Differences between repeated lipid profile measurements in a tertiary hospital over a short time period

Abstract

Background: Measurement of the plasma lipid profile, mainly low-density lipoprotein cholesterol (LDL-C), is widely used in the management of hospitalized patients as part of their cardiometabolic risk assessment. In common practice, LDL-C is calculated indirectly by the Friedewald equation. For many years, fasting of 8–14 h is needed to obtain an accurate lipid profile measurement, although recent guidelines do not necessitate it. The aim of this study was to find patients with two consecutive LDL-C measurements taken over a short time period on the same admission to see if a significant difference exists and to suggest reasons that may explain it. We also aim to define whether the difference between LDL-C calculated by the Friedewald equation is diminished while using the newer Martin/Hopkins, de Cordova or Sampson/NIH equations. Methods: This was a retrospective cohort study performed in one medical center in Israel. In a five-year time period, 772 patients with two repeated LDL-C measurements taken on the same admission were found. The median time gap between tests was 2 days. Correlations between laboratory results and LDL-C measurements were determined. Results: A total of 414 patients (53.6%) had a difference greater than the acceptable total error of 8.9% in LDL-C calculation using the Friedewald equation, with a mean 25.8% difference between the two tests. Newer LDL-C calculations showed less diversity. Non-HDL-C was found as the only variable with a major correlation with LDL-C results in all equations. A weaker correlation was found with HDL-C. Triglycerides showed an even weaker correlation, and glucose differences had no correlation with LDL-C differences. Conclusions: Repeated LDL-C measurements can vary widely, even during a short period of hospitalization. In this study, more than half of the patients had a significant difference between their consecutive LDL-C results. This wide difference between two consecutive tests was diminished using newer calculations, yet not well explained. The fasting state likely has no effect on LDL-C levels. The results of this study might emphasize that many factors influence LDL-C calculation, especially in the disease state. Further research is needed, especially in looking for a more accurate LDL-C calculation from existing formulas.