2026 年 13 巻 p. 485-491
Brain abscess developing at the site of intracerebral hemorrhage has rarely been reported. We report herein a case of brain abscess following intracerebral hemorrhage that mimicked a chronic encapsulated intracerebral hematoma. An 80-year-old man presented with right hemiparesis caused by a left frontoparietal subcortical intracerebral hemorrhage, which was managed conservatively. During hospitalization, progressive motor aphasia developed. Two weeks after admission, magnetic resonance imaging demonstrated worsening perihematomal edema and ring enhancement along the hematoma margin. Diffusion-weighted imaging demonstrated hyperintensity with heterogeneous restricted diffusion on the apparent diffusion coefficient map. Because there were no systemic signs of infection, chronic encapsulated intracerebral hematoma was initially suspected. However, the neurological and radiological deterioration progressed, prompting neuroendoscopic evacuation of the hematoma cavity. Microbiological examination revealed Enterococcus faecalis within the cavity, confirming the diagnosis of brain abscess. Targeted antimicrobial therapy led to clinical improvement. The possibility of a secondary brain abscess should be considered if delayed-onset neurological deterioration and worsening perihematomal edema are observed after intracerebral hemorrhage even in the absence of systemic signs of infection or typical diffusion restriction. Early microbiological evaluation of the hematoma cavity is essential to establish the diagnosis and determine the optimal treatment.
Brain abscess developing at the site of intracerebral hemorrhage (ICH) has rarely been reported. In the subacute phase after ICH onset, worsening perihematomal edema and delayed seizures cause delayed-onset neurological deterioration.1-4) However, rare complications, such as a brain abscess or chronic encapsulated intracerebral hematoma (CEIH), may also cause delayed-onset neurological deterioration, and because they often present with similar clinical and radiological findings, differentiating between them preoperatively can be challenging.5) Herein, we report a brain abscess arising within a pre-existing hematoma cavity following ICH that lacked typical clinical and radiological features and closely mimicked a CEIH.
An 80-year-old man, who had undergone a percutaneous coronary intervention for an acute myocardial infarction 1 month earlier and was receiving dual antiplatelet therapy (consisting of aspirin 100 mg and prasugrel 3.75 mg), presented with right-sided weakness, which initially developed in the right upper extremity before progressing to right lower extremity paralysis on the following day.
Upon admission, he was alert and oriented. A neurological examination revealed mild dysarthria associated with right facial palsy, right-sided sensory impairment, and severe right hemiparesis (manual muscle testing grade 0/5 in the right upper extremity and 1/5 in the right lower extremity). He had no febrile episodes before admission, and his body temperature was normal on admission. Laboratory findings revealed no evidence of systemic inflammation. He had no history of diabetes mellitus, steroid use, or immunosuppressive therapy. Non-contrast computed tomography (CT) demonstrated a bilobar, subcortical ICH in the left frontoparietal region, but CT angiography revealed no vascular abnormalities (Figure 1A and B). Magnetic resonance imaging (MRI) demonstrated slight hyperintensity on diffusion-weighted imaging (DWI) without a corresponding reduction on the apparent diffusion coefficient (ADC) map; hyperintensity within the hematoma on fluid-attenuated inversion recovery (FLAIR) imaging; and hypointensity surrounding the lesion on T2*-weighted imaging (Figure 1C-F). T2*-weighted imaging demonstrated no cerebral microbleeds suggestive of cerebral amyloid angiopathy. Based on the medication history and imaging findings, the hemorrhage was attributed to antiplatelet therapy and was managed conservatively with strict blood pressure control and discontinuation of dual antiplatelet therapy, which was subsequently switched to clopidogrel 75 mg/day for the secondary prevention of ischemic heart disease. Follow-up CT on hospital days 1 and 7 demonstrated stabilization of the perihematomal edema and no hematoma expansion (Figure 2A).
Neuroimaging findings at the initial presentation. A: Non-contrast axial CT on admission demonstrated a bilobar subcortical ICH in the left frontoparietal region. B: CT angiography demonstrated no vascular abnormalities. C-F. MRI findings. C: Diffusion-weighted imaging demonstrated slight hyperintensity within the lesion. D. The apparent diffusion coefficient map showed no corresponding reduction, indicating no definite restricted diffusion. E: Fluid-attenuated inversion recovery imaging demonstrated hyperintensity within the hematoma. F: T2*-weighted imaging demonstrated hypointensity surrounding the lesion consistent with hemorrhagic components.
CT: computed tomography; ICH: intracerebral hemorrhage; MRI: magnetic resonance imaging
Neuroimaging findings during clinical deterioration. A: Non-contrast CT on hospital day 1 demonstrated a bilobar subcortical intracerebral hemorrhage without hematoma expansion and with stable, perihematomal edema. B: Non-contrast CT approximately 2 weeks after admission demonstrated worsening of the perihematomal edema without evidence of rebleeding. C: Diffusion-weighted imaging demonstrated hyperintensity within the lesion. D: The apparent diffusion coefficient map showed heterogeneous signal reduction within the lesion. E: FLAIR imaging demonstrated exacerbation of the perihematomal edema. F. Contrast-enhanced T1-weighted imaging demonstrated distinct ring enhancement along the hematoma margin.
CT: computed tomography; FLAIR: fluid-attenuated inversion recovery
Approximately 2 weeks after admission, the patient experienced progressive motor aphasia despite having stable systemic status. CT found no evidence of rebleeding but demonstrated worsening perihematomal edema (Figure 2B), thus prompting the start of osmotic therapy. MRI demonstrated hyperintensity on DWI (Figure 2C), and the ADC map suggested heterogeneous restricted diffusion (Figure 2D). FLAIR imaging demonstrated exacerbation of the perihematomal edema (Figure 2E), and contrast-enhanced T1-weighted imaging revealed distinct ring enhancement along the hematoma margin (Figure 2F). In the absence of fever and elevated inflammatory markers, CEIH was considered the most likely diagnosis based on the clinical course and imaging findings. However, because preoperative differentiation between CEIH and brain abscess has been reported to be challenging, the possibility of a brain abscess cannot be ruled out.
The neurological and radiological deterioration prompted neuroendoscopic hematoma evacuation on hospital day 21. Intraoperatively, the hematoma was found to contain serous components and to be surrounded by a brownish wall structure indicating hemosiderin deposition (Figure 3A), but no frank purulent material was observed. Samples of the hematoma capsule and its content were collected for histopathological and microbiological analysis. Preoperative imaging had demonstrated 2 adjacent compartments; therefore, the internal septum was endoscopically fenestrated to allow maximum evacuation and thorough irrigation with saline.
Intraoperative and histopathological findings. A: Neuroendoscopic view of the brownish wall structure indicating chronic hemosiderin deposition and containing the content of the serous hematoma. B: Photomicrograph of the hematoma wall showing prominent neutrophil-dominant inflammatory cell infiltration and marked tissue destruction (hematoxylin and eosin staining, original magnification ×200). C: Elastica van Gieson staining (original magnification ×200) of the hematoma wall demonstrated fragmentary connective tissue composed predominantly of collagen fibers and fibroblasts, suggesting an abscess wall. However, no well-formed, mature, fibrous capsule characteristic of a chronic encapsulated intracranial hematoma was observed. D: Gram staining (original magnification ×200) confirmed clusters of Gram-positive cocci (arrowheads) identical with the Enterococcusfaecalis isolated from the cultures.
On postoperative day 4, cultures of the intraoperative specimen obtained from the hematoma cavity grew Enterococcus faecalis, which prompted targeted antibiotic therapy with ampicillin 2 g every 4 hours. After the pathogen was identified, a comprehensive evaluation, including urinalysis and gastrointestinal assessment for potential sources of infection, was performed. Blood cultures were negative, and transthoracic echocardiography revealed no findings suggestive of infective endocarditis. A dental evaluation and tests for paranasal sinusitis and otogenic infection were performed; possible contiguous sources of infection were negative.
Histopathological examination of the hematoma wall demonstrated prominent neutrophil-dominant inflammatory cell infiltration and marked tissue destruction on hematoxylin and eosin staining (Figure 3B). Elastica van Gieson staining revealed fragmentary connective tissue composed predominantly of collagen fibers and fibroblasts, raising suspicion of an abscess wall; however, the well-formed, mature, fibrous capsule characteristic of CEIH was not observed (Figure 3C). Gram staining confirmed clusters of Gram-positive cocci identical with the cultured organism (Figure 3D). On the basis of these findings, secondary brain abscess following ICH was diagnosed.
After completing a 6-week course of intravenous antibiotics, the patient experienced a gradual improvement of the motor aphasia, and follow-up MRI demonstrated no evidence of a recurrence (Figure 4A and B). He was transferred to a rehabilitation facility on postoperative day 30 with residual right hemiparesis. At 3 months after the transfer, the patient was followed up in the outpatient clinic. Although the right hemiparesis remained, his general condition was stable. MRI demonstrated complete resolution of the ring enhancement on contrast-enhanced T1-weighted imaging and improvement of the edema on FLAIR (Figure 4C and D). In addition, magnetic resonance angiography demonstrated no findings suggestive of an infectious intracranial aneurysm.
Postoperative follow-up imaging. A: Gadolinium-enhanced T1-weighted imaging at the completion of antibiotic therapy demonstrated marked improvement of the abscess wall enhancement, with only slight residual enhancement remaining. No evidence of abscess recurrence was observed. B: FLAIR imaging at the same time point demonstrated a marked reduction in the perihematomal edema and significantly greater shrinkage of the abscess cavity than seen preoperatively. C: Gadolinium-enhanced T1-weighted imaging at 3 months after transfer demonstrated complete resolution of the ring enhancement, with no evidence of a recurrence. D: FLAIR imaging at 3 months after transfer demonstrated further improvement of the edema.
FLAIR: fluid-attenuated inversion recovery
This case of secondary brain abscess, which developed during the clinical course of an ICH, was difficult to differentiate from CEIH. Both brain abscesses and CEIH may present with ring enhancement on contrast-enhanced MRI, making it challenging to differentiate them radiologically.5) In the present case, despite the hyperintensity on DWI and the ring enhancement, the ADC values demonstrated heterogeneous reduction rather than the typical restricted diffusion seen in brain abscesses. Moreover, the neurological deterioration occurred approximately 2 weeks after ICH onset, which was earlier than the typical clinical course of CEIH involving gradual progression over several weeks to months.6-8) However, because there were no clinical findings suggesting infection, such as a fever or elevated inflammatory markers, CEIH was initially considered the most likely diagnosis. The imaging and pathological findings in this case can be explained by the process of brain abscess formation.
Brain abscesses typically demonstrate restricted diffusion due to the high viscosity and cellularity of purulent material.9,10) However, in the present case, although DWI demonstrated hyperintensity, the ADC values were heterogeneous and did not indicate typical restricted diffusion. The lesion developed within the hematoma cavity approximately 2 weeks after the onset of the ICH. Subacute hematomas may display high signal intensity on T2-weighted imaging.11,12) Therefore, the hyperintensity observed on DWI may have partly reflected a T2 shine-through effect rather than true restricted diffusion. Moreover, brain abscess formation is known to progress through 4 pathological stages: early cerebritis, late cerebritis, early capsule formation, and late capsule formation. The typical imaging features of a mature abscess generally appear after approximately 2 weeks, during the late capsule formation stage.13,14) Previous reports of secondary brain abscess following ICH have generally described abscess formation occurring 4-20 weeks after the hemorrhagic event.15) In contrast, neurological deterioration and radiological progression became apparent approximately 2 weeks after ICH onset in the present case, representing a relatively early presentation. This relatively early clinical course may explain why the abscess had not fully matured and why the lesion demonstrated atypical diffusion findings.
A hemorrhagic brain abscess should be included in the differential diagnosis. However, several findings in this case did not support this interpretation. Although the possibility of a pre-existing microscopic abscess obscured by the hemorrhage cannot be completely excluded, there were no clinical findings suggestive of a brain abscess at the initial presentation, such as fever, headache, or elevated inflammatory markers. In addition, the initial MRI demonstrated no significant surrounding edema, making the presence of a pre-existing abscess less likely. Furthermore, the histopathological findings were more consistent with the early capsule formation stage of a brain abscess than with a long-standing pre-existing abscess. Taken together, these findings supported the interpretation that the abscess developed secondarily within the hematoma cavity rather than before the hemorrhage.
Ring enhancement in either a brain abscess or a CEIH is generally considered to reflect capsule formation; however, the underlying pathological structures differ between the 2 conditions. In a brain abscess, the enhancing rim corresponds to a fibrous capsule surrounding the necrotic cavity and is typically accompanied by an inner layer of granulation tissue.13,16) In contrast, CEIH is characterized by a thick fibrous capsule containing numerous thin-walled vessels.6,7,17) In the present case, histopathological examination demonstrated neutrophil-dominant inflammatory cell infiltration accompanied by tissue destruction, indicating a brain abscess. Elastica van Gieson staining revealed only fragmented collagenous connective tissue consistent with developing granulation tissue but demonstrated no evidence of a mature fibrous capsule. Although secondary infection during CEIH development cannot be ruled out, no histopathological findings typical of a CEIH were observed, and the lesion was considered more consistent with the early capsule formation stage of a brain abscess.
Secondary brain abscess following ICH has rarely been reported, and Enterococcus spp. do not commonly cause brain abscesses,14,18) which generally arise either through direct extension of an adjacent infection or through hematogenous dissemination from a distant, infectious focus.14) Previous reports found that most brain abscesses caused by Enterococcus spp. were associated with direct spread from an adjacent infection, such as chronic suppurative otitis media.18) In the present case, however, there was no history of direct spread from an adjacent infection. Previous reports of brain abscesses developing after stroke have suggested that bacteremia may result in bacterial seeding of regions where the blood-brain barrier has been disrupted by a preceding vascular event. Under this condition, microorganisms may colonize the injured brain tissue and form an abscess.19,20) In the present case, a similar hematogenous mechanism may have contributed to the formation of the abscess, and transient bacteremia may have occurred at some point during the clinical course. Enterococcus faecalis, which was identified in this case, is a commensal organism of the gastrointestinal tract.21) Although examinations of the gastrointestinal tract and urinary system were performed, no definite infectious source was identified.
Treatments for brain abscesses differ fundamentally from those for CEIH. Brain abscesses require not only surgical drainage but also lengthy, targeted antimicrobial therapy.14,22) In contrast, a CEIH can generally be managed by hematoma evacuation, with capsule removal being performed if necessary.6) Therefore, early and accurate diagnosis is essential to guide treatment appropriately.
In the present case, neuroendoscopic hematoma evacuation enabled removal of the hematoma while simultaneously allowing tissue sampling and microbiological examination. Although CEIH was initially considered the most likely diagnosis based on the clinical and radiological findings, previous reports have indicated that preoperatively differentiating between CEIH and brain abscess can be challenging. Therefore, the possibility of a secondary brain abscess could not be ruled out, prompting histopathological and microbiological analyses of intraoperative specimens. This approach allowed early microbiological assessment and prompt initiation of targeted antimicrobial therapy, which likely contributed to the favorable clinical outcome.
Secondary brain abscesses may develop within a hematoma cavity following intracerebral hemorrhage and can present with delayed neurological deterioration during the subacute phase. Because their clinical and radiological features can closely mimic those of CEIH, early microbiological evaluation of the hematoma cavity is essential to establish the diagnosis and determine the optimal antimicrobial therapy.
All authors have no conflict of interest. All the authors who are members of the Japan Neurosurgical Society have registered their COI status online.
The patient gave his informed consent for the publication of this study.
