PEDIATRIC PHARMACOLOGY

AND THERAPEUTICS

Acute toxic reaction to carbamazepine'

Clinical effects and serum
concentrations
J a m e s Tibballs, MBBS,BMedSci(Hons)
From the Intensive Care Unit, Royal Children's Hospital, Melbourne, Australia
The clinical spectrum of toxic effects a n d serum concentrations after ingestion
of c a r b a m a z e p i n e were studied in 82 pediatric patients. Serum c a r b a m a z e p i n e
level was related to the depth of c o m a (p <0.001), convulsions (p = 0.002), hypotension (p <0.001), and the requirement for m e c h a n i c a l ventilation (p <0.001).
In 10 patients in d e e p c o m a with a Glasgow Coma Scale (GCS) of 3-4, the mean
serum level was 213/~mol/L (range 143 to 343); seizures, ventilatory failure, or hypotension caused by m y o c a r d i a l failure and c o n d u c t i o n defects were observed. In four of these, large doses of inotropic agents were required, one patient was treated with plasmaphaeresis, and two d i e d - - o n e of c a r d i a c failure
and one of aspiration pneumonitis. In 27 patients with moderate c o m a (GCS 5-8),
the mean serum level of c a r b a m a z e p i n e was 112/~mol/L (range 63 to 176); convulsions were observed in two patients in this group. In 45 patients whose conscious state was mildly depressed or normal (GCS 9-15), the mean serum level
was 73/~mol/L (range 37 to 128); a d d i t i o n a l effects were drowsiness (80%), a t a x i a
(53%), nystagmus (38%), vomiting (17%), and dystonia (7%). I c o n c l u d e that patients with serum c a r b a m a z e p i n e levels of a p p r o x i m a t e l y 100 ~mol/L require
close observation, whereas those with levels greater than 150 ~mol/L may
require intensive life support. (J PEDIATR1992,121:295-9)

Carbamazepine has become the most frequently prescribed

antiepileptic medication. The drug has many additional
pharmacologic actions, including sedative, anticholinergic,
antiarrhythmic, muscle relaxant, antidiuretic, and antidepressant. Its structure has a close resemblance to that of the
tricyclic antidepressants, it is generally regarded as a relatively safe drug; only a few deaths and cases of severe toxic
reaction have been reported.
This report is of a group of children poisoned with carbamazepine. The aim was to examine the spectrum of toxic
effects and to relate clinical effects and outcomes to serum
concentrations of the drug.

Submitted for publication July 5, 1991; accepted Feb. 28, 1992.

Reprint requests: James Tibballs, MBBS, Intensive Care Unit,
Royal Children's Hospital, Flemington Rd., Parkville, Victoria,
Australia 3052.
9/25/37479

METHODS
The study was conducted for a 10-year period (1981 to
1990) at the Royal Children's Hospital, Melbourne. Details
of presentation, management, and outcome of patients with
carbamazepine poisoning admitted to the intensive care
unit were recorded prospectively. Details of other patients
with carbamazepine poisoning admitted to the hospital, but
[

GCS

Glasgow Coma Scale

not to the ICU, were studied retrospectively. Patients poisoned with additional substances were excluded. The study
was approved by the hospital ethics committee.
The neurologic, respiratory, and cardiovascular effects of
poisoning were recorded. Coma was graded according to the
Glasgow Coma Scale. Coincident blood samples were
assayed for serum carbamazepine levels on admission to the
hospital and after admission to the ICU, at intervals during
management. However, regular blood sampling was not

295

296

Tibballs

The Journal of Pediatrics

August 1992

40O

4OO

3OO

8
o

iii
Z

LU
"7

&

&

lOO 5:

,oo
CO

69

YES

NO
~

VEHTI_.A'I"kDt,I

NO

YES

HYPOTENSION

Figure. Serum carbamazepine concentration (in micromolesper liter) and the occurrence of convulsions,requirement for
mechanical ventilation, and occurrence of hypotension.

undertaken, nor was measurement of levels of metabolites.

Statistical data were compiled with EDIT SYSTATsoftware
(Baysville, Ontario, Canada) and analyzed with SYSTAT
software (Systat, Inc., Evanston, Ilk).
RESULTS
Eighty-two patients (aged 1 to 17 years) with carbamazepine poisoning alone were admitted to the hospital. Of
these, 36 were admitted to the ICU. During the same period, the number of admissions to the hospital for all intoxications was 3186, of whom 314 were admitted to the ICU.
The total number of deaths from poisoning was four. Of
these, two were due to carbamazepine poisoning and one
each to imipramine and paracetamol poisoning.
Kruskal-Wallis one-way analysis of variance revealed
that the serum carbamazepine level was related to the depth
of coma (p <0.001), as was also the occurrence of convulsions (p = 0.002), the requirement for mechanical ventilation (p <0.001), and the occurrence of hypotension (p
<0.001) (Figure). The Lilliefors test confirmed that fhe
logarithm of the serum carbamazepine concentration was
normally distributed (maximum difference 0.059; p =
0.635).
Forty-five patients had a GCS of 9-15. Among this group,
the predominant neurologic signs were drowsiness (80%),
ataxia (53%), nystagmus (38%), and dystonia (7%). Vomiting occurred in 17% of patients. Occasionally, hallucinations were present. The mean (geometric) carbamazepine
level was 73 #mol/L (range 37 to 128)'.~Nine of these patients were treated with induced emesis or with gastric lavage and activated charcoal. All survived without sequelae.
Twenty-seven patients taking carbamazepine had a GCS
of 5-8 (moderate coma). Convulsions occurred in two

patients. Tachycardia was observed frequently. In this

group the mean serum level was 112/xmol/L (range 63 to
176). Treatment consisted of intensive nursing alone or of
gastric lavage and installation of activated charcoal (nine
patients). All survived without sequelae.
Ten patients taking carbamazepine had a GCS of 3 or 4
(deep coma). Of this group, six had seizures and eight required mechanical ventilation for respiratory failure and
loss of protective airway reflexes. Four of seven who had
hypotension required inotropic-vasopressor support for
treatment of low cardiac output caused by poor myocardial
contractility. The mean peak serum level in this group was
213 /~mol/L (range 143 to 343). Two patients died--one
with left ventricular failure and the other with septicemia
during extracorporeal membrane oxygenation instituted for
the treatment of aspiration pneumonitis. The details of the
clinical presentation, management, and outcome of the 10
severely intoxicated patients are presented in the Table.
DISCUSSION
Carbamazepine is a frequently prescribed anticonvulsant
agent with a large therapeutic index. However, overdose or
ingestion may cause serious morbidity or death. Forty-five
percent of patients admitted to the hospital in this series of
patients with carbamazepine poisoning had manifestations
of moderate or severe toxic effects. Two deaths occurred
among a total of 82 cases, one from cardiac failure and one
from aspiration pneumonitis with septicemia. Only six other
deaths from acute poisoning have been reported. Two of
these have been attributed to aspiration pneumonitisI and
one to cardiovascular collapse.2 In the remaining three, the
mode of death was not specified. 3, 4
Only a few other attempts have been made to correlate

Volume 121
Number 2

Acute toxic reaction to carbamazepine

297

Table. Clinical presentation, peak serum carbamazepine levels, treatment, and outcome in 10 patients with severe toxic
effects

Patient Age
No.
(yr)

Ingested
dose

Clinical
presentation
Dystonia, myoclonus
Choreoathetosis
Coma (GCS 4), seizures
Respiratory failure
Opisthotonus
Coma (GCS 4), seizures
Respiratory failure
Coma (GCS 4), seizures
Respiratory failure
Hypotension (mean BP
35 mm Hg)
AV dissociation, ileus
Coma (GCS 3-4), seizures
Respiratory failure
Hypotension (mean BP
40 mm Hg)
Bradycardia
CI 2.8, PCWP 35 mm Hg
Pulmonary edema
Coma (GCS 4), seizures
Respiratory failure
Hypotension (mean BP
50 mm Hg)
Coma (GCS 3)
Respiratory failure
Hypotension
Bradycardia
AV dissociation
Pulmonary edema
Coma (GCS 4), seizures
Respiratory failure
Hypotension (mean BP
40 mm Hg)
LVSF 15%, PCWP
20 mm Hg
Pulmonary edema
Coma (GCS 3-4)
Respiratory failure
Hypotension (mean BP
47 mm Hg)
Coma (GCS 4)

3.5

1.3 gm

8.0

Unknown

2.5

Unknown

3.0

Unknown

5.0

Unknown

13.0

8-10 gm

2.0

Unknown

4.0

Unknown

2.5

800 mg

10

15.0

8 gm

Peak serum
carbamazepine
(#mol/L)

Coma (GCS 4)
Hypotension (mean BP
51 mm Hg)

Treatment

Outcome

143

IPPV
Gastric lavage
Activated charcoal

Recovered

169

IPPV
Gastric lavage
Activated charcoal
IPPV
Dopamine, 20 #g/kg/min
Adrenaline, 0.8 #g/kg/min
Noradrenaline, 0.5 ~g/kg/min

Recovered

308

Recovered

343

IPPV
Dopamine, 35/zg/kg/min
Adrenaline, 2 izg/kg/min
Gastric lavage
Activated charcoal
Plasmapheresis (3.5 vol)

Recovered

221

IPPV
Colloid infusion
Gastric lavage
Activated charcoal
IPPV
Dopamine, 20 izg/kg/min
Gastric lavage
Activated charcoal
Transvenous pacing

Recovered

280

IPPV
Dopamine, 15 izg/kg/min
Epinephrine, 1.5 izg/kg/min

Died (septicemia)

145

IPPV
Gastric lavage
Activated charcoal

Recovered

150

Gastric lavage
Activated charcoal
Colloid infusion

Recovered

300

182

Died (cardiac failure)

Recovered

IPPV, Intermittent positive-pressureventilation;AV, atrioventricular; BP, blood pressure; CI, cardiac index;PCWP, pulmonary capillary wedgepressure; LVS~,

left ventricular shortening fraction.

severity of clinical toxic effects with serum levels. In one

prospective study 5 of clinical status and serum levels of 62
patients, the authors concluded that coma, seizures, and
respiratory depression requiring mechanical ventilation
were poorly related to serum levels. However, restratifica-

tion of their data and reanalysis suggest otherwise: subjects

with serum levels up to 19.9 # g / m l (0 to 84 ~ m o l / L ) had
a 4% incidence of coma, a 4% incidence of seizures, and a
4% incidence of mechanical ventilation; subjects with serum
levels 20 to 29.9 ~zg/ml (85 to 168 # m o l / L ) had a 27% in-

298

Tibballs

cidence of coma, a 21% incidence of seizures, and a 12% incidence of mechanical ventilation; subjects with serum
levels 40 to 59.9 #g/ml (169 to 252 #mol/L) had a 100%
incidence of coma, a 33% incidence of seizures, and a 100%
incidence of mechanical ventilation. Subjects with higher
serum levels also remained longer in the ICU and in the
hospital. Sinus tachycardia was noted frequently (33%), but
no patient was hypotensive and only one had a cardiac conduction defect. Another study6 of five episodes of overdose
in four subjects revealed that serum levels of 11 to 15 #g/ml
(45 to 65 izmol/L) were associated with drowsiness and
ataxia, and levels of 15 to 25/zg/ml (65 to 105 #mol/L) with
"combativeness," hallucinations, and choreiform movements, whereas levels greater than 25 #g/ml (105 #tool/L)
were associated with coma and seizures.
An apparent lack of correlation between clinical effects
and serum levels of carbamazepine may be attributed to
additional effects of carbamazepine metabolites, particularly of 10,11-epoxide, which may have intrinsic anticonvulsant activity7 and which has a longer half-life of
elimination than the parent compound after overdose. 8
Moreover, coma after carbamazepine overdose may be cyclic9, 10 because of delayed absorption.
Carbamazepine is not generally regarded as a cardiotoxic
drug, but there are several case reports in which therapeutic doses caused bradycardia and complete heart block.11, 12
Congestive cardiac failure 13 and aggravation of heart
block 14 have also been recognized as a consequence of regular therapy. In several cases of deliberate self-poisoning,
disturbances of conduction have been recognized as prolongations of the PR interval, QRS complex, and QT interval.15, 16 Severe hypotension has been reported on several
occasions,15, ~6 and in one fatal case hypotension was
refractory to infusion of large doses of catecholamines2; the
mechanism of hypotension was probably left ventricular
failure, because elevated pulmonary capillary wedge pressure was also observed. In another case, acute pulmonary
edema in association with severe hypotension was reported. 17 In several cases in this series, hypotension and pulmonary edema were associated with a reduced left ventricular
shortening fraction, low cardiac output, and elevated pulmonary capillary wedge pressure, indicating that hypotension was indeed due to depression of myocardial contractility. Conduction defects were also observed.
Treatment of an acute toxic reaction is largely supportive, with the use of gastric lavage and of activated charcoal
to adsorb the drug and to hasten elimination from the enterohepatic circulation and intestinaY~cosa. 18 For adults
intoxicated with carbamazepine and other drugs, intermittent doses of activated charcoal are used, 19 whereas for
children a continuous instillation may be associated with a
lower incidence of vomiting.2~ If intestinal ileus is n o t
present, we administer 1 to 2 gm/kg initiallyl followed by

The Journal of Pediatrics

August 1992

0.25 gm/kg every 4 hours for 16 hours 21 or 0.25 gm/kg

hourly22 via nasogastric tube.
Invasive methods to remove carbamazepine from the
body have been reported. Charcoal hemoperfusion is moderately effective.l 5, 16, 23-25However, hemoperfusion may be
associated with thrombocytopenia and disturbances of
serum electrolyte and glucose levels. For the removal of
drugs such as carbamazepine, which are highly protein
bound26 and have a low volume of distribution,27 we prefer
the use of plasmapheresis, which is as efficacious as charcoal
hemoperfusion28 and, moreover, devoid of the problems of
thrombocytopenia, although a balanced solution of electrolytes, proteins, and glucose must be replaced in a volume
identical to that of the filtrate removed.
I recommend that patients who have serum carbamazepine concentrations of approximately 100/zmol/L be
monitored closely and treated with gastric lavage or activated charcoal or both. Patients with serum levels of
approximately 150 #mol/L or greater may requireintensive
life support and may need to undergo invasive methods to
remove the toxin if conventional supportive therapy is inadequate.
The assistance of Dr. Frank Shann with the statistical analysis
is gratefully acknowledged.
REFERENCES

1. Denning DW, Matheson L, Bryson SM, Streete J, Berry D J,

Henry JA. Death due to carbamazepine self-poisoning:remedies reviewed. Hum Toxicol 1985;4:255-60.
2. Fisher RS, Cysyk B. A fatal overdose of carbamazepine: case
report and review of literature. Clin Toxicol 1988;26:477-86.
3. Hundt HKL, Aucamp AK, Muller FO. Pharmacokiaetic aspects of carbamazepine and its two major metabolites in
plasma during overdosage. Hum Toxicol 1983;2:607-14.
4. Baselt RC. Disposition of toxic drugs and chemicals in man.
2nd ed. Davis, California: Biomedical Publications, 1982:
113-7.
5. SpillerHA, Krenzelok EP, Cookson E. Carbamazepine overdoses: a prospee~ive study of serum levels and toxicity. Clin
Toxicol 1990;28:445-58.
6. Weaver DF, Camfield P, Fraser A. Massive carbamazepine
overdose: clinical and pharmacologicalobservationsin fiveepisodes. Neurology 1988;38:755-9.
7. Eichelbaum M, Bertilsson L, Lund L, Palmer L, Sjoqvist F.
Plasma levels of carbamazepine and carbamazepine-10,11epoxide during treatment of epilepsy. Eur J Clin Pharmacol
1976;9:417-21.
8. Pynnonen S, Sillanpaa M, Frey H, et al. Carbamazepine and
10,11-epoxy carbamazepine levels in children. Proc Eur Soc
Toxicol 1977;18:192-4.
9. Sullivan JB, Rumack BH, Peterson RG. Acute carbamazepine
toxicity resulting from overdose. Neurology 1981;31:621-4.
10. De Zeeuw RA, Westenberg HGM. An unusual case of
carbamazepine poisoning with a near-fatal relapse after two
days. Clin Toxicol 1979;14:263-9.
11. Hamilton DV. Carbamazepine and heart block [Letter]. Lancet 1978;1:1365.

Volume 121
Number 2

A c u t e toxic reaction to carbamazepine

12. Herzberg L. Carbamazepine and bradycardia [Letter]. Lancet 1978;1:1097-8.

13. Terrence CF, Fromm G. Congestive heart failure during carbamazepine therapy. Ann Neurol 1980;8:200-1.
14. Bearmann B, Edhag O, Vallin H. Advanced heart block
aggravated by carbamazepine. Br Heart J 1975;37:668-71.
15. Leslie P J, Heyworth R, Prescott LF. Cardiac complications of
carbamazepine intoxication: treatment by haemoperfusion.
BMJ 1983;286:1018.
16. Gary NE, Byra WM, Eisinger RP. Carbamazepine poisoning:
treatment by hemoperfusion. Nephron 1981;27:202-3.
17. Kitson GE, Wauchab TD. Pulmonary oedema following carbamazepine overdose. Anaesthesia 1988;43:967-9.
18. Boldy DAR, Heath A, Ruddock S, Vale JA, Prescott LF. Activated charcoal for carbamazepine poisoning [Letter]. Lancet
1987;1:1027.
19. McLuckie A, Forbes AM, Ilett KF. Role of repeated doses of
oral activated charcoal in the treatment of acute intoxications.
Anaesth Intensive Care 1990;18:375-84.
20. Ohning BL, Reed MD, Blumer JL. Continuous nasogastric
administration of activated charcoal for the treatment of theophylline intoxication. Pediatr Pharmacol 1986;5:241-5.

299

21. Lilley B, Nolan T, Tibballs J. Paediatric pharmacopoeia. 10th

ed. Melbourne: Royal Children's Hospital, 1989.
22. Shann F, Duncan A. Drug doses in paediatrics. 6th ed. Melbourne: Royal Children's Hospital, 1991.
23. Chart KM, Aguanno J J, Jansen R, Dietzler DN. Charcoal hemoperfusion for treatment of carbamazepine poisoning. Clin
Chem 1987;27:1300-2.
24. De Groot G, van Heijst ANP, Maes RAA. Charcoal hemoperfusion in the treatment of two cases of acute carbamazepine
poisoning. Clin Toxicol 1984;22:349-62.
25. Nilsson C, Sterner G, Idvall J. Charcoal hemoperfusion for
treatment of serious carbamazepine poisoning. Acta Med
Scand 1984;216:137-40.
26. Rawlins MD, Collste P, Bertilsson L, Palmer L. Distribution
and elimination kinetics of earbamazepine in man. Eur J Clin
Pharmacol 1975;8:9l-6.
27. Rey E, d'Athis P, deLauture D, et al. Pharmacokinetics of
carbamazepine in the neonate and in the child. Int J Clin
Pharmacol Biopharm 1979;17:90-6.
28. Laussen P, Shann F, Butt W, Tibbatls J. Use of plasmapheresis in acute theophylline toxicity. Crit Care Med 1991; 19:28890.

Clinical and laboratory observations

Randomized, controlled trial of antibiotic therapy for
Escherichia coil O 57:H7 enteritis
Frangois Proulx, MD, Jean P. Turgeon, MD, FRCP(C), Gilles Delage, MD, MSc,
FRCP(C), L u c e t t e

Lafleur, MD, FRCP(C), a n d Luc Chicoine, MD, FRCP(C)

From the Departments of Pediatrics and of Microbiology and Immunology, Sainte-Justine Hospital, Universit~ de Montr6al, Montreal, Canada

We undertook a prospective, controlled study to evaluate the effect of trimethoprim-sulfamethoxazole in children with proven Escherichia coli O157:H7
enteritis on the duration of symptoms, on fecal excretion of pathogen, and on
the risk of progression to hemolytic-uremic syndrome. There was no statistically
significant effect of treatment on progression of symptoms, fecal pathogen excretion, or the incidence of HUS (2/22 vs 4/25; p = 0.67). Our results suggest that
a multicentric trial using rapid diagnostic methods to permit early randomization should be carried out. (J PEDIATR1992;121:299-303)
Submitted for publication Dec. 10, 1991; accepted Feb. 25,
1992.
9/26/37450

Reprint requests: Francois Proulx, MD, Department of Pediatrics,

Sainte-Justine-Hospital, 3175, chemin C6te Sainte-Catherine,
Montr6al (Quebec), Canada H3T IC5.