In Reply:-Anderson and Holford disagree with the values for Vd/f that we estimated for rectally administered acetaminophen. To support their claim, they cite two values for the relative bio-availability of rectal versus oral acetaminophen:
1. They claim that “bioavailability of rectal compared to oral acetaminophen formulations has been reported as 0.52 (range, 0.24–0.98),” referring to a manuscript by Montgomery et al. [1]Unfortunately, those data were not obtained by Montgomery et al. but are reported in those authors' introduction as the results of “an unpublished adult study,” in which a SmithKline Beecham preparation (rather than the Upsher-Smith preparation used in our study) was examined. We question the relevance of a study performed in adults, the citation of “unpublished data,” whose accuracy cannot be verified, and data from a different preparation.
2. They cite a rectal-oral bioavailability ratio of 0.3. This is based on a study in which acetaminophen concentrations peaked at 3 h after rectal administration, yet the final (of four) samples was obtained at 4 h. [2]It is likely that those investigators underestimated the area under the plasma concentration versus time curve, thereby underestimating the relative bioavailability of rectally administered acetaminophen.
Anderson and Holford simulate plasma acetaminophen concentrations that might occur with a 20 mg/kg rectal dose. We agree that the mean concentrations observed with this dose do not overlie the simulated values. However, figure 2 in our manuscript demonstrates that mean concentrations for the three doses differ and that our 10-mg/kg dose yields a peak concentration of 4.0 micro gram/ml at approximately 200 min and that our 30-mg/kg dose (normalized to a dose of 10 mg/kg) yields a peak concentration of 3.7 micro gram/ml at approximately 220 min. [3]These times-to-peak concentration are consistent with Anderson and Holford's simulations. Doubling these peak concentrations (to predict the peak concentration attained with a 20-mg/kg dose) yields values of 8.0 and 7.4 micro gram/ml, slightly less than the values predicted by Anderson and Holford's simulations. This difference is expected in that concentrations for each patient should peak at different times so that the average concentration at the median peak time should be less than the average of the individual peak concentrations (Figure 1). Note also that Anderson and Holford use only our most discrepant data (the data from the 20-mg/kg dose) to criticize our model.
Anderson and Holford agree with our claim that suppository size may affect absorption characteristics but are concerned that there is no consistent pattern in the dissolution times. We agree and note in our manuscript that additional studies are needed to determine “factors influencing differences in dissolution.” Our model was developed because, with the traditional first-order adsorption model, we observed that “the pharmacokinetics of acetaminophen varied as a function of the dose administered as smaller-… versus larger-dose suppositories.” Allowing for a more complex absorption model markedly improved the quality of the fit. Readers are referred to our manuscript for additional detail.
In their pharmacokinetic analysis of rectally administered acet-aminophen, Anderson et al. [4]used the traditional first-order absorption model. Unfortunately, their manuscript provides no graphics that demonstrate whether their absorption model fits the early plasma concentration data. In addition, their first sample was obtained 1 h after drug administration (whereas our first sample was obtained at 30 min), limiting their ability to determine whether their absorption model fit the plasma concentrations that occurred during the initial absorption phase. If they do not look, they will never know if their pharmacokinetic model misspecifies the early absorption phase.
In summary, we appreciate Anderson and Holford's interest in our analysis. However, we contend that their simulations are compared selectively, rather than with our entire dataset. In addition, their claim about relative bioavailability of rectal versus oral preparations of acetaminophen is based on questionable data.
Dennis M. Fisher, M.D.
Professor of Anesthesia and Pediatrics; University of California; San Francisco, California 94143–0648
Patrick K. Birmingham, M.D.
Assistant Professor of Anesthesiology
Charles J. Cote, M.D.
Professor of Anesthesiology and Pediatrics; Northwestern University Medical School; Children's Memorial Hospital; Chicago, Illinois 60614
(Accepted for publication December 17, 1997.)