Drs. Puig, Pol, and Warner  studied the interactive effect of morphine and clonidine on gastrointestinal transit. They determined the ED20to ED80of each drug separately and of mixtures of the two in three proportions: 1:1 (equal fractions of the ED sub 50 of each), 1:0.33, and 1:3. They concluded that, with the 1:3 and 1:1 mixtures, the interaction between morphine and clonidine was synergistic at 20% and 50% inhibition but antagonistic at 60% and 80% inhibition. We congratulate the authors on the comprehensiveness of their experimental work, but we are disconcerted by the number of inconsistencies between the tables and figures-and even within a table.
In their figures 2 and 3, the SEMs on the ED20, ED50, and ED80values for morphine and clonidine on their own are mostly much smaller than those given in their table 1, whereas the SEMs for the mixtures are sometimes smaller or sometimes larger than those in their table 2.
Table 2 also showed an internal anomaly: with the “1:1” mixture, the ratios of doses, morphine:clonidine, are fairly close to the 16:1 of the ED50values of table 1. However, for the 1:3 mixture, the ratios should be approximately (16/3):= 5.3:1, whereas in the table, they range from 3.7:1 to 0.9:1. Similarly, for the 1:0.33 mixture, the ratios should be approximately (16/0.33):1 = 48:1; in fact they are all about 240:1.
In the graph for ED60(their fig. 3), the coordinates of the “(1:1)” interaction point appear to be the ED60values for morphine and clonidine individually, obtained by interpolation in their table 1. Correspondingly, the doses of morphine and clonidine at the ends of the ED60isobole line appear to be derived, not by interpolation for each drug in table 1, but from the “1:1” mixture line in figure 1 (and the 16:1 ratio of actual doses from table 1). In other words, the authors appear to have swapped the two items of data.
When the graph for ED60(their fig. 3) is correctly plotted, it shows a probably nonsignificant synergism. Also, in the graph for ED80(their fig. 3), if the SEM for the clonidine in the 1:3 mixture is as large as given in their table 2 (0.35 mg/kg), the error bar will overlap the isobole line. Thus, even on their own, lenient criterion ("points were considered to differ significantly from additivity if their SEMs did not overlap [the isobole line]"), the authors do not appear to have demonstrated antagonism by the isobologram method.
Takashi Asai, M.D.
Honorary Research Follow
W. W. Mapleson, D.Sc., F.Inst.P., F.R.C.A.(Hon)
Professor Emeritus; Department of Anaesthetics and Intensive Care Medicine University of Wales College of Medicine; Cardiff, Wales; United Kingdom
Electronic mail: firstname.lastname@example.org
(Approved for publication July 7, 1997.)