Fusobacterial liver abscess: a case report and review of the literature

Fusobacteriae are arguably unique amongst non-spore forming anaerobes in their ability to cause severe, clinically distinct, mono-microbial infections in a range of organs, often with severe sequelae as originally described by Lemierre and others in the first half of the twentieth century [1]. Fusobacterial infections can affect the head and neck, the lung and pleura, the gastro–intestinal organs, and the female genital tract [2]. Despite their ability to cause severe infections, fusobacterium species are uncommonly identified as pathogens in clinical practice, accounting for less than 1% of anaerobic bacteremias. Regional population-based studies have described the incidence of Fusobacterial bacteraemia to be around 0.55 per 100,000 people per year [3, 4], and that of clinically significant infections to be 0.99 per 100,000 people per year [2]. Whilst the clinical presentations associated with fusobacterial infections are well defined, little is known about the clinical progress and outcomes of specific syndromes [1, 5, 6]. Fusobacteriae are a particularly uncommon cause of pyogenic liver abscesses. Therefore, our objective is to describe a case of this condition and to summarise the presentation and outcomes of Fusobacterial liver abscesses through a review of the relevant literature.

A 51-year-old previously fit and well woman who recently returned from Samoa presented with a four-day history of watery non-bloody diarrhoea, associated with generalised abdominal pain, fevers, chills, and lethargy. Family members reported similar symptoms of milder severity. On admission, she was febrile with a temperature of 38 °C, with otherwise unremarkable vital signs. Examination of the abdomen found a palpable mass in the right upper quadrant. There were no signs of periodontal disease on inspection of the oral cavity. Laboratory tests showed a white cell count of 26.7 × 10E9/L, deranged liver enzymes (alanine transaminase 263 U/L, aspartate transaminase 177 U/L, bilirubin 15 mg/L), and a C-reactive protein of 407 mg/L. A Computed Tomography (CT) scan of the abdomen with contrast revealed a 90 × 92 × 93 mm mixed density lesion in segment II and III of the liver The lesion was percutaneously drained under ultrasound guidance. Cultures of the aspirated abscess fluid and blood both grew anaerobic gram-negative bacilli. Following abscess drainage and treatment with ceftriaxone and metronidazole, the patient continued to experience fevers and diarrhoea and her inflammatory markers did not improve. A follow-up ultrasound scan of the liver revealed an emerging focal abscess in segment IV-b of the liver. The initial CT scan was then retrospectively reviewed to reveal subtle heterogeneous changes in this segment; however, no demarcated fluid collection was present at the time. The two abscesses were not contiguous with but in close proximity to each other. This abscess was also percutaneously drained. Samples from the blood culture were further identified as Fusobacterium nucleatum via Biomerieux Vitek (Knowledge database version 2.0) MALDI-TOF (Matrix assisted Laser Desorption/Ionisation - Time of Flight) mass spectrometry with a confidence of 99.9%. Cultures of the initial aspirate were not identifiable using MALDI-TOF and were referred to a quaternary level laboratory in Auckland and later identified as Prevotella pleuritidis through 16 s ribosomal Ribonucleic Acid identification. Both specimens underwent susceptibility testing using Clinical Laboratory Standards Institute (CLSI) – recommended media and breakpoint determination using MIC gradient strips. The Prevotella pleuritidis was tested on Brucella agar with sheep blood, haemin, and vitamin K, while the F. nucleatum required Brain Heart Infusion agar to achieve satisfactory growth.

Given the growth of a second organism common to the periodontal flora and in the absence of symptoms, signs, or imaging evidence of bowel pathology, a periodontal, rather than a gastrointestinal, source was considered more likely and no further evaluation of the large bowel was undertaken.

Treatment was changed on day nine to a two-week course of IV ertapenem, prior to susceptibilities being known, followed by four further weeks of oral amoxicillin-clavulanate. The patient recovered fully.

Although Fusobacterium was only grown in the blood culture and not in the aspirate culture, we were confident in treating this case as a Fusobacterial pyogenic liver abscess. The aspirate sample was collected afterhours and was not processed until 7 hours later. This delay could have significantly affected our ability to identify this anaerobic organism using standard culture methods, especially in the setting of antimicrobial therapy.

Whilst the essential features of disease caused by Fusobacteriae have been recognised for over a century, presentations remain uncommon and widely distributed across medical specialties [1]. Thus the historical background, aetiopathogenesis, and likely clinical outcomes of fusobacterial disease are poorly understood by most clinicians. This report demonstrates the common features of fusobacterial liver abscess and by combining individual reports and case series provides insights into the associated spectrum of underlying illness.

Loeffler identified a thin, gram-negative rod present in necrotic tissue in calf diptheriae in 1884 [7]. The term ‘necrobacillosis’ was coined by Bang, who isolated the organism from diseased cattle and pigs [8]. Koch’s postulates were satisfied accidentally when Schmorl and an assistant suffered pyogenic infection of their fingers whilst working with diseased rabbits, samples from both of which yielded Fusobacteriae [5]. Cases of de novo necrobacillosis were described in humans by the turn of the twentieth century. Lemierre gave his classic dissertation on anaerobic bacteremia in 1936, earning the right to an eponymous syndrome of tonsillar abscess progressing to internal jugular vein thrombosis, distant abscess formation, and overwhelming sepsis [9].

In the twentieth century it was axiomatic, based on very little evidence, that Fusobacteriae were part of the normal human oropharyngeal, gastro-intestinal (GI) tract and female genital tract flora. The apparent rarity of illness caused by these organisms reinforced the notion that they were opportunists, causing illness only when natural barriers to infection were compromised [1, 10, 11]. Contemporary studies have shed light on the true environmental niche and pathogenic potential of Fusobacteriae in general and Fusobacterium nucleatum specifically. Fusobacterium nucleatum is a major constituent of the gingival, oropharyngeal and together with F. necrophorum, the appendiceal flora. Neither organism is a component of the healthy colonic flora nor do they form a major component of the healthy vaginal micro-biome [1, 12, 13]. F. necrophorum was found to exist in the healthy oropharyngeal flora in only one polymerase chain reaction (PCR) based study reported and in none of the culture-based studies [1]. Gingivo-periodontitis, endodontic infection and suppurative appendicitis occur in the setting of clonal expansion and invasion of Fusobacteriae, and it has been postulated that adverse obstetric outcomes arise from haematogenous spread of F. nucleatum to placental tissue [14, 15]. Prevotella pleuritidis was identified in the case we described and the gingival/oropharyngeal flora were identified as co-pathogens in 18% of liver abscess caused by F. nucleatum in our review of the literature.

Indeed, our review of fusobacterial pyogenic liver abscesses showed a low prevalence of general risk factors for liver abscess including malignancy, dialysis treatment, and older age [4]. Most cases occurred in young or middle-aged, immunocompetent individuals with risk factors for haematogenous spread of Fusobacteriae such as periodontal disease and recent pharyngitis.

Attribution of the lower GI tract as a source of liver abscess was relatively uncommon and the grounds for doing so were often unclear. The utility of investigating for a presumed source is unknown at this point due to the lack of research assessing this particular question. Our findings support the contention that the majority of disseminated fusobacterial infections originate from the oral cavity even without symptoms of a dental disease. Dental X-Rays to identify possible dental abscesses and cavities or more sophisticated imaging modalities of the oral cavity and the sinuses may be of the most benefit, should clinicians choose to investigate the underlying sources of bacteremia in patients without localising symptoms [16].

Among cases where F. necrophorum were cultured, only three of the 23 liver abscesses were associated with a co-pathogen. None of these cases had clinical findings consistent with classic Lemierre syndrome. Whilst the environmental niche of F. necrophorum is less understood than that of F. nucleatum, it is among the most commonly identified cause of tonsillar abscess after Lancefield Group A beta-hemolytic streptococci [17, 18]. This potentially points to a more evolved role for F. necrophorum as a human pathogen.

We observed good outcomes with appropriate source control and antimicrobial therapy. The only case involving a fatality was a post-mortem diagnosis and treatment had not been initiated. There was significant variation between studies with regards to the choice and duration of antimicrobial therapy. Current evidence suggests the need for prolonged antimicrobial therapy for patients with liver abscess and risk factors for treatment failure. These include malignancy, older age, presence of septic shock at presentation, certain biochemical derangements (e.g. anaemia, azotaemia, and hyperbilirubinaemia), and an Acute Physiology and Chronic Health Evaluation (APACHE) II score of more than 15 [19‐21].

Routine prolonged broad-spectrum antimicrobial therapy may not be warranted in cases of fusobacterial liver abscess with a presumed oral or periodontal origin, especially in the setting of abscess drainage. Fusobacteriae remain highly susceptible to beta-lactams. Studies on susceptibility patterns have shown no increase in amoxicillin-resistant Fusobacteriae over the past decade [22‐24]. They are unlikely to be accompanied by highly resistant co-pathogens in polymicrobial infection of oral or periodontal origin. Metronidazole has been associated with good clinical responses in the treatment of fusobacterial and mixed anaerobic infections [25]. Where source control is achieved, and with the aid of reliable clinical biomarkers such as C-reactive protein, antimicrobial therapy may be shortened [26, 27]. Further research into shorter course therapy for highly sensitive organisms such as Fusobacteriae could better delineate optimal antimicrobial treatment duration for patients with liver abscesses.

The strengths of this review are that it employed an extensive search strategy to identify a greater number of cases than previous reviews on the same topic. Previous literature reviews of case reports have focused mainly on only one Fusobacteriae species (F. necrophorum or F. nucleatum) [28, 29]. To our knowledge, this is the first review to include all Fusobacteriae spp., to demonstrate their homogeneity in clinical presentation, clinical findings, co-pathogens and outcomes.

Our study has several weaknesses. The main weakness is that in the case presented, Fusobacterium nucleatum was identified in the blood culture but not in the abscess aspirate. We remain confident in attributing F.nucleatum as a causative co-pathogen for a number of reasons. Firstly, delays in sample processing have been known to increase the difficulty in identifying anaerobes through conventional culture methods [30], especially in the setting of antimicrobial therapy, which was commenced in advance of abscess drainage. Studies using pyrosequencing of direct samples have shown that pyogenic abscesses are frequently polymicrobial, with the number of aetiologic agents identified significantly exceeding what is achieved using standard culture-based methods [30, 31]. Secondly, many population-based studies accept blood cultures as an appropriate standard in diagnosing causative organisms and in guiding antimicrobial treatment in pyogenic liver abscess [32]. Some cases included in our review identified Fusobacteriae by blood culture alone [33, 34]. Therefore, we have grounds to believe that F. nucleatum was indeed a causative agent.

In terms of other weaknesses, we reported only published observations; publications may have been biased towards those of specific interest to a journal’s readership or those that were associated with good clinical outcomes. Cases in languages other than English were excluded. Attribution of underlying source was presumptive rather than definitive in a significant number of case series. Furthermore, there was minimal information on microbiologic methods. This may have led to an over-reporting of mono-microbial infection, for example, where culture-based methods lacked sensitivity or where growth of the anaerobic flora at low quantities was under-reported.

In this case report and review of the literature, we found that fusobacterial liver abscesses resulted in good outcomes with contemporary medical and surgical care.