LAUSR

To the Editor: I have read the article by Nyberg et al. with interest, and would like to comment on their approach to model fluid volume kinetics. The two-volume model they used is well known to become unstable if the plasma dilution time curve has a flat appearance postinfusion. To prevent this problem, the customary procedure is to equalize the elimination to the urinary excretion. The authors quantified the excreted urine by weighing but did not use this information. Their choice resulted in an unstable model, which is evidenced by a coefficient of variation as large as 123% for the elimination rate constant (k e ). Recent articles using the same population kinetic model report a coefficient of variation for k e of only 5% to 15%. The key result requires clarification. The study compared the plasma dilution in bled volunteers who did and did not receive isoflurane anesthesia. The Abstract says that the maximum plasma dilution was 35% higher, and that the area under the curve for the plasma dilution was 99% larger, in the group that received isoflurane anesthesia. However, the observed data plotted in figs. 6 and 7, as well as my own simulation based on table 1, show that the plasma dilution was similar between both groups and was even slightly lower among those who received isoflurane. I still assume that the Abstract is correct because previous studies show that induction of epidural, spinal, or general anesthesia increase the plasma dilution resulting from infused crystalloid fluid. The magnitude of this dilution depends directly on the decrease in arterial pressure. The reason is retarded distribution. No excessive dilution occurs if the pressure is unchanged. Nyberg et al. established arterial access and measured the pressure, but they did not consider the anesthesia-induced hypotension in their model. Finally, the mean arterial pressure was the strongest predictor of k e in a population volume kinetic analysis of 78 conscious and anesthetized humans receiving crystalloid fluid, as well as in another cohort of anesthetized patients. This potential covariate does not seem to have been considered either. Competing Interests