Review

The Neurohospitalist
2014, Vol. 4(4) 196-204
Bacterial Brain Abscess The Author(s) 2014
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1941874414540684
nho.sagepub.com

Kevin Patel, MD1, and David B. Clifford, MD2

Abstract
Significant advances in the diagnosis and management of bacterial brain abscess over the past several decades have improved the
expected outcome of a disease once regarded as invariably fatal. Despite this, intraparenchymal abscess continues to present a
serious and potentially life-threatening condition. Brain abscess may result from traumatic brain injury, prior neurosurgical pro-
cedure, contiguous spread from a local source, or hematogenous spread of a systemic infection. In a significant proportion of
cases, an etiology cannot be identified. Clinical presentation is highly variable and routine laboratory testing lacks sensitivity.
As such, a high degree of clinical suspicion is necessary for prompt diagnosis and intervention. Computed tomography and mag-
netic resonance imaging offer a timely and sensitive method of assessing for abscess. Appearance of abscess on routine imaging
lacks specificity and will not spare biopsy in cases where the clinical context does not unequivocally indicate infectious etiology.
Current work with advanced imaging modalities may yield more accurate methods of differentiation of mass lesions in the brain.
Management of abscess demands a multimodal approach. Surgical intervention and medical therapy are necessary in most cases.
Prognosis of brain abscess has improved significantly in the recent decades although close follow-up is required, given the poten-
tial for long-term sequelae and a risk of recurrence.

Keywords
abscess, bacteria, fungi, imaging, infection, brain

Introduction epidemiology, diagnosis, treatment, sequelae, and ultimate

outcome of this uncommon condition.
Brain abscess is a serious and life-threatening clinical entity.
Pyogenic infection of brain parenchyma begins with a loca-
lized area of inflammatory change referred to as cerebritis. Epidemiology
This early stage of infection is characterized by increased
blood vessel permeability without angiogenesis. When unrec- The incidence of brain abscess has decreased in the United
ognized, this process will progress to an immature capsular States through the 20th century. A single-center study from
stage and then to brain abscess, a condition defined by an area Olmstead County Minnesota showed a change in incidence
of parenchymal infection containing pus encapsulated by a of 2.7 cases per 100 000 in 1935 to 1944 to 0.9 in 1965 to
vascularized membrane. 1981.1 More recent investigations of incidence of abscess are
Intraparenchymal abscess is the result of spread from a lacking. Brain abscess may occur at any age, but the majority
contiguous focus of infection, hematogenous seeding from of cases occur between the third and fifth decades of life.
a distant source, the sequelae of head trauma or neurosurgical Recent case series demonstrate a significant variance in mean
procedure. In some instances, an etiology cannot be identified. age at presentation ranging from 24 to 57 years.2-12 Mean age
Diagnosis can be challenging, as abscess presentation is of onset is higher in series from centers in developed nations
highly variable and routine studies frequently lack specificity. than in developing nations. This relationship has not been
High clinical suspicion is necessary for early recognition and
initiation of appropriate treatment. In recent years, there have 1
Department of Neurology, Washington University in St Louis, St Louis, MO,
been significant refinements of neurosurgical technique and USA
advances in the development and use of new antibiotics. 2
Departments of Neurology and Medicine, Washington University in St Louis,
Despite this, there is controversy regarding the role of surgical St Louis, MO, USA
management and the optimal use of medications. The out-
Corresponding Author:
comes associated with brain abscess are guarded but have David B Clifford, Department of Neurology, Washington University in St
improved remarkably over the prior several decades. We pres- Louis, Campus Box 8111, 660 S. Euclid Avenue, St Louis, MO 63110, USA.
ent a review of the literature of the clinically relevant Email: cliffordd@neuro.wustl.edu
Patel and Clifford 197

formally investigated but may reflect the greater proportion of fungal brain abscess mostly related to Aspergillus although
older immunosuppressed individuals, greater life expectancy, other species of fungi including Candida, dematiaceous
and lower rates of traumatic brain injury in younger individu- fungi, or phaeohyphomyctes may produce localized infec-
als in developed areas. There is a significant male predomi- tion as well. Nocardia-related infection is not uncommon
nance reported consistently through the literature which is in solid organ transplant and bone marrow transplant recipi-
independent of geography2,4-6,8-16 with male to female ratios ents and appears most often after cardiac transplant with
ranging from 1.5:1 to 4.5:1.2,4,5,8-15 incidence reported as high as 37.5%. Central nervous system
spread resulting from secondary dissemination from a pri-
mary pulmonary infection may produce single or multiple
Etiology abscesses.21 Tuberculous abscess is uncommon but will
Pathogenesis. Etiology can be identified in many cases, but the occur in about 1% of solid organ transplant recipients.22
source of infection remains unclear in a significant proportion Other bacterial pathogens are uncommon causes of abscess
of those with abscess even after a thorough investigation. In formation in transplant recipients accounting for less than
recent case series, cryptogenic abscess occurs between 4.6% 1% of abscess in liver transplant and bone marrow transplant
and 43.4% of cases.2,3,6-8,11,13,14,16,17 Brain abscesses are often recipients.21 Abscesses must be distinguished from toxo-
attributed to hematogenous spread, contiguous spread, recent plasma encephalitis which is the most common multifocal
neurosurgical procedure, or penetrating head trauma. Endo- infectious process encountered in advanced HIV. Individuals
carditis or pulmonary infections (pneumonia, empyema, and with AIDS having brain abscess are more likely to have mul-
abscess) are the most common sources of hematogenous tiple abscesses and tuberculous abscess.8 This population is
spread 2,6-11,13,14 Cyanotic heart disease and pulmonary arter- also more susceptible to intracranial infection from Listeria,
iovenous malformation are consistently reported in associa- Cryptococcus, and Nocardia.23
tion with brain abscess.6-8,10,11,15,16 The pulmonary
circulation represents a potential filtering apparatus for sys- Microbiology. A single organism is isolated in the majority of
temic bacterial pathogens. In patients with right to left cardiac bacterial brain abscess. However, isolation of multiple
shunt bypass, this mechanism and seeding to the central ner- pathogens from abscess materials is not uncommon (4%-
vous system (CNS) is thought to occur more readily. Further 23%).6-11,14,16,24,25 Cultures are negative in 14%-34% of
some authors hypothesize that ischemic injury from hypoxe- samples.9-11,14,24,25 Administration of antibiotics prior to
mia and polycythemia may act as a nidus for infection.18 collection of abscess material is often cited as the explana-
Infection from any systemic source may lead to bacteremia tion for sterile culture. The authors of 1 case series note that
and subsequent spread to brain parenchyma even in the abscesses drained within 3 days of antibiotic administration
absence of cyanotic heart disease. Rarely, authors report bac- had much greater yield than otherwise (84% and 32%,
terial meningitis as a source of hematogenous spread.3,7 respectively).5 All individuals with positive cultures after
Hematogenous spread accounts for 9% to 43% of brain this period had large abscesses. Aerobic organisms are more
abscesses.2,6,7,9-12,14 Contiguous infection may result from commonly identified than anaerobes. Streptococci are most
primary dental, sinus, ear infections, or mastoiditis 2,3,6-16 and often identified among aerobic pathogens. Bacteroides fragi-
represents 14% to 58% proportion of brain abscess. Invasive lis and Peptostreptococcus spp are the most common anaero-
neurosurgical procedures are a known risk factor for develop- bic organisms isolated.6,8,16,24 Organisms vary significantly
ment of brain abscess. They account for 3% to 18% of brain with the etiology of abscess (Table 1). Otogenic abscess is
abscess.2,3,7,8,11-15 A recent series that compiles infectious most often associated with Proteus, Streptococcus milleri
complications of neurosurgery at a single center demonstrates group organisms, and Streptococcus pneumoniae.6,7,20
that surgical risk of abscess is low (.2%).19 Mixed and negative cultures are more common in otogenic
Abscess occurs most commonly in the frontal lobe but abscess.12,20 Anaerobic organisms are isolated more fre-
may occur in any location.2,4,7,10-15 Location is closely asso- quently in this population.16 Paranasal sinusitis is most often
ciated with source. Otogenic abscess occurs almost exclu- associated with intracranial complications from Streptococ-
sively in the temporal lobe and cerebellum, while abscess cus sp, Staphylococcus sp, and less commonly by the Enter-
associated with sinus infection is predominantly fron- obacteriaceae.7,26 Staphylococcus aureus, Staphylococcus
tal.2,3,8,12,16,20 Abscess has been found to consistently occur epidermidis, and Enterobacteriaceae (with Pseudomonas
more often on the left than on the right.7,12 Several authors aeruginosa being the most common pathogen in this group)
have hypothesized that this is the result of the observed are the most common organisms found in abscess related to
greater relative incidence of left-sided penetrating trauma traumatic brain injury. Similarly, S aureus, S epidermidis,
which in turn has been attributed to the handedness of attack- P aeruginosa, and Propionibacterium acnes are associated
ers. A significant proportion of individuals develop multiple with abscess related to neurosurgical procedure.8,12,13 Sta-
abscesses (9.3%-28%).3,4,6,8-14,16 phylococcus aureus, Streptococcus viridans, and Klebsiella
Immunocompromise raises risk of CNS infection. Solid pneumonaie are the most common organisms isolated in
organ transplantation leads to an increased incidence in abscess attributed to hematogenous seeding.6,8
198 The Neurohospitalist 4(4)

Table 1. Spectrum of organisms differs by anatomic source. Laboratory data. Laboratory data are of limited utility in diag-
nosis. Leukocytosis and elevation in erythrocyte sedimenta-
Source Most Commonly Cultured Organisms
tion rate are common, but absence of these laboratory
Paranasal sinus Streptococcus spp abnormalities will not exclude the diagnosis.3,4,6,11,14,15
infection Staphylococcus spp A significant minority of patients present with leukopenia.
Enterobacteriaceae (especially Hemophilus spp, Blood cultures should be performed early in all patients with
Pseudomonas aerugonisa) suspected abscess, given their relative ease in collection. Their
Otogenic Proteus mirabilis
reported yield is modest (14%-50%), but the potential value
infection Streptococcus milleri group organisms
Streptococcus pneumoniae for identification of the organism is substantial in circum-
Staphylococcus aureus stances where collection of abscess material cannot be per-
Dental infection Streptococcus spp formed promptly or is not advisable due to the associated
Bacteroides fragilis risks.4,6,9,10,13,14
Traumatic brain Staphylococcus aureus Cerebrospinal fluid (CSF) analysis may reveal pleocytosis,
injury Staphylococcus epidermidis elevated protein, and decreased glucose but will be normal in a
Enterobacteriaceae (most commonly P
significant proportion of individuals.3,6,7,10,11,13,14
aerugonisa, Enterobacter spp)
Neurosurgical Staphylococcus aureus Cerebrospinal fluid culture is infrequently positive (0%-
procedure Staphylococcus epidermidis 43%). Lumbar puncture may be complicated by rapid neuro-
Pseudomonas aeruginosa logic deterioration typically due to downward brain herniation
Propionibacterium acnes in as many as 20% of patients with brain abscess.7,9 Routine
Streptococcus spp. CSF collection is often discouraged in circumstances where
Hematogenous Staphylococcus aureus abscess is suspected because of the perception of low yield
spread Streptococcus viridans
and significant risk. This is a controversial topic, however, and
Klebsiella pneumoniae
lumbar puncture may be considered in instances where there is
limited mass effect, and the organism cannot be identified
from an alternative source.
Diagnosis
Clinical presentation. The presenting signs of brain abscess are Imaging. Brain imaging is critical to diagnosis and management
variable and nonspecific. Patients most commonly present and is critical to improving outlook for brain abscess. Typical
with headache (49%-93%), fever (14%-88%), altered mental characteristics of imaging studies are illustrated in Figure 1.
status (33%-70%), focal neurologic symptoms (29%-71%), The early stages of cerebritis are characterized on noncontrast
and nausea and vomiting (26%-71%).2-7,9-16,24 Seizures are computed tomography (CT) by localized hypoattenuation. Con-
less common (2%-49%) and may have either focal or general- trast enhancement is variable in this stage and when present
ized presentation. Rarely, brain abscess will present with sta- may demonstrate a nodular or ring-like pattern.27 This pattern
tus epilepticus (0.3%).7 Neck stiffness and meningismus have remains unchanged or progresses on delayed images performed
been reported (4%-23%).2,3,6,7,11,13-16 These symptoms may 30 to 60 minutes after contrast bolus administration. In the late
suggest a temporal or cerebellar location6 but often occur in cerebritis phase, noncontrast CT again demonstrates an area of
the context of a concomitant meningitis or prior intraventricu- hypoattentuation but contrasted images demonstrate a thick
lar rupture. Papilledema should be sought on physical exami- ring-like or nodular enhancement that persists on delayed
nation but is an insensitive sign (1%-19%).3,7,14,16 The classic images. As a capsule begins to form, a round or ovoid area of
clinical triad of fever, headache, and focal neurologic deficit is hypoattenuation will present with ring enhancement that dissi-
suggestive of abscess, but recent reports indicate that this con- pates on delayed scans.
stellation occurs in a minority of cases (2%-34%).6,10,11,13,14,16 Magnetic resonance (MR) imaging has greater sensitivity
Since most patients with brain abscess present with nonspe- and specificity than CT in identifying pyogenic infection
cific and unclear symptoms, high clinical suspicion is neces- (Figure 1). Early cerebritis is characterized by poorly defined
sary for prompt diagnosis. The diagnosis should be hyperintensity on T2-weighted sequences and hypointensity on
considered in all patients with new progressive headache, T1-weighted sequences. As the infection matures, localized col-
signs of increased intracranial pressure, or gradual onset focal lections of fluid will be apparent on imaging. On T1-weighted
neurologic deficit. Higher clinical suspicion may be necessary imaging, these areas will be hyperintense relative to CSF and
in immunosuppressed individuals, as they are at greater risk hypointense relative to the surrounding white matter. Fluid
and their limited ability to mount an immune response may appears isointense to hyperintense to CSF on T2-weighted
shroud typically associated infectious signs. images. Vasogenic edema appears hypointense on T1 images and
Mean times from symptom onset to presentation at a medical hyperintense on T2 sequences. A smooth contrast-enhancing
center range from 7 to 25 days. Most reports indicate that the time capsule which is isointense to hyperintense to white matter on
to presentation in individual cases is quite variable and may occur T1 images and isointense to hypointense on T2 images will be
hours to more than 60 days from symptom onset.3,4,6,7,10,11,14 evident. Imaging in immunocompromised patients may lack
Patel and Clifford 199

Figure 1. A 56-year-old gentleman with a history of type 2 diabetes mellitus presented with a 5-day history of progressive fatigue, malaise, and
subjective fever. On the day prior to presentation, he developed severe headache and had significant change in mental status. His examination
was notable for confusion and mild right arm weakness. Noncontrast CT (A) demonstrates a left frontal mass at the gray-white junction with
surrounding vasogenic edema. On magnetic resonance imaging (MRI), there is subfalcine herniation. T2 (B) and T1 postcontrast (C and D) maps
demonstrate a heterogeneous ring enhancing, T1-hypointense, T2-hyperintense fluid collection. There is thinning of the periventricular rim.
The central nonenhancing portion demonstrates restricted diffusion (E) and is hypointense on apparent diffusion coefficient (ADC) images (F).

ring-enhancing regions and vasogenic edema may be less promi- proteolysis (cytosolic amino acids). A recent study investi-
nent than in immunocompetent patients. gated the added value of spectroscopy and diffusion weighted
Although characteristic of abscess, the MR constellation of imaging (DWI) to MR imaging in the diagnosis of intracranial
a ring-enhancing lesion with surrounding evidence of vaso- cystic lesions in 50 patients.30 Spectroscopy and DWI raised
genic edema is nonspecific. Diffusion-weighted imaging sensitivity and specificity of MR for abscess from 61.9% and
(DWI) may aid in diagnosis. The central nonenhancing por- 60.9% to 95.2% and 100%, respectively.
tion of an abscess has diffusion restriction, appearing hyperin- Positron emission tomography (PET) investigations using
tense on diffusion weighted sequences and hypointense on the radiotracers 18F-fluoro-2-deoxyglucose and [methyl-
apparent diffusion coefficient maps. Infrequently, diffusion 11C]-L-methionine consistently show increased tracer accu-
restriction may not be present in previously treated abscess. mulation in brain abscess but may show increased uptake in
Diffusion restriction is an uncommon finding in ring- neoplastic lesions as well. A new radiotracer, O-(2-18Ffluoro-
enhancing neoplasms (Table 2).28,29 ethyl)-L-tyrosine, has significant diagnostic value in the eva-
Recently, 1H nuclear MR spectroscopy has been shown to luation of glioma but will show uptake in some cases of
be of value in the differentiation of abscess from cystic tumor. abscess.31 As such currently PET imaging is of limited utility
Spectroscopy allows for the detection of products of bacterial in differentiating the underlying etiology of a ring-enhancing
metabolism (lactate, acetate, and succinate) and neutrophil mass and will not spare biopsy.
200 The Neurohospitalist 4(4)

Table 2. Magnetic Resonance Imaging Features Differentiating Abscess From Neoplasm.

Postcontrast T1 Imaging T2 Imaging Diffusion Imaging

Abscess Ring enhancement Hyperintense center Hyperintense on DWI

Thinning of internal border of capsule Hypointense border Hypointense on ADC
Low-grade glioma Variable solid enhancement Hyperintense Hypointense on DWI
Hyperintense on ADC
High-grade glioma Nonnecrotic lesions Hyperintense Nonnecrotic lesions
Solid enhancement in most Nonnecrotic lesions may Isointense to hyperintense on DWI
A minority will not enhance appear isointense Hypointense to isointense on ADC
Necrotic lesions
Necrotic lesions Hypointense on DWI
Hyperintense on ADC
Very rarely necrotic lesions are
Ring enhancement hyperintense on DWI,
hypointense on ADC
Lymphoma Solid enhancement Isointense to hyperintense Typically hyperintense on DWI
Hypointense on ADC
Metastases Ring or solid enhancement Variable Typically hypointense on DWI
Hyperintense on ADC
Abbreviations: ADC, apparent diffusion coefficient; DWI, diffusion weighted imaging.

Management Surgical management. The ideal method of surgical intervention

The appropriate management of brain abscess typically neces- in brain abscess remains a matter of controversy. En bloc
sitates a multimodal approach including both medical and sur- abscess excision had historically been the approach of choice,
gical therapies. All patients should receive prompt empiric but in the preceding several decades, CT-guided aspiration has
antibiotic coverage. In most cases, surgical drainage of puru- become the more commonly practiced initial intervention. The
lent material is necessary for abscess resolution. Stereotactic improved outcome and prognosis of brain abscess in the CT era
drainage by CT guidance is typically considered the interven- is largely credited to the feasibility and availability of stereotac-
tion of choice, but en bloc excision may be considered as ini- tic drainage. Despite this, excision has several advantages and a
tial therapy in particular circumstances. The choice of surgery minority of authors still prefer this method at the outset.
remains a matter of significant debate. Computed tomography-guided aspiration is simpler than
Although most patients will require surgical intervention, in open excision and spares the patient the morbidity associated
a carefully selected subset of patients, a purely medical with the complications of extensive surgical trauma.32 It is
approach may be attempted. Patients with intact mental status favored in abscess in eloquent locations or in deep-seated
and without signs of increased intracranial pressure in whom abscess.32,35 It is preferred also in patients with multiple
there is little doubt of diagnosis32 may be considered. This strat- abscesses necessitating drainage. In cases where there is signif-
egy is rarely effective for abscesses larger than 3 cm in dia- icant uncertainty in diagnosis, abscess wall biopsy is possible
meter.33 Most authors advocate a more conservative threshold through a limited approach and may yield important informa-
of 2.5 cm.34 Further, this approach should not be considered tion.35 Abscess recurrence or failure to improve is not uncom-
in high-risk patients such as those in which structural factors mon after aspiration; this constitutes the major limitation to this
make prompt evacuation necessary such as infratentorial approach. To ensure abscess resolution, weekly imaging is rec-
abscess, developing hydrocephalus, or significant mass effect. ommended after aspiration. Patients failing to improve after ini-
A surgical procedure may be considered when an organism tial aspiration may require repeat procedures or may ultimately
cannot be identified from an alternative source even in the demand excision. Inadequate aspiration, chronic immunosup-
absence of other indications for intervention.10,32 As abscess pression, and inadequate antibiotic therapy are the clinical fac-
material culture is frequently unrevealing, and it is rare that tors most commonly associated with failure.35
culture leads to the discovery of a resistant organism, these Open craniotomy with excision is associated with a lower
decisions should take into account surgical risk and be made rate of recurrence and reaccumulation.32,36 In situations where
on a case-by-case basis. a structural abnormality underlies the development of abscess,
Computed tomography should be repeated on a weekly basis open craniotomy is required for definitive treatment. For
in medically managed patients for at least 2 weeks and more instance, this applies in cases in which the abscess has resulted
often if there is clinical deterioration. Active growth of the from a contiguous primary source (eg, osteomyelitis from
abscess, clinical deterioration, or a lack of improvement in size sinus infection), a fistulous connection, or retained foreign
on imaging at 3 to 4 weeks signify medical failure. Surgical man- bodies following trauma.17,32,36 Similarly, aspiration is often
agement should not be further delayed in these instances.17,35 insufficient in the treatment of multiloculated abscess.35,36
Patel and Clifford 201

Abscess in high-risk locations such as the posterior fossa or are indicated, given the uncertainty of required duration of
with significant clinical consequence from mass effect bene- therapy in any particular case.
fits from open procedure as well.32,35,36 Certain microbiologi- Corticosteroids are not routinely used in brain abscess and
cal factors may necessitate an open procedure. Gas-containing should be reserved for cases in which abscess-related edema
abscesses, actinomycotic abscesses, nocardial abscesses, and is severe and has led to significant clinical mass effect. In early
fungal abscesses respond less well to aspiration and require animal experiments, antibiotic concentration within abscess
open intervention.17,34,36-38 artificially produced by inoculation of brain tissue was found
Even in the absence of these factors, some authors favor open to be significantly and consistently lower in those animals
excision to aspiration. Excision allows for the thorough irriga- receiving corticosteroids.41,42 As such corticosteroid adminis-
tion and removal of purulent material as well as the subsequent tration is thought to lead to a reduction in the penetration of
verification of complete evacuation with ultrasound. As such, it antimicrobials into the abscess.36 Although some retrospective
is thought to provide a definitive therapy. It abbreviates the case series show a relationship between steroid administration
length of hospitalization and lowers the rate of reoperation.32,36 and poor outcome,3 this is not a consistent finding.43 Further
It may shorten the necessary duration of antimicrobial treatment as steroids are typically given only to those with clinically sig-
and as such may be preferable in situations where outpatient nificant mass effect, the consequence of the underlying illness
compliance with oral therapy is a concern.36 cannot be reliably separated from the effect of corticosteroid
administration. In abscesses adjacent to a ventricular wall, it
Medical management. Classically, penicillin G and chloram- is thought that steroid administration may increase the risk of
phenicol were the antimicrobials of choice in the treatment intraventricular rupture although this is of unclear basis.36
of abscess. The emergence of antibiotic resistance and the
development of agents with improved tolerability have led Common complications. Intraventricular rupture represents a
to a shift in preferred agents over the past several decades. potentially preventable complication of deep-seated abscess.
A recent retrospective analysis based on a prospectively Rupture clinically manifests as sudden-onset headache, menin-
designed antibiotic treatment protocol concluded that the com- geal irritation, and an abrupt deterioration in mental status.
bination of cefotaxmine and metronidazole may be a safe and Imaging is notable for hydrocephalus, ependymal enhancement,
effective regimen for empiric coverage.5 Antibiotic adjustment septation of the ventricle, meningeal enhancement, or the pres-
based on antimicrobial resistance was necessary in 2 patients, ence of ventricular debris. Short distance between the ventricle
one of whom was discovered to have a methicillin-resistant Sta- and the abscess walls (<1mm) and multiloculated abscess are
phylococcus epidermidis species. Methicillin resistance has associated with rupture.44 Some report hematogenous source
become more widespread since the publication of that analysis. of abscess as a potential risk factor, but this is an inconsistent
Currently, most authors recommend the routine addition of finding.40,44 Small abscesses abutting the ventricular space are
vancomycin to a third-generation cephalosporin and metro- no less likely to rupture than larger abscesses and should be
nidazole. In patients with risk factors for pseudomonal infec- approached with similar caution.44 There is no standardized
tion such as abscess associated with recent neurosurgical treatment protocol for rupture, as prospective data are lacking.
procedure, a cephalosporin with the appropriate coverage Rupture is approached often by urgent surgical evacuation of the
such as ceftazidime or cefepime or alternatively a carbape- abscess by open craniotomy or aspiration. This is often accom-
nem with pseudomonal coverage such as meropenem is rec- panied by lavage and ventriculostomy for drainage and antibiotic
ommended. Empiric treatment of the patient with known administration. It is unclear whether these measures are of ben-
immunocompromise must take into account the particular efit and patients often progress poorly despite these interven-
immunodeficiency. Patients with HIV are commonly started tions. As such, patients with significant structural risk factors
on pyrimethamine and sulfadiazine or alternatively clinda- for rupture should be treated aggressively at the outset.
mycin for Toxoplasma coverage.39 Patients with neutropenia An obstructive hydrocephalus may result from occlusion of
are at significant risk of infection from fungal agents. Some the ventricular system from mass effect. As of any space-
practitioners recommend the routine addition of empiric occupying lesion of the posterior fossa, this is not an uncommon
amphotericin in this population. complication of brain abscess and represents a source of signif-
Parenteral antibiotics should be continued for a minimum icant morbidity. A significant proportion of the morbidity asso-
of 6 to 8 weeks. A shorter course of parenteral therapy, poten- ciated with posterior fossa abscess is related to complications
tially as short as 2 weeks, has been suggested; the data for this related to hydrocephalus. Urgent CSF diversion with ventricu-
approach are limited.5 A 2- to 3-month course of oral antibio- lostomy is an uncontroversial intervention in the symptomatic
tics should follow the termination of intravenous therapy.39 patients. The literature suggests that intervening on asympto-
Duration of antimicrobial therapy in the immunocompromised matic hydrocephalus noted on imaging with urgent drainage
patients should be extended, although there is little agreement may be of benefit. A South African center, noting poor outcomes
regarding the optimal duration of treatment. Recommended in cerebellar abscess associated with hydrocephalus and also
length of treatment of parenteral antibiotics ranges from 12 unrecognized hydrocephalus discovered on postmortem exami-
weeks to 1 year.14,40 Careful clinical and imaging follow-up nation, instituted a policy mandating urgent CSF diversion in all
202 The Neurohospitalist 4(4)

with cerebellar abscess and imaging signs of hydrocephalus, in significant morbidity in up to 50% of patients and commonly
regardless of clinical presentation. They observed a significant heralds fatal outcome. Preoperative hydrocephalus may also
reduction in both morbidity and mortality.7 result from deep abscess and has been associated with morbid-
ity.14 Other structural factors affecting outcome include multilo-
culated abscess11 and number of lesions.7 Outcome in the
Outcome and Prognosis immunocompetent patients is typically unrelated to the respon-
The prognosis of brain abscess has improved considerably sible microbe. Abscess from organisms, such as Listeria, Nocar-
since the advent of CT. All-cause mortality in patients hospi- dia, and Actinomyces, affecting the immunocompromised
talized with abscess varies from 5% to 32%.2,3,5-11,13-16,24,25 patients presents a greater challenge in management and indi-
Degree of compromise in neurologic conditions on initial cates poorer prognosis. An immunocompromised host is found
evaluation, in particular alteration in mental status, is consis- to be an independent factor of worse outcome.10,11
tently found to be predictive of ultimate prognosis.3,9,10,14-16,25,43 A significant proportion of patients with appropriately
Patients with rapid neurologic decline and those with shorter treated abscess recover completely and can survive without sig-
disease duration prior to hospitalization have worse out- nificant neurologic sequelae.5,24 Symptomatic epilepsy from
come.3,6,14,40 Other early clinical factors potentially indicative abscess is the most common persistent deficit.15,25 Focal neuro-
of poor outcome include meningismus, leukocytosis, fever, or logic symptoms such as hemiparesis, loss of vision, and dys-
sepsis at presentation.3,10,12 The physical characteristics of the phagia3,5,15,24 are not uncommon as are global cognitive
abscess may be predictive of outcome. Several series report deep deficits.5 Routine follow-up of patients with abscess is advisa-
location as a poor prognostic marker.4,11,40 This may be for a ble, as abscess recurrence may occur many months to years
variety of reasons including a potential predisposition of these after the initial event.5,6,15,25 One series reports a case of
lesions to result in complications as well as limitations in terms abscess recurrence 7 years after initial treatment.25 Often recur-
of ease of surgical accessibility. Intraventricular rupture is con- rence is related to inadequate duration or choice of antimicro-
sistently reported to portend poor outcome.3,4,6,7,9,11,40 It results bial agent, but in other cases no explanation is found.

Appendix

Summarizing points

Epidemiology  The incidence of brain abscess has decreased in the United States in the 20th century
 Brain abscess may occur at any age but is most common between the third and fifth decades of life
 There is a significant male predominance in incidence that is independent of geography
Etiology  Brain abscess may result from hematogenous spread, contiguous spread, surgical procedure, or head trauma
 In a significant proportion of cases an etiology cannot be identified
 Bacterial culture typically yields a single organism but multiple pathogens are often isolated. Culture-negative cases are
not uncommon
 Streptococci are the most common pathogens isolated although flora vary with pathogenesis (see Table 1)
Diagnosis  Headache, fever, altered mental status, focal neurologic deficits, and nausea and vomiting are the most common
presenting symptoms, but the absence of these findings will not rule out abscess
 In cases of mature abscess, magnetic resonance (MR) imaging demonstrates a T1 hyperintense, T2 isointense lesion with a
smooth T1 iso- to hyperintense and T2 iso- to hypointense contrast enhancing capsule. Abscess will typically diffusion
restrict and appear hypointense on apparent diffusion coefficient
 1H nuclear MR spectroscopy shows significant potential in allowing for differentiation of abscess from other intracranial
mass lesions
Management  All patients should receive empiric antibiotic therapy with extended coverage for gram positive, gram negative, and
anaerobic organisms. Coverage may be narrowed based on culture data
 Surgical therapy is recommended in most cases, although in a carefully selected subset of patients medical therapy alone
may be sufficient for resolution
 En bloc excision and computed tomography-guided aspiration may be considered for surgical treatment. The choice of
procedure is a multifaceted decision that must weigh factors such as abscess etiology, location, and microbiology
 An extended course of parenteral antibiotics followed by oral therapy is recommended post procedure
Outcome and  All-cause mortality ranges from 5% to 32%
Prognosis  Degree of neurologic compromise at initial evaluation is a strong predictive factor of ultimate outcome
 Other clinical factors associated with poor outcome include rapid neurologic decline, shorter disease duration prior to
hospitalization, meningismus, leukocytosis, fever and sepsis at presentation
 Structural factors include deep abscess, multiloculated abscess, and multiple lesions affect outcome
 A significant proportion of appropriately treated patients recover completely and survive without residual neurologic
symptoms
 Routine follow-up is necessary, as abscess recurrence is a known complication and may occur years after the initial event
Patel and Clifford 203

Declaration of Conflicting Interests consecutive case series study from Pakistan. World Neurosurg.
The authors declared no potential conflicts of interest with respect to 2011;76(1-2):195-200; discussion 179-183.
the research, authorship, and/or publication of this article. 16. Menon S, Bharadwaj R, Chowdhary A, Kaundinya DV, Palande
DA. Current epidemiology of intracranial abscesses: a prospec-
Funding tive 5 year study. J Med Microbiol. 2008;57(pt 10):1259-1268.
The authors received no financial support for the research, authorship, 17. Sharma BS, Gupta SK, Khosla VK. Current concepts in the manage-
and/or publication of this article. ment of pyogenic brain abscess. Neurol India. 2000;48(2):105-111.
18. Lumbiganon P, Chaikitpinyo A. Antibiotics for brain abscesses
References in people with cyanotic congenital heart disease. Cochrane
1. Nicolosi A, Hauser WA, Musicco M, Kurland LT. Incidence and Database Syst Rev. 2013;3:CD004469.
prognosis of brain abscess in a defined population: Olmsted 19. McClelland S, 3rd, Hall WA. Postoperative central nervous sys-
County, Minnesota, 1935-1981. Neuroepidemiology. 1991; tem infection: incidence and associated factors in 2111 neurosur-
10(3):122-131. gical procedures. Clin Infect Dis. 2007;45(1):55-59.
2. Faraji-Rad M, Samini F. Clinical features and outcome of 83 adult 20. Sennaroglu L, Sozeri B. Otogenic brain abscess: review of 41
patients with brain abscess. Arch Iran Med. 2007;10(3):379-282. cases. Otolaryngol Head Neck Surg. 2000;123(6):751-755.
3. Hakan T, Ceran N, Erdem I, Berkman MZ, Goktas P. Bacterial 21. Singh N, Husain S. Infections of the central nervous system in
brain abscesses: an evaluation of 96 cases. J Infect. 2006; transplant recipients. Transpl Infect Dis. 2000;2(3):101-111.
52(5):359-366. 22. Singh N, Paterson DL. Mycobacterium tuberculosis infection in
4. Song L, Guo F, Zhang W, et al. Clinical features and outcome solid-organ transplant recipients: impact and implications for
analysis of 90 cases with brain abscess in central China. Neurol management. Clin Infect Dis. 1998;27(5):1266-1277.
Sci. 2008;29(6):425-430. 23. Cunha BA. Central nervous system infections in the compro-
5. Jansson AK, Enblad P, Sjolin J. Efficacy and safety of cefotaxime mised host: a diagnostic approach. Infect Dis Clin North Am.
in combination with metronidazole for empirical treatment of 2001;15(2):567-590.
brain abscess in clinical practice: a retrospective study of 66 con- 24. Roche M, Humphreys H, Smyth E, et al. A twelve-year review of
secutive cases. Eur J Clin Microbiol Infect Dis. 2004;23(1):7-14. central nervous system bacterial abscesses; presentation and
6. Kao PT, Tseng HK, Liu CP, Su SC, Lee CM. Brain abscess: clin- aetiology. Clin Microbiol Infect. 2003;9(8):803-809.
ical analysis of 53 cases. J Microbiol Immunol Infect. 2003; 25. Sharma R, Mohandas K, Cooke RP. Intracranial abscesses:
36(2):129-136. changes in epidemiology and management over five decades
7. Nathoo N, Nadvi SS, Narotam PK, van Dellen JR. Brain abscess: in Merseyside. Infection. 2009;37(1):39-43.
management and outcome analysis of a computed tomography 26. Germiller JA, Monin DL, Sparano AM, Tom LW. Intracranial
era experience with 973 patients. World Neurosurg. 2011; complications of sinusitis in children and adolescents and their out-
75(5-6):716-726; discussion 612-717. comes. Arch Otolaryngol Head Neck Surg. 2006;132(9):969-976.
8. Sarmast AH, Showkat HI, Bhat AR, et al. Analysis and manage- 27. Rath TJ, Hughes M, Arabi M, Shah GV. Imaging of cerebritis,
ment of brain abscess; a ten year hospital based study. Turk Neu- encephalitis, and brain abscess. Neuroimaging Clin N Am.
rosurg. 2012;22(6):682-689. 2012;22(4):585-607.
9. Tattevin P, Bruneel F, Clair B, et al. Bacterial brain abscesses: a 28. Chang SC, Lai PH, Chen WL, et al. Diffusion-weighted MRI
retrospective study of 94 patients admitted to an intensive care features of brain abscess and cystic or necrotic brain tumors: com-
unit (1980 to 1999). Am J Med. 2003;115(2):143-146. parison with conventional MRI. Clin Imaging. 2002;26(4):227-236.
10. Tseng JH, Tseng MY. Brain abscess in 142 patients: factors 29. Faehndrich J, Weidauer S, Pilatus U, Oszvald A, Zanella FE,
influencing outcome and mortality. Surg Neurol. 2006;65(6): Hattingen E. Neuroradiological viewpoint on the diagnostics
557-562; discussion 562. of space-occupying brain lesions. Clin Neuroradiol. 2011;
11. Xiao F, Tseng MY, Teng LJ, Tseng HM, Tsai JC. Brain abscess: 21(3):123-139.
clinical experience and analysis of prognostic factors. Surg Neu- 30. Lai PH, Hsu SS, Ding SW, et al. Proton magnetic resonance
rol. 2005;63(5):442-449; discussion 449-450. spectroscopy and diffusion-weighted imaging in intracranial
12. Lu CH, Chang WN, Lin YC, et al. Bacterial brain abscess: cystic mass lesions. Surg Neurol. 2007;68(suppl 1):S25-S36.
microbiological features, epidemiological trends and therapeutic 31. Floeth FW, Pauleit D, Sabel M, et al. 18F-FET PET differentia-
outcomes. QJM. 2002;95(8):501-509. tion of ring-enhancing brain lesions. J Nuclear Med. 2006;47(5):
13. Carpenter J, Stapleton S, Holliman R. Retrospective analysis of 776-782.
49 cases of brain abscess and review of the literature. Eur J Clin 32. Cavusoglu H, Kaya RA, Turkmenoglu ON, Colak I, Aydin Y.
Microbiol Infect Dis. 2007;26(1):1-11. Brain abscess: analysis of results in a series of 51 patients with
14. Radoi M, Ciubotaru V, Tataranu L. Brain abscesses: clinical a combined surgical and medical approach during an 11-year
experience and outcome of 52 consecutive cases. Chirurgia. period. Neurosurg Focus. 2008;24(6):E9.
2013;108(2):215-225. 33. Rosenblum ML, Hoff JT, Norman D, Edwards MS, Berg BO.
15. Manzar N, Manzar B, Kumar R, Bari ME. The study of etiologic Nonoperative treatment of brain abscesses in selected high-risk
and demographic characteristics of intracranial brain abscess: a patients. J Neurosurg. 1980;52(2):217-225.
204 The Neurohospitalist 4(4)

34. Mamelak AN, Obana WG, Flaherty JF, Rosenblum ML. Nocar- 40. Takeshita M, Kagawa M, Izawa M, Takakura K. Current treatment
dial brain abscess: treatment strategies and factors influencing strategies and factors influencing outcome in patients with bacterial
outcome. Neurosurgery. 1994;35(4):622-631. brain abscess. Acta Neurochirurgica. 1998;140(12):1263-1270.
35. Moorthy RK, Rajshekhar V. Management of brain abscess: an 41. Quartey GR, Johnston JA, Rozdilsky B. Decadron in the treat-
overview. Neurosurg Focus. 2008;24(6):E3. ment of cerebral abscess. An experimental study. J Neurosurg.
36. Gadgil N, Patel AJ, Gopinath SP. Open craniotomy for brain 1976;45(3):301-310.
abscess: a forgotten experience? Surg Neurol Int. 2013;4:34. 42. Kourtopoulos H, Holm SE, Norrby SR. The influence of steroids
37. Valarezo J, Cohen JE, Valarezo L, et al. Nocardial cerebral on the penetration of antibiotics into brain tissue and brain
abscess: report of three cases and review of the current neurosur- abscesses. An experimental study in rats. J Antimicrob Che-
gical management. Neurol Res. 2003;25(1):27-30. mother. 1983;11(3):245-249.
38. Adeyemi OA, Gottardi-Littell N, Muro K, Kane K, Flaherty JP. 43. Seydoux C, Francioli P. Bacterial brain abscesses: factors influen-
Multiple brain abscesses due to Actinomyces species. Clin Neu- cing mortality and sequelae. Clin Infect Dis. 1992;15(3):394-401.
rol Neurosurg. 2008;110(8):847-849. 44. Lee TH, Chang WN, Su TM, et al. Clinical features and predic-
39. Bernardini GL. Diagnosis and management of brain abscess tive factors of intraventricular rupture in patients who have bac-
and subdural empyema. Curr Neurol Neurosci Rep. 2004; terial brain abscesses. J Neurol Neurosurg Psychiatry. 2007;
4(6):448-456. 78(3):303-309.