Clinical features of
status epilepticus in
patients with HIV
Abstract—The authors reviewed the records of 42 patients with HIV infection
and status epilepticus (SE). Brain tumor and infection were the most common
etiologies. The median duration of SE was 2.0 ? 10 hours. Most patients (37
[88%]) responded to IV benzodiazepine or phenytoin treatment. Nevertheless,
12 (29%) patients died and 15 (36%) developed new neurologic deficits. In
patients with HIV infection, aggressive management of seizures may limit the
risk of SE.
Kelly C. Lee, PharmD, BCPP; Paul A. Garcia, MD; and Brian K. Alldredge, PharmD
HIV infection and AIDS are risk factors for seizures
with the incidence between 4 and 11%.1,2Although
seizures commonly occur late in HIV disease, pa-
tients may present with seizures at any time.1-3Sta-
tus epilepticus (SE) occurs in 8 to 14% of HIV
patients with new-onset seizures; however, little is
known regarding the causes, clinical features, and
outcomes of this serious complication.2,4
admitted to the University of California, San Francisco (UCSF)
Medical Center and San Francisco General Hospital (SFGH). Po-
tential cases were identified initially by selecting patients with 1)
concomitant diagnoses of HIV/AIDS (International Classification
of Diseases, Revision 9 [ICD-9] code: 042) and SE (ICD-9 code:
345.3) and 2) discharge date between January 1, 1989, and Au-
gust 31, 2000. For UCSF, we also identified all cases involving
convulsions, not otherwise specified (ICD-9 code: 780.3) in addi-
tion to HIV codes to detect possible code errors for SE. However,
less than 3% of patients were undetected using ICD-9 codes for SE
and HIV. Therefore, the second search strategy was not repeated
We classified a patient as having SE if he or she had a seizure
lasting 15 minutes or longer or at least two seizures occurring
over a 15-minute period without recovery of consciousness be-
tween events. Demographic information, clinical SE details, his-
tory of HIV illness (HIV/AIDS diagnoses, most recent CD4 counts,
opportunistic infections), treatment response to antiepileptic
drugs (AEDs), changes in neurologic function, and overall outcome
were recorded for each patient. A FileMaker Pro 5.0 Version 3
database was used for data collection.
Outcome was categorized as death, neurologic/functional
change, or no change from baseline functioning before the index
SE episode. Outcome was based on the patient’s condition at the
time of hospital discharge. We classified function at discharge as
independent (no neurologic deficit and able to live independently),
impaired, independent (neurologic deficit but able to live indepen-
dently), or impaired, dependent (requiring assistance with activi-
ties of daily living).
This study was approved by the UCSF and SFGH Committees
on Human Research.
We retrospectively reviewed the records of patients
with HIV infection were identified using ICD-9 codes.
Forty-two patients met inclusion/exclusion criteria. Demo-
graphic and clinical characteristics of these patients are
A total of 203 potential cases of SE in patients
shown in table 1. Nine patients (21%) received antiretrovi-
ral drug therapy. Etiologies and provocative factors are
listed in table 2. In approximately one-third of patients,
CNS infection was the likely cause of SE. Toxoplasma in-
fections accounted for approximately 40% of CNS infec-
tions. Other common potential etiologies were CNS tumors
(two patients with confirmed CNS lymphoma) and AED
withdrawal. In seven patients, there was no reasonable
laboratory, radiologic, or historical evidence to explain
the cause of SE, and their etiologies were classified as
The median interval from the onset of SE to initiation of
drug treatment was 0.53 hours (range 0 to 6.3). The me-
dian interval from treatment initiation to cessation of SE
was 1.3 hours (range 0 to 57). The median (? SD) total
duration of SE was 2.0 ? 10 hours (range 0.25 to 57). In 21
(50%) patients, the total duration of SE was 2 hours or
SE was terminated with IV benzodiazepine with or
without phenytoin in 36 patients (86%). In five of the re-
maining patients, seizures were terminated with pheno-
barbital with or without continuous infusion of midazolam
or pentobarbital. SE ceased spontaneously in one patient
before drug treatment. Thirty of the 42 patients (71%)
were admitted to the intensive care unit for management.
The figure shows the change from baseline neurologic
function to functional status at the time of discharge. At
baseline, patients were relatively evenly distributed be-
tween the three function groups (independent, impaired
independent, impaired dependent). However, after the in-
dex SE episode, there was a decline in neurologic function
in 15 patients (36%). Twelve patients (29%) died during
the acute hospitalization.
Patients who died did not differ in age from those who
survived. The median duration of SE in those who died
was longer (4.4 ? 18 hours; mean 11 ? 17.6) than in those
who survived (1.3 ? 4.0 hours; mean 3.1 ? 4.0) (p ?
0.0240). There was a trend toward lower average CD4
counts in patients who died (18.25 ? 20.20/mm3) compared
with those who survived (151.95 ? 297.73/mm3), but the
difference was not significant (p ? 0.1311). The most com-
mon etiology of the SE in both groups was CNS infection.
Autopsy information was available for only one patient
whose death was due to a CNS lymphoma involving the
caudate and parietal and frontal cortex.
Fourteen patients had multifocal or diffuse neuroimag-
ing abnormalities and 13 had normal studies. Fifteen pa-
tients (36%) had focal imaging abnormalities (seven left,
eight right). All but one were in extratemporal regions.
School of Pharmacy (Dr. Lee), Loma Linda University, Loma Linda, CA;
Schools of Medicine (Drs. Garcia and Alldredge) and Pharmacy (Dr. All-
dredge), University of California, San Francisco.
Received December 23, 2004. Accepted in final form April 7, 2005.
Address correspondence and reprint requests to Dr. Kelly C. Lee, School of
Pharmacy, Loma Linda University, 11262 Campus Street, West Hall 1333,
Loma Linda, CA 92350; e-mail: firstname.lastname@example.org
Copyright © 2005 by AAN Enterprises, Inc.
Localization and lateralization of the focal findings did not
predict response to therapy or neurologic outcome.
port on the treatment response and outcomes of SE
in the HIV/AIDS population. Previous studies in the
general population have shown that advanced age
and acute etiologies are associated with greater mor-
bidity and mortality.5,6Additionally, studies have
consistently shown an association between SE dura-
tion and neurologic outcome. The type and severity
of the underlying pathology undoubtedly confound
this observation. Nevertheless, animal studies sug-
gest that SE of longer duration may result in poorer
response to initial treatment.7Although we found an
association between SE duration and mortality, our
study design does not allow us to determine whether
there is an independent association. Although there
To our knowledge, this is the first re-
was no clear difference in etiology for our patients
who died compared with those who survived, the
trend toward lower CD4 counts in patients who died
is suggestive of a more severe underlying disease in
these patients. Perhaps due to our more homoge-
neous population, age did not predict mortality in
Despite the fact that most patients received effec-
tive treatment promptly, many patients died or de-
veloped new neurologic disability. The mortality rate
was approximately twice that found in a previous
study of unselected patients with SE from the same
Table 1 Characteristics of HIV-infected patients with status
epilepticus (n ? 42)*
Sex, n (%)
Male 38 (91)
Median age SD (y)39.5 ? 7.4 (range 30–61)
Ethnicity, n (%)
Asian/Pacific Islander1 (2)
White 21 (50)
Other 2 (5)
Seizure history, n (%)
Previous seizures19 (45)
Location at time of SE onset,
Out of hospital 28 (67)
In hospital 14 (33)
HIV disease status, n (%)
HIV infected, non-AIDS6 (14)
CD4 count at the time of SE†Mean 103.9/mm3? 181.9
?200/mm3: 28 patients
200–500/mm3: 5 patients
?500/mm3: 1 patient
* Percentages of patients may not equal 100% due to rounding.
AIDS defined as HIV-infected and CD4 ?200/mm3and/or
presence/history of AIDS-defining diagnosis.10
† Represents only 34 patients in whom CD4 counts were
SE ? status epilepticus, CD4 ? CD4? lymphocyte.
Table 2 Etiologies of status epilepticus in 42 patients with HIV
Subtotal for CNS infections
Antiepileptic drug withdrawal or noncompliance6
Nonspecific HIV disease2
Total for all etiologies
* Patients may have had more than one etiology.
† History of medically refractory epilepsy and admission antiepi-
leptic drug blood levels within the usual therapeutic range.
Figure. Overall neurologic function at discharge compared
July (2 of 2) 2005
hospitals.8This may be due to the underlying HIV- Download full-text
related disease process being more likely to cause
permanent neurologic injury. It is also possible that
SE may cause lasting neurologic dysfunction in pa-
tients with severe HIV-related brain injuries. Re-
gardless, it suggests that patients with HIV/AIDS
who survive SE will likely need discharge plans ap-
propriate to a new functional level.
SE often occurs in patients with advanced HIV
disease. Thus, patients and caregivers may have pre-
viously preferred less aggressive medical treatment.
Although the idea of hospitalization in the intensive
care unit may be controversial for this high-risk pop-
ulation, we found no factors that absolutely pre-
dicted severe morbidity or mortality. Based on this
observation and the fact that SE was quickly and
easily treated in this group, we believe that nonag-
gressive management of SE would be inappropriate
for most patients. Indeed, many patients were dis-
charged without new neurologic problems.
Because even a single, unprovoked seizure is asso-
ciated with a high recurrence rate in patients with
HIV/AIDS, many patients with new seizures are
treated with AEDs.9Nevertheless, patients with pre-
vious seizures comprised nearly half of our patients
with SE, suggesting that suboptimal seizure treat-
ment may contribute to SE in HIV-infected patients.
Although AEDs do not provide absolute protection
against SE as evidenced by two patients who had
recurrent seizures and SE despite adequate drug lev-
els, most patients with previous seizures were either
untreated or noncompliant with their treatments.
Perhaps more aggressive medical management, pa-
tient counseling, and the use of rescue medicines
would decrease the incidence of SE in these patients
who can ill afford added neurologic disability.
1. Wong MC, Suite ND, Labar DR. Seizures in human immunodeficiency
virus infection. Arch Neurol 1990;47:640–642.
2. Van Paesschen W, Bodian C, Maker H. Metabolic abnormalities and
new-onset seizures in human immunodeficiency virus-seropositive pa-
tients. Epilepsia 1995;36:146–150.
3. Garg RK. HIV infection and seizures. Postgrad Med J 1999;75:387–390.
4. Sagduyu A, Tarlaci S, Sirin H. Generalized tonic-clonic status epilepti-
cus: causes, treatment, complications and predictors of case fatality. J
5. Logroscino G, Hesdorffer DC, Cascino G, Annegers JF, Hauser WA.
Short-term mortality after a first episode of status epilepticus. Epilep-
6. Walton NY. Systemic effects of generalized convulsive status epilepti-
cus. Epilepsia 1993;34(suppl 1):S54–S58.
7. Walton NY, Treiman DM. Response of status epilepticus induced by
lithium and pilocarpine to treatment with diazepam. Exp Neurol 1988;
8. Lowenstein DH, Alldredge BK. Status epilepticus at an urban public
hospital in the 1980s. Neurology 1993;43:483–488.
9. Holtzman DM, Kaku DA, So YT. New-onset seizures associated with
human immunodeficiency virus infection: causation and clinical fea-
tures in 100 cases. Am J Med 1989;87:173–177.
10. 1993 revised classification system for HIV infection and expanded sur-
veillance case definition for AIDS among adolescents and adults.
MMWR Recomm Rep 1992;41:1-19.
WWW.NEUROLOGY.ORG OFFERS IMPORTANT INFORMATION
TO PATIENTS AND THEIR FAMILIES
The Neurology Patient Page provides:
— a critical review of ground-breaking discoveries in neurologic research that are written especially for patients and
— up-to-date patient information about many neurologic diseases
— links to additional information resources for neurologic patients.
All Neurology Patient Page articles can be easily downloaded and printed, and may be reproduced to distribute for
educational purposes. Click on the Patient Page icon on the home page (www.neurology.org) for a complete index of
July (2 of 2) 2005