Anesth Analg 2000;91:1443-1449
© 2000 International Anesthesia Research Society
NEUROSURGICAL ANESTHESIA
Phenytoin, Midazolam, and Naloxone Protect Against Fentanyl-Induced Brain Damage in Rats
Elizabeth H. Sinz, MD,
W. Andrew Kofke, MD, FCCM, and
Robert H. Garman, DVM
Department of Anesthesiology, West Virginia University, Morgantown, West Virginia
Address correspondence and reprint requests to W. Andrew Kofke, MD, West Virginia University, Department of Anesthesiology, 3618 HSC, PO Box 9134, Morgantown, WV 26506-9134. Address e-mail to akofke{at}wvu.edu
 |
Abstract
|
|---|
In previous studies, large-dose fentanyl produced electrographic seizure activity and histologically evident brain damage. We assessed whether fentanyl-induced brain damage is attenuated by using anticonvulsant drugs. Using halothane/nitrous oxide anesthesia, 40 Sprague-Dawley rats underwent tracheal intubation, arterial and venous cannulation, and insertion of biparietal electroencephalogram electrodes and a rectal temperature probe. Halothane was discontinued. The dose of IV fentanyl shown previously to cause maximal brain damage was given to all animals and N2O was discontinued. Control rats were given fentanyl only. Rats in the three study groups also received midazolam, phenytoin, or N2O/naloxone. After characteristic seizure activity began with fentanyl loading the study drug was started. After a 2-h infusion, wounds were closed, and animals recovered overnight and underwent cerebral perfusion-fixation. Neuropathologic alterations were ranked on a scale of 0–5 for both neuronal death (0 = normal, 5 = more than 75% neuronal death) and for malacia. Significantly fewer rats in the N2O/Naloxone, Phenytoin, and Midazolam Groups sustained any brain damage compared with controls. Protection against opioid neurotoxicity is achieved with midazolam, naloxone, and phenytoin. If opioid neurotoxicity is clinically relevant, a small change in anesthetic practice might reduce any potential neurologic morbidity.
Implications: Narcotics in large doses can cause brain damage in rats. This brain damage is attenuated by a narcotic antagonist, a sedative, and an antiepileptic drug.
|
Introduction
|
|---|
Previous studies have demonstrated that large doses of narcotics can produce epileptiform activity in rats (1–3) and humans (4–6). Prolonged seizures can lead to histologically evident brain damage, particularly in the limbic system (7). Opioid-induced seizures in rodents lead to brain damage in the limbic system and other brain areas (1–3). The clinical relevance of these observations has not yet been established. However, observations that large-dose opioids produce epileptiform activity in humans (4–6) and that patients sometimes sustain postoperative cognitive and neuropsychiatric deficits (8,9) referable to limbic system lesions (10) indicates a possibility that opioids may, in some situations, increase the risk of limbic system injury in humans.
After cardiac surgery, the incidence of neurologic deficits is variously estimated at 5–24% of patients (9). Although the etiology of these deficits is likely multifactorial, microembolic ischemia, and global hypoperfusion (11) have been suggested as contributing factors. It is possible that opioid use might contribute to this.
The pathogenesis of opioid neurotoxicity is not known. Cortical interneurons, which express 
-aminobutyric acid (GABA), are an important element of modulation of brain excitation. The interneurons release GABA onto glutamatergic neurons, thus attenuating the propensity for glutamate-induced excitation. On the basis of in vitro studies (12), inhibition of GABA-ergic interneurons is thought to be one likely contributor to opioid-mediated excitation and neurotoxicity. In addition, µ-opiate receptors also are likely contributors to µ-opioid-induced seizure activity and brain damage (13). Finally, the role of electrographic seizure in the genesis of opioid-induced brain damage has not been examined. In previous work (1), large-dose fentanyl produced seizure activity on electroencephalogram (EEG) with apparent relation to histologically evident brain damage.
The purpose of this study was to determine whether opioid-induced seizures and brain damage could be prevented with clinically available anticonvulsant drugs, and, in the process, test the hypotheses in vivo that opioid neurotoxicity is produced by: (a) GABA-ergic inhibition, (b) µ-opiate receptors, and (c) EEG seizure activity.
|
Methods
|
|---|
Approval for this study was obtained from our animal care and use committee. Forty fed male Sprague-Dawley rats were anesthetized with halothane in N2O/O2 (60%/40%) and ventilated to maintain an ETCO2 of 35–45 mm Hg. Vascular access was obtained via cutdown procedures using an aseptic technique and included a femoral artery cannula for continuous mean arterial pressure monitoring and for arterial blood gas samples, and femoral and internal jugular vein cannulae for drug administration. Biparietal stainless steel EEG electrodes were used to evaluate epileptiform activity. Rectal temperature was maintained between 37.5–38.5°C by a servo-controlled homeothermic blanket. Halothane was discontinued when all surgery was complete and allowed to wash out for 1 h. Pancuronium (0.4–0.6 mg) was used as needed to maintain relaxation. One hour after halothane was discontinued, nitrous oxide was also discontinued, at which time all animals were given IV fentanyl (800 mcg/kg initial dose followed by a continuous infusion of 32 mcg/kg/min for 2 h). Animals were randomly assigned to one of four groups: controls given fentanyl only, and three study groups additionally receiving midazolam, phenytoin, or naloxone. Two minutes after characteristic seizure activity had begun with fentanyl initiation the study drug was started. Because naloxone reverses the sedative effects of fentanyl, additional N2O was administered in the Naloxone Group to avoid an unsedated paralyzed animal. Nitrous oxide was chosen because it has demonstrated no effect in prior studies on opioid-induced seizures (2). Dosing was as follows: midazolam 25 mg/kg IV initial dose followed by a continuous infusion of 9.7 mg/kg/hr for 2 h (14), phenytoin 90 mg/kg IV as a single dose (15) over 20 min; naloxone 50 mcg/kg IV initial dose (empirically derived for naloxone) followed by a continuous infusion of 50 mcg/kg/min for 2 h. Arterial blood gas and glucose were measured 20 min before and approximately every half-hour during the 2-h infusions. Once the animal showed initial signs of recovery, manifest with spontaneous, purposeful movement, wounds were closed under brief halothane anesthesia. The animals were then given 100% O2 and when ventilating and moving spontaneously they were extubated. Animals had access to food and water ad lib postrecovery.
On the next day, the rats were reanesthetized with halothane, intubated, and underwent cerebral perfusion-fixation with buffered formalin. The calvaria were removed and the exposed brains were immersed in neutral buffered formalin for at least 24 h before removal. Each brain was sliced at 2.5 mm intervals and the slices were processed by routine techniques and embedded in paraffin. Six-micrometer sections of each brain block were prepared, stained with hematoxylin and eosin, and evaluated by light microscopy. Degrees of neuropathologic alteration (primarily eosinophilic neuron degeneration and malacia) were subjectively graded within all brain sections using a five-tiered scale ranging from 1+ (minimal) to 5+ (severe). "Eosinophilic neuron degeneration" was used to indicate neuron necrosis. Grades for eosinophilic neuron degeneration were subjective but based on the approximate percentage of the neuron population affected within each neuroanatomic area as follows: 0 = none, 1 = minimal (<5%), 2 = mild (6–25%), 3 = moderate (26–50%), 4 = marked (51–70%), and 5 = severe (76–100%). The term "malacia" was assigned when there was such extensive cell loss as to lead to pallor and general loss of brain architecture. Grades for malacia were also subjective but were based primarily on the sizes of the foci of malacia within each neuroanatomic region. The neuropathologic scores for the individual neuroanatomic areas (10 of which are reported here) were weighted according to the approximate size (viz. neuronal population) within each area, and an overall score of neuropathologic damage assigned to each brain. Histopathologic evaluations were performed in a blinded fashion.
Statistical analysis was performed by using SPSS software, version 6.0.1TM (SPSS, Chicago, IL). Physiologic data and animal weights were compared by using a one-way analysis of variance (ANOVA) for each variable with post hoc least significant difference testing. To assess for overall significant pathologic effects, pathologic scores of all brain areas for each rat were summed producing an overall eosinophilic degeneration and a malacia score. Overall scores then underwent Kruskal-Wallis ANOVA followed by Mann-Whitney U-test of treated groups versus control using P < 0.05 to indicate a significant difference. Pathologic data from ten brain areas were analyzed by using the Kruskal-Wallis ANOVA with the significance level adjusted to P = 0.05/10 (=0 0.005) to adjust for multiple comparisons (Bonferroni correction). Within each brain area, between-group comparisons with the control group were made with the Mann-Whitney U-test. The Cochran Q test was used to compare the presence or absence of brain damage among groups.
|
Results
|
|---|
No significant intergroup differences were found in the initial weights, blood gas, or blood glucose values (Table 1). The Fentanyl Control Group had significantly more (P 
0.001) weight loss than the other three groups on the day after the experiment (average 
= 48 g vs 27 g), indicating a lack of appropriate feeding and drinking in these animals overnight. The Fentanyl-Only Control Group maintained a significantly increased mean arterial pressure than each of the study groups (P 0.0125) during the drug infusion period. The Phenytoin Group developed a significant increase in blood glucose concentration that persisted throughout the remainder of the experiment (P 0.001). In addition, the Midazolam Group developed significantly decreased blood glucose concentration for the first hour of drug infusion (P 0.001), although the values were still within physiologic range (mean = 78). Mean time to tracheal extubation was significantly shorter for the Naloxone Group (159 min) than the other three groups (295–303 min) (P < 0.0015).
Analysis of neuropathologic data (Table 2) revealed that all three of the treated groups had a decreased incidence of any brain lesions than controls (Figure 1) (P < 0.015). Assessment of overall pathologic scores revealed a significant effect for eosinophilic neuron degeneration (P = 0.006) with all three treatment groups different than control (P < 0.01). An overall significant effect for malacia occurred (P = 0.02) with all three treatment groups different from control (P = 0.007 for midazolam and naloxone, P = 0.02 for phenytoin). Analysis of changes within ten regions, given the constraints of multiple comparisons, showed no significant effects, although Kruskal-Wallis ANOVA of the septal nucleus suggested an effect with P = 0.006.
View larger version (25K):
[in this window]
[in a new window]
|
Figure 1. Percentage of rats sustaining any brain damage are depicted for each of the experimental groups. All three of the treated groups had a decreased incidence of brain damage than controls (P < 0.015).
|
|
EEG revealed a characteristic epileptiform pattern soon after the administration of the fentanyl bolus in all animals studied. Similar to previous reports (2,3,12) this pattern persisted and intensified throughout the duration of the infusion in the control animals. In the N2O/Naloxone Group, the seizure activity was completely eliminated quickly after the naloxone bolus dose was given, and typically the pattern looked similar to baseline tracings obtained before fentanyl administration. The rats that received midazolam had much less seizure activity than controls; however, their EEG pattern did not return to baseline. The rats in the Phenytoin Group appeared to develop a rhythmic spiking pattern with suppression between spikes and occasional bursts resembling burst-suppression, but also consistent with seizure bursts (Figure 2).
View larger version (25K):
[in this window]
[in a new window]
|
Figure 2. Baseline and treated electrencephalograms for fentanyl-infused rats treated with saline (A, B), midazolam (C, D), naloxone (E, F), and phenytoin (G, H).
|
|
|
Discussion
|
|---|
Large-dose opioids in rats produce hippocampal hypermetabolism (2,16), epileptiform activity (1–3,16), and neuropathologic lesions (1–3). These doses in this species are comparable in potency to a large-dose regimen that might be used in humans. This has been discussed in more detail previously (1,3). It is thus tempting to speculate that these physiologic and electrographic observations in rodents might be clinically relevant. Using bitemporal epidural electrodes, temporal lobe seizure activity has been noted in epileptic humans with induction doses of fentanyl (4) and moderate doses of alfentanil (5). In a prospective human study by Kearse et al. (6), 19 of 20 nonepileptic patients undergoing coronary artery revascularization were found to develop epileptiform activity when given typical doses of fentanyl or sufentanil for this procedure. Qualitative fluorodeoxyglucose positron-emission tomography scan studies in three large-dose fentanyl-treated monkeys have demonstrated the same pattern of limbic system activation as that seen in rats with large-dose fentanyl (17). And, finally, Kofke et al. (17) have recently administered large doses of remifentanil briefly to physiologically supported volunteers during quantitative fluorodeoxyglucose positron emission tomography. They observed a significant increase in temporal lobe glucose use with remifentanil administration.
These data indicate that opioids, unlike other anesthetics, in large doses have the potential to produce brain hypermetabolism. As this might be deleterious in situations of compromised cerebrovascular reserve we chose to assess the effects of potential anticonvulsant drugs midazolam and phenytoin to identify potentially ameliorative concurrent therapies. In addition, we assessed naloxone to identify a potential role for opiate receptors in opioid neurotoxicity.
For this experiment, the two antiseizure medications that were chosen are already commonly used. Midazolam, a benzodiazepine often used as an adjunct in narcotic-based anesthesia for its anxiolytic and amnestic properties, attenuates seizures caused by mercaptopropionic acid or inhaled flurothyl and decreases postseizure maturation of brain damage in rats (14). In the study by Kearse et al. (6) in humans, opioid-induced epileptiform activity was rapidly and consistently attenuated with midazolam administration. Therefore it may be reasonable to routinely add benzodiazepines to large-dose narcotic anesthetics, not only for their known amnestic qualities, but for brain protection as well. There is experimental evidence in animals that suggests that opioids are epileptogenic, partly because they excite neurons by releasing the inhibition of GABA-ergic interneurons (12,18). Midazolam is a potent anticonvulsant that potentiates the actions of GABA on neurons (19) and attenuates opioid-induced seizures. In this study midazolam significantly reduced the brain damage seen in the large-dose fentanyl animals; the protective effect of midazolam may be because of attenuation of seizure activity or secondary to its GABA-ergic effects opposing interneuron disinhibition.
Phenytoin is another commonly used antiseizure medication that greatly increases seizure threshold in rats (15). Although the exact mechanism of action of phenytoin is not known, it is believed that phenytoin at the usual therapeutic dose does not have immediate or direct GABA-ergic effects (20,21). With the usual anticonvulsant dose phenytoin, in vitro, diminishes repetitive firing. However, as the concentration increases, concomitant with decrements in consciousness, phenytoin has other relevant effects that include decreased rate of rise of action potential, decreased spontaneous neuronal firing, decreased convulsant-induced paroxysmal depolarization, and augmented iontophoretically applied GABA (21). We did not measure the serum or cerebrospinal fluid phenytoin concentration, therefore we cannot with certainty indicate where in this spectrum of dose-response our experiments were. The doses chosen were based on anticonvulsant doses used in a kindling model (15). In that study, the phenytoin dose did not produce overt toxicity manifest by coma. It is thus unlikely, in our experiments, that the larger-dose effects occurred. Nonetheless, it is possible that a minor element of GABA-ergic effect contributed to our observations.
As it is likely that midazolam and phenytoin have somewhat different mechanisms for attenuating seizures in these experiments, it is interesting that both drugs were protective against narcotic-induced brain damage. Even more intriguing is the observation that protection occurred with phenytoin, despite continued epileptiform activity as assessed qualitatively by EEG. This suggests that EEG seizure activity may be dissociated from the occurrence of opioid neurotoxicity.
Naloxone, a fairly specific µ-opiate receptor antagonist (22), is not generally viewed as an antiseizure medication. Because fentanyl is relatively µ-receptor specific, and its effect on the µ receptor is antagonized by naloxone, our data support the notion that both opioid-induced seizures and opioid-induced brain damage in our experiments were mediated by µ-opiate receptors. Other µ-opioids and opiates, including morphine, met-enkephalin, and ß-endorphin in laboratory animals produce limbic system seizures (23) and kainic acid seizures that are similar in appearance and outcome to µ-opioid-induced seizures involving endogenous µ-opiates (24). Moreover naloxone completely halts seizure activity caused by endogenous opiates when given in adequate doses (25), an effect observed in our studies with exogenously administered fentanyl. It is of interest that Myer et al.1 reported an epilepsy syndrome in humans that responded to µ-receptor antagonist therapy.
It is important to note, however, that naloxone-treated rats also received N2O. N2O has some naloxone-reversible analgesic effects mediated by opiate receptors (26,27). It also has an N-methyl D-aspartate-like effect in rats that can be either neurotoxic or cerebroprotective (28). At the concentrations used in our experiments, protection would have been the predominant effect. Thus, the necessary concomitant use of N2O with naloxone administration does introduce an element of complexity that makes it more difficult to make conclusive determinations about a naloxone effect on opioid neurotoxicity.
In summary, our data confirm previous studies demonstrating epileptiform activity and hippocampal brain damage caused by large-dose fentanyl in rats. Furthermore, we observed that significant protection could be achieved by using a variety of clinically available drugs. The data indicate that µ-opiate receptors and GABA-ergic mechanisms may be important in the pathogenesis of opioid neurotoxicity, and that protection can be produced despite continuing EEG epileptiform activity.
|
Acknowledgments
|
|---|
The authors acknowledge the technical assistance of Marie Rose and Rosalyn Garman and the assistance of Julie Yantz and Mary Eddy in manuscript preparation.
|
Footnotes
|
|---|
1 Myer EC, Trepathi HL, Dewey WL: Cerebrospinal fluid ß-endorphin immunoreactivity in epilepsy and the response to naltrexone [abstract]. Epilepsia 1990;31:611. 
|
References
|
|---|
-
Kofke WA, Garman RH, Stiller RL, et al. Opioid neurotoxicity: fentanyl dose-response effects in rats. Anesth Analg 1996; 83: 1298–306.[Abstract]
-
Kofke WA, Garman RH, Tom WC, et al. Alfentanil-induced hypermetabolism, seizure, and neuropathology in rats. Anesth Analg 1992:953–64.
-
Kofke WA, Garman RH, Janosky J, Rose ME. Opioid neurotoxicity: neuropathologic effects in rats of different fentanyl congeners and the effects of hexamethonium-induced normotension. Anesth Analg 1996; 83: 141–6.[Abstract]
-
Tempelhoff R, Modica PA, Bernardo KL, Edwards I. Fentanyl-induced electrocorticographic seizures in patients with complex partial epilepsy. J Neurosurg 1992; 77: 201–8.[ISI][Medline]
-
Cascino GD, So EL, Sharbrough FW, et al. Alfentanil-induced epileptiform activity in patients with partial epilepsy. J Clin Neurophysiol 1993; 10: 520–5.[ISI][Medline]
-
Kearse LA Jr, Koski G, Husain MV, et al. Epileptiform activity during opioid anesthesia. Electroencephalogr Clin Neurophysiol 1993; 87: 374–9.[ISI][Medline]
-
Ben-Ari Y, Tremblay E, Riche D, et al. Electrographic, clinical and pathological alterations following systemic administration of kainic acid, bicuculline or penetrazole: metabolic mapping using the deoxyglucose method with special reference to the pathology of epilepsy. Neuroscience 1981; 6: 1361–91.[ISI][Medline]
-
Moller JT, Cluitmans P, Rasmussen LS, et al. Long-term postoperative cognitive dysfunction in the elderly: ISP0CD1 study. Lancet 1998; 351: 857–61.[ISI][Medline]
-
Slogoff S, Girgis KZ, Keats AS. Etiologic factors in neuropsychiatric complications associated with cardiopulmonary bypass. Anesth Analg 1982; 61: 903–11.[Abstract/Free Full Text]
-
Penfield W, Maxhieson G. Memory: autopsy findings and comments on the role of hippocampus in experiential recall. Arch Neurol 1974; 31: 145–54.[ISI][Medline]
-
Prough DS, Rogers AT, Stump DA, et al. Cerebral blood flow decreases with time whereas cerebral oxygen consumption remains stable during hypothermic cardiopulmonary bypass in humans. Anesth Analg 1991; 72: 161–8.[ISI][Medline]
-
Lupica CR, Dunwiddie TV. Differential effects of mu- and delta-receptor selective opioid agonists on feedforward and feedback GABAergic inhibition in hippocampal brain slices. Synapse 1991; 8: 237–48.[ISI][Medline]
-
Snead OC III, Bearden LF. The epileptogenic spectrum of opiate agonists. Neuropharmacology 1982; 21: 1137–44.[ISI][Medline]
-
Kofke WA, Towfighi J, Garman RH, et al. Effect of anesthetics on neuropathologic sequelae of status epilepticus in rats. Anesth Analg 1993; 77: 330–7.[ISI][Medline]
-
Rundfeldt C, Hönack D, Löscher W. Phenytoin potently increases the threshold for focal seizures in amygdala-kindled rats. Neuropharmacology 1990; 29: 845–51.[ISI][Medline]
-
Tommasino C, Maekawa T, Shapiro HM. Fentanyl-induced seizures activate subcortical brain metabolism. Anesthesiology 1984; 60: 283–90.[ISI][Medline]
-
Kofke WA, Gupta N, Sinz E, et al. Mu-opioid temporal lobe activation in human and subhuman primates [abstract]. Anesthesiology 1999; 91: A171.
-
Werz MA, Macdonald RL. Opiate alkaloids antagonize post-synaptic glycine and GABA responses: correlation with convulsant action. Brain Res 1982; 236: 107–19.[ISI][Medline]
-
Olsen RW, Stauber GB, King RG, et al. Structure and function of the barbiturate-modulated benzodiazepine/GABA complex. Adv Biochem Psychopharmacol 1986; 41: 21–32.[ISI][Medline]
-
Katzung, BG. Basic and clinical pharmacology. East Norwalk: Appleton & Lange, 1987:264–6.
-
McLean MJ, MacDonald RL. Multiple actions of phenytoin on mouse spinal cord neurons in cell culture. J Pharm Exp Ther 1983; 227: 779–89.[Abstract/Free Full Text]
-
Takemori AE, Portoghese PS. Comparative antagonism by naltrexone and naloxone of mu, kappa, and delta agonists. Eur J Pharmacol 1984; 104: 101–4.[ISI][Medline]
-
Cain DP, Corcoran ME. Epileptiform effects of met-enkephalin, --endorphin and morphine: kindling of generalized seizures and potentiation of epileptiform effects by handling. Brain Research 1985; 338: 327–36.[ISI][Medline]
-
Sztriha L, Lelkes Z, Benedek G, Joó F. Potentiating effect of morphine on seizures induced by kainic acid in rats. Arch Pharmacol 1986; 333: 47–51.
-
Tortella FC, Robles L, Mosberg HI. Evidence for mu-opioid receptor mediation of enkephalin-induced electroencephalographic seizures. J Pharm Exp Ther 1987; 240: 571–7.[Abstract/Free Full Text]
-
Guo T-Z, Poree L, Golden W, et al. Antinociceptive response to nitrous oxide is mediated by supraspinal opiate and spinal
2 adrenergic receptors in the rat. Anesthesiology 1996; 85: 846–52.[ISI][Medline]
-
Fang F, Guo T-Z, Davies MF, Maze M. Opiate receptors in the periaqueductal gray mediate analgesic effect of nitrous oxide in rats. Eur J Pharmacol 1997; 336: 137–41.[ISI][Medline]
-
Jevtovic-Todorovic V, Todorovic SM, Mennerick S, et al. Nitrous oxide (laughing gas) is an NMDA antagonist, neuroprotectant and neurotoxin. Nature Medicine 1998; 4: 460–3.[ISI][Medline]
Accepted for publication August 23, 2000.
This article has been cited by other articles:
|
|
|
|
 
L. J. Voss, J. W. Sleigh, J. P. M. Barnard, and H. E. Kirsch
The Howling Cortex: Seizures and General Anesthetic Drugs
Anesth. Analg.,
November 1, 2008;
107(5):
1689 - 1703.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
G. Blaise
Should we use naloxone epidurally?/Devrait-on utiliser la naloxone dans l'espace peridural ?
Can J Anesth,
November 1, 2003;
50(9):
875 - 878.
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Okutomi, M. Saito, J. Mochizuki, and K. Amano
Prophylactic epidural naloxone reduces the incidence and severity of neuraxial fentanyl-induced pruritus during labour analgesia in primiparous parturients
Can J Anesth,
November 1, 2003;
50(9):
961 - 962.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
