Pyogenic liver abscesses in peritoneal dialysis patients: A single-centre retrospective case series

Introduction

Peritoneal dialysis (PD) and kidney function impairment are unique risk factors for the development of liver abscesses. ¹ The incidence of liver abscesses in both prevalent chronic kidney disease and dialysis patients, especially in Asian populations, is reportedly high. ^{2 –4} However, intra-abdominal abscesses in PD are poorly and uncommonly described in literature. A single-centre review of 1345 episodes of peritonitis in 537 patients between 1982 and 1992 at a single centre identified only 10 patients (0.7% of all peritonitis) to have intra-abdominal abscess, of whom only one out of the 10 patients developed liver abscess. ⁵ In the largest retrospective review data from Taiwan, the incidence of liver abscesses in individuals on dialysis was higher than the general population, but this included both PD and haemodialysis (HD) patients. ² In addition, liver abscesses in dialysis patients carry high mortality rate of >30%. ³

Liver abscesses can result from infection originating from biliary tract, haematological spread through the portal venous system related to intra-abdominal infection, systemic haematological spread and rarely penetrating injury. ⁶ In individuals receiving PD, peritonitis can lead to liver abscesses through the portal venous system or rarely abscess rupture leading to peritonitis as reported in non-PD patients. ⁷ The management of liver abscesses in the general population is well described involving antibiotics and drainage ⁶ ; however, data in PD patients are limited. ³ In addition, challenges posed in PD patients include the risk of peritonitis development post aspiration of abscesses ³ and the consideration for PD catheter removal. However, no clear guidelines regarding the indications for PD catheter removal during treatment of liver abscesses exist.

Hence, this study aims to describe the characteristics, management and outcomes of liver abscesses in PD patients in our local cohort.

Materials and methods

This is a retrospective review of all prevalent PD patients on follow-up with the Department of Renal Medicine, Tan Tock Seng Hospital, Singapore between 1st January 2016 and 30th June 2021 to identify patients who were hospitalised and managed for liver abscesses as diagnosed on radiological imaging. This study was approved by the National Healthcare Group (NHG) Domain-Specific Review Board (DSRB) under reference number 2021/00443. Waiver of informed consent was granted, as the risk to the patient is deemed low with precautionary measures, and data collected relate to the standard of care which they have already received.

Clinical, biochemical and radiological characteristics of the liver abscess and microbiology were collected. The relevant treatment received including antibiotics use, percutaneous aspiration and/or surgical drainage of liver abscess and PD catheter removal were also collected. Clinical outcomes such as permanent transfer to HD (more than 30 days) and mortality within 6 months of liver abscess were also collected.

Statistical analysis was performed using STATA version 14 (Texas 77845 USA).

Results

A total of 383 prevalent PD patients were screened, and 11 (2.9%) were treated for liver abscess. Majority were diabetic and hypertensive (90.9% and 81.8%, respectively), with a mean age of 60.8 years ± 9 years (Online Supplemental Table S1). Only one patient had a history of PD-related infection within 3 months of diagnosis of liver abscess (Table 1). The median PD vintage at the time of liver abscess diagnosis was 541 days (interquartile range (IQR): 310–931). Of the 11 patients, 9 patients were treated inpatient at our hospital, while 2 patients were treated at another hospital and followed up as an outpatient at our centre. At presentation, fever (n = 7/10, 70%) and abdominal pain (n = 4/10, 40%) were the most common symptom. Elevated inflammatory markers were prominent; median C-reactive protein at 191.8 µg/mL (IQR: 174.5–263.2) and mean procalcitonin of 34.5 ng/L ± 43.9. Most patients also presented with bacteraemia (n = 7/11, 63.6%). At initial presentation, three had clinical evidence of peritonitis, with another four patients only subsequently diagnosed to have peritonitis at the same time of liver abscess diagnosis (7/11 with peritonitis as defined by 2 out of 3 characteristics; clinical symptoms, dialysis effluent white cell >100/µL and positive dialysate effluent culture). Amongst those without PD peritonitis (n = 4/11, 36.4%), none developed peritonitis during their treatment, and three (27.2%) of these patients remained peritonitis-free despite radiological aspiration under antibiotic coverage. Most of the abscesses were solitary (n = 6/10, 60%), and the mean size of the abscess was 4.3 cm ± 3.2 cm. There was no predilection for either lobe of the liver – four patients had abscesses in both lobes, four in the right lobe and two in the left lobe. Klebsiella pneumoniae was the most common pathogen (n = 7/11, 63.6%), followed by Enterococcus faecium (n = 3/11, 27.2%) isolated on abscess culture. One patient had no positive microbiology.

OPEN IN VIEWER

Table 1.

Summary of patient’s clinical history, presentation, management and outcomes.

Patient biodata	Significant medical history	PD-related infections prior to liver abscess diagnosis	Symptoms	Bacteraemia	Indication for radiological imaging	Peritonitis at presentation	Peritonitis during treatment	Radiological drainage of abscess	Antibiotic route at the time of liver abscess	Removal of PD catheter	Indication for removal	HD at the time of last observation	PD at the time of last observation	Death
A, 59 Male	DM, multifocal hepatocellular carcinoma on a background of alcoholic liver cirrhosis and cholecystitis	No	Abdominal pain	No	Severe abdominal pain	Yes	N.A.	No	IV	No	N.A.	Yes	No	Yes
B, 74 Female	Lung adenocarcinoma status post curative resection and chemotherapy, hepatitis B carrier	Yes, history of MSSA PD peritonitis and MSSA exit site infection, 1 year and 3 months prior	Diarrhoea, abdominal pain and cloudy effluent	No	Polymicrobial peritonitis	Yes	N.A.	No	IV	Yes	Refractory PD peritonitis	Yes	No	No
C, 42 Male	DM	No	Fever and cloudy effluent	Yes, Klebsiella pneumoniae	Bacteraemia	Yes	N.A.	Yes	IV	Yes	Empirical removal	Yes	No	No
D, 55 Female	DM	No	Abdominal pain	No	Persistent fever with slow to respond to antibiotics	No	No	Yes	IV	Yes	Empirical removal	Yes	No	No
E, 60 Male	Hypertension	No	Fever and vomiting	Yes, Klebsiella pneumoniae	Bacteraemia	No	No	Yes	IV	No	N.A.	No	Yes	No
F, 58 Female	DM	Coagulase negative staphylococcus exit site infection more than 3 months prior	Fever	Yes, Klebsiella pneumoniae and Enterococcus faecium	Bacteraemia	Yes	N.A.	Yes	IV	No	NA	Yes	No	Yes
G, 73 Male	DM, iron deficiency anaemia with colonoscopy performed 4 months before liver abscess diagnosis, cholecystitis treated conservatively	No	Fever and vomiting	Yes, Enterococcus faecalis	Bactereamia	Yes	N.A.	No	IV	Yes	Empirical removal	Yes	No	No
H, 59 Male	DM	No	Fever, diarrhoea, abdominal pain and cloudy effluent	Yes, Escherichia coli and Streptococcus bovis	Bacteraemia	Yes	N.A.	Yes	IV	Yes	Empirical removal	Yes	No	No
I, 61 Male	DM	No	Fever, diarrhoea and vomiting	No	Severe sepsis with septic shock to evaluate for source of sepsis	Yes	N.A.	Yes	IV	Yes	Septic shock	Yes	No	Yes
J, 58 Female	DM	MSSA PD peritonitis 2 year and 8 months prior	Fever	Yes , Klebsiella pneumoniae	Bacteraemia	No	No	Yes	IV	No	N.A.	No	Yes	No
K, 70 Female	DM	No	Missing data	Yes, Klebsiella pneumoniae	Bacteraemia	Missing data	Missing data	Missing data	IV	No	N.A.	No	Yes	No

PD: peritoneal dialysis; HD: haemodialysis; DM: diabetes mellitus; IV: intravenous; MSSA: Methicillin-sensitive Staphylococcus aureus.

All patients received intravenous antibiotics and most underwent ultrasound-guided aspiration (n = 7/11, 63.6%), with the remaining four too small for aspiration and none underwent surgical drainage. Mortality was high with three patients (27.2%) passing between 2 days and 43 days of presentation. Most patients required interim HD (n = 9, 72.7%) with five permanently transferred to HD after censoring for death (n = 5/8, 62.5.%). Reasons necessitating conversion to interim HD included empirical conversion to HD in view of intra-abdominal abscess (n = 6/9, 66.7%), septic shock (n = 2/9, 22.2%) and refractory PD peritonitis (n = 1/9, 11.1%).

At the end of the observation period, only three patients (37.5%) remained on PD. All three of these patients did not develop PD peritonitis throughout the course of their treatment. Of these, two patients remained on PD throughout the course of their treatment, even while undergoing radiological aspiration of the liver abscess. One patient was temporarily converted to interim HD empirically due to need for radiological aspiration and was successfully transitioned back to PD after the abscess was treated.

Discussion

This review adds to the literature on PD patients with liver abscesses. Amongst our 11 patients, we observe similar characteristics such as high rates of diabetes, increased mortality risks and high rates of K. pneumoniae as the causative organisms such as in previous reports. ^2,3 However, it is interesting to note that most of these PD patients presenting with liver abscess often present with bacteraemia and have higher than usual inflammatory markers than we would usually expect in PD peritonitis.

Apart from intravenous antibiotics, percutaneous drainage of the liver abscess is the main stay of treatment. An earlier report of over 27 patients with kidney failure and liver abscess (8 of whom were on PD) noted high risks of peritonitis following invasive aspiration in the patients on PD (80%). ⁸ However, in our cohort, most patients presented with peritonitis at presentation, and amongst those without peritonitis, three underwent aspiration with no peritonitis episodes reported. Similar experience in PD patients with liver transplant undergoing liver biopsy reported no complications peri-procedure. ⁹ It is unclear what accounts for the difference in the risk of peritonitis post-aspiration, but possibly differences in peri-procedural care such as temporary transfer to HD and duration of PD free period peri-procedure may account for this difference. All patients who underwent aspiration were covered with intravenous antibiotics and had PD fluid drained peri-procedure in our centre.

The overall mortality in our cohort is high. The rate of permanent transfer to HD is also high. The high rates of transfer to HD may be related to the perceived high risk of refractory PD peritonitis amongst treating physicians leading to early empirical PD catheter removal reported in our cohort. Unfortunately, there is insufficient evidence in the literature to guide the optimal management of the PD catheter during treatment of liver abscess. It is, however, encouraging to observe that patients without PD peritonitis at the time of liver abscess can be managed with optimal antibiotics and drainage without the need for PD catheter removal in our cohort. Therefore, catheter removal should be reserved for cases of refractory, relapsing, repeat and fungal PD peritonitis in line with International Society of Peritoneal Dialysis guidelines in 2022. ¹⁰

It is recognised that this is a single-centre retrospective case series and significant potential bias exists. However, it is still one of the few studies looking at liver abscess in a PD-only population and brings to light the need for more contemporary larger studies to guide the optimal treatment of liver abscesses in PD patients.

Supplemental material