Professional Documents
Culture Documents
CURRENT
OPINION Epidemiology, diagnosis, and treatment of
brain abscesses
Matthijs C. Brouwer and Diederik van de Beek
Purpose of review
This article describes the epidemiology, diagnosis, and treatment of brain abscesses focusing on studies
published in the past 2 years.
Recent findings
In the recent literature, advances have been made in describing the disease, ancillary investigations, and
treatment, mostly by combining previously available literature in meta-analyses. These studies identified
Staphylococcus and Streptococcus species as the most frequent cause of brain abscess. New developments
include the analysis of genetic risk factors for brain abscess, evaluation of shorter antibiotic courses, and
the use of hyperbaric oxygen treatment. However, many studies in this field are limited by methodology
and results are less helpful for clinical practice. Nevertheless, there has been a gradual improvement in the
outcome of patients with brain abscess over the past 50 years, which might be driven by improved brain
imaging techniques, minimally invasive neurosurgical procedures, and protocoled antibiotic treatment.
Multicenter prospective studies and randomized clinical trials are needed to further advance treatment and
prognosis in brain abscess patients.
Summary
Our understanding of brain abscesses has increased by meta-analysis on clinical characteristics,
ancillary investigations, and treatment modalities. Prognosis has improved over time, likely due to
improved brain imaging techniques, minimally invasive neurosurgical procedures, and protocoled
antibiotic treatment.
Keywords
antibiotic treatment, brain abscess, magnetic resonance imaging, neurosurgery
INTRODUCTION EPIDEMIOLOGY
Brain abscesses are focal infections of the brain that Few studies [3,4] on the incidence of brain abscess
can present with a wide variety of symptoms and representing population-based estimates have been
signs, depending on the number, location, and size published. A study [3] from Minnesota, USA showed
&& &&
of the abscess [1 ,2 ]. Brain abscess can be caused a declining incidence from 2.7 to 0.9 per 100 000
by diverse micro-organisms including bacteria, population between 1935 and 1981. A study [6] from
mycobacteria, protozoa, helminths, and fungi. Copenhagen, Denmark, performed between 1994
The incidence has previously been estimated at and 2009, showed an incidence of 0.4 per
0.3 and 0.9 per 100 000 population, but accurate 100 000. More recently an overview study [7] from
figures are not available for large parts of the world Finland, describing patients from a tertiary referral
[3,4]. In immunocompromised patients due to HIV,
use of immunosuppressant or organ transplan-
Department of Neurology, Centre of Infection and Immunity Amsterdam
tation, the incidence is thought to be substantially
(CINIMA), Academic Medical Centre, Amsterdam, the Netherlands
higher [4,5]. In these patients, the causative organ-
Correspondence to Dr Matthijs C. Brouwer, MD, PhD, Department of
isms are more likely to be opportunistic, and empiric Neurology, Centre of Infection and Immunity Amsterdam (CINIMA),
treatment and ancillary investigations should take Academic Medical Centre, University of Amsterdam, PO Box 22660,
this into account. 1100DD Amsterdam, the Netherlands. Tel: +31 20 566 4042;
In this review, we describe the recent publi- fax: +31 20 566 9374; e-mail: m.c.brouwer@amc.uva.nl
cations on epidemiology, diagnosis, and treatment Curr Opin Infect Dis 2017, 30:129–134
of brain abscess. DOI:10.1097/QCO.0000000000000334
0951-7375 Copyright ß 2017 Wolters Kluwer Health, Inc. All rights reserved. www.co-infectiousdiseases.com
&&
stable over the past 60 years [2 ]. symptoms together are identified in only 20% of
&&
patients [2 ]. Patients may however also present
with only progressive changes in behavior or cog-
RISK FACTORS nitive defects, without focal neurologic deficits or
The majority of patients with brain abscesses have fever (Fig. 1). Therefore, the identification of a brain
predisposing conditions, which may provide clues abscess on cranial imaging may sometimes be an
&&
toward the causative micro-organism [1 ,4]. Con- unexpected event because of the low initial clinical
tiguous spread of bacteria occurs in half of the cases suspicion. Cranial imaging is crucial in the diagnosis
and can result from penetrating trauma, neurosur- of brain abscesses and should be performed in all
gery, or infections of the ear and paranasal sinuses &&
patients in whom it is suspected [1 ,9]. Computed
&& &&
[1 ,2 ]. The pathogens associated with contiguous tomography (CT) with intravenous contrast
spread are Staphylococcus aureus, anaerobic and aer- enhancement can be used for determining the size
obic Streptococcus species, but also enterobacteria, and number of abscesses but cannot accurately dis-
Klebsiella, Proteus, and Salmonella species have been criminate between metastases or primary brain
&&
described [1 ]. Hematogeneous spread of bacteria tumors and brain abscesses. Magnetic resonance
can occur in patients with endocarditis, congenital imaging (MRI) however, is a valuable diagnostic tool
heart disease, lung infections, or dental infections in differentiating brain abscess from primary, cystic,
&&
and is identified in one-third of patients [2 ]. Brain or necrotic tumors using diffusion-weighted and
abscesses associated with endocarditis are typically apparent-diffusion coefficient images [9]. A prospec-
caused by S. aureus and Streptococcus spp., whereas tive study [10] in 115 patients with 147 space
(a) (b)
FIGURE 1. Brain MRI scans showing large abscess in patient presenting with an isolated change in behavior. Axial T2-
weighted (a) and gadolinium enhanced T1-weighted (b) MRI image of large abscess in the right frontal lobe. MRI, magnetic
resonance imaging.
occupying lesions of the brain including 97 patients brain shift caused by the puncture of which 5%
with brain abscesses showed that diffusion- died [14].
weighted imaging (DWI) had a 96% sensitivity Identification of underlying infectious foci that
and specificity for the differentiation of brain disseminated to the brain, resulting in brain abscess
abscesses from tumors. A 2014 meta-analysis of 11 formation, should be a priority in all patients in
studies including 504 patients with 519 abscesses on whom the cause is not evident. This includes echo-
the value of DWI in brain abscess discrimination cardiography to rule out endocarditis, chest X-ray to
showed a combined sensitivity of 95% (95% confi- identify lung infections, evaluation of ear, nose, and
dence interval 0.87–0.98) and specificity of 95% throat foci of infection (sinusitis, mastoiditis), and
& &&
(95% confidence interval 0.88–0.97) [11 ]. Several the teeth [1 ]. In selected patients, fludeoxyglucose-
studies [12,13] assessed novel MRI techniques PET imaging may be indicated to identify distant
including magnetic resonance spectroscopy and foci of infection [15]. Tooth extraction, cardiac valve
cerebral blood volume measurement in the differ- replacement, or sinus surgery may be indicated to
entiation between brain abscesses and malignan- remove the focus of infection and prevent further
cies. However, it is unclear whether these spread of the bacteria [16].
parameters perform better than common MRI with Identification of the pathogen is crucial to deter-
DWI. Cranial imaging shows multiple abscesses in mine antibiotic susceptibility patterns and tailor
21% of cases recently summarized in the meta- antibiotic treatment. To achieve pathogen identifi-
&&
analysis of cohort studies (Fig. 2) [2 ]. cation neurosurgery is essential. With the use of
Other diagnostic methods used in brain abscess volumetric CT or MRI combined with stereotactic
patients can provide evidence of an infection in navigation a trajectory can be planned to aspirate
&&
general, such as erythrocyte sedimentation rate, the purulent center of the abscess [1 ,17]. Aspira-
blood leukocytosis or elevation in CRP, or of con- tion of the abscess for diagnostic purposes should be
comitant bacterial meningitis by means of cerebro- performed unless it is contraindicated because of the
&&
spinal fluid examination [2 ]. However, the blood suspected pathogen or the clinical condition of the
&&
parameters of infection are within the normal patient [1 ]. For instance, in HIV-infected patients
range in 30–40% and normal cerebrospinal fluid with suspected toxoplasmosis, presumptive therapy
leukocyte count is observed in 30%. Therefore, is justified when tests for antitoxoplasma immuno-
these cannot be used to rule out that a lesion in globulin G antibodies are positive without further
the brain is infectious. A lumbar puncture should tissue-based diagnosis. When a single pathogen has
be performed with caution as in one large case been cultured from blood or cerebrospinal fluid,
series 19% of patients deteriorated due to increasing stereotactic aspiration of the abscess for diagnostic
0951-7375 Copyright ß 2017 Wolters Kluwer Health, Inc. All rights reserved. www.co-infectiousdiseases.com 131
(a) (b)
FIGURE 2. Examples of brain MRI scans showing multiple abscesses due to Streptococcus milleri (a) and Nocardia farcinica
(b). Axial gadolinium enhanced T1-weighted (a) and fluid-attenuated inversion recovery weighted MRI (b) showing multiple
abscesses, with perifocal edema. MRI, magnetic resonance imaging.
&&
purposes may still be considered if polymicrobial patients [2 ]. In immunosuppressed patients such
infection is suspected, for instance when a patient as transplant recipients and HIV-infected patients,
does not respond to antibiotic treatment. Diagnos- this needs to be supplemented with voriconazole,
tic aspiration should always be aimed at achieving and trimethoprim-sulfamethoxazole or sulfadiazine
maximal drainage to speed up recovery and some to cover fungi, yeasts, and toxoplasmosis while
&&
suggest continuous drainage of the abscess post- awaiting further diagnostics [1 ]. A total of 16%
operatively through a catheter in the abscess may of patients included in the meta-analysis received
decrease the reoperation rate [18]. Few studies, how- only antibiotic treatment without neurosurgical
&&
ever, are available to assess the benefit of this aspiration of the abscess [2 ]. The advised duration
approach. of intravenous antimicrobial therapy in bacterial
&&
The microbiological investigation of the aspirate brain abscess patients is 6–8 weeks [1 ]. Observa-
of the abscess and, when acquired, cerebrospinal tional data have been published on shorter
fluid should include a Gram stain and both aerobic durations of treatment, including a recent retrospec-
&&
and anaerobic cultures [1 ]. In immunocompro- tive study of 55 patients [19]. The authors used
mised patients or those with specific risk factors thermal curve and MRI DWI imaging to determine
such as a history of pulmonary tuberculosis or when to stop antibiotic treatment [19]. This resulted
opportunistic infection, smears and cultures should in cessation of antibiotics after an average of 21 days
be obtained for mycobacteria, fungi, and Nocardia (range 10–64) after which abscesses recurred in two
species, and a PCR for T. gondii should be performed. patients and 66% of patients had a good outcome.
The British Society for Antimicrobial Chemotherapy
recommends just 1–2 weeks of intravenous therapy
TREATMENT for patients with bacterial brain abscess and when
Because of the wide range of potential pathogen the patient has a good clinical response, a change to
brain abscesses, both Gram positive and Gram nega- an appropriate oral regimen can be considered [20].
tive bacteria should be covered with the empiric Oral treatment regimens can consist of ciprofloxa-
treatment. Therefore, all patients need to receive cin, metronidazole, and amoxicillin [21]. Retrospec-
an extended-spectrum cephalosporin (cefotaxime tive studies showed that this approach can be
or ceftriaxone) combined with metronidazole. The successfully applied in selected patients, but it
meta-analysis of cohort studies showed that this should not be considered as standard therapy [22].
regimen was used as empiric treatment in the Prolongation of antibiotic treatment is generally
majority (53%) of patients, while all other combi- advised as long as the abscess cavity is visible on
nations of treatment were used in less than 10% of cranial MRI.
0951-7375 Copyright ß 2017 Wolters Kluwer Health, Inc. All rights reserved. www.co-infectiousdiseases.com 133