Clostridioides difficile Infection-Related Hepatic Abscess: A Review of the Literature

Abstract

Hepatic abscesses (HAs) are suppurative cavities caused by the infiltration of liver parenchyma by microbes. The etiology of HAs can include biliary disease, intra-abdominal infections, bile duct ischemia following pancreaticoduodenectomy, liver transplantation, and rarely septicemia, among others. While HA of a bacterial origin is rare, intra-abdominal biliary infections can be seen after appendicitis or proctitis. Empiric treatment generally involves antibiotic coverage and percutaneous drainage. Due to a high health care association, Clostridioides difficile infection (CDI) is becoming more frequent and has increasing antibiotic resistance. While generally limited to the gastrointestinal tract, extra intestinal CDIs are generally related to bacteremia, and related HAs are even rarer, with their prognostic courses not well described. A literature search was performed evaluating case reports and manuscripts relating to CDI-related HA. Articles with relevant information were analyzed for quality. Those fitting predetermined selection criteria and the level of evidence were included in this study. While most affected patients reported recent hospitalization and significant comorbidities, some of them reported no apparent risk factors at all. While all included individuals had HA, only 33.3% had CD-related bacteremia. Treatment of HA was managed with drainage as well as antibiotics and was generally successful, with one death. As CDI becomes more prevalent, studies such as this will be needed to address the management of associated HA.

1. Introduction

Hepatic abscesses (HAs) are suppurative cavities caused by the infiltration of liver parenchyma by microbes. The etiology of HAs can include biliary disease, intra-abdominal infections, bile duct ischemia following pancreaticoduodenectomy, liver transplantation, and rarely septicemia, among other causes [1]. Very often, intra-abdominal biliary infections contaminate the biliary tree or seed the portal venous system and can commonly be seen after appendicitis or proctitis [1]. HA of a bacterial origin is even more rare with a mortality of 15% [2]. Overall management includes early antibiotic initiation after blood cultures are taken and before drainage to limit systemic spread but to obtain accurate culture data [1]. Generally, HA diagnostic imaging includes sonography or a CT scan, and in up to 90% of cases, imaging may also assist in describing the etiology, whether it is related to diverticulitis-related seeding, other intra-abdominal infection, or post instrumental seeding. This article will describe non-parasitic-related HA. General management of HA that is non-parasitic includes antibiotics, percutaneous drainage, and treatment of the underlying disease or cause [1]. Empiric coverage generally includes Gram-negative bacilli (such as Escherichia coli and Enterobacter), Gram-positive cocci (including Staphylococcus aureus and Streptococcus), and anaerobes. While HA is rare overall, its incidence is hard to define as it varies by geographic location, related pathogens, and etiology of the HA itself and has differing clinical severity. The duration of treatment varies per clinical picture, but generally speaking lasts two to six weeks [1].

Clostridioides difficile infection (CDI) is one of the most frequent health care-associated infections. Due to its prevalence, antibiotic resistance is high and effective treatment has become a challenge worldwide [3]. Its clinical severity ranges from diarrhea to colitis and death, with symptoms primarily being related to the intestinal tract. Antibiotic resistance development of CDI is variable and antibiotic selection can differ from region to region; thus, resistances to each antibacterial vary and there have been reported resistances to clindamycin as well as fluoroquinolones [3]. Extra intestinal CDI (Ex-CDI) is a rarely reported manifestation, with even fewer case studies being reported related to Ex-CDI bacteremia. As CDI affects the intestinal tract, there is potential for seeding into the portal venous system, the likely pathophysiology of the majority of CDI-related HA [1]. These occur mainly in hospitalized individuals with significant comorbidities and have a one-year survival rate of 36% correlating with severity of the underlying CDI [4]. CDI-related hepatic abscesses are rare with only six cases reported to our knowledge. While general management of HA and CDI have been separately described, there are few cases described of post CDI-related HAs and their prognostic course. This review hopes to identify and analyze the gaps associated with the etiology and treatment for CDI-related HA.

2. Materials and Methods

We performed a comprehensive literature search across three databases (Pubmed/Medline, Embase, and Cochrane) using variations of the keywords “Liver” or “hepatic”, “abscess”, and “Clostridioides difficile” to identify original studies published from inception to October 19, 2023. Results were limited to human studies published in English that provided patient data. Studies pertaining to animal subjects or studies without patient data were excluded. There was a total of 647 studies for review. Two authors (T.V. and K.T.) independently reviewed these titles and abstracts, after which seven case studies were deemed relevant. Full texts were pulled and then reviewed by at least two of the following authors (T.V., M.C., K.T.). One article was removed as it was unable to be accessed. In case of disagreement, a senior reviewer (K.T.) arbitrated the final decision for inclusion. In the final selection process, six articles were deemed to fit with all inclusion criteria and only studies with at least a moderate, fair and above, level of evidence were included. Due to scarcity of data on the subject, a study describing a perihepatic liver abscess secondary to CDI as well as case reports were included. Quality appraisal was performed by at least two of the following authors (T.V., M.C., K.T.). If there was any disagreement, a senior reviewer (K.T.) evaluated the article and achieved a consensus through discussion. All data were retrieved via public open access data review; thus, all patient data had already been de-identified and thus further patient permission was not required. See

Table 1. NOS Quality Assessments per Study [5].

Author/Year	Morioka, 2017 [7]	Toprak, 2016 [8]	Roy, 2017 [9]	Sakurai, 2000 [10]	Tun, 2022 [11]	Acharya, 2022 [12]
1. Was the study question or objective clearly stated?	1	1	1	1	1	1
2. Was the study population clearly and fully described, including a case definition?	1	1	1	1	1	1
3. Were the cases consecutive?	0	0	0	0	0	0
4. Were the subjects comparable?	0	0	0	0	0	0
5. Was the intervention clearly described?	1	1	1	1	1	1
6. Were the outcome measures clearly defined, valid, reliable, and implemented consistently across all study participants?	1	1	1	1	1	1
7. Was the length of follow-up adequate?	1	1	1	1	1	0
8. Were the statistical methods well described?	0	0	0	0	0	0
9. Were the results well described?	1	1	1	1	0	0
Total quality (total number of 1)	6	6	6	6	5	4
Quality (7–9: Good; 4–6: Fair; 1–3: Poor)	Fair	Fair	Fair	Fair	Fair	Fair

for NOS quality assessment scores for each study [5]. Study selection was completed via the Preferred Reporting Items for Systematic Re-views and Meta-Analyses (PRISMA) statement and is detailed in Figure 1 [6].

3. Results

Overall, the patients included had a mean age of 74.5 (range of 63–83) and reported recent hospitalizations or significant comorbidities such as GI malignancy, liver carcinoma, or intestinal perforation [7,8,9,10,11,12]. A summary of the included studies is provided in

Table 2. Summary of Included Studies.

Author	Study Type	Country	Sex	Age	Risk Factors	Antibiotics Used for CDI Treatment	Procedural Intervention	Patient Outcome	Bacteremia
Morioka, 2017 [7]	Case Report	Japan	M	74	liver cirrhosis, HCC, antibiotic exposure, proton pump inhibitor use, TACE	Metronidazole then Vancomycin	Percutaneous drainage	Clinical and radiographic resolution	Y
Toprak, 2016 [8]	Case Report	Turkey	F	80	cholecystitis, diabetes mellitus, age	Ceftriaxone and Metronidazole	Laparotomy	Expired	N
Roy, 2017 [9]	Case Report	USA	M	63	antibiotics for prior CDI, sigmoid perforation, recent hospitalization	Metronidazole then Vancomycin	Percutaneous drain	Clinical and radiographic improvement over 3 months	N
Sakurai, 2000 [10]	Case Report	Japan	F	75	none	Imipenem/Cilastatin then Flagyl	Percutaneous drainage	Clinical and radiographic resolution at 14 months	Y
Tun, 2022 [11]	Case Report	USA	M	72	sigmoid adenocarcinoma with partial obstruction with exploratory laparotomy and diverting loop ileostomy placement	Fidaxomicin, Vancomycin, Tigecycline	Percutaneous drainage	Clinical improvement at 5 months	N
Acharya, 2022 [12]	Case Report	USA	F	83	none	Flagyl then Vancomycin	None	NR	N

CDI: Clostridioides difficile infection, TACE: transarterial chemoembolization, HCC: hepatocellular carcinoma, NR: not reported.

. The general reported management of HA generally included a percutaneous drain as well as a course of antibiotics for the underlying CDI and only two individuals had associated bacteremia [7,10]. Apart from Toprak et al.’s case of patient demise following a prolonged hospitalization course, following treatment, the rest of the patients had clinical and radiologic improvement during their respective follow-up periods [8].

Morioka et al. described a recurrent case of CDI-related bacteremia and a liver abscess in a 74-year-old male. He had hepatocellular carcinoma (HCC) complicated by primary biliary cirrhosis (PBC) and was hospitalized for a transarterial chemo-embolism (TACE) to the lesion. This patient had previously been treated with an unknown antibiotic and proton pump inhibitor (PPI). Post-operatively, he had new-onset diarrhea and fevers. At the initial presentation of symptoms, a CT abdomen scan was completed and stool cultures were negative. Ten days later, the patient had worsening symptoms and was found to have two blood cultures with C. difficile growth as well as a CT abdomen and pelvis scan notable for an edematous colon. He was transitioned to intravenous (IV) vancomycin and oral metronidazole and then transitioned later to oral vancomycin for discharge. He returned one month later for recurrence of a fever and was found to have a liver abscess at the same site of the previous TACE treatment. Aspiration was completed at the site of the newly noted abscess, and he was started back on oral metronidazole and eventually he had a resolution of symptoms. Ex-CDI as a complication of TACE without bilioenteric anastomosis lends to the uniqueness of this patient presentation [7].

In Toprak et al.’s study, an 80-year-old woman with a known history of carotid angioplasty completed for the treatment of facial hemiparesis and insulin-dependent diabetes, and no history of diarrhea or antibiotic use, presented with abdominal pain, fever, and chills [8]. During her workup, she was noted to have elevated liver enzymes leading to evaluation with a hepatobiliary ultrasound. Ultrasound showed a hydropic gallbladder with sludge and stones and within the liver were several hypoechoic, heterogenous, well-demarcated lesions with irregular walls. CT of the abdomen showed multiple hypodense lesions in the left and right lobes of the liver, suggesting malignancy or an abscess. She was started on empiric piperacillin/tazobactam and underwent cholecystectomy and laparotomy and was transferred to the ICU for further evaluation and management. A gram stain was obtained and showed toxin-positive C. difficile growth. Stool studies were performed at that time and were also positive for C. difficile. The patient was transitioned to ceftriaxone and metronidazole. On post-operative day 11, her liver enzymes normalized, clinical improvement was noted, and antibiotic treatment was stopped after day 14. Her status acutely worsened the next day, and she was started on imipenem and metronidazole; however, she continued to rapidly deteriorate despite medical intervention and ultimately succumbed to illness on day 18 of hospitalization.

Roy et al. described a 63-year-old man with medical history of acid reflux, hypertension, and hypothyroidism with a presentation of acute-onset abdominal pain [9]. He was noted to have elevated lipase without alcohol use. A CT abdomen and pelvis scan showed acute pancreatitis. During his hospitalization, he had increasing abdominal pain as well as an acute drop in hemoglobin, so a repeat CT was completed. The imaging showed necrotizing pancreatitis with concerns for hemorrhagic pancreatitis, so he was started on piperacillin-tazobactam. He then developed diarrhea and a stool PCR taken was positive for C. difficile. He was started on metronidazole then transitioned to vancomycin as symptoms did not resolve with the prior treatment. He had an even further complicated hospital course with pleural effusions, acute respiratory distress syndrome, acute splenic vein thrombosis, and pseudocyst formation. Eventually with treatment the diarrhea improved, and the patient stabilized and was able to be discharged. He returned a week later with fevers, chills, and right-sided back pain. Stool PCR was taken; however, it was indeterminate for active toxin production versus the carrier state. A repeat CT showed fluid air collection in the pancreas and a large subcapsular hepatic fluid collection. He was started on meropenem and had percutaneous drains placed into the lesions in the pancreas as well as the liver. Unfortunately, due to a worsening clinical status, the patient did require emergent exploratory laparotomy with stool found in the abdominal cavity secondary to a perforated sigmoid colon. Fluid cultures from the liver and pancreas fluid grew heavy C. difficile and Candida glabrata. Blood cultures did remain negative during the course. Metronidazole, vancomycin, and micafungin were added to the antibiotic regimen. Repeat body cultures were negative, clinical status improved, and he was eventually discharged with monthly CT scans for monitoring.

Sakurai and colleagues’ case describes a 75-year-old woman with a previous history of a sterile nonmalignant liver cyst that had been treated with a drainage catheter. Fluid studies taken at that time were consistent with a benign cyst. Eventually, the cyst had clinically resolved, leaving an asymptomatic hyperechoic area [10]. Two years after the initial cyst was found, she had fevers and malaise without any history of diarrhea or antibiotic use. She was noted to have palpable hepatomegaly as well as elevated liver enzymes. At the workup, an abdominal ultrasound showed an enlarged hepatic cyst. CT showed two areas interpreted as an infected cyst and an HA. They started the patient on imipenem with cilastatin. The larger “infected cyst” was percutaneously drained and aspirates taken grew C. difficile, and thus she was transitioned to vancomycin on hospital day 14. The HA was drained on hospital day 16 and its aspiration was also positive for C. difficile growth. Vancomycin was administered locally to the cyst and abscess sites. She was treated for 4 weeks and the clinical picture improved; however, repeat drainage samples from both lesions still grew C. difficile. Drainage catheters were withdrawn as the patient declined operative management and she was discharged on oral metronidazole. At a six-month follow-up, the lesions had disappeared. Then, 11 months later, she was readmitted with recurrence of both the cyst and abscess. Percutaneous drainage was performed on day two of hospitalization and recurrent C. difficile growth was noted, although stool cultures were negative. She received imipenem with cilastatin and improved after three weeks. Aspirates from her drain sites continued to grow C. difficile. Thus, when they were removed, she continued on four more weeks of metronidazole. At a 14-month follow-up, she was clinically well, and the lesions had disappeared on CT.

Tun et al. describe a 72-year-old male with sigmoid adenocarcinoma complicated by a partial obstruction requiring sigmoid colectomy and loop ileostomy [11]. He developed an anastomotic leak that required exploratory laparotomy and was subsequently found to have a perihepatic abscess anterior to the left hepatic lobe. A drainage catheter was placed. Culture from the abscess grew C. difficile with concomitantly positive stool studies for the same organism. He was treated with fidaxomicin and rectal vancomycin due to the presence of the loop ileostomy. He also received tigecycline, and the drain was removed. A few days later, repeat imaging showed fluid in the location of the prior drain consistent with hematoma. CT-guided drainage revealed bloody material without the growth of organisms.

In Acharya’s case, an 83-year-old woman with previous cholecystectomy, appendicitis, and renal stones presented with worsening right-upper-quadrant pain for four months [12]. At the workup, she was found to have a large liver abscess. The abscess was biopsied and cultured, resulting in growth of C. difficile. She had no history of previous C. difficile infection and was treated with seven weeks of metronidazole. Two to three months later, she presented with pleurisy and a cough and underwent a pleural biopsy. The biopsies were notable for organizing pneumonia and she was noted to have developed right-lower-lobe empyema. With the significant history of recent pulmonary intervention, this case describes a possible needle track seeding of CDI during access using a trans-pleural route of aspiration.

4. Discussion

All in all, Ex-CDI is a rare presentation of CDI with about 45% of Ex-CDI associated with the inoculation of the pathogen through surgical intervention [11]. In a couple of the reported cases, the patients had no previous history of diarrhea or known antibiotic use [8,9,10,12]. In other reports, some patients had previous unrelated hospitalizations, so it is theorized that they were colonized at a point during their hospitalization or had already been colonized at some point in the community. Regarding the exact etiology and timeline of the development of C. difficile-related infections leading to a hepatic abscess, it is likely that the method of colonization, whether history of antibiotic use or previous hospitalization, has less to do with the occurrence or method of inoculation itself, and more so with underlying health and comorbidities of the affected individual and the increased possibility of seeding. HAs are generally polymicrobial with commonly isolated species being Streptococcus, Escherichia coli, and Klebsiella pneumoniae. As there is not a current empiric treatment method for CDI-related HA, the timely administration of antibiotics and drainage can still generally lead to mortality in 5–30% of known cases [13]. Risk factors for higher mortality include a large abscess size, anaerobic bacterial pathogen, gas formatting bacterium, and need for open surgical intervention [13]. Some included patients had surgical intervention previously that may have exposed them to inoculation via translocation through the intestinal lining or ascent into the biliary tract such as the total hysterectomy reported in Sakurai et al.’s case, possibly during the TACE procedure in Morioka et al.’s case, and in Acharva et al.’s case during either the patient’s previous appendectomy or cholecystectomy [7,10,12]. Or perhaps this applies to direct inoculation such as in Tun et al.’s study due to surgical dehiscence or Roy et al.’s patient with complicated pancreatitis with possible exposure during surgical intervention [9,11]. There is also a possibility of intestinal drainage into the portal vein itself in those individuals with a history of intestinal infection. It is likely that the recurrence or presentation of symptomatology itself was due to each patient’s underlying liver disease or comorbidities causing relative immunosuppression. While most patients had prolonged hospitalization related to their HA management, they generally had improvement, with the exception of the patient in Toprak et al.’s case, who ultimately did perish [8]. While she did not have a noted source for seeding, it is possible that she was inoculated during her previous angioplasty and that an overall decreased robustness of immune response allowed the bacterium to inoculate the abscesses. Her acute decompensation may possibly be due to the presence of toxin-positive species causing more tissue denaturation and thus spread. Notably, two cases reported the presence of C. difficile in the blood cultures of their patients, both of which survived. In another case, a patient underwent other surgical intervention of the hepatic or colonic systems and they were also associated with severe underlying comorbidities.

All in all, it is important to note that while this paper discusses cases of isolated CDI-related HA, there have been other reported incidences of C. difficile being noted as part of a polymicrobial HA [14]. As the incident of CDI becomes more prevalent and resistant worldwide, and the use of frequent antibiotics continues, further investigation will be needed into common risk factors, empiric treatment, and overall prevention of CDI-related HA.

5. Limitations and Future Directions

As CDI-related HA is an uncommon pathology, information on the subject remains scarce. There are a limited number of publications on the subject, leading to the potential of over-reporting of severity and publication bias. It is possible that the data presented here represent a more acute population, with presenting symptoms and a clinical picture severe enough for an actual sampling of the HA to occur, lending to the CDI HA diagnosis. Unfortunately, as CDI becomes more prevalent, and more pathogenic, there will be a need to investigate how to slow its spread. Likely, more education into the reduction in the spread as well as antibiotic stewardship will be needed, but while CDI remains a prevalent issue, future studies will be needed regarding the severity of infection, and a standard methodology for the diagnosis and treatment of the disorder.

6. Conclusions

As the incidence of CDI worldwide increases, so may the incidence of Ex-CDI. While the availability of data is currently scarce, the articles included in this study illustrate the vast possibility of etiologies for HA with positive cultures for C. difficile. While patients with previous CDI infection are at an increased risk, those with no reported CDI previously and further no reported antibiotic use or even diarrheal symptoms concurrently may still be at risk for colonization during other unrelated hospitalization or even in the community. For our interpretation, treatment should include timely initiation with antibiotics and abscess drainage as well as follow-up imaging dependent on clinical status, and frequent monitoring as patients who decompensate tend to do so quickly. There is no recommended empiric antibiotic regimen for abscess-related Ex-CDI, but it should potentially include broad-spectrum, anaerobic, and C. difficile coverage.