600
GASTROENTEROLOGY
CORRESPONDENCE
The following
replies
refer to the above
five letters
Reply to Dr. Fick et al. We have not systematically studied the effects of infusing various solutions into animals in a preconvulsive state. We did, however, infuse an octanoic acid solution into an ear vein of hyperammonemic rabbits, that were receiving an ammonium chloride infusion into a femoral vein, but no seizures were observed. Thus, although it is possible that the phenomenon observed by Fick et al. is relevant to our observations, we believe that this possibility is unlikely. Reply to Dr. Tomes. We did not attempt to compare precise latencies and amplitudes of the visual evoked potential (VEP) waveform in rabbits and rats. One would not expect them to be the same in different species. We did, however, attempt to compare similarities and differences in the overall pattern of the VEP waveform in response to specific experimental treatments in our rabbits (1,Z) and in rats studied by others (3,~). We agree with Torres that the variability of VEPs in humans is appreciable. In humans, however, the intraindividual variation of VEPs is less than the interindividual variation of VEPs. Accordingly, serial studies of VEPs in a single individual may be clinically useful (5,6). In addition, flash evoked VEPs may be useful in defining qualitative, rather than quantitative, changes in the VEP, particularly in comatose patients (7). Furthermore, serial evoked potential studies, especially studies involving multimodality evoked potentials (i.e., a battery of visual, auditory, and somatosensory evoked potentials] may be useful in assessing latent hepatic encephalopathy or the effects of therapies on hepatic encephalopathy in humans (6) (Pappas SC, et al., unpublished observations). Reply to Dr. Zeneroli. We did state that the abnormal pattern of the VEP in barbiturate-induced coma (in rabbits) is different from that in aminooxyacetic acid-induced encephalopathy (in rats) (1). This statement was part of a more general statement, the purpose of which was to point out that the abnormal VEP waveforms in rabbits with coma induced by drugs that mediate activation of the raminobutyric acid (GABA) neurotransmitter system are similar to each other but do not resemble the abnormal VEP waveforms associated with encephalopathies induced by various other agents (1). In an attempt to provide support for this statement we quoted two relevant studies by Zeneroli et al. in which VEPs had been recorded in rats (3,4). We have reexamined the figure that shows aminooxyacetic acid-induced changes in the VEP in rats (3) and confirm that these changes do not resemble the changes in the VEP induced in rabbits by pentobarbital (1). We have not studied the effects of aminooxyacetic acid on the VEP in rabbits. In the Materials and Methods section of the article by Zeneroli et al. which addresses the synergistic toxins hypothesis in rats (4], it is stated that the EEG was monitored on an oscilloscope but, as stated by us (Z), there is no statement in their article that the EEC was “continuously” monitored or recorded. To be confident that we had not induced seizures during an experiment in which ammonium chloride (NH&l), dimethyldisulfide (DMDS), and octanoic acid (OA) were administered to rabbits, we recorded the EEG on paper and subsequently scrutinized the trace obtained throughout the entire experiment for evidence of seizure activity. We appreciate the additional information provided in the letter by Zeneroli et al. Clearly the interpretation of the data on VEPs in rats (3,4) depends critically on whether seizures were or were not induced by the toxins administered. Reply to Dr. Zieve. Zieve states that we “apparently made no effort to reproduce the conditions of [his] previous studies of combinations of these toxins [NH,Cl, DMDS, and OA] designed to simulate hepatic encephalopathy.” In fact, we did. When we
Vol.
88, No.
2
administered NH&l, DMDS, and OA to rabbits in the same doses (per kilogram body weight) that had previously been shown to induce coma in rats (4), the rabbits exhibited no obvious behavioral changes or changes in the EEG or VEP. We did not “use just any old mixture of toxins.” A major problem for us was to find a combination of NH,Cl, DMDS, and OA that induced encephalopathy in the absence of seizures. Like Zieve’s experience, “after many trials” we ultimately found doses of the three toxins (NH,Cl 1 mmol intravenously over 10 min; DMDS 2 mmol intraperitoneally; OA 3 mmol intraperitoneally) that did not induce overt encephalopathy alone, but when given together induced deep coma and changes in both the EEG and the VEP in the absence of EEG evidence of seizures. We found that “the doses of the three neurotoxins necessary to produce deep coma without seizures are almost as large as those that induce seizures” (2). When we administered DMDS (2 mmol intraperitoneally) and OA (3 mmol intraperitoneally) alone, no encephalopathy or VEP changes were observed, so the VEP changes that occurred when all three of the toxins were administered could not have been attributable to the doses of DMDS and OA alone. Attempts were made to induce encephalopathy in rabbits by administering combinations of the three toxins in which the ratio of the dose of NH,Cl to that of DMDS and the ratio of the dose of NH&l to that of OA were higher than in the mixture mentioned above. In these experiments, depending on the absolute doses of the toxins administered, either no change in the mental status of the animals was observed or seizures occurred; encephalopathy or coma in the absence of seizures was not induced. It is possible that the differences between the VEP waveforms in rabbits and rats after the administration of the three toxins may be attributable, at least in part, to the different routes of administration of NH,Cl, the different doses used, and the different species studied. Clearly, to test the synergistic toxins hypothesis it was necessary for us to select doses of the three toxins that would induce encephalopathy in the absence of seizures. As indicated, great care was taken in selecting the doses that were used to achieve this goal. Our VEP data are therefore relevant to the issue whether the three toxins studied could be implicated in the pathogenesis of hepatic encephalopathy in rabbits. As Zeneroli et al. (4) addressed the same issue in rats, it would be expected that we would discuss the findings in that study in relation to our own findings. Reply to Dr. Raabe. Raabe believes that we were amiss in not pointing out that pentobarbital “significantly affects excitatory synaptic transmission mediated by glutamate and acetylcholine.” In our first article (1) it was not appropriate to provide a comprehensive review of the various effects of pentobarbital on the central nervous system and we made no attempt to do so. We have examined the three references concerning the modes of action of pentobarbital that Dr. Raabe cites. The first (8) describes neurophysiologic studies on several in vitro preparations of cells: R, and R,, cells of the marine mollusk, Apiysia cafifornica; neurosecretory cells of the land snail, Otala lactea; and muscle cells of the walking leg of the lobster, Homarus americanus, and the crayfish, Oreonectes virulis. Of these preparations only the cells from the marine mollusk were neurons. In these preparations from invertebrates pentobarbital was shown to depress excitatory postsynaptic events. The second reference (9) provides data gathered using a preparation of cells from an amphibian, specifically the ninth and tenth sympathetic paravertebral ganglia of the bullfrog, Nana catesbeiana. Using this preparation it was found that pentobarbital mediated two distinct electrophysiologic actions, a depression of nicotinic acid-induced neural excitation and “a postsynaptic enhancement of GABA-mediated synaptic inhibition.” Thus, in these studies, one action of pentobarbital was to block excitation of the target cells. These observations would only be relevant to