Comparison of LDL-Cholesterol Estimate using Various Formulae with Directly Measured LDL-Cholesterol in Indian Population
Nishtha Wadhwa1, Radhika Krishnaswamy2
1 Postgraduate, Department of Biochemistry, St. John’s Medical College, Sarjapur Road, Bangalore, Karnataka, India.
2 Associate Professor, Department of Biochemistry, St. John’s Medical College, Sarjapur Road, Bangalore, Karnataka, India.
NAME, ADDRESS, E-MAIL ID OF THE CORRESPONDING AUTHOR: Dr. Nishtha Wadhwa, Postgraduate, Department of Biochemistry, St. John’s Medical College, Sarjapur Road, Bangalore-560034, Karnataka, India.
E-mail: nishthawadhwa@gmail.com
Introduction
Low-Density Lipoprotein Cholesterol (LDLc) is widely recognized as an established cardiovascular risk marker. Different formulae have been proposed for calculation of LDLc but have not been validated in Indian population and over a wide range of Triglycerides (TG). Friedewald formula is most commonly used which has various shortcomings.
Aim
To calculate LDLc using various formulae and compare it with directly measured LDLc at various ranges of TG concentration in Indian population.
Materials and Methods
One year lipid profile data of 21,503 samples was taken. Calculation of LDLc was done by the following formulae: Friedewald; Cordova and Cordova; Vujovic; Ahmadi; Anandaraja; Puavillai and Hattori. Comparison of calculated LDLc with directly measured LDLc was done at following TG ranges: <150mg/dL; 151-199mg/dL; 200-399mg/dL and >400mg/dL using Pearson’s correlation coefficient and two-paired t-test.
Results
For TG range <150mg/dL, Puavillai formula correlated best with direct measurement (r = 0.978). For TG range 151-199mg/dL, Vujovic formula correlated best with direct measurement (r = 0.977and mean difference of -1.2 mg/dL). For TG range 200-399mg/dL, Vujovic formula correlated best with direct measurement (r = 0.968). For TG range >400 mg/dL, Vujovic formula correlated best with direct measurement (r = 0.791).
Conclusion
Vujovic formula appears to be more accurate than any other formula when applied to Indian population.
Introduction
Cardiovascular diseases are the leading cause of death in the world. The concentration of circulating Low Density Lipoprotein cholesterol (LDLc) is a predictor for assessing the risk for Coronary Heart Disease (CHD) [1]. It is considered as the primary basis for accurate classification into risk categories [2]. β-quantification is the reference method for the quantitative estimation of LDLc in circulation. It requires ultracentrifugation, uses large volumes of samples and is a time consuming and expensive technique. Therefore, this method is not suitable for routine laboratory testing [3].
The other recommended methods include homogeneous direct measurement [4,5]. The direct methods require expensive automation and are not affordable by most laboratories in the developing countries. Because of these limitation many clinical laboratories throughout the world use a less expensive and easy approach for the estimation of LDLc i.e., Friedewald formula [6,7].
National Cholesterol Education Programme (NCEP) Adult Treatment Panel III (ATP III) guidelines [2] recommend the use of LDLc calculated by Friedewald formula for determination of LDLc treatment goals for prevention of cardiovascular diseases. However, there are several shortcomings of this formula, mainly the underestimation of LDL cholesterol at high Triglyceride (TG) levels and overestimation at low TG levels [7].
Several other formulae have been proposed over the years for calculation of LDLc such as Cordova and Cordova; Vujovic; Ahmadi; Anandaraja; Puavillai; Hattori; Chen; Saiedullah; Planella and Wagner [7–16] but have not been validated in varied populations.
A survey of the College of American Pathologists (CAP) reported more than 2,200 laboratories use, several different assays for direct LDLc measurements and more than 3,300 laboratories reported calculating LDLc using the Friedewald’s calculation [17,18].
The present study was designed to compare the LDLc calculated by several formulae which use High Density Lipoprotein cholesterol (HDLc), Total Cholesterol (TC) and TG to calculate LDLc with directly measured LDLc over a wide range of TG levels in Indian population with the assumption that the results obtained by direct assays are the most accurate.
Materials and Methods
This study is a retrospective analytical study. One year lipid profile data of 21,503 samples, measured by direct homogenous method on Siemens Dimensions RxL Max and EXL were taken from the laboratory database from September 2014 to August 2015. Calculation of LDLc was then done using the following seven formulae mentioned in [Table/Fig-1] [6,9–13,16].
Different formulas for Calculation of LDLc [6,9–13,16].
| Proposed by: | Formula: |
|---|
| Friedewald et al., [6] | LDLc=TC-HDLc-TG/5 |
| Cordova and Cordova [9] | LDLc =3/4 (TC-HDLc) |
| Vujovic et al., [10] | LDLc=TC-TG/6.85-HDLc |
| Ahmadi et al., [11] | LDLc=TC/1.19+TG/1.9-HDLc/1.1 |
| Anandaraja et al., [12] | LDLc=0.9TC-0.9TG/5-28 |
| Puavillai et al., [13] | LDLc=TC-HDLc-TG/6 |
| Hattori et al., [16] | LDLc=0.94TC-0.94HDLc-0.19TG |
Statistical Analysis
Mean and SD were calculated by various formulae. The mean values obtained from various formulae were compared with the mean value of LDLc obtained by direct measurement using Pearson’s correlation coefficient. Paired t-test was also performed to compare the means. A p-value of <0.05 was considered as statistically significant. Statistical analyses were performed using SPSS version 16.0.
Results
A total of 21,503 lipid profiles were grouped into four TG ranges i.e., <150mg/dL (N=13,982), 151-199mg/dL (N=3449), 200-399mg/dL (N= 3515) and >400mg/dL (N=557). For TG range <150mg/dL, presented in [Table/Fig-2], the mean value obtained from all formulae showed good correlation with value of LDLc obtained by direct method, but Puavillai formula correlated best with direct measurement with r = 0.978 and mean difference of 0.08 which was statistically insignificant (p= 0.284). For TG range 151-199mg/dL, presented in [Table/Fig-3], the mean value obtained from all formulae showed good correlation with value of LDLc obtained by direct method, but Vujovic formula correlated best with r = 0.977 and mean difference of -1.2 mg/dL (p= <0.001).
Calculation by different formulas, At TG range <150mg/dl; N=13,982.
| Method | Mean±SD(mg/dl) | Mean Difference(mg/dl) | Correlation (r) | p-value obtained by paired t-test |
|---|
| Direct | 104.24±39.03 | | | |
| Friedewald | 101.14±38.87 | 3.1 | 0.977 | <0.001 |
| Cordova | 90.15±30.53 | 14.09 | 0.975 | <0.001 |
| Vujovic | 106.29±39.29 | -2.05 | 0.978 | <0.001 |
| Ahmadi | 110.45±42.79 | -6.21 | 0.883 | <0.001 |
| Anandaraja | 98.77±39.08 | 5.47 | 0.937 | <0.001 |
| Puavillai | 104.32±39.12 | 0.08 | 0.978 | 0.284 |
| Hattori | 94.88±36.53 | 9.36 | 0.977 | <0.001 |
Calculation by different formulas, At TG range 151mg/dl to 199 mg/dl; N= 3449.
| Method | Mean±SD(mg/dl) | Mean Difference(mg/dl) | Correlation (r) | p-value obtained by paired t-test |
|---|
| Direct | 120.44±40.78 | | | |
| Friedewald | 112.37±41.82 | 8.07 | 0.977 | <0.001 |
| Cordova | 110.04±31.49 | 10.40 | 0.976 | <0.001 |
| Vujovic | 121.64±41.85 | -1.2 | 0.977 | <0.001 |
| Ahmadi | 173.32±36.48 | -52.88 | 0.945 | <0.001 |
| Anandaraja | 104.08±42.08 | 16.36 | 0.965 | <0.001 |
| Puavillai | 118.09±41.84 | 23.51 | 0.977 | <0.001 |
| Hattori | 105.28±39.31 | 15.16 | 0.977 | <0.001 |
For TG range 200-399mg/dL, presented in [Table/Fig-4], the mean value obtained from Vujovic formula correlated best with the mean value of LDLc obtained by direct method (r = 0.968). For TG range >400 mg/dL, presented in [Table/Fig-5], all formulae showed high significant mean difference. Among the various formulae, the mean value obtained by Vujovic formula correlated best with direct measurement r = 0.791 and mean difference of 5.89mg/dL.
Calculation by different formulas. At TG range 200mg/dl to 399 mg/dl; N= 3515.
| Method | Mean±SD(mg/dl) | Mean Difference(mg/dl) | Correlation (r) | p-value obtained by paired t-test |
|---|
| Direct | 120.56±45.74 | | | |
| Friedewald | 107.43±47.83 | 13.13 | 0.966 | <0.001 |
| Cordova | 119.78±35.89 | 0.78 | 0.957 | 0.003 |
| Vujovic | 121.55±47.63 | -0.99 | 0.968 | <0.001 |
| Ahmadi | 231.67±50.26 | -149.11 | 0.736 | <0.001 |
| Anandaraja | 95.98±47.92 | 24.57 | 0.960 | <0.001 |
| Puavillai | 116.14±47.69 | 4.42 | 0.967 | <0.001 |
| Hattori | 100.46±44.97 | 20.10 | 0.966 | <0.001 |
Calculation by different formulas, At TG range >400mg/dl; N= 557.
| Method | Mean±SD(mg/dl) | Mean Difference(mg/dl) | Correlation (r) | p-value obtained by paired t-test |
|---|
| Direct | 106.99±46.95 | | | |
| Friedewald | 67.87±79.97 | 39.12 | 0.701 | <0.001 |
| Cordova | 143.18±43.18 | -36.19 | 0.761 | <0.001 |
| Vujovic | 101.10±66.69 | 5.89 | 0.791 | 0.001 |
| Ahmadi | 444.46±217.37 | -337.47 | 0.019 | <0.001 |
| Anandaraja | 56.28±75.88 | 50.71 | 0.727 | <0.001 |
| Puavillai | 88.38±71.38 | 18.61 | 0.756 | <0.001 |
| Hattori | 62.57±75.72 | 44.42 | 0.697 | <0.001 |
Discussion
Friedewald formula is the formula of choice for LDLc calculation in most laboratories across the world. Many studies have shown its limitation and some have shown that other equations perform better for certain groups of populations.
This study compared several formulae including Friedewald formula with direct LDLc measurement. Our results indicate that Friedewald formula fails to provide a good result at TG> 400mg/dL. This is contradictory to the study done by Sha MFR et al., in Bangladeshi population which concluded that Friedewald formula can be used up to serum TG concentration of 700mg/dL [7].
Cordova and Cordova suggested a new formula that performed better than Friedewald formula in Brazilian population over a wide TG range [9]. Our results show that Cordova formula does not provide any significant advantage over Friedewald formula in Indian population.
Ahmadi formula has been validated in Iranian subjects at TG<300mg/dL [11]. In our study, it performed well only at TG ranges <150mg/dL. At all other TG ranges it performs poorly. Hence, it is not suitable to be used in Indian population.
This study supports the study done by Gupta S et al., which concluded that Anandaraja formula does not provide any advantage over Friedewald formula for LDLc estimation in Indian population [19]. This is contradictory to the study done by Anandaraja et al., which found their formula more accurate than Friedewald formula for TG <350mg/dL [12]. The modified Friedwald equation developed by Puavillai et al., also correlated well with direct measurement and performed better than Friedewald formula at TG range >200mg/dL in Indian population [13].
Hattori formula developed by Hattori et al., shown to perform better than Friedwald formula in Japanese population does not provide any advantage over Friedewald formula in Indian population [16].
Our study supports the study done by Vujovic et al., which validated a modified formula in Serbian population with TG<400mg/dL [10]. They concluded that there is no significant difference between LDLc calculated by Vujovic formula and directly measured LDLc. Our results show that Vujovic formula shows good correlation at TG range <150mg/dL. It performed better than any other formulae at TG ranges 151-199mg/dL, 200-399mg/dL and TG> 400mg/dL. Hence, in this group of Indian population, it performed better than any other formula.
Limitation
This study compares calculated LDLc with direct LDLc assay and not with the reference method i.e., ultracentrifuge and precipitation for comparison. Also, the study uses only one assay for TG, TC, LDLc and HDLc and other assay methods have not been considered. Another limitation is that the number of samples with TG>400mg/dL was small. Finally, several other equations for LDLc calculation besides the ones used here have been described which have not been taken into consideration.
Conclusion
We propose that Vujovic formula is most suitable for estimation of LDLc in Indian population. It can be used over a wide TG range and should be preferred over other formulae for calculation of LDLc in resource- poor settings. However, more studies using larger sample sizes, from different ethnic and geographical populations and under different settings and preferably compared with the other reference method are recommended.
Conversion factors to SI units: To convert TG from mg/dL to mmol/L multiply by 0.01129. To convert total cholesterol, LDLc and HDLc from mg/dL to mmol/L multiply by 0.02586.
[1]. Cheng AY, Leiter LA, Implications of recent clinical trials for the National Cholesterol Education Program Adult Treatment Panel III guidelines Curr Opin Cardiol 2006 21:400-04. [Google Scholar]
[2]. Expert panel on detection, evaluation, and treatment of High Blood Cholesterol in adultsExecutive summary of the third report of the national cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of High Blood Cholesterol in adults (Adult Treatment Panel III) J Am Med Assoc 2001 285:2486-97. [Google Scholar]
[3]. Bachorick PS, Ross JW, National Cholesterol Education Program recommendations for measurement of low density lipoprotein cholesterol: executive summary Clin Chem 1995 41:1414-20. [Google Scholar]
[4]. Bairaktari ET, Seferiadis KI, Elisaf MS, Evaluation of methods for the measurement of low-density lipoprotein cholesterol J Cardiovasc Pharmacol Therapeut 2005 10:45-54. [Google Scholar]
[5]. Nauck M, Warnick GR, Rifai N, Methods for measurement of LDL-cholesterol: a critical assessment of direct measurement by homogeneous assays versus calculation Clin Chem 2002 48:236-54. [Google Scholar]
[6]. Friedewald WT, Levy RI, Fredrickson DS, Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultra centrifuge Clin Chem 1972 18:499-502. [Google Scholar]
[7]. Sha MFR, Siddique AH, Saiedullah M, Khan MAH, Evaluation of recently developed regression equation with direct measurement of low-density Lipoprotein Cholesterol in a Bangladeshi population J Enam Med Col 2015 5(2):75-79. [Google Scholar]
[8]. Nigam PK, Calculated low density lipoprotein-cholesterol: friedewald’s formula versus other modified formulas International Journal of Life Science and Medical Research 2014 4(2):25-31. [Google Scholar]
[9]. Cordova CM, Cordova MM, A new accurate simple formula for LDL-cholesterol estimation based on directly measured lipids from a large cohort Ann Clin Biochem 2013 50:13-19. [Google Scholar]
[10]. Vujovic A, Kotur-Stevuljevic J, Spasic S, Bujisic N, Martinoric J, Vujovic N, Evaluation of different formulas for LDLc calculation Lipids Health Dis 2010 9:27-35. [Google Scholar]
[11]. Ahmadi SA, Boroumand MA, Moghaddam KG, Tajik P, Dibaj SA, The impact of low serum triglycerides on low density lipoprotein cholesterol estimation Arch Iranian Med 2008 11:318-21. [Google Scholar]
[12]. Anandaraja S, Narang R, Godeswar R, Laksmy R, Talwar KK, Low density lipoprotein cholesterol estimation by a new formula in Indian population Int J Cardiol 2005 102:117-20. [Google Scholar]
[13]. Puavillai W, Laorugpongse D, Is calculated LDLc by using the new modified Friedewald equation better than the standard Frieldewald equation J Med Assoc Thai 2004 87:589-93. [Google Scholar]
[14]. Planella T, Cortes M, Martinez-Bur C, Gonzalez-Sastre F, Ordonez-Llanos J, Calculation of LDL-cholesterol by using apolipoprotein B for classification of nonchylomicronemic dyslipemia Clin Chem 1997 43:808-15. [Google Scholar]
[15]. Wagner AM, Zapico E, Boner R, Parez A, Ordonez-Llanos J, The effect of VLDL particles on the accuracy of a direct LDL-cholesterol method in type 2 diabetic patients Clin Bio Chem 2003 36:177-83. [Google Scholar]
[16]. Hattori Y, Suzuki M, Tsushima M, Yoshida M, Tokunaga Y, Wang Y, Development of approximate formula for LDL-cholesterol, LDL-apo B and LDL-chol/LDL-apo B as indices of hyperapobetalipoproteinemia and small dense LDL Atherosclerosis 1998 138:289-99. [Google Scholar]
[17]. College of American Pathologists 2009 C-B proficiency survey for Chemistry/Therapeutic drug monitoring [Google Scholar]
[18]. Agrawal M, Spencer HJ, Faas FH, The method of LDL Cholesterol measurement influences classification of LDL Cholesterol to treatment goals: Clinical research study J Investig Med 2010 58(8):945-49. [Google Scholar]
[19]. Gupta S, Verma M, Singh K, Does LDL-C estimation using Anandaraja’s Formula give a better agreement with direct LDL-C estimation than the Friedewald’s Formula? Ind J Clin Biochem 2012 27(2):127-33. [Google Scholar]