Incidence of spontaneous fungal peritonitis in patients with liver cirrhosis in a Mexico City population

Abstract

BACKGROUND

Individuals with liver cirrhosis (LC) are likely to experience multiple infectious processes due to the immune dysfunction caused by the disease. Our hypothesis is that this group of patients is predisposed to fungal infections. To date, the incidence of spontaneous fungal peritonitis (SFP) has not been determined in Mexico; this endeavor is of great importance because many patients may be suffering from this condition without receiving targeted treatment, which may increase mortality.

AIM

To report the incidence of SFP in patients presenting with decompensated LC with ascites.

METHODS

This was a prospective, single-center, descriptive, observational and cross-sectional study where patients presenting with decompensated LC with ascites were evaluated from November 2023 to May 2024 in Mexico City. Fungal cultures of ascites were performed and the samples kept in an incubator for 10 days to 14 days, and molecular tests (the API 20 C AUX test) were used for molecular characterization.

RESULTS

Of the 48 patients included, 54.2% were women, 77.1% had a comorbidity, 47.9% had LC secondary to metabolic dysfunction, 43.8% were classified as Child-Pugh C with a model for end-stage liver disease 3.0 median score of 22, and 10.4% were in secondary prophylaxis for spontaneous bacterial peritonitis (SBP). Only four patients had positive cultures where Candida parapsilosis and Candida albicans were isolated, with two of the four patients being positive for Rhodotorula minuta; an SBP incidence of 8.3% was thus calculated. Chronic kidney disease [P = 0.012 and relative risk (RR) = 15] and secondary prophylaxis for SBP (P = 0.049 with RR = 8.6) were statistically significant and associated with a high mortality risk (P = 0.001 with RR = 33).

CONCLUSION

The presence of infection of fungal origin in ascites in patients presenting with cirrhosis increases short- and medium-term mortality; therefore, it is recommended that fungal culture tests are performed in those patients who visit the emergency room or experience continuous admission with acute decompensation and no bacteria identified in ascites cultures, and even more so in patients with chronic kidney disease and a history of antibiotic use as prophylaxis for SBP. Further studies are needed for the identification of clinical and biochemical data that can help to define SFP so that its presence may be assessed without the need to wait for a positive fungal culture. Thus, treatment may be initiated early in the hope of having a positive impact on the prognosis in this group of patients.

Key Words: Spontaneous fungal peritonitis; Liver cirrhosis; Ascites; Decompensation; Death

Core Tip: Liver cirrhosis is a condition that causes immune dysfunction, which predisposes the patient to infections. In this study, we investigated the presence of mycotic microorganisms in ascites to calculate their incidence in patients in a Mexico City hospital. For this purpose, ascites samples were taken, and mycotic cultures were performed, revealing an incidence of 8. The microorganisms identified were Rhodotorula minuta, and Candida parapsilosis and Candida albicans. It was found that chronic renal disease and secondary prophylaxis for spontaneous bacterial peritonitis are associated with spontaneous fungal peritonitis and that patients with positive mycotic cultures have a high mortality risk.

INTRODUCTION

Liver cirrhosis (LC), where the normal liver architecture is replaced by regenerative nodules, is a consequence of chronic hepatic inflammation and can have multiple etiologies[1,2]. Two main stages are recognized in LC: The compensated and decompensated stages. In the former, the patient is usually asymptomatic, presents with a portal venous pressure of less than 10 mmHg and can be diagnosed incidentally. In the decompensated stage, the patient shows some clinical manifestations; the decompensations related to the progression of the disease are mainly due to the increase in portal hypertension and immune deficiency associated with cirrhosis and become more accentuated as the disease progresses[3]. Ascites is the most frequent decompensation and marks the beginning of the decompensated phase. It is defined as the pathological accumulation of intra-abdominal fluid in the peritoneal space and occurs in 60% to 70% of patients 10 years after diagnosis; it is associated with a poor prognosis, with alteration in quality of life and an average survival of 2 years[4,5].

Patients with LC are predisposed to multiple infectious processes due to immune system dysfunction and are considered non-neutropenic immunosuppressed patients, because neutrophils have decreased function while presenting in adequate or even increased quantities in some infectious processes[3]. In addition, CD4 Lymphocytes in this group of patients are decreased in 65% of cases[6], which, together with an increase in the permeability of the intestinal barrier in the intestine-liver axis, allows for the translocation of the intestinal microbiota and predisposes patients to infections that cause decompensation, such as spontaneous bacterial peritonitis (SBP)[1,3,7-9].

In all patients presenting with decompensated LC with ascites, we must rule out an infectious process, for example, SBP, the exclusion of which is performed according to international well-established clinical and biochemical criteria, such as abdominal pain and tenderness, nausea, vomiting and ileus, with systemic signs and symptoms such as hyperthermia or hypothermia, tachycardia or tachypnea, renal failure, alteration in liver function reflected by hepatic encephalopathy and biochemical alterations such as leukocytosis. SBP is confirmed by performing a diagnostic paracentesis to determine the neutrophil count, which should be more than 250 cells/mm[3,4].

The composition of the intestinal microbiota includes fungi. It is possible that the presence of these microorganisms leads diagnosticians to suspect SBP in patients who do not have a positive culture for bacteria[7,8]. Therefore, the aim of this study was to determine the incidence of spontaneous fungal peritonitis (SFP) in ascites samples from patients diagnosed with LC. The secondary objectives were to evaluate whether secondary antibiotic prophylaxis for SBP confers a higher risk of SFP, compare different Child-Pugh stages to determine which stage is associated with higher incidence of SFP, identify the main fungal agents present in ascites cultures, determine whether hospitalization confers a higher risk of presenting SFP, and assess whether patients with acute over chronic liver failure have higher incidence of SFP.

MATERIALS AND METHODS

A prospective, single-center, descriptive, observational and cross-sectional study was conducted at Hospital de Especialidades Dr. Bernardo Sepulveda, Centro Médico Nacional Siglo XXI, from November 2023 to May 2024 in Mexico City following authorization by the local ethics committee with registration No. R-2023-3601-240. Patients with LC of any etiology, in the decompensated stage and presenting ascites were included according to the following inclusion criteria: Patients older than 18 years and an ascites grade ≥ II. The exclusion criteria were as follows: Patients under 18 years, patients from whom it was not possible to take blood and/or ascites samples and ascites not caused by LC.

Cytological, cytochemical, bacterial and fungal cultures were performed on the ascites samples obtained. The ascites samples were placed in 15 mL Falcon tubes and centrifuged (TierraR centrifuge model 002) at 4000 revolutions per minute for 10 minutes, which resulted in 2 mL of sediment. The samples were cultured in simple Sabouraud medium with cycloheximide and kept in an incubator for 7 days to 10 days at 28 °C, with culture revision every 48 hours to evaluate the growth of colonies phenotypically compatible with fungal agents. According to the culture result, a microscopic examination was performed; subsequently, cultures were performed in Candida agar, potato agar and/or urea agar to phenotypically characterize any colony obtained from the cultures in Sabouraud medium. Molecular tests (API 20 C AUX test) were used to identify carbohydrate reduction according to the fungal agents grown in the initial culture for molecular characterization.

RESULTS

Of the fifty-one patients evaluated, three were excluded: One due to cardiac ascites, and the other two due to the lack of biochemical parameters of ascites. Therefore, the total number of patients evaluated was 48. The mean age was 59.06 ± 12.09 years, and 77.1% had a comorbidity, with type 2 diabetes mellitus being the most frequent condition (45.8%), followed by systemic arterial hypertension (43.8%) and chronic kidney disease (16.7%) (Table 1).

Table 1 Demographic data (n = 48).

Characteristic	n (%)
Age (year), mean ± SD	59.96 ± 12.09
Sex
Female	26 (54.16)
Male	22 (45.8)
Comorbidity	37 (77.08)
Body index mass
Low weight	8 (16.66)
Normal	16 (33.33)
Overweight	15 (31.25)
Obesity	9 (18.74)
Diabetes mellitus type 2	22 (45.83)
Systemic arterial hypertension	21 (43.75)
Chronic kidney disease	8 (16.66)
Hepatocarcinoma	2 (4.16)
Etiology of cirrhosis
MASLD	23 (47.91)
Alcohol	2 (4.16)
Hepatitis B virus	1 (2.08)
Hepatitis C virus	3 (6.25)
Autoimmune hepatitis	2 (4.16)
Primary biliary cholangitis	4 (8.33)
MetALD	9 (18.75)
Cryptogenic	4 (8.33)
Child-Pugh class
A	1 (2.08)
B	21 (43.75)
C	26 (54.16)
MELD 3.0
< 14	5 (10.41)
≥ 14	43 (89.58)
SBP prophylaxis	5 (10.41)
Acute-on-chronic liver failure
Yes	7 (14.58)
No	41 (85.41)

MASLD: Metabolic liver disease associated with metabolic dysfunction; MetALD: Hepatic steatotic liver disease associated with metabolic dysfunction and alcohol consumption; MELD: Model for end-stage liver disease; SBP: Spontaneous bacterial peritonitis.

Among the causes of LC, the leading cause was secondary steatotic hepatopathy associated with metabolic dysfunction (found in 47.9% of patients), followed by alcohol consumption disorder and metabolic dysfunction (18.8%), and the third cause was primary biliary cholangitis (8.3%). In total, 54.2% of patients were classified as Child-Pugh-Turcotte scale grade C, 43.8% as grade B and 2.1% as grade A. We further determined a model for end-stage liver disease 3.0 median score of 22, with an interquartile range of 18 to 28 points. Additionally, 10.4% of the sample underwent prophylaxis for SBP; regarding hospital admissions, the median was one, with an interquartile range of one to three, and 14.6% of patients had acute or chronic liver failure at hospital admission.

The total of ascites sediment samples from the 48 patients, from whom samples were taken multiple times, was 119. Six patients had positive bacterial cultures that did not coincide with a positive mycotic culture of that same sample collection, but in four patients, positive mycotic cultures were identified as follows: Candida Parapsilosis (in one patient), Rhodotorula minuta (in two patients) and combination of Candida albicans and Rhodotorula minuta (in one patient) (Table 2). The biochemical and cytological characteristics of ascites compatible with internationally known criteria for SBP were not evidenced; the four patients with positive fungal cultures died in less than 20 days. A cumulative incidence of 8.3% was calculated for the presence of fungal agents in ascites.

Table 2 Microorganisms identified (n = 4).

Characteristic	n (%)
Candida albicans	1 (20)
Candida parapsilosis	1 (20)
Rhodotorula spp.	1 (20)
Rhodotorula minuta	2 (40)

Of the 48 included patients was performed by dividing them into two main groups according to the positivity or negativity of the mycotic culture and to whether they had had less than two paracentesis or equal to or greater than three paracentesis. The group with equal to or greater than three paracentesis with positive cultures had a P value significance of 0.43, with an relative risk (RR) of 3.66 (95% confidence interval: 1.62-8.26).

The secondary objective of evaluating the number of hospital admissions was not achieved, since most patients underwent evacuative paracentesis in continuous medical admission and were discharged because they did not meet the hospitalization criteria. There was also no statistical significance in terms of the Child-Pugh functional stage and the presence of fungal agents in the ascites samples. In the assessment of the risk of mortality and fungal presence in the ascites samples, a P value of 0.001 was obtained, with an RR of 33 (95% confidence interval: 4.38-248.38).

DISCUSSION

The determination of the presence of fungal agents in ascites from patients with LC is of great importance because of the immunocompromised state these patients develop as the disease progresses and because of the risk of translocation of not only bacteria, but also fungal agents belonging to the intestinal microbiota, which makes this population more susceptible to infections and increases mortality[3].

In addition to bacteria, fungi are also present in the microbiota. Among them, Candida species have been reported as the main agents responsible for infectious processes in immunocompromised patients; however, for some time now, the appearance of other types of fungi that have been associated with pathological states has been reported. Such is the case of Rhodotorula species, a type of yeast belonging to the family of basidiomycetes fungi, whose presence has been found in human feces. Eight species of Rhodotorula have been documented, but the most relevant ones in humans are R. glutinis, R. minuta and R. mucilaginosa, which are peculiarly resistant to dryness; survive in air, soil and water environments; have an affinity for plastic materials such as plastic catheters and dental equipment, where they grow in the form of biofilms; and can also colonize the hands of medical staff[10-12].

Rhodotorula species have been reported to be associated with catheter-associated fungemia, peritonitis, endocarditis, meningitis and eye infections worldwide. In a systematic review by Tuon and Costa[13] published in 2008, the authors evaluated 128 cases, 87% of which were associated with immunosuppression or cancer; moreover, 5% of patients with peritoneal dialysis presented with peritonitis due to Rhodotorula mucilaginosa. In the review, they did not find the presence of Rhodotorula species in patients with decompensated LC with ascites[13].

There have been retrospective studies of cases of SFP in patients with LC where diagnostic criteria or aggregate definitions were not established, as the findings could also have indicated fungiascitis. An example is the study by Gravito-Soares et al[14] in Portugal from 2006 to 2015, in which they compared patients with bacterial and fungal peritonitis with a positive culture of ascites fluid. They found that 6.2% of patients had SFP compared with 93.7% of patients with SBP; they isolated mainly Candida species (87.5%), where 37.5% were Candida albicans, 25% Candida Krusei, 12.5% Candida lusitaniae, 12.5% Candida tropicalis and 12% other species. The demographic, clinical and biochemical characteristics with statistical significance were days of hospital stay, lactate dehydrogenase levels in ascites fluid, urea nitrogen and blood leukocytes, and a mortality rate of 50% at 30 days was determined in patients with SFP[14].

Hwang et al[15], in South Korea, evaluated the risk factors for developing SFP, the presence of fungi in ascites fluid and the prognosis of patients with SFP vs SBP in the period from January 1, 2000 to December 31, 2005. They evaluated 416 patients with spontaneous peritonitis, 3.6% of which had a positive fungal culture; of these patients, 86.7% were classified as Child-Pugh class C, 80% were diagnosed with nosocomial SFP, and 60% had a history of antibiotic use 3 months prior to the study. Four fungal species were isolated, with Candida species being found in 10 patients (66.6%) and Cryptococcus neoformans in 5 patients (33.3%); of the Candida species, C. albicans was isolated in 8 patients (80%), and the remaining patients presented C. Glabrata and C. Tropicalis. A total of 73.3% of patients had a history of antibiotic use in the previous 3 months. Staphylococcus epidermidis, Klebsiella pneumoniae, Klebsiella oxytoca, Escherichia coli and Enterococcus faecium co-infection was present in 73.3% of patients with SFP. Mortality was also assessed: At 1 month, the mortality rates were 73.3% and 28.7% for patients with SFP and SBP, respectively, and at 6 months, they were 93.3% and 56.1%, respectively[15].

Our work produced similar results to those of the studies previously discussed. For example, in the systematic review of infections by Rhodotorula species, chronic kidney disease was a component that predisposed patients to infections by this type of fungus. It is worth mentioning that this was found in patients with peritoneal dialysis, and it is not yet known with certainty if there is any peculiarity in the peritoneum that causes such predisposition in patients or facilitates the colonization of that area by Rhodotorula species. Regarding mortality, the retrospective studies by Gravito-Soares et al[14] and Hwang et al[15] share the finding of very high rates, which we also found in our study.

This is the first study in Mexico in which the presence of fungal agents in ascites samples from patients with LC is intentionally determined, which lays the foundation to investigate this type of microorganisms for providing timely treatment and improving the prognosis in this population. It is worth mentioning that we also evaluated the presence of hepatocarcinoma in association with SFP, and although the P value was not significant, the confidence interval was significant, with a relative risk of 7.66, which could be due to the small sample size. The history of secondary prophylaxis for SBP was evaluated and shown to increase the risk by 8.6 times. A weakness of this study is the sample size, which was only 48 patients, so it would be worthwhile to increase the sample in future efforts. The RR was determined because this was a prospective study; however, the RR result should be analyzed according to the sample size, since it does not indicate the magnitude of the absolute risk but only the relationship between the risk values of the two groups. Therefore, the attributable risk and the proportion of attributable risk were calculated. The result for renal disease was proportion of attributable risk = 85. 3%, which means that in 85% of cases, chronic renal disease was the underlying cause of LC and a positive culture for fungi. In the same way, this metric was calculated for antibiotic prophylaxis for SBP and hepatocarcinoma, but the results of the multivariate study showed no statistical significance; this could be justified by the sample size, so it will be necessary to conduct further research with the inclusion of more patients, as well as more centers. The attributable risk scores are shown in Table 3.

Table 3 Variables associated with the development of spontaneous fungal peritonitis, bivariate and multivariate analysis.

Characteristic	Bivariate analysis, P value; RR (95%CI)	RA%	Multivariate analysis P value; RR (95%CI)
Chronic kidney disease	0.012; 15 (1.77-126.48)	85.3	NS
Antibiotic prophylaxis for SBP	0.049; 8.6 (1.53-48.38)	87%	NS
Hepatocarcinoma	0.161; 7.67 (1.312-44.8)	93.3%	NS

RR: Relative risk; RA: Attributable risk; SBP: Spontaneous bacterial peritonitis; CI: Confidence interval; NS: No significance.

Currently, there is no international definition for the diagnosis of SFP. The Asia-Pacific Liver Study Association’s Management of Ascites in Liver Disease Guidelines, in its 2023 update, mentions the term SFP but does not provide diagnostic criteria such as those internationally recognized for SBP. Therefore, in our study, we defined SFP as the presence of a positive fungal culture without a clear cause of infection, such as a ruptured organ or malignancy, with a negative bacterial culture with a polymorphonuclear count < 250 cells/mm[3,16,17].

In this study, it was not possible to assess whether the number of hospital admissions confers an increased risk of the presence of fungal agents in ascites in this group of patients, since as previously mentioned, some patients received outpatient care without requiring hospitalization. This is considered a weakness of this study, as are the time during which it was carried out and its being a single-center study. Despite the existence of tests for the early diagnosis of fungal infection, they are not available to us. This leads to limitations on a timely diagnosis of fungal infection in ascites, due to the time needed to evaluate fungal growth, which delays the initiation of treatment.

CONCLUSION

Infections in patients with liver disease are a major cause of decompensation and lead to increased mortality in the short and medium term. Sometimes, this group of patients is repeatedly admitted with acute decompensation without bacteria being identified in ascites cultures, and they do not meet the international criteria for spontaneous bacterial peritonitis, which is why fungal culture is indicated. It is important to look for and diagnose fungal peritonitis as a cause of decompensation in order to start treatment in a timely manner and thus improve the prognosis for these patients.

ACKNOWLEDGEMENTS

Our thanks to the resident physicians of the Gastroenterology service of the Hospital de Especialidades, Centro Médico Nacional for taking samples. Special thanks for their collaboration and invaluable support for: Patricia Manzano-Gayoso, MD. Francisca Hernández-Hernández, Micology Unit, Department of Microbiology and Parasitology, School of Medicine, National Autonomous University of Mexico, Mexico City.