Non-candidal Fungal Peritonitis in Far North Queensland: A Case Series

Abstract

♦ Background: Fungal peritonitis is a recognized complication in patients with end-stage renal failure treated with peritoneal dialysis (PD). Most infections are attributable to Candida species. In approximately one third of cases, the causative fungus is a non-Candida species. Recent reports in the literature show a rising incidence of non-candidal fungal peritonitis (NCFP). We report a case series of NCFP, together with two hitherto unreported species of fungi causing peritonitis, from a tropical geographic area (Far North Queensland).

♦ Methods: This series of 10 cases of NCFP was identified from the PD peritonitis database in Far North Queensland between 1998 and 2010. All 10 patients were from the Aboriginal and Torres Strait Islander ethnic group, 8 of whom lived in remote locations. All but 1 patient had type 2 diabetes mellitus. Of the 10 cases, 7 occurred while the patients received continuous ambulatory PD. Only 1 patient avoided catheter removal, and 5 patients were permanently transferred to hemodialysis. No patient died as a result of the fungal infection. All 10 fungi represented different species. Most (6 of 10) were saprophytic; only 2 were normal skin flora. Two of the causative species (Chaetomium and Beauveria) have rarely been associated with any form of human infection. In 7 patients, the infection occurred during the wet season (November - April). All cases met clinical criteria for peritonitis.

♦ Discussion and Conclusions: The NCFP cases described in this series involved a variety of previously known fungal species and also two new species that have not been reported to cause disease in humans. Indigenous patients from Far North Queensland are particularly predisposed to infection with these exotic fungi as a result of environmental and social factors. Further understanding is desirable to help devise preventive strategies to avoid the consequences of catheter failure.

Fungal peritonitis constitutes up to 10% of all peritoneal dialysis (PD)-related peritonitis episodes and is more prominent in the tropical states of Australia (1). Its presentation is identical to that of bacterial peritonitis (2). Episodes are more likely to occur after a period of antibiotic use, although such use is not a prerequisite, and they may be related to yeasts rather than to filamentous organisms (3-6). Colonization with fungi is resistant to antifungal therapy alone, and therefore removal of the catheter together with antifungal treatment is now considered standard therapy (6). Permanent transfer to hemodialysis may result, and mortality rates are reported to range between 9% and 44% (3,6,7).

Peritonitis rates are known to be relatively high in the tropics because of the hot and humid environment, which favors fungal growth of increased virulence and diverse species (8,9). Rates of infection with unusual, and often otherwise nonpathogenic, fungi are increasing (2). Environmental moisture levels reflect the bimodal seasonal variation, with the wet season occurring during summer (November - April in our region), leading to an increased potential for fungal organisms to be identified.

In published case series of fungal peritonitis throughout the world, Candida species predominate. The reported incidence of non-candidal species was 8.5% in a series ending in 2000 (10), with a rising trend (17% - 40%) being evident in later series (7,11-15). Despite that increase, there is a paucity of data in the literature regarding the potential pathogenicity of these organisms in the tropical environment. The present case series describes 10 cases of PD-related non-candidal fungal peritonitis (NCFP) in which specific causative organisms have been isolated, some of which are new to the peritonitis literature.

Further knowledge of the causative organisms will help in a number of ways. Antifungal therapy should be individualized in NCFP (7). In PD peritonitis management, the common protocol is for fungal prophylaxis therapy to be given at the time of giving antibiotics for bacterial infections. The specific prophylaxis is chosen based on knowledge of local organisms. Guidance for the appropriate choice of empiric treatment is also important because fungi can be difficult to grow, and the lag period before samples become positive can be many days.

Patients living in remote settings are particularly affected if catheter removal becomes necessary as a result of fungal peritonitis. The catheter usually cannot be salvaged with antifungals, but instead must be removed (16). The patient then requires hemodialysis, and the advantages of PD are lost.

Methods

All cases of PD-related peritonitis treated at Cairns Base Hospital during 1998 - 2010 were reviewed. A retrospective chart and pathology review was undertaken for episodes of NCFP. Approval was obtained from the local human research ethics committee.

Results

Over the entire study period, 379 cases of peritonitis occurred in 184 patients. Between 2005 and 2010 (a representative period with the most complete data available), our area had an average peritonitis rate of 1 episode in 14.5 patient-months (0.83 episodes per patient-year). The overall Australian peritonitis rate was 1 episode in 20.7 patient-months (0.58 episodes per patient-year) for the same period (1).

Of all the peritonitis cases in this study from Far North Queensland, 31 (8.2%) were fungal. Among the fungal cases, 9 patients had 10 episodes of peritonitis (32%) caused by non-candidal fungi. The remaining fungal episodes were caused by Candida albicans or non-albicans species (68%). Tables 1 and 2 respectively describe the relevant attributes of the patients and the organisms in the NCFP episodes.

TABLE 1.

Patient Details of Ten Episodes of Dialysis-Related Non-Candidal Fungal Peritonitis

TABLE 2.

Organism and Pathology Details of Ten Episodes of Dialysis-Related Non-Candidal Fungal Peritonitis

All patients fit the accepted criteria for a diagnosis of peritonitis (abdominal pain, cloudy peritoneal effluent, and elevated effluent white cell count). Effluent samples were consistent with peritonitis, containing 280 - 80 000×10⁶/L white cells where counts were available. The proportion of polymorphonuclear cells exceeded 50% except in 1 episode (patient 4, episode 1), who had two separate episodes with different organisms (episodes 1 and 2). Each episode in the series involved a different fungal species (no species occurred in more than 1 episode). Isolation of the fungus was possible from both PD effluent and the catheter tip specimen in all samples that were available for testing. In only 2 of the 10 episodes was sensitivity or minimal inhibitory concentration (MIC) data available (patient 4, episode 1, and patient 8)

Seven episodes occurred in the wet season (November - April), and three in the dry season (May - October). Most of the patients (7 of 9) were using continuous ambulatory PD; the others were using automated PD. Diabetes was the cause of end-stage renal failure in 8 of the 9 patients. Eight episodes occurred in a remote or very remote setting [defined according to the Accessibility/Remoteness Index of Australia (17)], most of which were in the Torres Strait Islands. Ethnicity as an Aboriginal or Torres Strait Islander (ATSI) was overrepresented at 100% in a background dialysis population that is approximately 60% ATSI.

The current literature suggests that a combination of antifungal treatment and catheter removal is required to limit mortality in fungal peritonitis (6). Management varies somewhat because of patient or physician preference. In this retrospective series, both treatments were applied in 5 episodes. Patient 8 was treated with fluconazole until catheter removal; the others received between 3 weeks and 3 months of standard-dose oral fluconazole or itraconazole. In only 1 patient (patient 3) was catheter removal not performed; to avoid hemodialysis, he refused removal. He was treated successfully with a 6-week course of oral itraconazole. The other 4 cases were successfully managed with catheter removal alone because all were clinically well at the time that the positive fungal culture was reported. No direct fatalities occurred as a result of the infections.

Permanent transfer to hemodialysis occurred in 5 episodes (50%). The only other complication potentially attributable to fungal peritonitis was that of recurrent sterile intra-abdominal fluid collections requiring intermittent drainage (patient 6). One patient experienced a second episode of NCFP after an interval of 9 months (patient 4, episodes 1 and 2). One patient (patient 7) was in fact able to remain dialysis-free for 19 months after removal of her catheter, suggesting that removal of the substantial systemic insult from the fungus led to some recovery of renal function.

Discussion

In Far North Queensland, the PD prevalence rate is similar to the overall Australian rate, with 22% of dialysis patients using this home-based therapy (17). The population is spread over Cape York from Cardwell in the south, to Normanton in the west, and to the Torres Strait Islands in the north. Remoteness is defined by the Accessibility/Remoteness Index of Australia, which categorizes locations as very remote, remote, moderately accessible, accessible, or highly accessible based on road distance from a locality to the closest service centers (17). Most of Far North Queensland falls into the remote or very remote categories. A large proportion of the PD population is indigenous (52%). Patients who are ethnically ATSI have a known increased risk of fungal peritonitis (3).

Unpublished local data show that PD practices in Far North Queensland over the period of the present study (1998 - 2010) remained relatively consistent. Automated PD is used in 90% of patients, and fungal prophylaxis during antibacterial use was not standard until the last 1 - 2 years of that period. Compared with automated PD, continuous ambulatory PD was overrepresented in the present case series, possibly because of the increased requirement for catheter access in continuous ambulatory PD (8 times compared with 2 times daily). The expected seasonal variation in peritonitis episodes was observed, with predominance in the wet season, suggesting that health carers should have a higher index of suspicion for the diagnosis during that season.

This series identified 10 fungal species other than Candida that caused peritonitis in our tropical environment. Most of the non-candidal fungi appear to be opportunistic species that ordinarily do not cause disease. However, in this cohort of relatively immunosuppressed patients (predominantly diabetic, with end-stage renal failure), these usually harmless fungi resulted in peritonitis. Diabetes is likely to be a major contributing factor, given its overrepresentation in this study. Further complications from these pathogens were uncommon, and death as a result was not encountered, despite suggestions in the literature that non-candidal fungal infections predict mortality (18).

The clinical impact of fungal peritonitis on individual patients on PD is dire. Catheter removal for patients living in remote areas in the tropics usually results in significant lifestyle disruption and psychosocial stress as a result of transfer to in-center hemodialysis.

Sensitivity patterns are better understood for Candida species because sensitivity testing is more commonly available and performed. Empiric data are available for other species, although specific sensitivity testing is suggested for individual cases (19). Nystatin prophylaxis during antibiotic use has been suggested to potentially limit Candida peritonitis rates (20,21). That approach is unlikely to apply to non-candidal organisms, which are mostly non-yeasts.

In most of the episodes reported here (6 of 10), the organisms were saprophytic; only 2 were normal skin flora: Trichosporon species (19). That finding suggests that the organisms introduced into the peritoneum are likely to have come from external sources rather than from the skin. Trichosporon species are known to cause common superficial mycosis, but can act in an opportunistic fashion to cause systemic disease. This opportunistic nature is also known for the other species found, except for those in patients 8 and 9: Chaetomium and Beauveria. Of the fungi found in the present series, those two are rarely associated with human infections (22,23). The present series includes the first reported cases of peritonitis attributable to those organisms. Further understanding of the organisms involved will help in adjusting prophylactic strategies to improve outcomes.

Fungal peritonitis is known to be more likely to occur after a course of an antibacterial agent (3), but previous antibiotics may not be necessary, as reported in a series by Rosa et al. (4), in which fungal peritonitis was preceded by recent antibacterial use [preceding 3 months (6)] in only 1 of 5 cases. Antibiotic treatment preceded NCFP in 3 of our 10 cases, and it is unclear whether prophylactic antifungal therapy would have prevented those episodes.

Conclusions

This case series supports the need for more understanding of NCFP in PD patients. It also supports the background literature, in that infection with such fungi can cause substantial morbidity and is best treated with prompt catheter removal and aggressive antifungal therapy. Fungal peritonitis can be attributed to a wide variety of fungi apart from the more common yeast genus Candida. The biodiversity found in tropical regions such as Far North Queensland is likely to contribute to the breadth of variation in isolates. Chaetomium and Beauveria can be opportunistic pathogens and cause peritonitis in the diabetic PD population. This descriptive series will help to further clarify management and prevention strategies when considering guidelines with respect to antifungal prophylaxis and treatment and to making appropriate choices in a tropical location.

Disclosures

All the authors declared no financial conflicts of interest.