Seizures in adults with bacterial

meningitis

E. Zoons, MSc ABSTRACT

M. Weisfelt, PhD Objective: To evaluate the occurrence and prognostic relevance of seizures in adults with
J. de Gans, PhD community-acquired bacterial meningitis.
L. Spanjaard, PhD
Methods: An observational cross-sectional study, in which patients with seizures are selected
J.H.T.M. Koelman,
from a prospective nationwide cohort of 696 episodes of community-acquired bacterial meningi-
PhD
tis, confirmed by culture of CSF in patients aged ⬎16 years. We retrospectively collected data on
J.B. Reitsma, PhD
EEGs.
D. van de Beek, PhD
Results: Seizures occurred in 121 of 696 episodes (17%). Death occurred in 41% of patients
with seizures compared to 16% of patients without seizures (p ⬍ 0.001). The median number of
Address correspondence and
seizures was 2 (interquartile range [IQR] 1 to 4). The median time between admission and the first
reprint requests to Dr. Diederik seizure was 1 day (IQR 0 to 3). Patients with in-hospital seizures were more likely to have a CSF
van de Beek, Department of
leukocyte count below 1,000 cells/mm3 (36% vs 25%; p ⫽ 0.01), had higher median CSF protein
Neurology, Academic Medical
Center, PO Box 22660, 1100 levels (4.8 g/L [IQR 3.4 to 7.6] vs 4.1 g/L [IQR 2.1 to 6.8]), and higher median erythrocyte sedi-
DD Amsterdam, The mentation rate (46 mm/hour [IQR 31 to 72] vs 36 mm/hour [IQR 18 to 69]; p ⫽ 0.02) than patients
Netherlands
D.vandebeek@amc.uva.nl without in-hospital seizures. Focal cerebral abnormalities developed more often in patients with
in-hospital seizures than in those without (41% vs 14%; p ⬍ 0.001). In a multivariate analysis,
seizures were significantly more likely in patients with predisposing conditions, tachycardia, a low
Glasgow Coma Scale score on admission, infection with Streptococcus pneumoniae, and focal
cerebral abnormalities. Neuroimaging was performed on admission in 70% of episodes with pre-
hospital seizures, with CT revealing a focal lesion in 32% of those episodes. Antiepileptic drugs
were administered in 82% of patients with seizures and EEG was performed in 31% of episodes;
a status epilepticus was recorded in five patients.
Conclusions: Seizures occur frequently in adults with community-acquired bacterial meningitis.
Seizures are associated with severe CNS and systemic inflammation, structural CNS lesions,
pneumococcal meningitis, and predisposing conditions. The high associated mortality rate war-
rants a low threshold for starting anticonvulsant therapy in those with clinical suspicion of a
seizure. Neurology® 2008;70:2109–2115

GLOSSARY
GCS ⫽ Glasgow Coma Scale; GOS ⫽ Glasgow Outcome Scale; IQR ⫽ interquartile range.

Bacterial meningitis is a serious and life-threatening disease. Streptococcus pneumoniae

Supplemental data at
www.neurology.org
Editorial, page 2095
and Neisseria meningitidis are the predominant causative pathogens in adults, causing 80
to 85% of all cases, with mortality rates varying from 28 to 30% in studies reporting on S
pneumoniae and 7 to 10% for N meningitidis.1,2 Previous studies have described seizures
as an individual predictor of poor outcome in bacterial meningitis.2-6 Recently, we de-
scribed clinical features and prognostic factors in 696 episodes of community-acquired
bacterial meningitis in adults.1 We now report on the frequency of seizures in this cohort
and compare the clinical features and outcomes in patients with and without seizures.
e-Pub ahead of print on February 27, 2008, at www.neurology.org.
From the Department of Neurology (E.Z., M.W., J.d.G., J.H.T.M.K., D.v.d.B.), Department of Medical Microbiology (L.S.), The
Netherlands Reference Laboratory for Bacterial Meningitis (L.S.), Department of Clinical Neurophysiology (J.H.T.M.K.), and Department
of Clinical Epidemiology, Biostatistics and Bioinformatics (J.B.R.), Center of Infection and Immunity Amsterdam (CINIMA), Academic
Medical Center, Amsterdam, The Netherlands.
Disclosure: The authors report no conflicts of interest.

Copyright © 2008 by AAN Enterprises, Inc. 2109


Table 1 Clinical and laboratory features in adults with and without prehospital
seizures among 666 episodes of bacterial meningitis*
Characteristic
History of meningitis
Predisposing conditions
Pneumonia
Immunocompromise†
Symptoms and signs on admission
Heart rate ⬎120 beats/min
GCS on admission
⬍14 (indicating change in mental status)
⬍8 (indicating coma)
CSF culture
Streptococcus pneumoniae
Neisseria meningitidis
*Data are number/number evaluated (%).
†Defined as the use of immunosuppressive drugs, presence of asplenia, diabetes mellitus,
alcoholism, or infection with HIV.
GCS ⫽ Glasgow Coma Scale.
2110
METHODS The Dutch Meningitis Cohort Study, a pro-
spective nationwide observational cohort study in the Neth-
erlands, included 696 episodes of community-acquired
bacterial meningitis in adults.1 Inclusion and exclusion crite-
ria are described more extensively elsewhere.1 In summary,
all patients were aged over 16 years and had CSF culture
proven bacterial meningitis. In total, 696 episodes of
community-acquired bacterial meningitis occurred in 671
patients; 25 patients had a second episode of bacterial men-
ingitis. The mean age of the group was 50 ⫾ 20 years and
50% were male. CSF culture yielded S pneumoniae in 352
episodes (51%), N meningitidis in 257 episodes (37%), and
other bacteria in 87 episodes (13%).1 The current study is an
observational cross-sectional study, in which patients with
seizures are selected from the database. The Dutch Meningi-
tis Cohort Study was approved by our ethics committee and
informed consent was obtained from all participating pa-
tients or their legally authorized representatives. Informa-
tion, including specific queries about seizures, was collected
by means of a case record form. EEG reports were collected
retrospectively from all patients in whom seizures were re-
corded. At discharge, all patients underwent a neurologic
examination performed by a neurologist, and outcome was
graded according to the Glasgow Outcome Scale (GOS).
This is a well-validated measurement scale with scores vary-
ing from 1 (indicating death) to 5 (good recovery). A favor-
able outcome was defined as a score of 5, and an unfavorable
outcome as a score of 1 to 4. We categorized the cause of
death in patients who died within 14 days after admission, as
death within this period is likely to be caused by direct con-
sequences of the meningitis.1,7 Two experienced clinicians
(M.W., D.v.d.B.) independently classified the cause of death
into systemic causes (e.g., septic shock, respiratory failure,
multiple-organ dysfunction, and cardiac ischemia) or neuro-
logic causes (e.g., brain herniation, cerebrovascular compli-
cations, intractable seizures, and withdrawal of care because
of poor neurologic prognosis). The kappa for interrater
Prehospital No prehospital
seizures seizures
(n ⫽ 33) (n ⫽ 633) p Value
6 (18) 27 (4) 0.004
10/33 (30) 67/633 (11) 0.002
11/33 (33) 92/632 (15) 0.004
8/31 (26) 61/591 (10) 0.02
29/33 (88) 421/631 (67) 0.01
12/33 (36) 78/631 (12) 0.001
25/33 (76) 301/633 (48) 0.002
4/33 (12) 252/633 (40) 0.001
Neurology 70 May 27, 2008 (Part 2 of 2)
agreement for the classification of cause of death was 0.60;
differences in classification were resolved by discussion.8,9
All EEG-reports were scored by a clinical neurophysiolo-
gist (J.K.) into categories on three subjects: background pat-
tern (normal, mildly abnormal, moderately abnormal, or
severely abnormal), the presence of epileptic discharges (epi-
leptic discharges absent/present or status epilepticus) and fo-
cal abnormalities (focal abnormalities absent/present,
asymmetric abnormalities, or multifocal abnormalities). The
CRF contained items concerning seizures, but did not specif-
ically evaluate status epilepticus. Therefore, status epilepti-
cus was EEG-defined.
Analyses were performed for prehospital seizures (de-
fined as seizures before admission), in-hospital seizures (de-
fined as seizures during admission), and all seizures.
Immunocompromise was defined as the use of immunosup-
pressive drugs, presence of asplenia, diabetes mellitus, alco-
holism, or infection with the HIV.
For the comparison of non-normally distributed vari-
ables nonparametric testing, Mann-Whitney U, ␹2, or Fisher
exact tests were used. We used logistic regression analysis to
calculate OR and 95% CI to assess the strength of the associ-
ation between potential risk factors and seizures. Based on
previous research and pathophysiologic interest, 13 poten-
tially relevant risk factors were selected.10 Although the me-
dian percentage of missing values for individual variables in
our study was low (2%), data were complete on all potential
predictors in 408 out of 696 episodes (59%). Complete case
analysis would seriously hamper the power of the multivari-
ate models and, in addition, might lead to biased results.
Therefore, we used the multiple imputation method to re-
place each missing value with a set of plausible values that
represent the uncertainty about the right value to impute. By
repeating this process several times, the uncertainty in the
precision is properly taken into account.11 The final esti-
mates of the multivariate model were obtained by combining
the results of 10 rounds of imputations. Analyses were also
performed on the non-imputed dataset. As a next step we
evaluated whether the prognostic value of these risk factors
could be attributed to the causative pathogen, by adjusting
the analysis with the inclusion of this variable into the prog-
nostic model. Finally, we examined whether the prognostic
value of individual independent risk factors was different for
the occurrence of early seizures (defined as seizures within 48
hours after admission) compared to late seizures (⬎48 hours
after admission). This was done by performing multinomial
response logistic regression analysis using three different
outcomes: no seizures, early seizures, and late seizures. In
this model separate coefficients (e.g., OR) are estimated for
each risk factor comparing early vs no seizures and late vs no
seizures. Population description is done with medians and
interquartile ranges (IQR). All analyses were performed us-
ing SAS software, version 9.11 (SAS Institute) and a p value
below 0.05 was considered as significant.
RESULTS Prehospital seizures occurred in 33 of
666 evaluated episodes (5%; table 1, significant
results only; see full table e-1 on the Neurology®
Web site at www.neurology.org). Information on
prehospital seizures was missing in 30 episodes,
and these episodes were more likely to have im-
paired mental status (score on the Glasgow Coma
Scale [GCS] ⬍ 14; 27/30 [90%] vs 421/631 [67%];
 more likely to have predisposing conditions, like
Table 2 Clinical and laboratory features in adults with and without in-hospital
seizures among 687 episodes of bacterial meningitis*
otitis/sinusitis or an immunocompromised state.
Again, CSF culture yielded S pneumoniae in a
In-hospital No in-hospital higher proportion of episodes (79% vs 46%; p ⬍
seizures seizures
Characteristic (n ⫽ 107) (n ⫽ 580) p Value 0.001). Patients with in-hospital seizures were
Age, y 58 ⫾ 19 48 ⫾ 20 ⬍0.001 more likely to a have a CSF leukocyte count be-
Predisposing conditions low 1,000 cells/mm3 (36% vs 25%; p ⫽ 0.01), had
Otitis or sinusitis 40 (37) 133 (23) 0.002
higher median CSF protein levels (4.8 g/L [IQR
Pneumonia 19 (18) 64 (11) 0.05
3.4 to 7.6] vs 4.1 g/L [IQR 2.1 to 6.8]; p ⫽ 0.02),
and higher median erythrocyte sedimentation rate
Immunocompromise† 36 (34) 77/579 (13) ⬍0.001
(ESR) (46 mm/hour [IQR 31 to 72] vs 36 mm/hour
GCS on admission
[IQR 18 to 69]; p ⫽ 0.01). Focal neurologic abnor-
⬍14 (indicating change in mental status) 93/106 (88) 376/579 (65) ⬍0.001
malities were more likely to develop during the clin-
⬍8 (indicating coma) 22/106 (21) 70/579 (12) 0.02
ical course of episodes with in-hospital seizures than
CSF culture in those without (41% vs 14%; p ⬍ 0.001).
Streptococcus pneumoniae 85 (79) 264 (46) ⬍0.001 Overall, in 121 of 696 episodes (17%) at least
Neisseria meningitidis 12 (11) 240 (41) ⬍0.001 one seizure occurred; both prehospital and in-
Indexes of CSF inflammation‡ hospital seizures occurred in 19 of 657 evaluated
White cell count episodes (3%). The median number of seizures in
⬍1,000/mm3 39 (36) 143 (25) 0.01 patients with seizures was 2 (IQR 1 to 4). In 91 of
ⱖ1,000/mm 3
68 (64) 437 (75) 0.01
121 episodes (75%) the first seizure occurred be-
fore or within 48 hours after admission; in the
Protein, g/L 4.8 (3.4–7.6) 4.1 (2.1–6.8) 0.02
remaining 30 episodes (25%) the first seizure oc-
Blood chemistry tests§
curred more than 48 hours after admission. In a
ESR, mm/hr 46 (31–72) 36 (18–69) 0.01
multivariate analysis, a distant focus of infection
*Data are number/number evaluated (%), mean ⫾ SD, or median (interquartile range). (sinusitis, otitis, or pneumonia), an immunocom-
†Defined as the use of immunosuppressive drugs, presence of asplenia, diabetes mellitus, promised state, tachycardia, and a low GCS score
alcoholism, or infection with HIV.
on admission were associated with a higher risk
‡The CSF leukocyte count was determined in 619 episodes.
§
The erythrocyte sedimentation rate (ESR) was determined in 523 episodes. for seizures (table 3). Results of this analysis on
GCS ⫽ Glasgow Coma Scale. the imputed and nonimputed dataset were similar
(table e-3); results did not change after inclusion
p ⫽ 0.008). The onset of seizures was marked as of hyponatremia in the multivariate analysis. Af-
focal in seven episodes with secondary generaliza- ter inclusion of bacterial cause in the model, the
tion in four. Patients with prehospital seizures effect of immunocompromised state, a low GCS,
were more likely to have a second episode of men- and infection with S pneumoniae remained ro-
ingitis (18% vs 4%; p ⫽ 0.004), and predisposing bust. Dichotomization of the cohort with respect
conditions like pneumonia (30% vs 11%; p ⫽ 0.002) to the time of the first seizure resulted in similar
or an immunocompromised state (33% vs 15%; risk estimates for most variables with wider CIs
p ⫽ 0.004) than those without prehospital sei- due to a decrease in power. However, in patients
zures. Patients with prehospital seizures were also with late seizures (more than 48 hours after ad-
more often admitted with a changed mental sta- mission) GCS on admission was not predictive for
tus (88% vs 67%; p ⫽ 0.01) and tachycardia seizures as in patients with early seizures (within
(heart rate ⬎120 beats per minute; 26% vs 10%; 48 hours after admission). In patients with sei-
p ⫽ 0.02). Lumbar puncture was performed in all zures more than 48 hours after admission focal
episodes and CSF culture yielded S pneumoniae in cerebral abnormalities (aphasia, monoparesis,
a higher proportion of episodes with prehospital hemiparesis, or quadriparesis) on admission
seizures (76% vs 48%; p ⫽ 0.002). Results of CSF emerged as a more predictive factor in this group
analysis and blood chemistry tests were similar. than for early seizures.
In-hospital seizures occurred in 107 of 687 Neuroimaging was performed on admission in
evaluated episodes (16%; table 2, see full table 288 of 696 episodes (41%) and consisted of CT of
e-2). A focal onset was present in 44 episodes with the brain in all. CT on admission was performed
secondary generalization in 23. Median time be- more often in episodes with than in those without
tween admission and first seizure was 1 day (IQR prehospital seizures (23/33 [70%] vs 265/633
0 to 3). Patients with in-hospital seizures were [42%]; p ⫽ 0.002). Imaging revealed abnormali-
older (58 ⫾ 19 vs 48 ⫾ 20 years; p ⬍ 0.001) and ties in 10 of 23 recorded episodes with prehospital

Neurology 70 May 27, 2008 (Part 2 of 2) 2111

 Table 3 Multivariate analysis of factors associated with seizures

Adjusted for Seizures within Seizures after

Characteristics All patients bacterial cause 48 hours* 48 hours

Age, y 1.09 (0.97–1.23) 1.03 (0.91–1.17) 1.07 (0.93–1.24) 0.93 (0.74–1.18)

Predisposing conditions

Distant focus of infection† 1.83 (1.19–2.83) 1.26 (0.79–2.02) 1.23 (0.72–2.09) 1.32 (0.57–3.05)

Immunocompromise‡ 2.64 (1.62–4.32) 2.54 (1.55–4.17) 2.68 (1.54–4.66) 2.61 (1.08–6.31)

Clinical characteristics on admission

Heart rate ⬎120 beats/min 1.85 (1.00–3.40) 1.81 (0.96–3.41) 1.97 (0.99–3.92) 1.39 (0.42–4.68)

Diastolic blood pressure ⬍60 mm Hg 1.18 (0.52–2.67) 1.37 (0.59–3.16) 1.56 (0.63–3.84) 0.71 (0.08–6.11)

Score on Glasgow Coma Scale 0.89 (0.83–0.96) 0.91 (0.85–0.98) 0.87 (0.80–0.94) 1.09 (0.93–1.26)

Cranial nerve palsies 1.31 (0.73–2.35) 1.14 (0.63–2.07) 0.99 (0.49–2.02) 1.72 (0.69–4.24)

Focal cerebral abnormalities§ 1.53 (0.96–2.44) 1.48 (0.92–2.38) 1.31 (0.77–2.23) 2.60 (1.12–6.08)

Laboratory features

CSF white-cell count ⬍1,000/mm3 1.24 (0.77–2.01) 1.21 (0.74–1.96) 1.09 (0.63–1.88) 1.54 (0.65–3.67)

Positive blood culture 1.35 (0.79–2.29) 1.25 (0.73–2.15) 1.14 (0.64–2.01) 1.85 (0.58–5.93)

ESR, mm/hour 1.07 (0.95–1.20) 1.08 (0.96–1.23) 1.08 (0.94–1.24) 1.11 (0.87–1.43)
3
Thrombocyte count, platelets/mm 1.13 (0.92–1.39) 1.07 (0.86–1.33) 1.06 (0.81–1.38) 1.11 (0.78–1.58)

CSF culture

Streptococcus pneumoniae 3.05 (1.72–5.42) 2.74 (1.45–5.19) 4.44 (1.43–13.80)

Other bacteria 1.00¶ 1.00¶ 1.00¶

Values are OR (95% CI). ORs are calculated in 10-year increments for age, per 20 mm per hour for erythrocyte sedimenta-
tion rate (ESR), per 100,000 per mm3, and per one-point decrease on the Glasgow Coma Scale.
*This group includes all patients with prehospital seizures or seizures within 48 hours after admission.
†Defined as otitis/sinusitis or pneumoniae.
‡Defined as the use of immunosuppressive drugs, presence of asplenia, diabetes mellitus, alcoholism, or infection with
HIV.
§
Defined as aphasia, mono-, hemi-, or quadriparesis.
¶
This group served as a reference category.

seizures (43%): sinusitis or otitis in 5 (50%), post-

traumatic lesions in 2 (20%), mastoiditis in 2
(20%), brain edema in 1 (10%), white matter le-
Table 4 Cranial CT scan in adults with and without seizures among 696
episodes of bacterial meningitis
sions in 1 (10%), and old infarction in 1 (10%).
Overall, episodes with seizures were more likely
Seizures No seizures to have abnormalities on CT (table 4; 53% vs
Characteristic (n ⫽ 121) (n ⫽ 575) p Value
40%; p ⫽ 0.01). A focal lesion (brain infarction,
CT scan performed 107 (88) 397 (69) ⬍0.001
empyema/cerebritis, posttraumatic lesion or
CT scan abnormal 57 (53) 158 (40) 0.01
bleeding) on CT was more common in episodes
Sinusitis/otitis 20 (19) 42 (11) 0.02
with seizures than in those without (34/107 [32%]
Infarction 21 (20) 43 (11) 0.02
vs 66/397 [17%]; p ⬍ 0.001).
Edema 15 (14) 42 (11) 0.32 Antiepileptic medication was administered in
Hydrocephalus 3 (3) 15 (4) 0.78 98 of 111 evaluated episodes (85%) with seizures
Cerebritis/empyema 9 (8) 17 (4) 0.09 and consisted of phenytoin in 36 of 98 evaluated
Brain swelling* 21 (20) 56 (14) 0.16 episodes (37%), valproinic acid in 25 episodes
Other† 8 (7) 22 (6) 0.17 (26%), and clonazepam in 6 episodes (6%); other
antiepileptic regimens (including combination of
Data are number/number evaluated (%). Percentages are calculated per number of episodes antiepileptics) were used in 23 episodes (23%). In
with cranial CT undertaken. Numbers do not add up to totals because of the presence of
multiple abnormalities in several patients.
the remaining 4 episodes (4%) antiepileptic ther-
*Brain swelling defined as cerebritis, empyema, or edema. apy was administered, but it is unknown what
†Cerebral atrophy in six, posttraumatic brain abnormalities in four, skull fracture in four, old drugs were used. In 2 of the 111 episodes (3%)
infarction in three, pneumocephalus in three, white matter lesions in two, meningioma in one,
only sedative drugs were administered; 11 of 111
arachnoid cyst in one, Dandy-Walker malformation in one, status after operation on hypophy-
seal malignancy in one, vascular aneurysm in one, small intracerebral bleeding in one, diffuse evaluated episodes with seizures (10%) remained
brain swelling in one, and subarachnoid hemorrhage in one. untreated for unknown reasons.

2112 Neurology 70 May 27, 2008 (Part 2 of 2)

 Outcome was unfavorable in 237 of 696 of ep-
Table 5 Outcome in adults with and without seizures among 696 episodes of
bacterial meningitis
isodes (34%). Death occurred in 143 of 696 epi-
sodes (21%). Episodes with seizures were more
Seizures No seizures likely to have an unfavorable outcome (64% vs
Characteristic (n ⫽ 121) (n ⫽ 575) p Value
28%; p ⬍ 0.001; table 5); 41% of the patients with
Glasgow Outcome Scale score ⬍0.001*
seizures died. Thirty-six out of 50 fatal episodes
1 (death) 50 (41) 93 (16)
with seizures (72%) resulted in death within 2
2 (vegetative state) 2 (2) 1 (⬍1)
weeks after admission and death was attributed
3 (severe disability) 12 (10) 12 (2)
to intractable seizures in 4 (8%) patients. In pa-
4 (moderate disability) 14 (12) 53 (9) tients who had their first seizure during hospital-
5 (mild or no disability) 43 (36) 416 (72) ization 44 out of 88 died (50%); median time
Cause of death‡ 0.11† between the first seizure and death was 5 days
Systemic 17/36 (47) 53/84 (63) (IQR 1 to 15). Patients with seizures tended to die
Neurologic 19/36 (53) 31/84 (37) more often of neurologic causes compared to
Neurologic findings at discharge§ those without, but this difference did not reach
Cranial nerve palsy 19/61 (31) 87/420 (21) 0.07
significance (53% vs 37%; p ⫽ 0.11). Neurologic
examination at discharge was performed in 550 of
Hearing loss 14/69 (20) 64/466 (14) 0.15
553 surviving patients (99%) and revealed focal
Focal cerebral abnormalities 17/70 (24) 29/472 (6) ⬍0.001
cerebral abnormalities in a higher proportion of
Aphasia 4/70 (6) 7/479 (1) 0.04
episodes with than without seizures (24% vs 6%;
Hemiparesis 14/69 (20) 10/479 (2) ⬍0.001
p ⬍ 0.001). Within the group of patients with ep-
Quadriparesis 2/70 (3) 4/476 (1) 0.17
ilepsy, EEG abnormalities were related to fatal
Data are number/number evaluated (%).
outcome: EEG (moderately or severely abnormal)
*Chi-square test for trend. background pattern (14/24 [58%] vs 2/13 [15%];
†Chi-square test, two-tailed. p ⫽ 0.01), epileptic discharges (9/11 [82%] vs 7/26
‡Evaluated in patients who died within 2 weeks after admission.
§
[27%]; p ⫽ 0.003). All five patients with status
Neurologic examination was performed at discharge in 550 of 553 surviving patients.
epilepticus died.
EEG was performed in 37 of 121 episodes with
seizures (31%; table E-4). Background patterns DISCUSSION Our study shows that seizures are
were mildly abnormal in 10 of 37 episodes (27%), common in adults with community-acquired bac-
and moderately or severely abnormal in 24 epi- terial meningitis (17%) and are associated with a
sodes (65%). Focal or multifocal abnormalities higher mortality rate (41%). Previous retrospec-
were present in 14 of 37 episodes (38%). Epileptic tive case series from tertiary hospitals reported
discharges were recorded in 6 of 37 episodes seizures in 15 to 23% of cases.2-6,10
(16%) and a status epilepticus was recorded in 5 Cortical inflammation has classically been de-
patients (14%). Episodes with an abnormal back- scribed as the main mechanism of seizures in bacte-
ground pattern were more likely to receive anti- rial meningitis.6 In our study seizures were related to
epileptics than those without these abnormalities CNS inflammation, reflected in high protein in CSF,
(21/24 [88%] vs 4/10 [40%]; p ⫽ 0.009). All 11 and associated meningoencephalitis, reflected in
episodes with epileptic discharges or a status epi- low levels of consciousness, but we identified several
lepticus on EEG received antiepileptics. other risk factors for seizures.
In total, 121 of 696 episodes (17%) were First, patients with seizures had higher ESR
treated with adjunctive steroids; 25 episodes with and patients with in-hospital seizures were also
seizures received steroids (21%). In the group of more likely to have a CSF leukocyte count below
patients who did not receive steroids the rate of 1,000 cells/mm3. Previous studies have identified
an unfavorable outcome was higher than in those low CSF leukocyte counts to be associated with
who did receive steroids (53 of 121 episodes severe systemic inflammation.12,13 A previous post
[44%] vs 184 of 575 [32%]; p ⫽ 0.01). However, hoc analysis of the Dutch Meningitis Cohort
in these patients steroids were often initiated after showed that a low CSF leukocyte count was re-
clinical deterioration. Episodes in which cortico- lated to the presence of signs of sepsis and sys-
steroids were administered before antibiotics temic complications.1,12 Systemic infection has
were less likely to have an unfavorable outcome been identified as an important risk factor for sei-
than episodes in which corticosteroids were ad- zures in other diseases.14,15
ministered after antibiotics (3 of 25 episodes Second, seizures were associated with focal
[12%] vs 50 of 96 episodes [52%]; p ⬍ 0.001). neurologic abnormalities and focal lesions on cra-

Neurology 70 May 27, 2008 (Part 2 of 2) 2113

 nial CT. The most common intracranial compli-
cation in patients with seizures was brain
infarction which was recorded in 20%. A previ-
ous retrospective cohort study on poststroke sei-
zures including 3,205 patients admitted for a first-
ever stroke showed that 5% of the patients with a
stroke developed seizures; 36% of these seizures
were early onset seizures, defined as a seizure
within 14 days of the stroke.16 Interestingly, our
multivariate analysis showed that focal abnor-
malities are related with seizures ⬎48 hours after
admission. Brain lesions that we classified as focal
lesions (infarction, cerebritis/empyema, posttrau-
matic lesions, and intracranial bleeding) have all
been associated with seizures in patients without
bacterial meningitis.17-22
Third, the causative organism had an indepen-
dent effect on the risk for seizures. The presence
of a distant focus of infection (e.g., ear or sinus
infections and pneumonia), which is indicative
for pneumococcal infections, was no longer sig-
nificant after adjustment for the bacterial cause.
The odds of a seizure were four times higher
among patients infected with S pneumoniae than
among patients infected with other bacteria, even
after adjustment for other clinical predictors.
This indicates an organism-specific effect, which
might be related to vasculitis, causing focal cere-
bral abnormalities after 48 hours. Finally, the pre-
dictive effect of an altered immune status
remained robust in the multivariate model, even
after inclusions of the causative organism. This
might be caused by the definition of altered im-
mune status, including alcoholism and diabetes.
Patients with alcoholism and diabetes might have
higher risk of seizures due to alcohol withdrawal
and hypoglycemia.23,24
Severe CNS and systemic inflammation, as
well as infection by S pneumoniae, have been de-
scribed previously as predictors of unfavorable
outcome in this cohort.1 Seizures can be regarded
as a quasi outcome measure such as mechanical
ventilation or intensive care unit admission.
However, prehospital seizures, focal neurologic
abnormalities, and an altered immune status were
not related with unfavorable outcome in our pre-
vious analysis.1 Nevertheless, it is impossible to
determine the driving force for the poorer prog-
nosis of patients with seizures.
Four characteristics of our study have to be
considered when interpreting the results. First,
only patients who had a positive CSF culture were
included. Negative CSF cultures are estimated to
occur in 11 to 30% of patients with bacterial men-
ingitis.1,2,25,26 This percentage is expected to be
2114 Neurology 70 May 27, 2008 (Part 2 of 2)
even higher in patients who have received antibi-
otics before lumbar puncture, which is recom-
mended in patients with new-onset seizures in
whom imaging should precede lumbar puncture
according to current guidelines.1,9,26 This might
have resulted in a relatively low rate of patients
with sepsis. In addition, patients with space-
occupying lesions on CT who are expected to be
prone to seizures may not undergo lumbar punc-
ture. Therefore, these patient groups were proba-
bly only partly represented in our study, which
could have resulted in an underestimated seizure
rate as well as the underestimation of the rate of
adverse outcome. Second, this study is an obser-
vational cohort study. Prehospital seizures were
not evaluated in 30 episodes. These episodes were
more likely to have impaired mental status, which
might be a sign of seizure activity. The study de-
sign may have led to an underestimation of the
rate of (prehospital) seizures. Third, most pa-
tients in this study did not receive steroid therapy.
In patients who received steroid therapy, treat-
ment was often started after clinical deteriora-
tion, for example onset of seizures; however, the
association between clinical deterioration and
start of high-dose steroids is unclear. The Euro-
pean Dexamethasone Study showed that treat-
ment with adjunctive dexamethasone, started
before or with the first dose of antibiotics, reduces
mortality in adults with bacterial meningitis.27
Dexamethasone treatment is now recommended in
most adults with suspected bacterial meningitis,28,29
which may affect clinical course and outcome in
these patients.30 And finally, lack of routine MRI
could have resulted in underestimation of the num-
ber of episodes with intracranial abnormalities.
How to treat patients with bacterial meningitis
and seizures? The high mortality rate in patients
with seizures warrants a low threshold for start-
ing anticonvulsant therapy in those with prior sei-
zure or clinical suspicion of a seizure. The
incidence of this complication does not justify
prophylactic treatment. The effect of antiepileptic
drugs in these patients has to be evaluated. A dif-
ficult group of patients are those admitted in
coma. Most patients with bacterial meningitis
and lowered consciousness are admitted to the in-
tensive care unit where they might experience
subtle, unnoticed seizures that might negatively
affect their outcome. Seizures might cause coma
and a rare cause of deterioration of consciousness
in meningitis is nonconvulsive status epilepticus.
If seizures have occurred and the patient does not
regain consciousness, or no seizures were re-
corded but the patient is in a coma or conscious-
ness fluctuates, an EEG is indicated and treatment
with anticonvulsant therapy should be initiated
accordingly.10
Received May 23, 2007. Accepted in final form August 8,
2007.
REFERENCES
1. van de Beek D, de Gans J, Spanjaard L, Weisfelt M,
Reitsma JB, Vermeulen M. Clinical features and prog-
nostic factors in adults with bacterial meningitis. N
Engl J Med 2004;351:1849–1859.
2. Durand ML, Calderwood SB, Weber DJ, et al. Acute
bacterial meningitis in adults: a review of 493 episodes.
N Engl J Med 1993;328:21–28.
3. Kastenbauer S, Pfister HW. Pneumococcal meningitis
in adults. Spectrum of complications and prognostic
factors in a series of 87 cases. Brain 2003;126:1015–
1025.
4. Hussein AS, Shafran SD. Acute bacterial meningitis in
adults. A 12-year review. Medicine (Baltimore) 2000;
79:360–368.
5. Bohr V, Rasmussen N, Hansen B, et al. Pneumococcal
meningitis: an evaluation of prognostic factors in 164
cases based on mortality and on a study of lasting se-
quelae. J Infect 1985;10:143–157.
6. Lu CH, Huang CR, Chang WN, et al. Community-
acquired bacterial meningitis in adults: the epidemiol-
ogy, timing of appropriate antimicrobial therapy, and
prognostic factors. Clin Neurol Neurosurg 2002;104:
352–358.
7. McMillan DA, Lin CY, Aronin SI, Quagliarello VJ.
Community-acquired bacterial meningitis in adults:
categorization of causes and timing of death. Clin In-
fect Dis 2001;33:969–975.
8. Weisfelt M, van de Beek D, Spanjaard L, Reitsma JB,
de Gans J. Clinical features, complications and out-
come in adults with pneumococcal meningitis: a pro-
spective case series. Lancet Neurol 2006;5:123–129.
9. Weisfelt M, van de Beek D, Spanjaard L, Reitsma JB,
de Gans J. Community-acquired bacterial meningitis in
older people. J Am Geriatr Soc 2006;54:1500–1507.
10. van de Beek D, de Gans J, Tunkel AR, Wijdicks EF.
Community-acquired bacterial meningitis in adults.
N Engl J Med 2006;354:44–53.
11. Little R, Rubin D. Statistical analysis with missing
data. New York: Wiley, 1987.
12. Weisfelt M, van de Beek D, Spanjaard L, Reitsma JB,
de Gans J. Attenuated cerebrospinal fluid leukocyte
count and sepsis in adults with pneumococcal meningi-
tis: a prospective cohort study. BMC Infect Dis 2006;6:
149E.
13. Ostergaard C, Konradsen HB, Samuelsson S. Clinical
presentation and prognostic factors of Streptococcus
pneumoniae meningitis according to the focus of infec-
tion. BMC Infect Dis 2005;5:93–103.
14. Haffey S, McKernan A, Pang K. Non-convulsive status
epilepticus: a profile of patients diagnosed within a ter-
tiary referral centre. J Neurol Neurosurg Psychiatry
2004;75:1043–1044.
15. Dunne JW, Summers QA, Stewart-Wynne EG. Non-
convulsive status epilepticus: a prospective study in an
adult general hospital. Q J Med 1987;62:117–126.
16. Berges S, Moulin T, Berger E, et al. Seizures and epi-
lepsy following strokes: recurrence factors. Eur Neurol
2000;43:3–8.
17. Arboix A, Comes E, Garcia-Eroles L. Site of bleeding
and early outcome in primary intracerebral hemor-
rhage. Acta Neurol Scand 2002;105:282–288.
18. Yang SY, Zhao CS. Review of 140 patients with brain
abscess. Surg Neurol 1993;39:290–296.
19. Bernardini GL. Diagnosis and management of brain
abscess and subdural empyema. Curr Neurol Neurosci
Rep 2004;4:448–456.
20. Yeates KE, Halliday W, Miyasaki J, Vellend H, Strauss
S. A case of “circling seizures” and an intratumoral ab-
scess. Clin Neurol and Neurosurg 2003;105:128–131.
21. Englander J, Bushnik T, Duong TT, et al. Analyzing
risk factors for late posttraumatic seizures: a prospec-
tive, multicenter investigation. Arch Phys Med Rehabil
2003;84:365–373.
22. Asikainen I, Kaste M, Sarna S. Early and late posttrau-
matic seizures in traumatic brain injury rehabilitation
patients: brain injury factors causing late seizures and
influence of seizures on long-term outcome. Epilepsia
1999;40:584–589.
23. Bugard M, McIntyre J, Hill KR, Woodside J Jr . Alco-
hol withdrawal syndrome. Am Fam Physician 2004;69:
1443–1450.
24. Vriesendorp TM, DeVries JH, van Santen S, et al. Eval-
uation of short-term consequences of hypoglycemia in
an intensive care unit. Crit Care Med 2006;34:2714–
2718.
25. Sigurdardottir B, Bjornsson OM, Jonsdottir KE, Er-
lendsdottir H, Gudmundsson S. Acute bacterial menin-
gitis in adults. A 20-year overview. Arch Intern Med
1997;157:425–430.
26. Tunkel AR. Bacterial meningitis. Philadelphia: Lippin-
cott Williams & Wilkins, 2001.
27. de Gans J, van de Beek D. Dexamethasone in adults
with bacterial meningitis. N Engl J Med 2002;347:
1549–1556.
28. Tunkel AR, Hartman BJ, Kaplan SL, et al. Practice
guidelines for the management of bacterial meningitis.
Clin Infect Dis 2004;39:1267–1284.
29. van de Beek D, de Gans J, McIntyre P, Prasad K. Ste-
roids in adults with bacterial meningitis: a systematic
review. Lancet Infect Dis 2004;4:139–143.
30. Auburtin M, Porcher R, Bruneel F, et al. Pneumococcal
meningitis in the intensive care unit: prognostic factors
of clinical outcome in a series of 80 cases. Am J Respir
Crit Care Med 2002;165:713–717.
Neurology 70 May 27, 2008 (Part 2 of 2) 2115
 Seizures in adults with bacterial meningitis
E. Zoons, M. Weisfelt, J. de Gans, et al.
Neurology 2008;70;2109-2115 Published Online before print February 27, 2008
DOI 10.1212/01.wnl.0000288178.91614.5d

This information is current as of February 27, 2008

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright . All rights reserved. Print ISSN: 0028-3878.
Online ISSN: 1526-632X.
 Updated Information & including high resolution figures, can be found at:
Services http://www.neurology.org/content/70/22_Part_2/2109.full.html

Supplementary Material Supplementary material can be found at:

http://www.neurology.org/content/suppl/2008/05/24/01.wnl.000028817
8.91614.5d.DC1.html
http://www.neurology.org/content/suppl/2009/11/06/01.wnl.000028817
8.91614.5d.DC2.html
References This article cites 28 articles, 5 of which you can access for free at:
http://www.neurology.org/content/70/22_Part_2/2109.full.html##ref-lis
t-1
Citations This article has been cited by 5 HighWire-hosted articles:
http://www.neurology.org/content/70/22_Part_2/2109.full.html##other
articles
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
All Epilepsy/Seizures
http://www.neurology.org//cgi/collection/all_epilepsy_seizures
Antiepileptic drugs
http://www.neurology.org//cgi/collection/antiepileptic_drugs
EEG
http://www.neurology.org//cgi/collection/eeg_
Meningitis
http://www.neurology.org//cgi/collection/meningitis
Status epilepticus
http://www.neurology.org//cgi/collection/status_epilepticus
Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in
its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions
Reprints Information about ordering reprints can be found online:
http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright . All rights reserved. Print ISSN: 0028-3878.
Online ISSN: 1526-632X.