Antagonism of imipramine poisoning by anticonvulsants in the rat

TOXICOLOGY

AND

APPLIED

Antagonism

PHARMACOLOGY

38, l-6 (1976)

of lmipramine

Poisoning in the Rat1

by Anticonvulsants

A. R. BEAUBIEN,~ D. C. CARPENTER, L. F. MATHIEU,~ M. MACCONAILL, AND P. D. HRDINA Drug Toxicology Division, Health Protection Branch, Tunney’s Pasture, Ottawa, Canada KlA OL2, and Department of Pharmacology, Faculty of Medicine, University of Ottawa, Ottawa, Ontario, Canada Received December 8,1975: accepted March 31,1976

Antagonism of Imipramine Poisoningby Anticonvulsants in the Rat. A. R., CARPENTER, D. C., MATHIEU, L. F., MACCONAILL, M., P. D. (1976). Toxicol. Appl. Pharmacol. 38, l-6. Doseresponsecurveswereobtained in maleWistar rats for imipramine-induced convulsionsand death. For ip administration, the medianconvulsive dose (CDSO)(64.7 mg/kg) wasonly slightly lower than the LD50 (85.8 mg/kg). The calculated LD90 (112 mg/kg) was used to test the effectivenessof diazepam,phenytoin, or phenobarbitalin preventing imipramineseizures. Imipramine convulsionscould be preventedby either diazepamor phenobarbital but not by phenytoin (5 to 200mg/kg ip). The ip medianpreventive dose(PD50) for diazepamwasestimatedto be0.32 mg/kg, and for phenobarbital, 17.6mg/kg.

BEAUBIEN, AND HRDINA,

With the increasing use of tricyclic antidepressants in the therapy of depression in adults and enuresis in children, poisonings with these drugs have been steadily increasing. The cardinal features of poisoning include severe hypotension, ECG abnormalities, and various CNS disturbances. Seizures almost invariably occur in severe poisonings, and the importance of controlling these has been emphasized since they put added stress on an already weakened heart (Giles, 1963; Steel et al., 1967). In fact, the primary cause of death in poisoning with tricyclic antidepressants has sometimes been attributed to a prolonged period of intense convulsions (Noack, 1960). At present, barbiturates (Rasmussen, 1965; Lloyd et al., 1967) and diazepam (Fouron and Chicoine, 1971; Young and Galloway, 1971; Freeman and Loughhead, 1973; Goel and Shanks, 1974) are the drugs most frequently used to arrest these convulsions. Phenytoin has also been used to treat the imipramine-induced seizures (Greenblatt et al., 1974), the cardiac arrhythmias (Freeman et al., 1969), or both (Brown et al., 1971; Price and Postlethwaite, 1974). Since no definitive study measuring the anticonvulsant effectiveness of these compounds in poisonings with tricyclic antidepressants has been reported, this work was aimed at determining which of the anticonvulsants, diazepam, phenobarbital, and phenytoin, would be most efficacious in reducing the occurrence of imipramine-induced convulsions in rats. 1A preliminaryreport of portionsof thiswork wasmadeelsewhere (Carpenteret al., 1975). ’ Drug ToxicologyDivision,HealthProtectionBranch,Ottawa,Canada. Copyright0 1976 by Academic Press, Inc. 1 All rights of reproduction in any form reserved. Printed

in Great Britain

2

BEAUBIEN ET AL. METHODS

Animals and Drugs

Male Wistar rats3 (180-235 g) were maintained on pellet food and water for 1 week prior to their experimental use. All drugs were dissolved in an appropriate vehicle and administered ip in a standard volume of 3.5 ml/kg. Imipramine hydrochloride4 and phenobarbital sodium5 were solubilized in 0.9% NaCl solution. Phenytoin sodium6 was dissolved in a solution composed of 0.01 N NaOH and 0.9 ‘A NaCl, and diazepam7, in a solution containing 0.0186 N HCl and 0.9 % NaCl. Imipramine Dose-ResponseDetermination

An 8 x 8 Latin square design was used to randomize the order in which animals were treated. The strength of the various dosing solutions was unknown to the experimenters until after the data had been collected. Normal saline was injected ip 20 min prior to imipramine administration to simulate injections of anticonvulsants which were to be given in subsequent experiments. Theoretical dose-response curves were generated by logit analysis Part I (Waud, 1972), and the 95 % confidence limits were obtained by the application of Eq. 4.35 from Finney (1971). The ratio + SE of the lethality dose-response curve to that for convulsions was calculated according to the technique (Part IV) of Waud (1972) for potency ratio determination. Computations were done on a NOVA 800 minicomputer (Datagen, Canada). The animals were kept in separate plastic cages to avoid effects due to aggregation (La1 and Brown, 1968). Seizure activity was defined as a burst of repetitive myoclonic jerks involving at least the head and forelimbs, and consisting of not less than three jerks in rapid succession. Most often all four limbs were involved and opisthotonus with the tail arched over the animal’s back was frequently observed. Several episodes of convulsive activity were not uncommon. Time to convulsions and time to death were measured from the time of the imipramine injection, and the animals were observed for a period of 150 min. Median time to convulsions was calculated and the range was expressed in terms of the 25th and 75th percentiles. Determination of Anticonvulsant Effectiveness

In order to test the effectiveness of the various anticonvulsant agents, the LD90 of imipramine (calculated from the results of the dose-response experiment above) was used, since this would probably be the highest dose that would normally be encountered in a clinical case of imipramine poisoning. Rats were pretreated with various doses of the anticonvulsant agent, with saline, or with the appropriate drug vehicle 20 min prior to imipramine. To avoid differences due to order of treatment, the animals were randomized within blocks containing one representative from each group. Time to convulsions were recorded as in the imipramine dose-response experiment above. Since the anticonvulsant experiments were run 2-3 weeks apart, records were kept of the number of deaths in saline and vehicle control groups. This provided an approximate measure 3 Woodlyn Farms, Guelph, Ontario, Canada. 4 Donated by Ciba-Geigy Canada Ltd., Montreal, Quebec, Canada. 5 British Drug House (Canada) Ltd., Toronto, Ontario, Canada. 6 Donated by Parke, Davis and Co., Brockville, Ontario, Canada. ’ Donated by Hoffman-La Roche Ltd., Vaudreuil, Quebec, Canada.

ANTICONVULSANTS

AND

IMIPRAMINE

POISONING

3

of altered responsiveness to irnipramine arising from either batch variation or the presence of an anticonvulsant drug vehicle other than saline. To determine if an anticonvulsant substance was effective in preventing seizures, the proportion of animals responding in the vehicle control group was compared by a Chi-square test (formula 8.10.3, Snedecor and Cochran, 1967) with the proportion responding in each of the groups given the higher doses of anticonvulsant. If an anticonvulsant agent wasjudged to be effective (i.e.,p < 0.01 for one dose group orp < 0.05 for at least two consecutive dose groups), a theoretical dose-response curve was generated as follows: (1) The proportion of animals responding in the absence of an effective dose of anticonvulsant (P,) was determined from the minimum value of the series PO, (P, + Pi)/27 (PO + PI + Pz)/3, . . 1(PO + PI + . . . + Pi)/(i + 1) . . .y (PO + PI + . . . + Ph)/(h f l), where PO was the proportion of animals responding in the vehicle control group, Pi was the proportion responding for the ith dose of anticonvulsant, and P,, was the proportion responding for the highest dose of anticonvulsant. (2) The value arPc (ni is the number of animals in the ith group) was subtracted both from the number of animals responding and from n, for each group (including the vehicle control group where ni = n,). The vehicle control group was assigned a dose value which was low3 times smaller than the lowest dose of anticonvulsant used, and the data obtained after subtracting out the contribution due to P, was used to generate a dose-response curve in the manner described for the imipramine dose-response determinations. However, all groups at doses lower than or equa1 to any group which gave a negative number of animals responding after subtracting n,P, were not used in the dose-response calculation. Potency ratio estimation was done according to Waud Part IV (1972). RESULTS CDS0 and L D50 of Imipramine

In order to properly design experiments to test the various anticonvulsant agents, a dose-response relationship for convulsive activity and lethality of imipramine was first determined. Figure 1 shows the results of this experiment in the form of log dose vs response curves as determined from logit analysis. The median convulsive dose (CD50) of imipramine was calculated to be 64.7 mg/kg with 95% confidence limits as 59.3 and 70.0. The LD50 was only slightly higher, 85.8 mg/kg with 95 oAconfidence limits at 78.4 and 95.3. The two curves can be considered to be parallel since their slopes (11 .O i 2.7 for convulsions and 8.23 & 1.99 for lethality) were not significantly different. The log dose ratio of lethality to convulsive activity for imipramine was 1.25 with 95 % confidence limits at 1.15 and 1.36. The LD90 of imipramine was calculated from the theoretical dose response curve to be 112 mg/kg (95 % confidence limits = 100 and 149). This dose of imipramine was used in subsequent experiments to test the activity of the anticonvulsant agents studied. The median time to convulsions was 7.5 min (PZs = 6.5 and PT5 = 9.5). As expected from pharmacokinetic considerations, animals receiving higher doses tended to convulse sooner than those given lower doses of imipramine. The median time to death was 17.5 min (Pz5 = 12.5 and PT5 = 32.5), and no meaningful relationship between the dose and time to death was observed, All animals, which died convulsed prior to death.

4

BEAUBIEN ET AL.

DOSE

OF

IMIPRAMINE

(me/kg)

1. Dose-response curves for imipramine-induced convulsions (0) and death (0). Abscissa is on log scale. The solid lines have been calculated by logit analysis, and the horizontal bars represent the 95 % cont%dence limits of the ED90. FIG.

Protection against Imipramine Seizures

The ability of phenobarbital, phenytoin, and diazepam to prevent imipramineinduced seizures was tested by administering appropriate varying doses of these anticonvulsants 20 min prior to imipramine, 112 mg/kg. The results are shown in Fig. 2. Chi-square tests, comparing the appropriate vehicle control group with the groups receiving the higher doses of anticonvulsant substances, showed that both diazepam and phenobarbital significantly protected against imipramine-induced seizures. Since virtualIy:no protection was observed for animals pretreated with phenytoin, theoretical doseresponse curves were generated (see Methods) only for diazepam and phenobarbital. 106

0 E ” r 0 %

9-

. DIAZEPAM 0 WENOBARB. A DPH

87-

DOSE

OF ANTICONVULSANT

(mg/kg)

FIG. 2. Protection against imipramine-induced convulsions achieved by administration of diazepam (o), phenobarbital (o), or phenytoin (A) 20 min prior to imipramine (112 mg/kg ip). Abscissa is on log scale. Theoretical dose-response curves were generated only for the data obtained with diazepam and phenobarbital.

ANTICONVULSANTS

AND

IMIPRAMINE

POISONING

5

The median protective dose (PDSO) value for diazepam was calculated to be 0.32 mg/kg (95% confidence limits = 0.05 and 0.70) and for phenobarbital, 17.6 mg/kg (95 % confidence limits = 12.5 and 23.3). The potency ratio calculated by Waud’s technique was 32.3 (95 y0 confidence limits = 15.5 and 67.2). However, since different batches of animals were used to test each anticonvulsant, the dose of imipramine used to approximate the LD90 (i.e., 112 mg/kg) cannot be considered to have exactly the same lethal properties in each experiment. The mortality figure for the saline control group was 5/10 in the phenobarbital experiment, lo/l0 in the diazepam experiment, and 9/10 in the phenytoin experiment. As well, the presence of the drug vehicle used for diazepam and phenytoin probably affected the lethal properties of imipramine since only 6/10 animals died in the vehicle control group for each of these two anticonvulsants. For these reasons, the PD50 and potency ratio figures together with their 95 % confidence limits can be regarded only as first approximations, and are probably more representative for an LD50 or LD60 than for an LD90 of imipramine. The median time to convulsions was 9.0 min (PZ5 = 7.0 and P,, = 11.O) for animals. pretreated with diazepam, 8.5 min (Pz5 = 6.5 and PT5 = 11.5) for the phenobarbitalpretreated rats, and 8.5 min (P,, = 6.5 and P,, = 9.5) for rats given phenytoin. These figures include the data from the appropriate control group for each anticonvulsant substance tested. No obvious relationship between time to convulsions and dose of anticonvulsant substance was observed. DISCUSSION

The results of our experiments show that both diazepam and phenobarbital are fully effective in preventing the occurrence of imipramine seizures in the rat. Clinical reports on the use of these two anticonvulsants in cases of poisoning with tricyclic antidepressants are in good agreement with our animal data (Rasmussen, 1965; Lloyd et al., 1967; Young and Galloway, 1971; Brown et al., 1972). Phenytoin has been recommended by some authors for the prevention of seizures in cases of overdosage with tricychc antidepressants (Brown et al., 1971; Price and Postlethwaite, 1974). To our knowledge, no good clinical or laboratory evidence exists to support this advice. Our experiments showed that phenytoin was completely ineffective in preventing imipramine-induced seizures in rats. This is not too surprising in that the etiology of the convulsions is an important factor in the usefulness of phenytoin as an anticonvulsant. For instance, this drug does not control seizures associated with barbiturate withdrawal (Essig, 1967) nor is it effective against febrile convulsions in children under 3 years of age (Melchior et al., 1971) or against convulsions induced by pentylenetetrazol (Toman and Goodman, 1948). However, before phenytoin could be excluded as a useful anticonvulsant in clinical casesof tricyclic antidepressant poisoning, our findings would have to be verified in other animal species, in other experimental situations, and with several other tricyclic antidepressants. Furthermore, phenytoin may be useful in treating the cardiac arrhythmias which can arise in these cases. Although diazepam and phenobarbital appear to be effective anticonvulsants for imipramine poisoning, this does not necessarily mean that they are safe drugs to use in situations of imipramine overdosage. The severe cardiovascular and respiratory disturbances seen during poisoning with tricyclic antidepressants require great caution

6

BEAUBIEN

ET AL.

in administering other drugs which could worsen the patient’s condition by some unforeseen mechanism. For instance, it has been suggested that phenobarbital potentiates the respiratory depression occurring in amitriptyline poisoning (Sunshine and Yaffe, 1963). Further studies are now in progress in our laboratories to examine the effects of various anticonvulsants on cardiovascular and respiratory parameters in animals poisoned with imipramine. The extremely small log dose ratio (1.25) of lethality to convulsions found for imipramine in this study appears to be species-dependent. In mice we have observed a log dose ratio of about 1.5 (unpublished observations). REFERENCES BROWN,

K.

G. E.,

MCMICHEN,

H. U. S., AND

BRIGGS,

D. S. (1972). Tachyarrhythmia

in severe

imipramine overdose controlled by practolol. Arch. Dis. Child. 47, 104-106. T. C. K., DWYER, M. E., AND STOCKS, J. G. (1971). Antidepressantoverdosagein children-A new menace.Med. J. Aust. 2, 848-851. CARPENTER, D. C., BEAUBIEN, A. R., MATHIEIJ,L. F., ANDHRDINA,P. D. (1975).Amitriptyline and imipraminepoisoningin children. Brit. Med. J. 1, 516-517. ESSIG, C. F. (1967).Clinical and experimentalaspectsof barbiturate withdrawal convulsions. Epilepsia 8, 21-30. FINNEY,D. J. (1971).Probit Analysis, p. 78. CambridgeUniversity Press,London. FOURON, J.-C., ANDCHICOINE, R. (1971).ECG changesin fatal imipramine(Tofranil) intoxication. Pediatrics 48, 777-781. FREEMAN, J. W., M~JNDY,G. R., BEATTIE,R. R., AND RYAN, C. (1969).Cardiacabnormalities in poisoningwith tricyclic antidepressants. Brit. Med. J. 2, 610-611. FREEMAN,J. W., ANDLOUGHHEAD, M. G. (1973).Beta blockadein the treatment of tricyclic antidepressantoverdosage.Med. J. Aust. 1, 1233-1235. GILES, H. McC. (1963).Imipramine poisoningin childhood. Brit. Med. J. 2, 844-846. GOEL, K. M., ANDSHANKS, R. A. (1974).Amitriptyline and imipraminepoisoningin children. BROWN,

Brit. Med. J. 1, 261-263.

D. J., KOCH-WESER, J., AND SHADER, R. I. (1974). Multiple complicationsand death following protriptyline overdose.J. Amer. Med. Assoc. 229,556-557. LAL, H., AND BROWN,R. M. (1968). Enhancedtoxicity of imipramineand desipraminein aggregatedmice.J. Pharm. Pharmacol. 20,581-582. LLOYD, T. W., HART, D. R., AND TORODE, S. A. (1967). Amitriptyline poisoning.Lancet 2, 716-717. MELCHIOR, J. C., BUCHTHAL, F., AND LENNOX-BUCHTHAL, M. (1971). The ineffectivenessof diphenylhydantoin in preventing febrile convulsionsin the ageof greatestrisk, under three years. Epilepsia 12, 55-62. NOACK, C. H. (1960).A death from overdosageof “Tofranil.” Med. J. Aust. 2, 182. PRICE, D. A., AND POSTLETHWAITE, R. J. (1974). Amitriptyline and imipramine poisoningin children. Brit. Med. J. 1, 575. RASMUSSEN, J. (1965).Amitriptyline and imipraminepoisoning.Lancet 2, 850-851. SNEDECOR, G. W., AND COCHRAN, W. G. (1967).Statistical Methods, p. 217. The Iowa State University Press,Ames. STEEL, C. M., O’DUFFY, J., AND BROWN, S. S. (1967).Clinical effectsand treatment of imipramineand amitriptyline poisoningin children. Brit. Med. J. 3, 663-667. SUNSHINE, P., AND YAFFE, S. J. (1963). Amitriptyline poisoning. Amer. J. Dis. Child. 106, 501-506. TOMAN, J. E. P., AND GOODMAN, L. S. (1948).Anticonvulsants.Physiol. Rev. 28,409-432. WAIJD, D. R. (1972). On biological assaysinvolving quanta1responses. J. Pharmacol. Exp. GREENBLATT,

Ther. 183,577-607. YOUNG,

J. A.,

AND

GALLOWAY,

Arch. Dis. Child. 46,353-355.

W. H. (1971). Treatment of severeimipramine poisoning.