting has minimal effect on results, leading to the endorsement of non-fasting lipid panels for routine evaluation of ASCVD risk in clinical practice guidelines. Furthermore, newer equations (e.g. Sampson-NIH) allow for more accurate estimation of LDL-C even in the presence of hypertriglyceridemia. Mayo Clinic has implemented the Sampson-NIH LDL-C calculation and removed fasting/non-fasting designation from lipid panel orders. We hypothesized that the change to fasting designation would allow more afternoon draws for routine lipid panels and that increased non-fasting collections may increase triglyceride levels, but that the Sampson-NIH equation would prevent bias in LDL-C estimation. In order to test these hypotheses, we analyzed patient data from 1 year prior and 1 year post-implementation. Methods Results of lipid panels performed for one year preceding (n=334,719) and one year after (n=381,804) removal of the fasting/non-fasting designation and implementation of Sampson-NIH equation were collected. Median, 10th, and 90th percentile of draw time, triglyceride, and LDL-C calculated by either Sampson-NIH or Friedwald formulas were compared by quantile regression. Results Draw times shifted post-implementation at the 10th percentile (7:06 vs. 7:10, p&amp;lt;0.0001), median (8:49 vs. 9:08, p&amp;lt;0.0001) and 90th (12:02 vs. 13:46, p&amp;lt;0.0001). No difference in median triglyceride measurement was observed between pre-and post-implementation (110mg/dL), but slight differences were observed in the 10th and 90th percentile values (60 vs 59mg/dL and 224 vs 229mg/dL, p&amp;lt;0.0001). No difference was observed in 10th, 50th, or 90th percentile LDL values when calculated by Sampson-NIH equation (56, 99, 153mg/dL); however, if the Friedwald formula is used, a slight shift is observed in the 10th and 50th percentile LDL values (54 vs. 53, 97 vs. 96; p&amp;lt;0.0001) but not the 90th (150mg/dL). Conclusions Changing to the Sampson/NIH formula for LDL-C and removing fasting vs. non-fasting designation from routine lipid panels had minimal effect on reported triglyceride and LDL-C values, while allowing increased flexibility in draw time and reporting of LDL-C on patients with triglycerides ≥400mg/dL. Patients with TG ≥400mg/dL represented 1.7% of the dataset overall, including 5,387 patients before the change in which LDL-C could not be reported and 6,697 patients in the year post-implementation for whom LDL-C was reported based on the Sampson-NIH formula. Using an equation that is robust in patients with elevated triglyceride levels and removing fasting requirements from routine lipid panels allows greater flexibility and convenience for patients while allowing for more even distribution of phlebotomy workload and reporting of LDL-C results for patients with hypertriglyceridemia without evidence of clinically significant bias in reported results.

Historically, triglycerides have been measured in the fasting state for 2 reasons. First, because of the marked increase in triglycerides after fat ingestion, the variability in triglyceride measurements is much smaller in the fasting state. Second, before the availability of direct assays for LDL cholesterol (LDL-C),1 estimation of LDL-C was performed in clinical practice almost exclusively by use of the Friedewald equation, which requires that both the HDL-C concentration and the fasting triglyceride concentration divided by 5 be subtracted from the total cholesterol concentration. The recommendations to measure triglycerides in the fasting state did not, however, derive from a consistent set of prospective cohort studies showing that fasting concentrations were superior to nonfasting concentrations for the detection of cardiovascular risk. Instead, following screening guidelines, most epidemiologic investigations simply relied on fasting triglycerides. Taken as a whole, such studies indicate that fasting triglycerides are a univariate predictor of vascular disease. Controversy exists, however, regarding the clinical usefulness of fasting triglycerides as an independent predictor of risk, because adjustment for other covariates—in particular HDL-C—markedly decreases both the magnitude and significance of observed epidemiologic effects (1). The extent to which investigators have sought to avoid nonfasting triglycerides as a method for risk detection is evident in a recent metaanalysis that limited evaluation only to those epidemiologic studies that measured fasting triglycerides, specifically “to exclude the possibility of postprandial effects” (2). Is it possible, then, that recommendations to measure triglycerides in the fasting state have systematically underestimated the impact of hypertriglyceridemia in clinical practice? Atherosclerosis has long been hypothesized to be a disorder influenced by postprandial effects. As early as 1950, J. R. Moreton, writing in the Journal of Laboratory and Clinical Medicine , suggested a linkage between chylomicronemia, fat tolerance, and atherosclerosis (3). A major source of circulating triglycerides is dietary …

Introduction: A long standing association exists between elevated serum LDL and cardiovascular disease. Studies have suggested that, increased LDL in serum is the major contributor to vascular complications of other diseases like diabetes mellitus, its measurement is recommended in routine clinical practice. Currently, estimation of LDL cholesterol is done in the clinical laboratories using Friedewald equation to make the lipid profile cost effective. However, it has been highlighted that calculated LDL is not reliable when serum triglyceride (TG) levels exceed 400mg/dl. Objective of the study: To compare estimated LDL by homogenous method with calculated LDL by Friedewald equation in lipid profile requests and to compare the same in different groups of triglyceride levels. Materials and Methods: About 260 lipid profile requests of both the genders aged between 25-75 years were considered for the study. Total cholesterol, triglycerides, HDL-cholesterol, LDL -cholesterol was estimated in the serum spectrophotometrically. LDL-C was also calculated using Freidewald formula. The Data thus collected was segregated based on triglycerides into three groups, Group I (TG ≤150mg/dl) Group II (TG 151-399 mg/dl) and Group III (TG ≥400 mg/dl). Results:LDL determined by direct assay correlated highly with calculated LDL in all subjects irrespective of the TG levels. Correlation coefficients being 0.96, 0.95 and 0.81 in group I, II and III respectively. Estimated LDL was significantlyhigher than the calculated LDL in group II and group III, suggestive of the fact that calculated LDL underestimates the true LDL levels in cases with TG levels above the normal range. Further, the differences in the means were significantly higher in hypertriglyceridemic groups (p < 0.001). Conclusion: it can be concluded that estimation of LDL needs a better accurate measurement technique than following calculations, considering the importance of patient care in management of life style disorders, aimed at lowering serum LDL levels.

Intercomparison study between the measurement of LDL cholesterol in the Atellica Solution® analyser and the estimation of LDL cholesterol using the Martin-Hopkins equation

Estimation of LDL cholesterol by the Friedewald formula during an oral glucose tolerance test

Intercomparison study between the measurement of LDL cholesterol in the atellica solution® analyser and the estimation of LDL cholesterol using the friedewald equation

There is significant inter-individual variability in the lipid-lowering effects of atorvastatin and simvastatin. Our goal was to investigate the impact of SLCO1B1 genetic polymorphism on the lipid-lowering effects of atorvastatin and simvastatin. We recruited 363 unrelated hyperlipidemic patients with the CYP3A4 1/1, CYP3A5 1/1, and CYP3AP1 1/1 genotypes: 189 of these were treated with atorvastatin and 174 were treated with simvastatin as a single-agent therapy (20 mg day(-1) orally) for 4 weeks. The genotyping of SLCO1B1 c.521T > C (p.V174A, OATP-C5) was performed with allele-specific polymerase chain reaction (AS-PCR), and PCR restriction fragment length polymorphism (RFLP) was performed to detect the carriers of SLCO1B1 c.388A > G (p.N130D, OATP-C1b). Serum triglyceride (TGs), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) levels were determined before and after treatment. The frequencies of the SLCO1B1 521T > C and 388A > G variant alleles in Chinese hyperlipidemic patients were found to be 16.2% and 72.1% respectively. After treatment with 20 mg simvastatin or atorvastatin daily for 4 weeks, TC, TG, and LDL-C concentrations were lower than at baseline, on average, by 18.1 ± 3.7%, 25.8 ± 9.7%, 27.7 ± 5.4% in the simvastatin-treated group, and 17.5 ± 3.7%, 22.6 ± 8.6%, 27.5 ± 5.5% in the atorvastatin-treated group respectively, and the mean relative reduction in serum HDL cholesterol did not reach statistical significance (-1.0 ± 10.9%, 0.5 ± 9.3%). However, no significant differences were observed in the lipid-lowering effects of atorvastatin and simvastatin between subjects with different SLCO1B1 genotypes. The SLCO1B1 521T > C and 388A > G variants were found to be relatively common in Chinese patients with essential hyperlipidemia. These frequencies were found to be similar to those observed in healthy Chinese and Japanese individuals, but significantly different from Caucasians and blacks. SLCO1B1 521T > C and 388A > G polymorphisms may not be associated with the lipid-lowering effects of atorvastatin and simvastatin.

Emerging evidence indicates that variants of alcohol-metabolizing enzymes may influence lipid metabolism. This study aimed to investigate whether the rs671 and rs1229984 variants affect lipid levels in East Asian individuals. PubMed, Foreign Medical Journal Service, Embase, Cochrane Library, Scopus, MEDLINE, Web of Science, Web of Knowledge, Wanfang, and Chinese Biomedical Literature databases were searched until December 31, 2021. Meta-analyses of studies that examined the effects of alcohol-metabolizing enzyme variants on lipid levels, as well as the interaction with alcohol intake, were selected. Data extraction was conducted independently by two investigators and confirmed by the third. In total, 86 studies (179 640 individuals) were analyzed. The A allele of rs671 (a functional variant in the ALDH2 gene) was linked to higher levels of low-density lipoprotein cholesterol (LDL-C) and lower levels of triglycerides and high-density lipoprotein cholesterol. In contrast, the A allele of the rs1229984 (a functional variant in the ADH2 gene) was associated only with lower levels of LDL-C. The effects of rs671 and rs1229984 on lipid levels were much stronger in Japanese than in Chinese individuals and in males than in females. Regression analysis indicated that the effects of rs671 on lipid levels were independent of alcohol intake in an integrated East Asian population (ie, Japanese, Chinese, and Korean individuals). Intriguingly, alcohol intake had a statistical influence on lipid levels when the sample analyzed was restricted to Japanese individuals or to males. The rs671 and rs1229984 variants of alcohol-metabolizing enzymes have significant effects on lipid levels and may serve as genetic markers for lipid dyslipidemia in East Asian populations. Circulating lipid levels in Japanese individuals and in males were modulated by the interaction between rs671 and alcohol intake.

BackgroundLow-density lipoprotein cholesterol is an important marker highly associated with cardiovascular disease. Since the direct measurement of it is inefficient in terms of cost and time, it is common to estimate through the Friedewald equation developed about 50 years ago. However, various limitations exist since the Friedewald equation was not designed for Koreans. This study proposes a new low-density lipoprotein cholesterol estimation equation for South Koreans using nationally approved statistical data.MethodsThis study used data from the Korean National Health and Nutrition Examination Survey from 2009 to 2019. The 18,837 subjects were used to develop the equation for estimating low-density lipoprotein cholesterol. The subjects included individuals with low-density lipoprotein cholesterol levels directly measured among those with high-density lipoprotein cholesterol, triglycerides, and total cholesterol measured. We compared twelve equations developed in the previous studies and the newly proposed equation (model 1) developed in this study with the actual low-density lipoprotein cholesterol value in various ways.ResultsThe low-density lipoprotein cholesterol value estimated using the estimation formula and the actual low-density lipoprotein cholesterol value were compared using the root mean squared error. When the triglyceride level was less than 400 mg/dL, the root mean squared of the model 1 was 7.96, the lowest compared to other equations, and the model 2 was 7.82. The degree of misclassification was checked according to the NECP ATP III 6 categories. As a result, the misclassification rate of the model 1 was the lowest at 18.9%, and Weighted Kappa was the highest at 0.919 (0.003), which means it significantly reduced the underestimation rate shown in other existing estimation equations. Root mean square error was also compared according to the change in triglycerides level. As the triglycerides level increased, the root mean square error showed an increasing trend in all equations, but it was confirmed that the model 1 was the lowest compared to other equations.ConclusionThe newly proposed low-density lipoprotein cholesterol estimation equation showed significantly improved performance compared to the 12 existing estimation equations. The use of representative samples and external verification is required for more sophisticated estimates in the future.

Low-density-lipoprotein cholesterol (LDL-C) is the main target in atherosclerotic cardiovascular disease (ASCVD). We aimed to validate and compare a new LDL-C estimation equation with other well-known equations. 177,111 samples were analysed from two contemporary population-based cohorts comprising asymptomatic Korean adults who underwent medical examinations. Performances of the Friedewald (FLDL), Martin (MLDL), and Sampson (SLDL) equations in estimating direct LDL-C by homogenous assay were assessed by measures of concordance (R2, RMSE, and mean absolute difference). Analyses were performed according to various triglyceride (TG) and/or LDL-C strata. Secondary analyses were conducted within dyslipidaemia populations of each database. MLDL was superior or at least similar to other equations regardless of TG/LDL-C, in both the general and dyslipidaemia populations (RMSE = 11.45/9.20 mg/dL; R2 = 0.88/0.91; vs FLDL: RMSE = 13.66/10.42 mg/dL; R2 = 0.82/0.89; vs SLDL: RMSE = 12.36/9.39 mg/dL; R2 = 0.85/0.91, per Gangnam Severance Hospital Check-up/Korea Initiatives on Coronary Artery Calcification data). MLDL had a slight advantage over SLDL with the lowest MADs across the full spectrum of TG levels, whether divided into severe hyper/non-hyper to moderate hypertriglyceridaemia samples or stratified by 100-mg/dL TG intervals, even up to TG values of 500–600 mg/dL. MLDL may be a readily adoptable and cost-effective alternative to direct LDL-C measurement, irrespective of dyslipidaemia status. In populations with relatively high prevalence of mild-to-moderate hypertriglyceridaemia, Martin’s equation may be optimal for LDL-C and ASCVD risk estimation.

Familial hypercholesterolemia (FH) is the most frequent genetic disorder seen clinically and is characterized by increased LDL cholesterol (LDL-C) (>95th percentile), family history of increased LDL-C, premature atherosclerotic cardiovascular disease (ASCVD) in the patient or in first-degree relatives, presence of tendinous xanthomas or premature corneal arcus, or presence of a pathogenic mutation in the LDLR, PCSK9, or APOB genes. A diagnosis of FH has important clinical implications with respect to lifelong risk of ASCVD and requirement for intensive pharmacological therapy. The concentration of baseline LDL-C (untreated) is essential for the diagnosis of FH but is often not available because the individual is already on statin therapy. To validate a new algorithm to impute baseline LDL-C, we examined 1297 patients. The baseline LDL-C was compared with the imputed baseline obtained within 18 months of the initiation of therapy. We compared the percent reduction in LDL-C on treatment from baseline with the published percent reductions. After eliminating individuals with missing data, nonstandard doses of statins, or medications other than statins or ezetimibe, we provide data on 951 patients. The mean ± SE baseline LDL-C was 243.0 (2.2) mg/dL [6.28 (0.06) mmol/L], and the mean ± SE imputed baseline LDL-C was 244.2 (2.6) mg/dL [6.31 (0.07) mmol/L] (P = 0.48). There was no difference in response according to the patient's sex or in percent reduction between observed and expected for individual doses or types of statin or ezetimibe. We provide a validated estimation of baseline LDL-C for patients with FH that may help clinicians in making a diagnosis.