Derivation and Validation of a Novel Method for More Accurate Estimation of LDL-C from the Standard Lipid Profile
Editor's Note: Commentary based on Martin SS, Blaha MJ, Elshazly MB, Toth PP, Kwiterovich PO, Blumenthal RS, Jones SR., Comparison of a novel method vs the Friedewald equation for estimating low-density lipoprotein cholesterol levels from the standard lipid profile. JAMA. 2013 Nov 20;310(19):2061-8.
Low density lipoprotein cholesterol (LDL-C) has been the primary atherogenic lipoprotein measure used in clinical practice since the paper by Friedewald et al1 where a simple method allowing estimation of LDL-C from total cholesterol (TC), high density lipoprotein cholesterol (HDL-C) and triglycerides (TG) was derived. The key finding in this original study was the fairly consistent 5:1 ratio in mg/dL between TG and very low density lipoprotein cholesterol (VLDL-C) allowing VLDL-C, and hence LDL-C to be estimated from the basic lipid panel as (TC) – (HDL-C) – TG/5. The Friedewald estimate performs very poorly at secondary prevention levels of LDL-C, underestimating actual LDL-C.2 These findings led us to develop a novel high accuracy algorithm circumventing many of these problems, building upon the basic concept of the Friedewald relationship.
We used the Very Large Database of Lipids (VLDL, NCT 01698489),3 a database of vertical spin ultracentrifugation lipids in over 1.3 million subjects. We generated TG/VLDL-C Friedewald-like factors in a derivation cohort of 900,605 in a TG by non-HDL-C quantile array of initially 2000 cells, simplifying the model without loss of predictive accuracy to 360 and ultimately 180 cell arrays. The novel estimates of LDL-C using adjustable TG/VLDL-C factor arrays were then compared to estimates using the original Friedewald equation and to the referent directly measured values in a validation cohort of 450,303 subjects.
Guideline classification by European and American LDL-C targets relative to directly measured LDL-C was highly (>88%) concordant using all 3 LDL-C estimates at LDL-C >130 mg/dL. At secondary prevention LDL-C levels 70-99 mg/dL, Friedewald estimates were concordant in 84.7%, while the 180 and 360 cell array novel estimates were concordant in 92.6% and 92.7% respectively. Results were similar at LDL-C levels <70 mg/dL with Friedewald estimates concordant in only 77.2% while the 180 and 360 cell array novel estimates were concordant in 94.0% and 94.1% respectively. Concordance of classification was highly dependent upon TG levels. All methods performed well at TG<100 mg/dL with concordance rates >95%. However, at TG 150-199 and 200-399 mg/dL, the Friedewald method performed poorly with only 61% and 40% concordance, whereas the novel estimates were concordant in 92% and 84% respectively, representing a substantial improvement over the Friedewald estimate.
A novel method for estimation of LDL-C using a variable TG/VLDL-C factor provides substantial improvement in accuracy of estimation over the Friedewald method.
Over the years since the original publication by Friedewald et al and the widespread adoption of the method for estimation of LDL-C, the limitations of this method were identified with various attempts to remedy the shortcomings of this attractively simple relationship by either changing the functional form of the TG – VLDL-C relationship or slightly altering the original TG/VLDL factor. None of these really solved the problem at its root cause: high variance of the TG/VLDL-C ratio around a central population value of about five. This is an inescapable consequence of lipoprotein biology and pathophysiology. The TG content of VLDL varies with prandial state, various diseases such as diabetes and other insulin resistant states, and genetic variation in activity of various lipases central to triglyceride-rich lipoprotein metabolism.4 To compound the problem, the original Friedewald model was based on a derivation set consisting of a relatively small number of subjects with few in the current range of LDL-C relevant to secondary prevention. At higher LDL-C and low TG, VLDL-C is small relative to LDL-C and resulting error in estimation of LDL-C is relatively small such as in patients with familial hypercholesterolemia. Yet these patients represent a relatively small fraction of the encounters in everyday practice where an accurate estimate of LDL-C is clinically important. Especially in those patients with a secondary prevention indication for measurement of LDL-C, the Friedewald estimates are unreliable are low LDL-C levels in the target range in the setting of hypertriglyceridemia. The novel method presented in this paper builds upon the original Friedewald concept, potentially obviating the need for additional laboratory measurements in many patients where the Friedewald estimate is inaccurate. The results of this study need to be validated in additional cohorts; however, the robust model based on very large derivation and initial validation cohorts lends a greater degree of confidence in the resulting novel LDL-C estimate compared with the original Friedewald estimate based on a small derivation cohort of only 448 subjects. Like Friedewald estimation, the novel method does not perform well in patients with type III dyslipidemia and should be viewed cautiously in hypertriglyceridemic patients with TG>200 mg/dL. Broader, more inclusive measures of atherogenic lipoprotein burden such as non-HDL-C5 or apolipoprotein B should be considered in these patients as treatment goals.
- Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18:499-502.
- Martin SS, Blaha MJ, Elshazly MB, Brinton EA, Toth PP, McEvoy JW, Joshi PH, Kulkarni KR, Mize PD, Kwiterovich PO, Defilippis AP, Blumenthal RS, Jones SR. Friedewald Estimated versus Directly Measured Low-Density Lipoprotein Cholesterol and Treatment Implications. J Am Coll Cardiol 2013;62:732-9.
- Martin SS, Blaha MJ, Toth PP, Joshi PH, McEvoy JW, Ahmed HM, Elshazly MB, Swiger KJ, Michos ED, Kwiterovich PO, Kulkarni KR, Chimera J, Cannon CP, Blumenthal RS, Jones SR. Very Large Database of Lipids: Rationale and Design. Clin Cardiol 2013; 36: 641-48.
- Wang T, Nakajima K, Leary ET, Warnick GR, Cohn JS, Hopkins PN, Wu LL, Cilla DD, Zhong J, Havel RJ Ratio of remnant-like particle-cholesterol to serum total triglycerides is an effective alternative to ultracentrifugal and electrophoretic methods in the diagnosis of familial type III hyperlipoproteinemia. Clin Chem 1999;45(11):1981-7.
- 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) final report. Circulation 2002;106:3143-421.
< Back to Listings