BackgroundThe objective of our study was to assess whether calculated low-density lipoprotein cholesterol (LDL-C) is inferior to direct LDL-C (dLDL-C) in identifying patients at higher risk of all-cause mortality, recurrent acute myocardial infarction (AMI), and major adverse cardiovascular event (MACE).MethodsA total of 9,751 patients with coronary artery disease (CAD) undergoing percutaneous coronary intervention (PCI) in the Fuwai PCI registry were included. DLDL-C was measured by the selective solubilization method (Kyowa Medex, Tokyo, Japan). Correct classification was defined as the proportion of estimated LDL-C in the same category as dLDL-C based on dLDL-C levels: less than 1.4, 1.4–1.8, 1.8–2.6, 2.6–3.0, and 3.0 mmol/L or greater.ResultsUnderestimation of LDL-C was found in 9.7% of patients using the Martin/Hopkins equation, compared with 13.9% using the Sampson equation and 24.6% with the Friedewald equation. Cox regression analysis showed compared the correct estimation group, underestimation of LDL-C by the Martin/Hopkins equation did not reduce all-cause mortality (HR 1.26, 95% CI: 0.72–2.20, P = 0.4), recurrent AMI (HR 1.24, 95% CI: 0.69–2.21, P = 0.5), and MACE (HR 1.02, 95% CI: 0.83–1.26, P = 0.9). Similarly, the overestimated group did not exacerbate all-cause mortality (HR 0.9, 95% CI: 0.45–1.77, P = 0.8), recurrent AMI (HR 0.63, 95% CI: 0.28–1.44, P = 0.3), and MACE (HR 1.07, 95% CI: 0.86–1.32, P = 0.6). The results of the diabetes subgroup analysis were similar to those of the whole population.ConclusionCompared with dLDL-C measurement, misclassification by the Martin/Hopkins and Sampson equations was present in approximately 20% of patients. However, directly measured vs. calculated LDL-C did not identify any more individuals in the PCI population with increased risk of all-cause mortality, recurrent AMI, and MACE, even in high-risk patients such as those with diabetes.