Study Examines Accuracy of Friedewald LDL-C Calculation
It has long been known that the Friedewald equation used to calculate LDL-C in clinical and trial settings is inaccurate at high triglyceride levels. A study published on March 12 in the Journal of the American College of Cardiology found that the error is large enough to adversely affect clinical decision making.
Researchers examined consecutive lipid profiles from 2009 to 2011 in 1,340,614 U.S. adults with directly measured cholesterol levels using vertical spin density gradient ultracentrifugation. Patients with triglycerides of 400 and higher were excluded due to known inaccuracies in Friedewald. That left a sample of 1,310,440 patients, roughly one of every 180 adults in the U.S.
The mean patient age was 59, and the sample was evenly distributed by gender. The median directly measured LDL-C was 109 mg/dL. The distribution of total cholesterol, HDL-C, triglycerides, and Friedewald LDL-C in the study cohort closely matched population values from the National Health and Nutrition Examination Survey.
In comparing Friedewald LDL-C with directly measured LDL-C, calculated LDL-C was typically lower than measured LDL-C, especially for Friedewald levels less than 100 mg/dL. Patients with higher triglycerides had greater median differences and within-group variance. If Friedewald LDL-C was less than 70, the median direct-measure LDL-C was 9.0 mg/dl higher when triglycerides were 150 to 199 mg/dL. When triglycerides were 200 to 399 mg/dl, the direct-measure LDL-C was 18.4 mg/dl higher than Friedewald LDL-C.
The Friedewald group cautioned that the calculation is not very accurate for very low-density lipoprotein cholesterol (VLDL-C), but the warning seemed to have little clinical impact because VLDL-C concentrations are small relative to LDL-C. But as therapeutic LDL-C goals have fallen to less than 70 and triglyceride levels have risen in the presence of obesity, glucose intolerance and diabetes, that calculation error could be more important.
"The Friedewald equation tends to underestimate LDL-C most when accuracy is most crucial," said lead author Seth Martin, MD, Johns Hopkins Ciccarone Center for the Prevention of Heart Disease, Baltimore. "Especially if triglycerides are greater than 150 mg/dl, the Friedewald estimation commonly classifies LDL-C as less than 70 despite directly measured levels greater than 70. Additional evaluation is warranted in high-risk patients."
The Friedewald equation was introduced into clinical practice in 1972 to work around the significant time and expense of ultracentrifugation to measure LDL-C directly. The equation calculates LDL-C as total cholesterol minus high-density lipoprotein cholesterol (HDL-C) minus triglycerides divided by five. The result is expressed in milligrams per deciliter.
In an editorial comment, John C. LaRosa, SUNY Downstate Medical Center, Brooklyn, New York, discusses how these findings leave clinicians in a conundrum and notes that it is not clinically appropriate to abandon Friedewald, but none of the alternatives have convincing evidence of superior ability to predict cardiovascular risk.
"Changing clinical parameters in clinical settings is a long, slow process, even when the data are more solid than they are today," LaRosa adds. "As a start, we might encourage clinicians and trialists to calculate and report non-HDL-C levels. That requires no additional cost. Non-HDL-C is less cumbersome than estimated LDL-C and would represent an evolutionary, not revolutionary, step forward."
< Back to Listings