I was attracted by the remarkable title of the paper by Kimand colleagues.[1] How can they do that? Measure cerebral concentrations of remifentanil? On reading, it became clear that they had not: the “effect site” concentration was predicted, as the authors stated later, in the discussion. Thus we have the attendant uncertainty of predicting what one patient’s steady state plasma concentration is going to be, on the basis of previous measurements. Perhaps the authors should indicate more clearly the errors in this prediction?

Reading their paper seemed to show a remarkably accurate relationship between effect site concentration and success (figure 2). Butagain: not real data, these dots are predicted. What may appear at the first glance to be data points on this plot are calculated values. Using this plot the EC50 for the predicted remifentanil concentration is 2.84 ng.ml^-1. However the Dixon method and even looking at the raw data doesn’t support this: when the predicted remifentanil value was either 3.0 ng.ml^-1or less, then there was success in 5 subjects and failure in 13. This is consistent with the Dixon method result obtained by the authors. What is the value of the alternative analysis, when it doesn’t seem to accord with the data?

Using a commonly quoted advertising slogan, perhaps the title of a scientific paper should be restrained to doing “what it says on the can”?

References

Kim MK, Lee JW, Jang DJ, Shin OY, Nam SB.Effect-site concentration of remifentanil for laryngeal mask airway insertion during target-controlled infusion of propofol. Anaesthesia. 2009; 64:136-40.