Culture-negative liver abscess identified with plasma microbial cell-free DNA sequencing: A case report

Highlights

• •We report, for the first time, the identification of Streptococcus intermedius in plasma by metagenomic sequencing, the Karius Test, as the etiological organism of pyogenic liver abscesses.
• •Deep sequencing of plasma may represent a less invasive laboratory diagnostic method for the etiological identification of liver abscesses.
• •It warrants further investigation whether elevated blood iron levels predispose individuals to an increased risk of developing multiple pyogenic liver abscesses.

Abstract

Pyogenic liver abscess (PLA) is a potentially life-threatening disease. Early diagnosis and appropriate treatment are crucial to ensure high-quality healthcare for patients with PLA. However, this is complicated by their non-specific clinical symptoms. In addition, the etiologic organisms responsible for PLA are frequently culture-negative, thus complicating clinical decision-making. Here, we report a case of PLA caused by Streptococcus intermedius, as identified via DNA metagenomic sequencing of plasma.

Graphical abstract

1. Case presentation

A 60-year-old Caucasian male with no significant past medical history presented to the emergency room with shortness of breath, headache, and right upper quadrant abdominal pain. He reported feeling fatigued, with headache, fever, chills, and diaphoresis after mowing his lawn a week prior and noted rust-colored urine. Despite receiving fluids for bilirubinuria after a visit to urgent care the previous day, his symptoms worsened. He denied chest pain, hematuria, nausea, rash, dysuria, vomiting, congestion, or sore throat. He also denied recent travel and reported owning four cats.

Vital signs revealed slightly elevated blood pressure (149/78 mmHg) and tachycardia (pulse 114 bpm). Laboratory tests showed elevated B-type natriuretic peptide (BNP), bilirubin, aspartate aminotransferase, and alanine aminotransferase levels. A BNP level of 4,230 pg/mL (about 30 times higher than normal, 125 pg/mL) indicated heart failure. Serum ferritin was significantly elevated at 2,020 ng/mL (normal: 30–400 ng/mL). He had leukocytosis (12,800 cells/μL) with increased granulocytes (10,900 cells/μL) and monocytes (1,600 cells/μL). Magnetic resonance imaging revealed numerous liver lesions in the right hepatic lobe, consistent with an infectious process (Fig. 1), which was consistent with the findings from a computed tomography (CT) scan (Fig. 2). Two sets of blood cultures were negative. He was admitted with a diagnosis of congestive heart failure of unknown cause and liver lesions likely related to infection.

Fig. 1.

Abdomen imaging by MRI. The white arrows indicate multiple irregular, slightly high-density mass in the right hepatic lobe (A and B), with indistinct borders, primarily suggestive of an infectious lesion.

Abbreviation: MRI, magnetic resonance imaging.

Fig. 2.

Abdomen imaging by CT with contrast. Innumerable low-density lesions in the liver on the hospital admission day (A), were decreased in size, after 30 days upon antibiotic treatment (B), as pointed by the white arrows, and continued decrease in size or coalesced together 68 days later (C).

Abbreviation: CT, computed tomography.

After consulting with cardiology and infectious disease specialists, as well as the informal diagnostic management team in Division of Clinical Microbiology, the primary healthcare team adjusted the empirical treatment regimen to target broad-spectrum organisms potentially causing the liver abscess. The revised regimen included ceftriaxone, doxycycline, and metronidazole, followed by an extensive laboratory workup to identify putative etiological agent(s). Both traditional and molecular assays were ordered to identify a broad range of pathogens, including viruses, parasites, bacteria, and fungi (Supplementary Table S1). Biopsy specimens, including liver aspirates and tissues obtained via fine needle aspiration, were submitted for cultures of aerobic and anaerobic bacteria, fungi, and acid-fast bacilli. Additionally, a transthoracic echocardiogram yielded normal results, reducing the likelihood of congestive heart failure exacerbation.

By day 10 of hospitalization, laboratory testing did not yield the isolation or identification of microorganisms. Therefore, a plasma sample was collected for performance of a Karius Test, which is a microbial cell-free DNA metagenomic next-generation sequencing (NGS) assay from Karius, Inc. (Redwood City, CA, USA). The following day, liver biopsy specimens were submitted for broad-range PCR and sequencing at the University of Washington to identify bacteria, fungi, and/or mycobacteria. Five days later, NGS results identified Streptococcus intermedius. This identification was subsequently confirmed by the targeted PCR plus sequencing performed at the University of Washington.

Histopathological analysis of liver biopsy specimens demonstrated consistent evidence of abscess formation, characterized by hepatic parenchyma with prominent portal expansion, fibroblastic proliferation, areas of granulation tissue formation, and patchy areas containing sheets of neutrophils and hemorrhage (Fig. 3). He was initially administered antibiotics, including doxycycline and metronidazole. Given the presence of intermittent fever, the patient was further empirically treated with cefepime. Three days later, the patient had a peripherally inserted central catheter line placed, and he was discharged with extended antibiotic treatment for 6 weeks, including cefepime (1,000 mg daily), doxycycline (100 mg every 12 hours), and metronidazole (500 mg every 8–12 hours). The patient was followed up with infectious disease and vascular disease specialists in an outpatient clinic for 2 months without any significant complaints. CT scans of the abdomen conducted 30- and 68-day post-discharge revealed a continuous decrease in the size of liver lesions (Fig. 2), with no other medical concerns.

Fig. 3.

Histopathological analysis demonstrates patchy areas of neutrophilic inflammation (A) and the inflammatory areas were surrounded by a myofibroblastic response with associated edema and fibrosis (B) after H&E stain. Bar = 50 µm.

2. Discussion

A PLA represents a severe disorder associated with significant morbidity and mortality. Diagnosis of PLA can be challenging due to its non-specific clinical presentations.¹ Patients with PLA are primarily managed with antibiotic treatment and percutaneous abscess drainage. Delayed or incorrect diagnosis, particularly the inability to identify an etiological agent, compromises the appropriate patient care.¹ Common pathogens responsible for PLA are Klebsiella pneumoniae and Escherichia coli, and members of Streptococcus anginosus group. In addition, anaerobes such as Bacteroides species and parasites such as Entamoeba histolytica can also cause PLA.²

Metagenomic next-generation sequencing (mNGS) is a high-throughput method that rapidly detects pathogens directly from clinical specimens, enhancing the identification of hard-to-detect pathogens. mNGS is a tool that computationally analyzes and compares sequences from a single specimen to identify mixed genomes.³ This methodology has been increasingly applied in laboratory diagnosis of etiological organisms via specimens such as blood, bronchoalveolar lavage, and cerebrospinal fluid.⁴ The Karius Test, a commercially available microbial cell-free DNA (mcfDNA) sequencing method using whole-blood samples, has received FDA approval.⁵ The test workflow involves collecting whole-blood samples, extracting DNA, deep sequencing, and analyzing the data using Karius's proprietary pathogen database. The results are reported as molecules per microliter values, which indicate the concentration of circulating cell-free DNA detected in the plasma for each pathogen. While the exact technical aspects of the Karius Test are not fully available, Blauwkamp et al.⁶ indicate that the test sequences approximately 24 million reads per sample using a single-end 75-base sequencing method. Sequencing depth influenced the limit of detection in the study; however, the processing methodology suggests that it does not affect overall sensitivity. Ninety-five percent of the validation samples had at least 181,000 unique whole-assay internal normalization control molecules. To minimize false positives, several strategies were employed. Both internal controls (unique synthetic DNA sequences) and external batch controls (environmental samples and known concentrations of specific bacterial organisms) are incorporated during the automated DNA extraction and NGS library preparation process. Bioinformatics techniques are also employed to reduce the false positive results arising from genetically similar organisms or the natural diversity of clinical isolates compared to reference sequences. Additionally, environmental control samples are included in each sequencing batch to establish a noise threshold, ensuring that only significant read levels are reported.

The Karius Test offers several advantages over traditional microbiological testing, including expedited diagnosis, tailored clinical management, enhanced sensitivity and specificity, broad-spectrum pathogen detection, and non-invasive sample collection methods. The clinical significance of the Karius Test is still debated, influenced by date of sample collection, the interpretation of the results, and its cost-effectiveness. Future considerations for the Karius Test include identifying a patient population, optimizing testing time, integrating this test into current algorithms, and broadening its application in more clinical contexts.⁷

In this case, mcfDNA sequencing identified Streptococcus intermedius in plasma five days after the specimen was received. Additionally, compared to blood drawing, fine needle aspiration for collecting liver tissues or fluids carries a potential risk of complications, such as infection or damage to surrounding organs. Therefore, mcfDNA sequencing offered not only a rapid but also a feasible diagnostic tool for determining PLA etiology. If metagenomic sequencing is implemented during the initial phase of the infection in this patient, the results may offer rapid and definitive guidance for the antibiotic regimen. Fortunately, in this case, the patient ultimately received successful empirical therapy. Furthermore, another limitation of the study is its small sample size. A broader clinical investigation is warranted to confirm the utility of the assay.

Streptococcus intermedius is a Gram-positive, non-motile, catalase-negative member of the Streptococcus anginosus group. Although S. intermedius is part of normal microbiota in the mouth and gastrointestinal tract,^8,9 it is also medically important as one of the most common pathogens associated with brain and liver abscesses.¹⁰ Before the availability of advanced molecular biology tests, S. intermedius was often neglected during polymicrobial infections as they are considered part of the human flora. With the implementation of molecular diagnostic methods in clinical settings, S. intermedius is now widely recognized as an opportunistic pathogen causing various types of empyema and abscesses in the head and neck, lung, heart, and liver.⁹ For the patient presented in this report, we were not successful in recovering S. intermedius from either the liver biopsy or blood specimens, under aerobic or anaerobic culture conditions. The possible reasons for the cultures being negative may be two-fold. First, both liver aspirate or tissues and blood were collected following empirical treatment with broad-spectrum antibiotics. Administering antibiotics prior to sampling is one of the most common factors that reduce the sensitivity of laboratory assays, especially bacterial cultures. Secondly, S. intermedius is infrequently grown as a pure culture from the infection site and may necessitate specialized anaerobic conditions due to its fastidious growth characteristics.11., 12., 13. This differs from other coagulase-negative staphylococci, which are frequently isolated.

There is currently no consensus on the treatment management for patients with confirmed S. intermedius infection.¹⁴ While intravenous antibiotics are generally considered first-line treatment, the specific regimen is typically guided by the antibiotic susceptibility profile of the isolated S. intermedius and the site of infection.^14,15 Recommendations for antibiotic treatment duration range from two to six weeks; however, no standardized guidelines currently exist.¹⁵ The choice of antibiotics varies widely and includes macrolides, aminoglycosides, cephalosporins, glycopeptides, lincosamide, penicillin, quinolones, rifamycin, streptogramins, and tetracyclines.¹⁶ A study by Moskatel et al.¹⁶ reported that, among 42 cases of S. intermedius infections, glycopeptides, specifically vancomycin, were the most commonly selected antibiotic class, followed by penicillins. In this case, the empiric antibiotic treatment regimen included coverage for S. intermedius and likely contributed to the observed reduction in the size of the liver lesions. One limitation of NGS-based diagnosis is the inability to determine the pathogen's antibiotic susceptibility unless there is a complete understanding of the antibiotic-resistance genes associated with S. intermedius.

In the present report, the patient presented with heart failure, but the direct cause remains unclear. This is possibly associated with the severity of PLA or bacteremia. Although the patient did not have a clear immunocompromised status, he had a significantly elevated ferritin level in the blood, nearly five times greater than normal. Ferritin is an iron-storage protein in cells and indicator of serum iron levels.¹⁷ Iron has been shown to be associated with increased pathogenicity of bacteria that cause PLA, potentially linked to the severity of the infection in this case. For example, Klebsiella pneumoniae has shown increased virulence, growth, and biofilm formation when cultured in an environment with iron compared to cultures without iron.¹⁸ Yersinia enterocolitica demonstrates a similar pattern of elevated serum ferritin, suggesting a potential connection between ferritin levels and the development of pyogenic liver abscesses.¹⁹ In our patient, numerous liver lesions were observed on CT and magnetic resonance imaging and could potentially have been associated with the high iron load. Future studies on the growth and virulence of S. intermedius with iron are required to further understand this potential association. Our case highlights that the ferritin level in blood is one of the possible risk factors for developing invasive bacterial infections, such as liver abscesses.

3. Conclusions

We present a case of a liver abscess caused by Streptococcus intermedius that was culture-negative and identified by metagenomic sequencing in plasma. To our knowledge, this is the first report of Streptococcus intermedius being identified by mNGS in plasma as the etiological organism for PLA. Multidisciplinary consultation as a team facilitated the diagnosis and management of the patient. Innumerable liver abscesses were successfully treated with antibiotics alone, without recurrence or sequelae.

CRediT authorship contribution statement

Yuanchao Xue: Writing – review & editing, Writing – original draft, Data curation, Conceptualization. Filipe M. Cerqueira: Investigation, Conceptualization. Heather L. Stevenson: Investigation. Natalie Williams-Bouyer: Investigation, Conceptualization. Rong Fang: Writing – review & editing, Writing – original draft, Supervision, Investigation, Formal analysis, Conceptualization. All authors have read and agreed to the published version of the manuscript.