Evaluation of the Novel Method and the Regression Equation for Calculation of Low-Density Lipoprotein Cholesterol

Background: Friedewald’s formula (FF) is used worldwide to calculate low-density lipoprotein cholesterol (LDL-chol). But it has several shortcomings: overestimation at lower triglyceride (TG) concentrations and underestimation at higher concentrations. In FF, TG to very low-density lipoprotein cholesterol (VLDL-chol) ratio (TG/VLDL-chol) is considered as constant, but practically it is not a fixed value. Recently, by analyzing lipid profiles in a large population, continuously adjustable values of TG/VLDL-chol were used to derive a novel method (NM) for the calculation of LDL-chol. Objective: The aim of this study was to evaluate the performance of the novel method compared with direct measurement and regression equation (RE) developed for Bangladeshi population. Materials and Methods: In this cross-sectional comparative study we used lipid profiles of 955 adult Bangladeshi subjects. Total cholesterol (TC), TG, HDL-chol and LDL-chol were measured by direct methods using automation. LDL-chol was also calculated by NM and RE. LDL-chol calculated by NM and RE were compared with measured LDL-chol by twotailed paired t test, Pearson’s correlation test, bias against measured LDL-chol by Bland-Altman test, accuracy within ±5% and ±12% of measured LDL-chol and by inter-rater agreements with measured LDL-chol at different cut-off values. Results: The mean values of LDL-chol were 110.7 ± 32.0 mg/dL for direct measurement, 111.9 ± 34.8 mg/dL for NM and 113.2 ± 31.7 mg/dL for RE. Mean values of calculated LDL-chol by both NM and RE differed from that of measured LDL-chol (p<0.01 for NM and p<0.0001 for RE). The correlation coefficients of calculated LDL-chol values with measured LDL-chol were 0.944 (p<0.0001) for NM and 0.945 (p<0.0001) for RE. BlandAltman plots showed good agreement between calculated and measured LDL-chol. Accuracy within ±5% of measured LDL-chol was 49% for NM, 46% for RE and within ±12% of measured LDL-chol was 79% for both NM and RE. Inter-rater agreements (κ) between calculated and measured LDL-chol at LDL-chol <100 mg/dL, 100–130 mg/dL and >130 mg/dL were 0.816 vs 0.815, 0.637 vs 0.649 and 0.791 vs 0.791 for NM and RE respectively. Conclusion: This study reveals that NM and RE developed for Bangladeshi population have similar performance and can be used for the calculation of LDL-chol.


Evaluation of the Novel Method and the Regression Equation for Calculation of Low-Density Lipoprotein Cholesterol
preparative ultracentrifugation, i.e., β-quantification. 3se of this reference method is limited for routine clinical practice due to the technical difficulties.To resolve these problems direct homogeneous methods have been developed and recommended for the measurement of LDL-chol as alternatives to the reference method. 4,5The direct methods are costly and require expensive automation and are not affordable by most of the laboratories in the developing countries.As a result Friedewald's formula 6 , the worldwide used formula, is generally used for the estimation of LDLchol by most of the laboratories in Bangladesh.][8][9][10][11] But there are several limitations in using this formula.There are underestimation in the measurement of LDL-chol at higher TG levels 12,13 and overestimation at low TG levels. 14Recently by analyzing lipid profiles from 1.34 million consecutive adult subjects referred for direct measurement of cholesterol subfraction by the Vertical Auto Profile (VAP, density gradient ultracentrifugation or vertical spin density gradient ultracentrifugation) Martin et al 15,16 also reported a meaningful underestimation of LDL-chol in US adults.These are related to the use of a fixed value of TG to VLDLcholesterol.

Materials and Methods
This cross-sectional comparative study was conducted in the Department of Biochemistry, Chevron Clinical Laboratory, Chittagong, Bangladesh during the period of July to December 2013.In this study, 1016 adult subjects, both male and female, from the outpatient department of Chevron Clinical Laboratory were included.Venous blood specimens were collected in tubes without anticoagulant for analysis of lipids from all the selected subjects after 12-hour fast.The specimens were allowed to clot at room temperature, and serum was obtained by centrifugation at 3000 rpm for 15 minutes.All blood lipid analyses were performed within 24 hours of specimen collection.Serum TG and TC were measured by enzymatic end-point method and HDL-chol and LDL-chol were measured by direct automated method using Olympus AU400 clinical chemistry analyzer (Japan

Results
A total of 1016 adult subjects were included in this study.Among them 61 (6%) subjects had serum TG above 400 mg/dL.The remaining 955 (94%) subjects had serum TG levels up to 400 mg/dL and we considered them as study subjects.The mean age of the study subjects was 47 ± 12 years.Among them 566 (59%) were male and 389 (41%) were female.

Discussion
Friedewald's formula is frequently used in clinical practice and population-based epidemiological studies.Underestimation of LDL-chol by Friedewald's formula is common. 12,13,15,168][19][20][21][22][23] All studies reported remarkable underestimation of LDL-chol by Friedewald's formula.Recently, Martin et al 24 derived a novel method (NM) for the calculation of LDL-chol.In this study, we compared LDL-chol calculated by NM and LDL-chol calculated by a regression equation (RE) developed for Bangladeshi population with measured LDL-chol.
Differences of mean values of calculated LDL-chol using NM and RE with measured LDL-chol were statistically significant (1.2 mg/dL for NM and 2.5 mg/dL for RE developed for Bangladeshi population), but possibly insignificant clinically.On the other hand it was meaningful and large for Friedewald's formula in US population 15,16 and >11 mg/dL in Bangladeshi population. 17,20,24Strong and similar correlation coefficients were observed for NM and RE (0.944 and 0.945).Bias of calculated LDL-chol was lower for NM compared with RE (1.1% vs 3.4%), but both are within good agreement whereas this is higher for Friedewald's formula. 23ccuracy of NM and RE within 5% and 12% of measured LDL-chol was similar and improved compared to Friedewald's formula. 23Furthermore, we observed good agreements (κ) between measured and calculated LDL-chol at LDL-chol values <100 mg/dL, 100-130 mg/dL and >130 mg/dL.Thus, underestimation of LDL-cholesterol by Friedewald's formula is significantly reduced using these methods (novel method and regression equation).
From the findings of our study we can conclude that good agreements exist between direct measurement and novel method and also between direct measurement and regression equation developed for Bangladeshi population.Therefore, both these formulas can be used for the calculation of LDL-cholesterol.
VLDL-chol against different combination of TG range and non-HDL-chol range.The 180-cell table produces an overall improvement of LDL-chol calculation.So, it is urgently needed to evaluate and also to compare the RE for Bangladeshi population and recently developed NM against measured LDL-chol simultaneously.In this context, this cross-sectional comparative study was designed to evaluate the performance of the NM and RE against measured LDLchol in this population.
by analyzing lipid profiles of 1.3 million US adults.They have generated a two dimensional table of different median values of TG/