Elsevier

Clinical Biochemistry

Volume 44, Issues 17–18, December 2011, Pages 1451-1456
Clinical Biochemistry

Extreme concentrations of high density lipoprotein cholesterol affect the calculation of low density lipoprotein cholesterol in the Friedewald formula and other proposed formulas☆

https://doi.org/10.1016/j.clinbiochem.2011.09.009Get rights and content

Abstract

Objectives

To investigate the effect of extreme levels of high density lipoprotein cholesterol (HDL-C) in the calculation of low density lipoprotein cholesterol (LDL-C) using Friedewald's formula (FF) and other formulas proposed recently.

Design and methods

Lipoprotein profile was performed in 2603 samples with HDL-C ≤ 20 mg/dL and 1953 samples with HDL-C ≥ 100 mg/dL.

Results

Wilcoxon's and Student's t-tests showed significant differences (p < 0.001) between calculated LDL-C by different formulas and direct determination in the two groups of HDL-C values. Passing–Bablok regression and Bland–Altman plot showed disagreement for the four formulas studied, except for Vujovic formula in the HLD-C ≥ 100 mg/dL group.

Conclusions

Our results suggested that none of the formulas under analysis should be used for estimating LDL-C in samples with extreme HDL-C concentrations due to absence of statistical correlation with LDL-C direct measurement.

Highlights

► Friedewald Formula is the most usual approach in clinical routine to calculate LDL-C. ► Different limitations of FF have been reported. ► We evaluated the effect of extreme low or high HDL-C levels on LDL-C estimation. ► Estimation formulas should not be used to calculate LDL-C with extreme HDL-C levels.

Introduction

Low density lipoprotein cholesterol (LDL-C) serum concentration is a well-established risk factor for diagnosing and evaluating atherosclerosis and cardiovascular disease (CVD). The US National Educational Cholesterol Program (NCEP) Adult Treatment Panel III (ATP III) suggests different LDL cut-off points for therapy depending on patient's pathologies and the presence of other risk factors [1].

The reference method for determining LDL-C is β-quantification (BQ), which is a laborious and time-consuming technique that requires an ultracentrifuge and trained staff. Thus, the most usual approach in clinical routine work is to calculate LDL-C indirectly from concentrations of total cholesterol (TC), high density lipoprotein cholesterol (HDL-C) and triglycerides (TG) using the Friedewald formula (FF) [2]. However, different limitations of FF that lead to erroneous results have been reported. These limitations, in which FF should not be used, include samples containing chylomicrons, patients with type III hyperlipidemia and serum TG > 400 mg/dL. Contradictory results have also been described in several pathologic states (diabetes, nephropathies or hepatopathies) [3], [4] and with TG between 200 and 400 mg/dL [5], low concentrations of TC [6] or LDL-C [7]. Therefore, some authors have published their own formulas in order to improve FF in diverse populations [8], [9], [10]. In addition, homogeneous assays have been developed for direct measurement of LDL-C concentration, which have become popular for determining LDL-C in samples with some of the FF limitations. They have the advantages of being completely automated, performing without any manipulation, showing better precision and reproducibility than the older heterogeneous assays and meeting the NCEP analytical goals [11], [12].

HDL-C serum levels are inversely correlated with CVD risk and consequently are used in many coronary risk algorithms. These levels are influenced by sex, genetic background, lifestyle, metabolic disorders and drug treatment, which may cause extreme HDL-C values. Usually, pharmacological therapy is established in patients with low HDL-C concentration although drug strategy also depends on the presence of high LDL-C levels [13].

Despite all the information published related to FF and its limitations, there are no data about FF validity in samples with low or high HDL-C levels. The objective of this paper is to assess the accuracy of FF and other recently proposed formulas for calculating LDL-C compared with a direct homogeneous method (DM) in a Spanish population with extreme HDL-C concentrations.

Section snippets

Study population

Population under analysis was formed by Spanish fasting-patients that underwent routine lipid determination during 2008 and 2009. Samples with HDL-C ≤ 20 mg/dL and ≥ 100 mg/dL were selected for further analysis. Samples with relevant dyslipidemia (TC ≥ 300 mg/dL, TG ≥ 300 mg/dL or TC ≥ 250 mg/dL with TG ≥ 200 mg/dL) were excluded. In addition and following the same criteria, 1000 samples with HDL-C between 21 and 99 mg/dL were selected randomly from our 2009-database and considered as “normal” or “non-extreme”

Results

1953 samples were found with HDL-C ≥ 100 mg/dL and 2603 with HDL-C ≤ 20 mg/dL, out of more than 300,000 analysis carried out during 2008–2009. In the evaluation of the AF with HDL-C ≤ 20 mg/dL, 98 samples with negative or zero LDL-C result were given a value of 1 for further analysis. Lipoproteins profile, CLDL-C, %ΔLDL-C and comparison between DM and CLDL-C using Passing–Bablok regression and Bland–Altman plot (Fig. 1) are shown in Table 1B–C.

First, Kolmogorov–Smirnov test was performed to study

Discussion

FF is the most common approach for measuring LDL-C in clinical laboratories and, in fact, its use is accepted by the NCEP and worldwide spread. Nevertheless, it is quite important to know when FF cannot be applied. Apart from its well-known limitations and its questionable use in hypertriglyceridemic patients [14], erroneous results for low concentrations ranges of TC, LDL-C and TG have also been reported [6], [7], [15]. However, there are no previous concrete studies about the influence of

Conclusions

The results shown here provide the first evidence that neither FF nor the other formulas considered in this study can be used for LDL-C calculation in samples with extreme HDL-C levels due to the significant statistical difference between CLDL-C and DM-LDL-C and the important percentage of misclassification into the upper NCEP risk categories. For the generalization of these results, they must be confirmed by using the reference BQ procedure and we recommend testing our findings in other

Competing interests

The authors stated that there are no conflicts of interest regarding the publication of this article.

Acknowledgments

The authors thank the laboratory technical staff for their valuable assistance.

References (27)

  • W.T. Friedewald et al.

    Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge

    Clin Chem

    (1972)
  • C. Matas et al.

    Limitations of the Friedewald formula for estimating low-density lipoprotein cholesterol in alcoholics with liver disease

    Clin Chem

    (1994)
  • J. Rubiés-Prat et al.

    Calculated low-density lipoprotein cholesterol should not be used for management of lipoprotein abnormalities in patients with diabetes mellitus

    Diabetes Care

    (1993)

    • Cited by (15)

    • Comparing calculated LDL-C with directly measured LDL-C in healthy and in dyslipidemic children

      2017, Clinical Biochemistry
      Citation Excerpt :

      In the present study the underestimation of cLDL-C by FF was positively correlated with HDL-C levels equal or above 60 mg/dl. In a study by Timón-Zapata et al., the accuracy of FF, AF as well as of two more formulas calculating LDL-C compared to direct measurement of LDL-C was evaluated in adults with extreme HDL-C levels [23]. In very low HDL-C (≤ 20 mg/dl) none formula was valid and more than 95% of misclassified cases were assigned at an upper cardiovascular risk category.

    • A multicentre analysis of four low-density lipoprotein cholesterol direct assays in samples with extreme high-density lipoprotein cholesterol concentrations

      2014, Clinica Chimica Acta
      Citation Excerpt :

      These types of LDL are quite different from normal LDL-C in terms of lipoprotein feature and structure, which results in a partial and incomplete recognition by chemical reagents that could vary between the distinct LDL-C homogeneous assays. We have previously described relevant discrepancies between FF and other proposed formulas in the comparison with a direct method (LDL-C Plus reagent from Roche) in a large group of samples with extreme HDL-C concentrations [6,23]. In concordance with us, Choi et al. found that HDL-C ≤ 40 mg/dL was one of the factors that cause significant differences between LDL-C calculated by the FF and directly measured LDL-C [24].

    • Development of a Modified Friedewald’s Formula to Calculate Low-Density Lipoprotein in an Iranian Population

      2023, Iranian Journal of Medical Sciences

    • Best practice for LDL-cholesterol: when and how to calculate

      2023, Journal of Clinical Pathology

    • High density lipoprotein as a therapeutic target: Focus on its functionality

      2023, Biocell

    • How should low-density lipoprotein cholesterol be calculated in 2022?

      2022, Current Opinion in Lipidology
    View all citing articles on Scopus
    ☆

    Conversion factors to SI units: To convert triglycerides from mg/dL to mmol/L multiply by 0.01129. To convert cholesterol, LDL-C and HDL-C from mg/dL to mmol/L multiply by 0.02586.

    1

    Both authors must be considered as main authors.

    View full text
    Copyright © 2011 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.