Article Text

Download PDFPDF

Predictors and long-term outcome of seizures after bacterial brain abscess
  1. Ming-Jung Chuang1,
  2. Wen-Neng Chang2,
  3. Hsueh-Wen Chang3,
  4. Wei-Che Lin4,
  5. Nai-Wen Tsai2,
  6. Mei-Jen Hsieh2,
  7. Hung-Chen Wang1,
  8. Cheng-Hsien Lu2
  1. 1Department of Neurosurgery, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan
  2. 2Department of Neurology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan
  3. 3Department of Biological Science, National Sun Yat-Sen University, Kaohsiung, Taiwan
  4. 4Department of Radiology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung, Taiwan
  1. Correspondence to Cheng-Hsien Lu, Department of Neurology, Chang Gung Memorial Hospital, 123 Ta Pei Road, Niao Sung Hsiang, Kaohsiung Hsien, Taiwan; chlu99{at}ms44.url.com.tw

Abstract

Background Seizures are one of the most important neurological complications of bacterial brain abscesses. A better understanding of the risk factors of seizures following bacterial brain abscesses is needed to predict those who will require treatment.

Methods A total of 205 patients were enrolled in this 22-year retrospective study. Prognostic variables were analysed based on Cox's proportional hazards model after a minimum of 18 months of follow-up.

Results Seizures occurred in 48 patients who had bacterial brain abscesses, including acute symptomatic seizures in 17% (35/205) and unprovoked seizures in 6.4% (13/205). Altogether, 27 patients had early seizures and 21 had late seizures. The overall mortality rate in the seizure patients was 23% (11/48) and seven patients progressed to epilepsy.

Conclusion Cox's proportional hazards model demonstrated that valvular heart diseases as the underlying diseases and the presence of a fronto-parietal distribution of bacterial brain abscess were independently predictive of seizures, and the presence of late seizures was predictive of developing epilepsy. Most first seizures occurred within 3 y after bacterial brain abscesses.

  • Outcome
  • risk factors
  • seizures
  • bacterial brain abscess
  • bacteriology
  • epilepsy
  • microbiology

Statistics from Altmetric.com

Introduction

Despite the advent of modern neurosurgical techniques, new antibiotics and new powerful imaging technologies, brain abscess remains a potentially fatal central nervous system (CNS) infection.1 2 Seizures are an important neurological complication of bacterial brain abscesses and may be an early or late complication of bacterial brain abscesses.2–5 Several researchers have examined the clinical features of seizures following bacterial brain abscesses.3–8 The majority of previous studies have had a relatively small number of patients,6–8 less strict selection criteria and included seizures that occurred only in the acute phase of bacterial brain abscesses or with a shorter follow-up period.2 6–8 The reported incidence of seizures varies, ranging from 9–70% in different series.2–4 6–8

Indications and duration of seizure treatment required after bacterial brain abscesses remain controversial.3 4 One early study demonstrated that the incidence of subsequent seizures after brain abscesses approached 70%,4 whereas another study demonstrated that if patients remained seizure-free at discharge, the chance of developing seizures was relatively small.3 Furthermore, some physicians are concerned about the potential adverse effects of anti-epileptic drugs (AEDs) and choose to treat seizures symptomatically.

This hospital-based study may provide accurate information about the relative frequency of seizures subtypes, their effect on fatalities, their effect on neurological and functional outcomes, and the relation to underlying cerebral pathological lesions. Because of the possible benefits of AEDs to reduce the degree of functional morbidity after seizures following bacterial brain abscesses, there is a need for better delineation of potential prognostic factors and outcomes in patients with bacterial brain abscesses admitted to hospital who should receive treatment. In this study, we attempt to analyse clinical features, neuroimaging findings, clinical scores and measurements to determine potential risk factors that are predictive of seizures following bacterial brain abscesses.

Materials and methods

Study population

We retrospectively reviewed the microbiological records for abscesses and blood cultures, medical records and neuroimaging findings using pre-existing standardised evaluation forms, for patients with bacterial brain abscesses admitted to Kaohsiung Chang Gung Memorial Hospital between 1986 and 2007. Kaohsiung Chang Gung Memorial Hospital, the largest medical centre in southern Taiwan, is a 2482-bed acute-care teaching hospital, which provides both primary and tertiary referral care of patients.

Diagnostic criteria of bacterial brain abscess

The criteria of bacterial brain abscess for inclusion were: (1) characteristic CT and/or MRI findings; (2) evidence of a brain abscess seen during surgery or histopathological examination; and (3) classical clinical manifestations, including headache, fever, localised neurological signs and/or consciousness disturbance.1 2 Brain abscesses were defined as nosocomial according to the 1988 guidelines of the Centers for Disease Control.9 Brain abscesses related to head trauma with skull fractures or neurosurgical procedures were classified as a post-neurosurgical form. Otherwise, patients who presented with no distinctive characteristics, and/or who had not undergone invasive procedures, were classified as having the spontaneous form.

Exclusion criteria

Patients were excluded if they (1) had a history of seizures; (2) had pre-existing neurological conditions with various neurological deficits (eg, stroke, head trauma and hypoxic encephalopathy) and (3) regularly took AEDs for epilepsy or other clinical indications (eg, trigeminal neuralgia or neuropathic pain). Thus, only 205 of the 210 patients were enrolled for analysis. The study protocol was approved by Chang Gung Memorial Hospital's Institutional Review Committees on Human Research.

Definition of seizures

Seizures were classified according to the recommendations of the International League Against Epilepsy.10 Status epilepticus (SE) was defined as more or less continuous behavioral seizures activity or repetitive seizures without full recovery of neurological function between seizures occurring over a period of more than 30 min.11 The definition of seizures was set according to those in previous studies: seizures occurring after a bacterial brain abscess are causally related to the bacterial brain abscess itself. A provoked seizure (acute symptomatic seizures) is one that occurs in close temporal relation with a bacterial brain abscess, which is the presumed aetiology. In contrast, an unprovoked seizure is a seizure occurring in the absence of one or more precipitating factors and includes events occurring in patients with antecedent stable (non-progressing) bacterial brain abscesses. Epilepsy is the occurrence of repeated unprovoked seizures.10 Based on seizures onset in relation to the clinical ictus of bacterial brain abscesses, patients who had seizures were divided into two subtypes: early seizures were those occurring within 2 weeks of the infections whereas late seizures were those occurring after 2 weeks.

Study protocol

All of the patients received brain CT scans soon after arrival at the emergency room. Follow-up brain CT scans and/or MRI were performed if clinical deterioration was noted, including acute onset of focal neurological deficits, seizures or status epilepticus, and a progressively disturbed consciousness, as well as post-neurosurgical procedures. Hydrocephalus was judged retrospectively by a dilated temporal horn of the ventricle without obvious brain atrophy and/or an Evan's ratio of >0.3 on initial CT scans. The Evan's ratio is the ratio of the ventricular width of the bilateral frontal horn to the maximum biparietal diameter. The ‘volume of brain abscess’ indicates the summation of all the volumes of the brain abscesses if at least two brain abscesses were found.

A combination of surgical intervention and antibiotic treatment were the mainstays of our treatment of bacterial brain abscesses. Our standard protocol was to administer AEDs only to those with acute symptomatic seizures in bacterial brain abscesses, whereas prophylactic AED treatments were not given to those without such symptoms in the acute stage of bacterial brain abscesses. AEDs were usually administered to patients who had acute symptomatic seizures during admission to hospital and were discontinued if there was no development of unprovoked seizures during follow-up at the out-patient department.

Evaluation of therapeutic outcomes after discharge used the Glasgow Outcome Score (GOS). Therapeutic outcomes and the degree of seizures control after discharge were determined by a Seizure Frequency Scoring System.12 The follow-up period was terminated by death or by the end of the study itself (December 2007). Most patients were followed-up at the out-patient department after discharge, while others were interviewed by telephone to identify neurological outcomes.

Statistical analysis

Two separate statistical analyses were performed. First, the demographic data between good and poor outcome groups were compared. Categorical variables were compared using the χ2 test or Fischer's exact test. Continuous variables within two groups were compared using the independent t test for parametric data and the Mann–Whitney U test for non-parametric data. Second, any imbalances between the seizure and non-seizure groups in baseline prognostic variables were considered. Analyses were repeated with adjustments based on Cox's proportional hazards model after a minimum of 18 months of follow-up. All statistical tests were two-tailed. Third, the association between the presence of late seizures and surgical intervention for bacterial brain abscesses, and the survival curve between the two patient groups (epilepsy and non-epilepsy) were assessed with Kaplan–Meier plots and compared by a log-rank test. All statistical analyses were conducted using SAS v13.0.

Results

Baseline data of study patients

The 205 patients with bacterial brain abscesses included 151 males (mean age 46.2 y) and 54 females (mean age 45.5 y). In this study, seizures occurred in 48 of the 205 patients with bacterial brain abscesses. The time intervals from the first clinical sign of brain abscess to first seizure ranged from 0–755 days (mean 116.4 days). The characteristics of the 205 patients with bacterial brain abscesses in terms of seizures are listed in table 1.

Table 1

Characteristics of 205 patients with bacterial brain abscesses in terms of seizures

Causative pathogens and neuroimaging findings

The causative pathogens between patients with or without seizures are listed in table 2. Gram-negative bacilli were the most common causative pathogens followed by Streptococcus species, anaerobic pathogens and Staphylococcus species in both patients with and without seizures. The interval between onset of symptoms to detection of a brain abscess ranged from 0–169 days (mean 13 days).

Table 2

Causative pathogens of 205 patients with bacterial brain abscesses

The locations of the bacterial brain abscesses in patients with or without seizures are listed in table 3. The most common sites for brain abscesses in the seizure patients were the frontal lobe followed by the fronto-parietal lobe, fronto-parieto-temporal lobe and basal ganglion. Other neuroimaging findings, including hydrocephalus, rupture into the ventricle, multiloculated character and the median (interquartile range (IQR)) volume of bacterial brain abscesses on admission are also listed in table 3.

Table 3

Neuroimaging findings

Clinical characteristics and outcome of seizures

Among the 48 cases with seizures, 35 were acute symptomatic seizures while the other 13 were unprovoked seizures. Regarding seizure onset in relation to the clinical ictus of bacterial brain abscesses, 27 had early seizures and 21 had late seizures (figure 1 and table 1). The mean time intervals from bacterial brain abscess to first seizure in the early and late seizure groups were 2.3 days, and 165.5 days, respectively. Regarding seizure subtypes among the 48 cases, 20 were generalised tonic–clonic, 19 were focal seizures with secondarily generalised seizures, 3 were myoclonic seizures, 4 were simple partial seizures and 2 were complex partial seizures. One progressed to status epilepticus. In total, 11 died of acute phase bacterial brain abscesses, while 37 survived. Regarding the degree of seizure control after discharge among the 37 survivors, 20 were seizure-free without AED treatment, 10 were seizure-free under AED treatment, 1 had non-disabling simple partial seizures, 3 had 1–3 seizures per year and the remaining 3 had 4–11 seizures per year. Of the 48 seizure patients, 32 were treated with phenytoin alone, 8 with vaproic acid, 2 with carbamazepine, 1 with both vaproic acid and levetiracetam, 1 with both vaproic acid and topiramate, 1 with both phenytoin and oxcarbazepine, and the remaining 3 with gabapentin. Four patients had phenytoin -related skin rash.

Figure 1

The frequency distribution of patients who had first seizures after bacterial brain abscesses during study periods (months).

There were 14 lost follow-up cases during the study period. The treatment of the bacterial brain abscesses between those patients with or without seizures is listed in table 1. In total, 162 of 205 patients survived. The death rate in patients with or without seizures was 23% (11/48) and 20% (32/157), respectively. Therapeutic outcomes after a minimum of 18 months of follow-up among the 37 survivors were determined by GOS as 87% having a normal life, 7% having moderate disabilities, 4% having persistent vegetative states and 2% deceased. The mean age at onset, median days of hospital admission and median GOS at discharge between the patients with or without seizures is listed in table 4.

Table 4

Comparisons of baseline clinical features and neuroimaging findings between patients with or without seizures following bacterial brain abscesses

Risk factors of seizures

To look at each factor individually, we calculated Kaplan–Meier estimates of the fraction developing epilepsy by different times for each subgroup and tested for differences by using a log-rank test. The presence of late seizures and surgical intervention or not for bacterial brain abscesses between the two patient groups (epilepsy and non-epilepsy) showed the following: presence of late seizures (p≤0.0001) and surgical intervention or not (p=0.088).

Comparisons of clinical features and neuroimaging findings between patients with and those without seizures following bacterial brain abscesses after a minimum of 18 months of follow-up are listed in table 1. Statistical analysis between the two patient groups revealed valvular heart diseases (p=0.012) and chronic otitis media (p=0.027) as the underlying diseases, the aetiologies of the bacterial brain abscesses (p=0.045) and presence of a fronto-parietal distribution of bacterial brain abscesses (p=0.005) were significant variables. Variables used in the Cox's proportional hazards model included valvular heart diseases as the underlying diseases, the aetiologies of bacterial brain abscesses and the presence of a fronto-parietal distribution of bacterial brain abscesses. After analysis of the aforementioned variables, only valvular heart diseases as the underlying diseases (p=0.033; OR 3.36, 95% CI 1.10 to 10.25) and presence of a fronto-parietal distribution of bacterial brain abscesses (p=0.028; OR 2.58, 95% CI 1.11 to 6.03) were independently associated with seizures.

Discussion

Although the frequency of seizures after bacterial brain abscesses is variously estimated at 12–86%, much of these data are based on retrospective studies with variable follow-up times.3 4 13–19 In these studies, most of the patients included were treated in the pre-CT era, and the seizures were reported to have manifested within 1 y after discharge in 86%,18 within 2 y in 78%16 or in 41% within 1 y and in 76% within 4 y4; however, intervals of more than 10 y have also been observed. By comparison, several recent studies, though with fewer patients and a much shorter follow-up period, have reported seizures in 10–16% of the survivors.13–15 In our study, seizures occurred in 48 out of 205 patients who had bacterial brain abscesses (23.4%), including acute symptomatic seizures in 17% (35/205) and unprovoked seizures in 6.4% (13/205). Among the 37 surviving seizure patients, 7 progressed to epilepsy. The incidence of epileptic seizures here is lower than in other reported series,4 16 18 which may be attributed to the exclusion criteria that excluded the high-risk group of seizure patients (eg, pre-existing neurological deficits and chronic epilepsy), the advent of modern neurosurgical techniques and treatment.

Study major findings

The present study examined predictive factors and outcomes of seizures after bacterial brain abscesses and produced three major findings. First, valvular heart diseases as the underlying disease and the presence of a fronto-parietal distribution of bacterial brain abscesses were independently predictive of seizures. Most cerebral emboli arise from the heart (eg, atrial fibrillation, valvular heart diseases with cardiac emboli, infectious endocarditis). Although emboli have access to the entire cerebral arterial tree, but branches of the middle cerebral arteries (MCA) are the most frequently involved. The MCA supplies most of the convexity of hemisphere except for its dorsal margin and the occipital pole. Embolism is inferred when the brain abscess is confined to either the cerebral surface territory of a single branch or a combination involving branches of different divisions of the MCA, which is responsible for those regions of frontal, temporal and parietal lobes. Second, the presence of late seizures was predictive of developing epilepsy. Third, seizures in bacterial brain abscess patients may have a delayed manifestation although most seizures occur during the acute phase of bacterial brain abscesses. None of the patients had the first seizure occurring more than 3 y following bacterial brain abscesses.

Study limitations

First, this was a retrospective analysis and, therefore, subject to bias of unmeasured factors (eg, a possible reporting bias due to patient selection presented to the hospital and a possible high association of ‘late symptomatic seizures’ with ‘spontaneous seizures’). It was also not possible to assess the effect of prophylactic AEDs after the acute stage of bacterial brain abscesses to prevent later epilepsy or draw conclusions. Second, patients who took AEDs for epilepsy or other clinical indications (eg, trigeminal neuralgia or neuropathic pain) or had pre-existing neurological deficits were excluded. Thus, continued uncertainty was present in assessing the incidence of unprovoked seizures after bacterial brain abscesses in non-selected patients. Third, the impact of different surgical methods on the frequency of seizures is not clear. Some studies have observed a reduced frequency of seizures after blurred-hole aspiration alone compared to craniotomy and excision of brain abscesses,16 17 whereas other studies have not.4 13 However, even though surgical interventions may cause potential brain insults and result in the possibility of unprovoked seizures during follow-up periods, the association between surgical intervention and presence of epilepsy was not statistically significant in our study. Finally, most patients in this study were treated with AEDs after their first acute symptomatic seizure in accordance with our study protocols. Thus, our findings might underestimate the ‘true’ frequency of seizures associated with the ‘natural history’ of untreated unprovoked seizures.

Several pathophysiological mechanisms, though requiring further elucidation, are implicated in the occurrence of seizures after bacterial brain abscesses, including a combination of the sudden development of a space-occupying lesion with mass effect, striking oedema and a surrounding zone of prominent perivascular inflammatory response, which might possibly account for seizures in the early phase of brain abscesses.20 Our study reveals that neither the volume of the brain abscess present in brain parenchyma (ie, lesion size) nor rupture into cisternal and ventricular spaces seem to be predictive of the occurrence of seizures.

The use of prophylactic AED treatment in the prevention of seizures in patients with bacterial brain abscesses remains unclear, although the possible benefit of AEDs is the reduction of functional morbidity after seizures following bacterial brain abscesses.1 21 Some authors have proposed seizures prophylaxis for all patients; the patients should then be re-evaluated by neurological and electroencephalography (EEG) examinations several months after the treatment of the brain abscess and the duration of AED treatment is dependent on the EEG results.1 To our knowledge, no formal recommendations can be made about the use of prophylactic AEDs in patients with bacterial brain abscesses based on currently available well-designed randomised controlled trials. Phenytoin was the most widely used AED in our study and 8% (4/48) of our cases had adverse effects after phenytoin use.

Conclusion

Valvular heart diseases as the underlying diseases and the presence of a fronto-parietal distribution of bacterial brain abscesses were independently predictive of seizures and the presence of late seizures was predictive of epilepsy. Our study also demonstrates that most first seizures occur within 3 y; therefore, prophylactic treatment, if considered, should not be performed for a period of more than 3 y.

References

Footnotes

  • Competing interests None.

  • Ethics approval The study was approved by Chang Gung Memorial Hospital's Institutional Review Committee on Human Research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.