To the Editor Dr Nicholls and colleagues found that adding evolocumab vs placebo to statin treatment among patients with coronary disease resulted in a 1% greater decrease in percent atheroma volume (PAV), measured by serial intravascular ultrasonography (IVUS) imaging, after 76 weeks. 1 Unfortunately, IVUS examinations have inherent variability. 2 Even in the best of laboratories, there are always measurement reproducibility errors but also intraoperator and interoperator variability in the actual plaque measurements. 2 The authors' goal was to detect a nominal treatment difference of 0.71%, assuming a 2.9% SD. 1 To justify this statement, the authors should have provided the within-subject SD (WSSD) in mm 3 for each individual IVUS study and converted to a percentage. The WSSD should not exceed 2.9% at any stage. It is possible that regression to the mean occurred for the SD, in which some individual studies were significantly greater than and some significantly less than the mean. The authors displayed CIs for the mean and median results, which is useful but not complete. The use of a repeatability coefficient (RCO) statistic 2 is the preferred way to show the real magnitude of measurement variability for all IVUS measurements. The RCO statistic can be thought of as a CI for error 2 and represents the value below which the absolute difference between 2 repeated test results may be expected to lie (with a probability of 95%). Thus, one can say that 95% of potential future measurement-related errors in this cohort should not be beyond the value calculated by the RCO. Therefore, any greater change is the true change in the measured plaque and could be attributed to the drug being investigated.Previous work looking at IVUS total plaque volume in mm 3 showed that with a WSSD of 1.2 mm 3 (0.56%) or below between independent measurements by an experienced single operator (intraobserver variability), the RCO was 3.32 mm 3 (1.6% of total plaque volume). When plaque volume was measured by multiple operators (interobserver variability), the WSSD increased to 2.6% of the total volume, with a subsequent RCO of 7.3%.Without considering the WSSD and RCO, it is not possible to confirm that evolocumab was responsible for the observed 1% difference, despite lowering all lipid parameters to historically low levels.