Utility of Metagenomic Next‐Generation Sequencing in Identifying Anaerobic Pathogens in a Pediatric Brain Abscess Secondary to Sinusitis: A Case Report

ABSTRACT

In this pediatric brain abscess case, metagenomic next‐generation sequencing (mNGS) identified an anaerobic bacterium that remained undetected using conventional culture methods, thus providing clinically relevant information for antimicrobial management. This indicates that mNGS may be a useful adjunct for optimizing treatment strategies.

1. Introduction

A brain abscess is an infection of the central nervous system that develops within the cerebral parenchyma. Its annual incidence in children is estimated to be 0.5 cases per 100,000 people, making it a very rare disease [1]. However, neurological sequelae remain in approximately 27.7%–39.2% of affected patients [2, 3], indicating that prompt diagnosis, identification of the causative pathogen, and initiation of appropriate antimicrobial therapy are essential for improving outcomes.

Identification of the causative microorganisms in brain abscesses is not always straightforward. The culture‐negative rate ranges from 14% to 34%, with prior administration of antimicrobial agents before specimen collection being a major contributing factor [4]. In particular, culture positivity decreases after more than 3 days of antimicrobial therapy [4]. Anaerobic bacteria are especially difficult to identify because of their strict growth requirements [5].

Metagenomic next‐generation sequencing (mNGS) is a technique that comprehensively analyzes all the nucleic acids present in patient samples to identify pathogenic microorganisms. mNGS has a higher detection rate than conventional culture‐based methods and is useful for identifying fastidious or emerging microorganisms that are difficult to detect using standard cultures [6]. Nevertheless, because of cost, turnaround time, and difficulties in clinical interpretation, mNGS has not yet been widely implemented in routine clinical practice [7]. In this report, we describe a pediatric case of a brain abscess in which mNGS identified the causative anaerobic bacteria that were not detected by culture and served as a useful adjunct to complement clinical decision‐making.

2. Case History, Differential Diagnosis, Investigations, and Treatment

A previously healthy 10‐year‐old boy developed an influenza A infection 2 weeks prior to presentation. He promptly became afebrile after receiving anti‐influenza medication. However, he remained lethargic and spent most of the day bedridden. Four days after the onset of influenza, fever recurred and headaches developed, with symptoms progressively worsening over time. He visited a local clinic three times within 1 week and, at the third visit, was diagnosed with sinusitis and prescribed clarithromycin. On the same day, after lunch, he vomited and became unresponsive to verbal stimulation from his parents, prompting emergency transportation to our hospital. At first contact with the emergency medical services, his level of consciousness was Glasgow Coma Scale (GCS) E3V1M5. There was no evidence of limb rigidity. He opened his eyes and turned his face in response to verbal stimulation but was unable to speak or follow commands. Although left‐beating nystagmus was observed, the bilateral pupillary light reflexes were preserved. Just before arriving at the hospital, he began to follow commands and engage in vague conversations. A prolonged seizure lasting 30–40 min was suspected. No focal motor deficits or neck stiffness was observed. The BioFire FilmArray Respiratory Panel 2.1 (RP2.1) detected influenza A(H1N1)pdm09. Blood tests revealed elevated inflammatory markers [C‐reactive protein level, 2.52 mg/dL; white blood cell count 22,400/μL (neutrophils 87%, lymphocytes 6%)]. No electrolyte abnormalities or hyperammonemia that could explain the impaired consciousness were identified. Non‐contrast head computed tomography revealed no evidence of intracranial hemorrhage or space‐occupying lesions; however, findings consistent with sinusitis were observed in the left maxillary sinus and bilateral ethmoid and frontal sinuses. The patient was admitted for observation of consciousness and further evaluation of seizure episodes. After admission, intravenous ampicillin–sulbactam was initiated for sinusitis. The following day, his level of consciousness improved; however, fever and headache persisted. Despite analgesic treatment with loxoprofen sodium and acetaminophen, the headache worsened and meningeal signs, including neck stiffness and Kernig's sign, became apparent. Cerebrospinal fluid examination revealed pleocytosis (2976 white blood cells/μL with 73% neutrophils) and hypoglycorrhachia (glucose level, 44 mg/dL), although no pathogens included in the BioFire FilmArray Meningitis Panel were detected. Empirical treatment for bacterial meningitis was initiated with cefotaxime and vancomycin. On hospital Day 3, his impaired consciousness worsened again (GCS E3V1M6) and inflammatory markers increased further (Figure 1). Brain magnetic resonance imaging (MRI) performed on the same day revealed a left frontal lobe abscess and a left subdural empyema, with ring enhancement on post‐contrast T1‐weighted images, hyperintensity on diffusion‐weighted imaging, and hypointensity on apparent diffusion coefficient maps (Figure 2). Sinusitis was considered the source of infection and endoscopic sinus drainage was performed on hospital Day 3. The antimicrobial therapy was changed to ceftriaxone, vancomycin, and metronidazole; levetiracetam was initiated as seizure prophylaxis. However, follow‐up brain MRI on hospital Day 5 revealed abscess progression (Figure 2) and craniotomy with abscess drainage was performed on hospital Day 6. Thereafter, no enlargement of the abscess was observed and antimicrobial therapy was continued. Postoperatively, the patient developed transient motor aphasia, which resolved within 1 week. Two sets of blood cultures and cerebrospinal fluid cultures failed to detect significant pathogens. Culture of the sinus aspirate yielded Streptococcus intermedius . No significant bacterial growth was observed in the brain abscess culture; however, Gram staining revealed streptococcal organisms. On hospital Day 21, the patient developed tingling paresthesia over the anterior aspect of the right lower leg without any objective sensory deficit or motor weakness. The patient had no history of prolonged positioning or focal compression that could account for the symptom. Owing to the absence of other new focal neurological deficits and 19 days of metronidazole therapy (cumulative dose 19.95 g), peripheral neuropathy due to metronidazole was considered more likely than brain abscess recurrence, and antimicrobial therapy was switched to meropenem. mNGS, performed to identify potential pathogens using a previously reported method [8], identified Fusobacterium vincentii in addition to S. intermedius on hospital Day 33 (Figure 3).

FIGURE 1.

Clinical course. Inflammatory marker levels improved after two sessions of drainage and antibiotic therapy. CTRX, ceftriaxone; CTX, cefotaxime; LZD, linezolid; MEPM, meropenem; MNZ, metronidazole; SBT/ABPC, sulbactam/ampicillin; VCM, vancomycin; WBC, white blood cell.

FIGURE 2.

Brain magnetic resonance imaging. (A) Brain magnetic resonance imaging on hospital Day 3. Magnetic resonance imaging revealed a ring‐enhancing lesion on post‐contrast T1‐weighted images, accompanied by restricted diffusion characterized by hyperintensity on DWI and low ADC values, suggestive of a brain abscess in the left frontal lobe and left subdural empyema. (B) Brain magnetic resonance imaging on hospital Day 5. These findings indicate interval enlargement of the left frontal lobe abscess and left subdural empyema compared to Day 3. (C) Follow‐up brain magnetic resonance imaging 1 month after the completion of antimicrobial therapy. This demonstrates the resolution of diffusion restriction on diffusion‐weighted imaging with normalization of apparent diffusion coefficient values and the absence of residual hyperintense lesions on fluid‐attenuated inversion recovery, indicating no evidence of recurrent inflammation after the completion of treatment. ADC, apparent diffusion coefficient; DWI, diffusion‐weighted imaging; FLAIR, fluid‐attenuated inversion recovery; Gd‐enhanced T1WI, Gadolinium‐enhanced T1‐weighted imaging.

FIGURE 3.

Results of metagenomic next‐generation sequencing analysis of (A) brain and (B) subdural abscess. Streptococcus intermedius and Fusobacterium vincentii were considered to be causative pathogens.

3. Outcome and Follow‐Up

Antimicrobial treatment was administered for a total duration of 6 weeks after craniotomy drainage (Figure 1). Electroencephalography revealed no abnormalities on hospital Day 49 and levetiracetam was discontinued on hospital Day 50. However, on hospital Day 55, he experienced an epileptic seizure beginning with a right conjugate deviation, followed by impaired consciousness, and levetiracetam was resumed. Follow‐up brain MRI 1 month after the completion of antimicrobial therapy showed no recurrence of the abscess (Figure 2). In the predischarge WISC‐IV assessment, the full‐scale IQ was 82, with no significant discrepancies among index scores, indicating overall performance in the slightly below‐average range. Although the possibility of neurological sequelae cannot be ruled out, he has returned to school and, 1 year after onset, is currently leading a normal social life without academic or interpersonal difficulties. During the 1‐year follow‐up period after discharge, the patient remained seizure‐free while receiving levetiracetam.

4. Discussion

We report a pediatric case of a brain abscess secondary to sinusitis in which anaerobic pathogens undetectable by conventional culture of the sinus aspirate and brain abscess specimens were identified as causative agents by mNGS. This suggests that mNGS may be a useful and adjunctive tool to guide therapeutic decisions.

In pediatric patients, approximately 40%–50% of brain abscesses arise from the contiguous spread of infection from adjacent foci, such as otitis media, mastoiditis, or sinusitis, or through disruption to cranial bone continuity caused by head trauma or neurosurgical procedures [9]. In 30%–40% of cases, infection spreads hematogenously from distant sites, including dental abscesses, infective endocarditis, pulmonary infections, or skin infections, whereas the remaining cases have no identifiable source [9]. In this case, direct extension from sinusitis was considered the source and, given the typical microbiology of sinusitis, polymicrobial infections, including anaerobic bacteria, were clinically suspected.

On the other hand, the reported isolation rates of anaerobic bacteria from brain abscesses are relatively low. A systematic review by Cobo et al. reported that anaerobes were isolated in only 715 of 6167 brain abscess cases (11.5%) [5]. This low detection rate may be attributed to the vulnerability of anaerobes to oxygen exposure during specimen collection and transport, as well as their specific nutritional requirements [10]. Additionally, brain abscesses are often polymicrobial and rapidly growing aerobic or facultative anaerobic bacteria may overgrow and obscure slow‐growing anaerobic colonies [10]. Anaerobic bacteria may also be difficult to detect using Gram staining. Similar to those used in culture systems, they are fragile and susceptible to oxygen exposure [11]. Moreover, some species are Gram‐variable, requiring special staining for visualization [12]. In particular, Fusobacterium often requires Warthin–Starry or Wayson staining [13], which was not performed in this case. Therefore, excluding the involvement of anaerobic bacteria based solely on culture and Gram staining results is difficult.

International guidelines for brain abscesses recommend third‐generation cephalosporin and metronidazole as an empirical therapy for community‐acquired brain abscess [14]. Although meropenem is listed as an alternative to metronidazole, they do not clearly define the clinical indications for its use [14]. In our case, the patient developed metronidazole‐induced peripheral neuropathy, creating a clinical dilemma regarding whether anaerobic coverage should be maintained despite negative anaerobic cultures. Although the guidelines recommend tailoring antimicrobial therapy according to culture results, they also support continued anaerobic coverage, based on expert opinion, when oral cavity bacteria are identified [14]. Since the brain abscess originated from sinusitis, and S. intermedius , an oral commensal, was isolated by culture, we decided to continue anaerobic coverage. Subsequently, F. vincentii detection using mNGS provided evidence supporting our decision despite negative anaerobic cultures. The risk of metronidazole‐induced peripheral neuropathy has been suggested to increase with higher cumulative doses [15]. This represents a potential complication in the treatment of brain abscesses, which typically require prolonged and high‐dose antimicrobial therapy. Our case suggests that, because identification of the causative microorganisms in brain abscesses is often challenging, mNGS can support antimicrobial decision‐making when conventional cultures cannot guide anaerobic coverage.

Our case suggests that mNGS may be a useful adjunctive diagnostic tool in the management of bacterial infections, particularly in situations where the involvement of anaerobic pathogens is suspected. In brain abscesses, specimens are often collected after the initiation of antimicrobial therapy because of anatomical or clinical constraints. The detection of anaerobic bacteria may be particularly challenging because of strict culture requirements, potentially leading to an underestimation of their prevalence. With the wider adoption of mNGS, causative pathogens can be identified even under such conditions. In addition, in this case, brain abscess diagnosis was straightforward based on typical brain MRI findings and the clinical course; however, differentiation from brain tumors based on imaging findings alone may occasionally be challenging [16]. In such cases, identifying the causative pathogen using mNGS itself may contribute to an accurate diagnosis. The accumulation of microbiological data is expected to clarify the etiological landscape of brain abscesses; support antimicrobial de‐escalation; and reduce the incidence of adverse effects, emergence of antimicrobial resistance, and unnecessary antibiotic use.

In this case, the clinical utility of mNGS was limited by delayed testing and turnaround time. Since mNGS was performed after metronidazole‐related adverse effects had developed and the specimen was batch‐processed with other specimens as a part of a research protocol, the results became available on hospital Day 33. By then, the treatment strategy had already been decided; thus, the mNGS results did not alter the ongoing management. Nevertheless, identification of the causative pathogen by mNGS reinforced continued anaerobic coverage and prevented inappropriate antimicrobial de‐escalation. Accordingly, this led to the minimum necessary duration of antimicrobial therapy, minimization of adverse effects, and prevention of antimicrobial resistance. In this case, even if mNGS results had been obtained immediately following brain abscess drainage, they would have been unlikely to alter the initial empirical therapy; however, they might have enabled earlier antimicrobial de‐escalation and earlier confirmation of the need for anaerobic coverage.

Moreover, mNGS has some limitations. Currently, mNGS is not widely available and remains costly [7]. Additionally, its specificity is relatively low, with the potential for contamination, and the inappropriate interpretation of results may lead to unnecessary or inappropriate antimicrobial use [7]. Unlike conventional culture methods, mNGS does not routinely provide antimicrobial susceptibility data [7]. In this case, although the abundance of Fusobacterium reads was lower than that of Streptococcus, Fusobacterium accounted for approximately 20% bacterial reads and showed broader mapping coverage than other Fusobacterium species. Owing to its established association with the oral cavity and anaerobic sinusitis, its detection was considered biologically plausible. Nevertheless, species‐level assignment based on short metagenomic reads should be interpreted cautiously in the context of the clinical course, imaging findings, the presumed source of infection, and other microbiological data. Further accumulation of cases is needed to clarify the clinical scenarios in which mNGS should be optimally utilized.

In conclusion, this case demonstrates that mNGS can detect anaerobic pathogens that are not identified by conventional culture in a pediatric brain abscess secondary to sinusitis and provide valuable information for therapeutic decision‐making. Because the management of brain abscesses requires high‐dose and prolonged antimicrobial therapy, decisions regarding de‐escalation are clinically important in light of potential adverse effects. In this case, F. vincentii detection using mNGS supported the continuation of anaerobic coverage, thereby reinforcing the rationale for the chosen treatment strategy.

Author Contributions

Risa Nagase: writing – original draft, writing – review and editing. Kunihiro Oba: supervision, writing – original draft, writing – review and editing. Mari Asakura: writing – review and editing. Masashi Ogasawara: writing – review and editing. Masahiro Noda: writing – review and editing. Yusuke Yamada: writing – review and editing. Asayo Furukawa: writing – review and editing. Naoki Kojima: writing – review and editing. Masanori Hashino: writing – review and editing. Kazuhiro Horiba: writing – review and editing.

Funding

The authors have nothing to report.

Consent

Written informed consent for publication of this case report and any accompanying images was obtained from the patient's mother. The nature and purpose of the case report were also explained to the patient in an age‐appropriate manner, and the patient's understanding and wishes were respected to the extent possible.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

This data supporting the findings in this study are available from the corresponding author upon request.