Apart from a misleading title, there are some methodological flaws worth commenting upon, in the work presented by Mazoit & Samii [1]. Our group has been researching and reporting on the binding properties of propofol for several years. We also have experience in pharmacokinetics/pharmacodynamics. The part of the title making reference to that field is inappropriate since the work presented was in vitro, on isolated proteins. Only in the discussion section there is a general mention of the implications that changes in the free fraction of a drug can have on pharmacokinetics and pharmacodynamics. We think scientific article titles (also keywords) should represent faithfully the true contents of the original research, not that of implied knowledge, and should be more carefully selected. Further, it is known that propofol adsorption to membranes has a time dependence demostrated by validated techniques (5 min centrifugation at 37 °C) [2–4]. In reference to the methodology of Mazoit & Samii on page 36, they are vague when mentioning that ‘…(found) an adsorption ≥ 20% on the dialysis and ultrafiltration membranes tested depending on the material and the membrane cut-off’. Then, discarding the established technique, they proceed to employ an alternative. In this instance, the authors fail to specify the conditions leading to their results (time or cut-off of their membrane system) and hence to justify their method. We certainly agree with the fact that propofol binds to human serum albumin (HSA). We have published similar results in isolated HSA [5]. But, in later reports we point out that this appears to be so when albumin is the only protein present [2–4]. So, in another study with serum from critically ill patients with very reduced albumin count (16.3 g l⁻¹vs 45.8 g l⁻¹ in healthy controls), this protein only accounted for 25% of the minor changes observed in the overall binding of propofol [3]. Furthermore, in patients with hepatic deficiency and significantly reduced albumin levels with respect to healthy volunteers (35.39 g l⁻¹vs 50.85 g l⁻¹), the binding of propofol was not altered [2]. In view of our results with patients we performed a study on isolated lipoproteins from human serum [4]. In this study the lipoprotein fractions were isolated by density gradient ultracentrifugation [6] because the fractions purchased from Sigma Chemical (as in the work of Mazoit & Samii) did not bind propofol. We found that binding was 88% to VLDL, 93% to LDL and 91% to HDL when physiological concentrations of these proteins were employed. Additionally, when we used serum extracted from hyperlipidaemic patients we found that the unbound fraction was significantly diminished with respect to healthy individuals. Regression analysis also showed that there was significant correlation between the bound-to-free ratio of propofol and lipoprotein levels whereas albumin and α₁-acid glycoprotein proved irrelevant [4]. Further work performed by our group involving patients with altered lipoprotein count also indicate that lipoproteins and not albumin (as is suggested by Mazoit & Samii) are the principal components affecting propofol binding in serum.