Journalof AnalyticalToxicology,Vol.
23, October 1999
AcuteZolpidem Overdose--Reportof Two Cases*
Susan B. Gock, Steven H.Y. Wongt, Naziha Nuwayhid, SusanE. Venuti, P. Douglas Kelley,
John R. Teggatz,and Jeffrey M. Jentzen
Milwaukee CountyMedical Examiner'sOffice and Departmentof Pathology,Medical Collegeof Wisconsin,
Milwaukee, Wisconsin 53233
[ Abstract I
This report describes two cases of acute zolpidem overdose. The
decedent in the first case was a 36-year-old female found dead in
bed in her secured home. She had a history of psychiatric illness,
including paranoid disorder, depression with panic episodes, and
post-traumatic stress disorder. She was treated with risperidone
and sertraline. Nine months prior to her death, the decedent was
also prescribed zolpidem (Ambien| The postmortem examination
revealed white foam within the larynx and upper trachea, which is
indicative of pulmonary edema. Toxicological analyses of the urine
showed the presence of caffeine, risperidone, and zolpidem.
Subsequent quantitation of postmortem iliac serum revealed
5.6 pg/L of 9-hydroxyrisperidone and the following zolpidem
concentrations: blood (subclavian), 4.5 rag/L; blood (iliac),
7.7 rag/t; vitreous humor, 1.6 rag/L; bile, 8.9 rag/L; urine,
1.2 rag/L; liver, 22.6 mg/kg; and gastric contents, 42 rag. The
second caseinvolved a 58-year old female, also found dead in bed,
with white foam around her mouth. The decedent had a 25-year
history of hypertension and mental illness--manic depression and
schizophrenia. She was medicated with carbamazepine, naproxen,
risperidone, and zolpidem. The postmortem examination revealed
cardiomegaly, pulmonary edema, hepatomegaly, mild coronary
atherosclerosis, and no signs of trauma. Toxicological analyses of
the urine showed the presence of zolpidem and carbamazepine
and metabolite. Zolpidem concentrations were as follows: blood
(iliac), 1.6 mg/L; vitreous humor, 0.52 rag/L; bile, 2.6 rag/L; liver,
12 mg/kg; and gastric contents, 0.9 rag. The zolpidem blood
concentrations of these cases are consistent with those of the
previously published fatalities. The blood/vitreous humor ratios of
zolpidem were 2.81 (subclavian) and 4.81 (iliac) in the first case
and 3.08 (iliac) in the second case. These ratios, along with the
sampling times of blood and vitreous humor for both cases, are not
conclusive to indicate a definitive presence or absence of
postmortem drug redistribution of zolpidem. The cause of death
for both cases was determined to be acute zolpidem overdose, and
manner of death was suicide.
Introduction
Zolpidem (Ambien, Searle, Chicago, IL) is used for the treat-
ment of insomnia (1-6). The chemical name is N,N-6-tfimethyl-
* Presented in part at the joint meeting of The International F~leration of Forensic Toxicologists and
Society of Forensic Toxicologists, October 1998, Albuquerque, New Mexico.
~" Auth~ t~ wh~ c~176 sh~ be addressed' Dr' Steven H' W~ Milwaukee C~
Medical Examiner's Office, 933 West Highland Avenue, Milwaukee, W153233.
Reproduction(photocopying)of editorialcontentof thisjournalis prohibitedwithoutpublisher'spermission.
Case Report I
2-p-tolyl-imidazol (1,2-alpyridine-3-acetamide L-(+)-tartrate),
and the drug is of the imidazopyridine class. The proposed mech-
anism of action involves modulation of GABAreceptor chloride
channel macromolecular complex--the benzodiazepine recept-
or. It is administered as 5- or 10-rag capsules.
Zolpidem undergoes rapid absorption from the gastrointestinal
tract (1,3,6). Peak plasma concentrations occur at about 0.5-1.6 h.
For 45 adults given 5-rag and 10-rag oral doses, the average con-
centrations were 0.059 mg/L (range 0.029--0.113 rag/L) and 0.121
mg/L (range 0.058-4).272 rag/L), respectively. Zolpidem, with an
elimination half-life of 2.6 h (range 1.4-4.5 h), displays simple,
linear first-order pharmacokinetics, which is consistent with a
one-compartment model. Zolpidem and its inactive metabolites
undergo renal elimination. Protein binding is 93%. Zolpidem does
not interact significantly with haloperidol, imipramine, and chlor-
promazine, as revealed by single-dose pharmacokinetic studies.
However, the combinations could result in additive side effects
such as decreased alertness and psychomotor performance. Co-
administration of alcohol with zolpidem would also result in an
additive sedative effect (4,6). Thus, caution is advised on adminis-
tering zolpidem with other central nervous system depressants.
This was further confirmed by recent reports involving the co-
administration of zolpidem with vinylbital (2), acepmmazine (7),
or cafisoprodol/meprobamate (8).
In a voluntary overdose case study involving ingestion of about
300 mg of zolpidem, 600 mg of prothipendyl (a neuroleptic), and
ethanol, the zolpidem plasma concentrations at 3, 4, and 5 h
postingestion were 0.550, 0.435, and 0.290 rag/L, respectively (9).
Accordingly, the estimated elimination half-life of zolpidem was
about 2 h, which is within the established range of 1.4 to 4.5 h.
The patient became comatose with hypotension and respiratory
depression. Pinpoint pupils were noted. Because the initial
administration of naloxone did not improve the clinical status,
the patient was subsequently treated successfully with flu-
mazenil.
In the review of 344 cases of acute zolpidem poisoning reported
to the Paris Poison Center (n = 247) and the Synthelabo (n = 97)
from 1987 to 1991 (2), doses of zolpidem ranged from 10 to 1400
mg. Females accounted for 70% of the cases. Antipsychotic medi-
cations and alcohol were concomitantly present in about half of
the patients. Signs of intoxication reported by 105 patients medi-
cated with doses of 140-440 mg included the following: drowsi-
ness (n = 89), coma (n = 4), and respiratory failure (n = 1).
Treatment of zolpidem intoxication included supportive therapy
and/or gastric lavage. Intoxication resulted in 6% fatalities; how-
559

ever, there was a lack of evidence to attribute the deaths solely to
zolpidem overdose. Among the 10 deaths from acute zolpidem
poisoning in the Synthelabo series, one case involved a 60-year-old
woman whose medications included 600 mg of zolpidem, 2 g of
vinylbital, and 4 g of meprobamate. The postmortem toxicological
analysis revealed 2.7 mg/L (3.5 IJmole/L) of zolpidem and 5.8 mg/L
(25.9 ]Jmole/L) ofvinylbital. The cause of death was vinylbital over-
dose. Another report of zolpidem-related fatality involved the
drowning of an 86-year-old woman living in a retirement commu-
nity (10). Zolpidem concentrations were as follows: blood, 7.9
mg/L; urine, 4.1 rag/L; and gastric contents, 7 rag. Sixty zolpidem
tablets were unaccounted for. Another case involved a 28-year-old
veterinary assistant who was found dead in her bathroom (7).
Although the ingested dose was unknown, zolpidem concentra-
tions were as follows:blood, 3.29 rag/L; urine, 2.54 mg/L; bile, 1.27
rag/L; and gastric contents, 34.27 mg/L. Acepromazine (a phe-
nothiazine) was also present at the following concentrations:
blood, 2.4 mg/L; urine, 0.37 rag/L; bile, 1.03 mg/L; and gastric con-
tents, 20.05 mg/L. The authors emphasized the additive sedative
effect of zolpidem in combination with acepromazine. The cause
of death was listed as fatal ingestion of zolpidem and acepromazine
and the manner of death as suicide. Another report of a 68-year-old
woman who ingested at least 300 mg of zolpidem showed that
zolpidem was present in the gastric contents. She had a blood con-
centration of 4.1 mg/L (8). Other drugs detected included cariso-
prodol, 2.4 rag/L, and its metabolite, meprobamate, 19.3 mg/L.
The authors cautioned that although previous reports showed
benign outcomes following zolpidem ingestion, co-medication
with central nervous system depressants such as carisoprodol
could result in death. However, these reports had not established
the fatal blood concentration range due to a single, acute ingestion
of zolpidem (9). Our report describes two suicidal cases of acute
zolpidem overdose, including toxicological analyses of blood, vit-
reous humor, bile, urine, gastric contents, and liver and the corre-
sponding blood/vitreous humor ratios.
Case Histories
The first case involved a 36-year-old white female who was
found dead in bed in her secured home. Her past psychiatric ill-
nesses included paranoid disorder, depression with panic
Table I. Toxicological Findings*
Cases Drug Elapsed time t Stage* Blood/serumS
Case I Zolpidem 0h I 4.5 mg/1. (Sub. BI)
Zolpidem +24 h II 7.7 mglL (lliac BI)
9-Hydroxy-
risperidone - 5.6 rngiL (Serum)
Case 2 Zolpidem +7 h Early II 1.6 mg/]. (lliac BI)
* Toxicological urine screen: Case 1, caffeine, resperidone, and zolpidem; Case 2, zolpidem and carbamazepine.
* Elapsed time between the initial collection of VH and subsequent blood collections.
* Postmortem blood movement time stagesas proposed by Anderson and Prouty (12).
Sub. BI,- Subclavian blood.
560
Journal of Analytical Toxicology, Vol. 23, October 1999
episodes, and post-traumatic stress disorder. She was treated with
risperidone, sertraline, and zolpidem. She was homebound
because of her paranoia. She had reportedly ingested an overdose
of medication two years previously. Nine months prior to her
death, she was prescribed 60 10-rag zolpidem tablets. One month
prior to her death, she complained about blackouts and stomach
problems. She was found dead at her residence between 8 and 9
a.m., but was seen alive the night before her death. She was pro-
nounced dead shortly after being discovered. During the initial
body examination at about 11 a.m., subclavian blood and vitreous
humor were collected for toxicological analyses. An autopsy was
performed 24 h later, and the following specimens were collected:
lilac blood, urine, bile, liver, and gastric contents. Serum was also
obtained from iliac blood.
The second decedent was a 58-year-old black female with a 25-
year history of bipolar disorder, hypertension, manic depression,
and schizophrenia. Her medications included carbamazepine,
naproxen, risperidone, and zolpidem. She was found dead in her
residence with the front door barricaded by chairs. She was lying
in bed with white foam around her mouth, with no evidence of
traumatic injury or foul play at the scene, l~vo prescription bot-
tles were found on the kitchen table: a bottle for risperidone and
an empty bottle for zolpidem that might have been refilled with
6010-rag tablets about 14 days prior to her death. No suicide note
was found at the scene. Four months before her death, she told
family members that she had discontinued her medications
because they were "poisoning her". She had also told family
members that she thought people were trying to break into her
residence. For almost 20 years, she had been compliant with her
medications. However, she had two previous suicide attempts by
pill ingestion. After she was conveyed to the medical examiner's
office, vitreous humor was collected. An autopsy was performed
7 h later, and the following specimens were collected: iliac blood,
liver, bile, and gastric content.
Analytical Procedures
A comprehensive toxicology analysis was performed on both
cases using the following protocols. Acid-dichromate test was used
for the detection of methyl, ethyl, isopropyl alcohol, acetone, and
acetaldehyde. Blood was analyzed by the IL 482 Co-oximeter for
Vitreous Bile Urine Liver Gastric
1.6 mglL - - -
- 8.9 mg/L 1.2 mg/L 22.6 mglkg 42 mg
. . . . .
0.52 mg/L 2.6 mg/L - 12 mg/kg 0.9 mg
Journal of Analytical Toxicology, Vol. 23, October 1999
the presence of carboxyhemoglobin.Routine color spot tests were
performed on the urine specimen to detect acetaminophen
metabolites, ethchlorvynol,salicylates,and phenothiazines. Urine
and gastric specimens were screened for the presence of acidic,
neutral, and basic drugs by the Toxi-LabTM AB thin-layer chro-
matography system by ANSYS (Irvine, CA). Urine EMITTM
immunoassay screens were performed to test for barbiturates,
benzodiazepines, cannabinoids, and cocaine metabolite. Blood
was also screened for zolpidem by gas chromatography (GC).
Zolpidemwas extracted using a copolymeric solid-phase extrac-
tion column (Clean Screen DAU,United Chemical Technologies,
Bristol, PA) by modifyingthe published forensic application (11).
The internal standard used was phenyltoloxamine.The extract was
then analyzed using a Hewlett Packard 5890 GC with splitless
injection,dual columns, and dual flame ionizationdetectors (FID).
Chromatographicseparation was achievedusing DB-5 and DB-17
capillary columns (J&WScientific,Folsom, CA).Temperature was
programmed to hold at 100~ for 1 min and to increase by
15~ Temperatures of the injector and detector were 200~
and 300~ respectively.Analytes detected on initial screening
techniques were subsequently confirmed by gas chromatog-
raphy-mass spectrometry (GC-MS) on a Hewlett-Packardmodel
5890 GC and model 5970 mass selectivedetector in the full-scan
mode. Chromatographicseparation was achievedusing a DB-5MS
capillary column (J&W Scientific). Chromatographic conditions
were as described here. Confirmation was achieved based on (1)
retention time at 14.4 min and (2) mass spectral match with drug
identificationlibrary, showing the followingmajor ions in order of
relative abundance:235, 307, 219, 92, and 65.
Other biological specimens including vitreous humor, gastric
contents, urine, bile, and liverwere analyzedby GC-FID for deter-
mination of zolpidem concentration. Calibration curves, estab-
lished by plotting the peak-area ratios of the drug to the internal
standards (phenyltoloxamineor protriptyline) versus drug stan-
dards concentrations, were used for analyte quantitation.
Results
Autopsy of the first decedent showed an obese individual with
a white foam cone around the bilateral nares and foam within the
tracheal lumen, which is indicative of pulmonary edema.
Autopsyof the second decedent showed an obese individualwith
Table II. Zolpidem Blood/Fluids and Liver Ratios, and Gastric Contents from this Study and Two Previously Published Reports
Cases Blood source Elapsed time ~ Stage*
Case 1 Subclavian 0 hr I 2.81
Iliac +24 hr II
Case 2 Iliac +7 hr Early II
Reference 10
Reference 7
* For Case I, ratios of blood/fluids or liver were estimated by using the subclavian blood concentrations.
Elapsed time between initial collection of vitreous humor and subsequent blood collections.
* Postmortem blood movement time stages as proposed by Anderson and Prouty (I 2).
cardiomegaly with a dilated left ventricle and focal interstitial
fibrosis of the myocardium. She had mild coronary atheroscle-
rosis with a 50% stenosis of the mid-left anterior descending
coronary artery. She had marked pulmonary edema and hep-
atomegaly with mild macrovesicular steatosis (microscopically).
Chronic pyelonephritis was noted. There was no evidence of
internal or external trauma. Toxicologicalfindings of both cases
are listed in Table I. The corresponding blood/fluids and liver
ratios are listed in Table II.
Discussion
This report presents two cases demonstrating zolpidem blood
concentrations of 1.6 to 7.7 mg/L, both resulting from acute
ingestion of higher than therapeutic doses of zolpidem. These
concentrations are consistent with those of previouslypublished
fatalities (2,7,8). For comparison, a single 10-mg dose of zolpidem
would result in an average blood concentration of about 0.121
mg/L (3). In both of the reported cases, zolpidem was present in
the gastric contents, indicating acute zolpidem ingestion. The
available decedents' histories and the collection times of blood
and vitreous humor, along with the blood/vitreous humor ratios
of this study do not provide conclusive evidence for the presence
or absence of postmortem redistribution of zolpidem.
The first decedent was found dead in bed between 8 and 9 a.m.
She was last seen alive the previous evening. Two years prior to
her death, she reportedly overdosed on prescription medication.
Nine months prior to her death, she was prescribed 60 10-mg
tablets of zolpidem. She complained about the blackouts and
stomach problems. Other medications included risperidone and
sertraline. The second decedent, who had a 25-year history of
mental illness, barricaded herself in her residence because of
paranoia about intruders. Previously,she had made two suicide
attempts by ingesting pills. More recently, she was suspected to
have discontinued some of her medications, which included
zolpidem, 10 mg; carbamazepine, 200 rag; risperidone, 2 mg; and
naproxen, 500 rag.
Autopsy of both decedents showed pulmonary edema, which
may have been associated with respiratory failure. Toxicological
analyses revealed high blood zolpidem concentrations of 1.6 to
7.7 mg/L. The corresponding blood/fluids and blood/liver ratios
are shown in Table II. Although "therapeutic" vitreous humor
Ratios of Blood/fluids or liver
Vitreous Bile Urine Liver Gastric contents
0.51 4.75 0.20
4.81 - 42 mg
3.08 0.62 0.13 0.9 mg
- 1.92 - 7 mg
- 2.60 1.30 - 34 mg/I.
561

concentrations and blood/vitreous humor ratios have not been
established for zolpidem, the combination of the high blood con-
centrations, blood/vitreous humor ratios of 2.81, 3.08, and 4.81
and the presence of zolpidem in gastric contents would indicate
acute zoipidem ingestion. Blood/vitreous humor ratios for the
first decedent were 2.81 and 4.81 for subclavian and iliac blood
concentrations of 4.5 and 7.7 rag/L, respectively. Subclavian
blood and vitreous humor were collected at the scene at about 11
a.m. on the date of her death. Because she was last seen the night
before, these collected samples were obtained within Stage I (i.e.,
within 24 h after death) of the postmortem blood movement
period as proposed by Anderson and Pmuty (12). This iliac blood
sample was drawn during the autopsy, which was performed
approximately24 h after the collection of subclavian blood. Thus,
the iliac blood sample was collected at Stage II. Although there
was an increase in the concentration, the magnitude of the
increase is less than the two- to eightfold increases observed in
the well-documented postmortem redistribution of antidepres-
sants established by Anderson (13). Furthermore, using the four
stages of postmortem blood movement, there was little difference
in collection time after death of blood sample collected in Stage I
for the first case and early Stage II for the second case. This might
account for the similar blood/vitreous ratios of the first and
second cases, 2.81 and 3.08, respectively.This represents a pre-
liminary finding of the distribution of zolpidern in blood and vit-
reous after an acute ingestion. Further reports and/or animal
model studies would be needed to confirm this finding.
The blood/bileratios of 0.51 and 0.62, though comparablewithin
these cases, are substantiallyless than the ratio of 2.6 reported pre-
viously by Tracqui et al. (7). In that report, 34 mg of zolpidemwas
present in gastric contents, indicating an acute ingestion. Further,
the blood/liver ratios were 0.20 and 0.13 for the first and second
cases, respectively.Blood/urine ratio of 4.75 in Case I is substan-
tially higher than both reported ratios of 1.92 and 1.30 (7,10).
For their diagnosed mental disorders, both decedents were
medicated with risperidone, an antipsychotic. For the first dece-
dent, a therapeutic concentration of 5.6 lag/L for 9-hydroxy-
risperidone, a metabolite of risperidone, was detected in serum,
and risperidone was not detected in the second decedent. This
might be due to her noncompliance as indicated in the patient
history. Although there is no report of a zolpidem-risperidone
interaction, a survey of previouslypublished reports indicated the
lack of significant pharmacokinetic interaction of zolpidem with
haloperidol, another antipsychotic, although an additive sedative
effect was indicated. Because the 9-hydroxyrisperidone concen-
tration was not toxic, a drug--drug interaction could not be estab-
lished for Case 1. It is possible that the decedent may have
experienced the additive sedative effects of zolpidem and risperi-
done. Other medications prescribed to the decedents included
sertraline in the first case and carbamazepine and naproxen in the
second case. Sertraline was not detected in the urine or blood
specimens from the first decedent, and carbamazepine and
metabolites were detected only in the urine specimen from the
second decedent. Both findings suggest noncompliance with pre-
scribed medication treatment. Again, there is a lack of published
evidenceof drug-drug interaction for these drugs with zolpidem.
562
Journalof Analytical Toxicology,Vol. 23, October 1999
Conclusions
The cause of death for both cases was determined to be acute
zolpidem overdose,and the manner of deathwas suicide. In addi-
tion, other significant autopsy findings for the first decedent
included pulmonary edema and pulmonary edema, cardiomegaly,
and left ventricular hypertrophy for the second decedent. From
the limited data of this study and the collection times of blood and
vitreous humor, the blood/vitreous humor ratios determined for
each case are not conclusive for the determination of the presence
or absence of postmortem redistribution of zolpidem.
Acknowledgments
The authors gratefully acknowledge the technical assistance of
Cora Sue Rozwadowski, Gwyn Doss, and Carol Kallie of our
ToxicologyLaboratory and the valuable support of Eileen Weller
and Karin Kussmaul in the preparation of the manuscript.
References
I. R. Guinebault, C. Dubrue, R Hermann, and J.R Thenot. HPLC deter-
minations of zolpidem, a new sleep inducer, in biological fluids with
fluorimetric detection. J. Chromatogr. 383:206-211 (1986).
2. R. Gamier, E. Guerault, D. Muzard, P. Azoyan, A. Chaumet-Riffaud,
and M. Efthymiou. Acute zolpidem poisoning--analysis of 344
cases. Clin. Toxicol. 32:391--404 (I 994).
3. R.C. Baselt and R.H. Cravey. Disposition of Toxic Drugs and
Chemicals in Man, 4th ed. Chemical Toxicology Institute, FosterCity,
CA, 1995, p 787.
4. D. Garside and B.A. Goldberger. Zolpidem. Ther. Drug Monit./
Toxicol. (AACC) 16:193-195 (1995).
5. M. Mercurio, F. DeRoos, and R.S. Hoffman. Zolpidem (Ambien):
exposure assessmentof a new nonbenzodiazepine GABA agonist.
Abstract. Vet. Hum. Toxicol. 36:371 (1994).
6. Physicians' Desk Reference, Medical Economics Co., Montvale, NJ,
1999, pp 2929-2933.
7. A. Tracqui, P. Kintz, and P. Mangin. A fatality involving two unusual
compounds---zolpidem and acepromazine. Am. J. Forensic Med.
Pathol. 14" 309-312 (1994).
8. C. Winek, W. Wahba, J. Jannssen,and L. Rozin. Acute overdose of
zolpidem. Forensic Sci. Int. 78:165-168 (1996).
9. G. Debailleul, F.Abi Khalil, and P.Lheureux. HPLC quantification of
zolpidem and prothipendyl in a voluntary intoxication. J. Anal.
ToxicoL 15:35-37 (1991).
10. M. Lichtenwalner and R. Tully. A fatality involving zolpidem. J. AnaL
ToxicoL 21: 567-569 (1997).
11. Forensicsample drug analysis for GC or GC/MS using 200 mg Clean
ScreenTM extraction column. United Chemical Technologies, Bristol,
PA, 1991.
12. W.H. Anderson and R.W. Prouty.Postmortem redistribution of drugs.
In Advances in Analytical Toxicology, R.C. Baselt, Ed. Year Book
Medical Publishers, Chicago, IL, 1989, pp 70--102.
13. W.H. Anderson. Antidepressants. In Principles of Forensic
Toxicology, B. Levine, Ed. AACC Press,Washington, D.C., 1999, pp
309-339.
Manuscript received April 1, 1999;
revision received June 8, 1999.