Temporal Trends of Fungal Infections in Cirrhotic Patients: A Retrospective Cohort Study 2016–2020

Abstract

Background

Patients with cirrhosis are susceptible to infections due to abnormalities in humoral and cell-mediated immunity. Fungal infections are associated with delayed diagnosis and high mortality rates, emphasizing the importance of performing fungal cultures and maintaining elevated levels of suspicion in this patient population.

Methods

This retrospective cohort study analyzes cirrhotic patients readmitted with bacterial and fungal infections and investigates outcomes, including in-hospital mortality and hospital resource utilization. Data was acquired from the Nationwide Readmission Database (NRD) from 2016 to 2020. Total hospital costs were calculated using HCUP Cost-to-Charge Ratio files and adjusted for inflation based on the Consumer Price Index (CPI) for medical care services in the U.S., with 2020 as the reference year. The NRD dataset lacks details like ascitic fluid cell counts, antifungal/antibacterial drugs used, and treatment responses, limiting the clinical insights that can be derived.

Results

The study analyzed 393,195 index hospitalizations. Among these, 102,505 account for 30-day and 157,079 account for 90-day readmissions. The 30-day and 90-day readmissions for spontaneous bacterial peritonitis (SBP) are 8478 and 15,690 respectively. The 30-day and 90-day readmissions for spontaneous fungal peritonitis (SFP) are 3106 and 5798 respectively. The mean age of patients was 57.9 years (standard deviation between 57.7 and 58.1). The mean length of stay (LOS) for SBP at 30 days is 9.4 days, while SFP has ranged from 14.9 to 32.3 days for various fungal infections. Aspergilloses have the longest LOS among SFP. There is an increased rate of mortality as well as hospital charges with SFP compared to SBP (P < 0.001). The 30-day index admission total charges for SBP are $42,258 and SFP are $51,739. The 30-day readmission total charges for SBP are 64, 266 and for SFP 89,913.

Conclusions

There is increased mortality, LOS, and hospital costs for SFP compared to SBP. It is important to consider SFP in the diagnostic workup for patients who do not respond to antibiotics. Early recognition and administration of antifungals can be associated with improved outcomes.

Spontaneous bacterial peritonitis (SBP) and Spontaneous fungal peritonitis (SFP) are a well-known complication of liver cirrhosis. The diagnostic criteria for SBP require a polymorphonucleated (PMN) cell count of ≥250/mm on fluid analysis.¹ A cirrhotic patient with ascites and tenderness to abdominal palpation or fever with PMN ≥250/mm3 in a community-acquired setting, in the absence of recent broad-spectrum antibiotic exposure, receives empiric antibiotic treatment for a possible bacterial infection, according to the American Association for the Study of Liver Disease (AASLD).² SBP occurs in 7–30% of hospitalized patients with cirrhosis and ascites and the mortality rate ranges between 40 and 60%.³ However, a more infrequent yet perilous complication is SFP, diagnosed with PMN ≥250/mm3 in ascitic fluid and a positive fungal culture.⁴ SFP is underdiagnosed and undertreated due to a lack of clinical suspicion and prolonged time for fungal cultures that lead to delays in treatment, and carries a high risk of mortality from 33% to 100%^4,5. The physiologic factors that contribute to SFP and SBP are similar. Both involve ascites with higher serum albumin gradient caused by elevated portal vein pressure and inflammation in the gut that allows bacteria and fungi to invade and infect the peritoneal fluid.⁵ Candida is the most common cause of SFP due to its commensal nature in our gut.⁶ However, it appears that SFP is often not considered on the initial differential due to its rarity, and the diagnosis is often missed.⁷ This study aims to analyze healthcare costs, morbidity, and mortality associated with SFP compared to SBP, with the overarching goal of enhancing awareness and facilitating early diagnosis and treatment.

Methods

Ethics approval and consent to participate

Not applicable.

The study analyzed hospital readmissions for cirrhotic patients with SBP and SFP in the US. The complete dataset was obtained from the Nationwide Readmission Database (NRD) spanning 2016 to 2020 using the ICD-10 codes (K70.3, K74, K74.1, K74.3, K74.4, K74.5, K74.6, K74.60, K74.69, K71, K71.7, K65.2, B37, B38, B45, B44). This extensive, comprehensive database encompasses national hospital discharges. It is meticulously maintained as an integral component of the healthcare cost and utilization project (HCUP) by the Agency for Healthcare Research and Quality (AHRQ). This sample includes community, general hospitals, and academic medical centers but excludes long-term facilities. Hospitals were divided based on geographic region, urban vs. rural location, teaching status, ownership, and number of beds. NRD accounts for around 60 percent of U.S. hospitalizations reported in the American Hospital Association (AHA) Annual Survey Database. NRD is limited to data from community hospitals that are not rehabilitation or LTAC facilities. Each record in the NRD represents a single hospital discharge and includes a unique identifier, demographic data (age, gender, and race), hospital transfer status, admission type (emergent, urgent, or elective), primary and secondary diagnoses (up to 15), primary and secondary procedures (up to 15), expected primary and secondary insurance payers, total hospital charges, length of stay (LOS) and hospital characteristics (region, urban vs. rural location, bed-size and teaching status).

The diagnosis of SBP is based on the presence of a polymorphonuclear leukocyte (PMN) count ≥250 cells/mm³ in the ascitic fluid without an evident intra-abdominal source of infection. If criteria for PMN is met, then bacterial cultures of ascitic fluid may support but are not required to confirm SBP. The diagnosis of SFP if fungi were identified in the fluid culture. The criteria for fungal infection require a positive fungal culture along with an elevated PMN count. Hospital costs refer to the total expenditure related to the hospitalization of patients with SFP or SBP, including costs of diagnostic testing, medications, procedures, and LOS. Mortality refers to in-hospital mortality which is defined as any death occurring during the hospitalization with a diagnosis of SFP or SBP.

A multivariate logistic regression analysis was used to calculate the predictors of in-hospital mortality, LOS, and total hospital costs (THC) following adjustment for age (included in the study are adults >18 years), sex, race, grouped with Charlson comorbidity index, insurance type, mean annual household income, and hospital characteristics. Calculating THC involves utilizing HCUP Cost-to-Charge Ratio files. Cost is adjusted for inflation using Consumer Price Index (CPI) for medical care services in the U.S., all urban consumers, chained, with 2020 as the reference point. Institutional review board approval was not needed for this database study as the de-identified data is publicly available on the HCUP website. Statistical significance was met with two-tailed P < 0.05 and all statistical analyses were conducted using STATA version 10.

Study design

Study Population

Inclusion Criteria: Cirrhotic patient's readmissions to the U.S hospitals nationwide between 2016 and 2020.

Exclusion Criteria: Patients without a diagnosis of cirrhosis.

Data Collection

Identify cirrhotic patients using ICD – 10 diagnostic codes.

Collect demographic data (age, gender, etc.)

Record data on fungal infections (types of infections, causative organisms, hospital costs, LOS).

Capture readmission data within the five-year period for SBP and SFP.

Study Variables

Independent Variable: Time period (2016–2020), demographic and clinical variables.

Dependent Variable: Readmission rates.

Statistical Analysis

Calculate the readmission rate of fungal and bacterial infections per year.

Analyze trends over the five-year period using statistical methods and multivariable logistic regression.

Adjust for confounding variables such as age, gender, and comorbidities.

Ethical Considerations

None.

Limitations

Potential biases due to retrospective study design.

Generalizability may be limited to the population studied.

Implications

Understanding the trends in fungal infections in cirrhotic patients can guide healthcare policies and resource allocation.

Identifying factors associated with readmission can help develop strategies for prevention and management.

Results

The study included 393,195 index hospitalizations including patients with SBP and SFP, with 102,505 for 30-day and 157,079 for 90-day readmissions from 2016 to 2020. The overall mortality for index admission and 30-day readmission has increased. The mortality for 30 days readmissions for SBP has stayed the same with a mean of 18% (OR: 0.99, 95% CI: 0.93, 1.04), and for SFP is 16%, but has increased over the years from 15% to 18% (OR:1.01, 95% CI: 0.91,1.11). The mortality for 90 days for SBP is the same with a mean of 17% (OR:0.99, 95% CI: 0.95,1.03), and for SFP is 16% (OR:0.98 95% CI: 0.97,1.04). The mean age of patients was 57.9 years (standard deviation between 57.7 and 58.1). The mean LOS for 30-day readmission for SBP has been 9.4 days and for 90 days 9.4. For SFP 30-day and 90-day readmissions, LOS has increased over the years from 13.5 to 15.8 with a mean of 14.9. The 30-day index admission total charges for SBP are $42,258 and SFP are $51,739. The 30-day readmission total charges for SBP are 64, 266 and for SFP 89,913. The 90-day index admission total charges for SBP are $40, 622 and SFP are $48,518. The 90-day readmission total charges for SBP are 65,472 and for SFP 88,093.

Trend Analysis for 30-day Readmissions

Trends for 30-day readmissions for SBP have increased from 2.0% in 2016 to 2.2% in 2020 (P = 0.042) (Table 1) (Table 2). This study included readmission rates for various fungal infections (Table 3). Trends for 30-day readmissions for candidiasis have stayed the same with a mean of 0.8% (P = 0.602), mortality increased from 14% to 17% (P = 0.605), LOS from 14.4 to 15.4, hospital charges from $52,045 to $58,238. For cryptococcosis, there is not much change in readmission over the years, with a mean of less than 0.1% (P = 0.416), but mortality increased from 24% to 100% (P = 0.112), LOS increased from 9.7 days to 15.3 days, hospital charges from $30,346 to $71,502. The sample size for cryptococcosis cases was extremely limited, between 3 and 7 people per year. For aspergillosis, mortality decreased from 53% to 32% (P = 0.203), LOS increased from 17 days to 27, and hospital charges decreased from $132,015 to $110,428. The sample size of cases per year ranged between 6 and 18. For coccidiosis mortality has remained constant with a mean of 20%, LOS decreased from 7.2 days to 3.3 days, hospital charges decreased from $35,961 to $14,684. The 30-day readmissions for cryptococcosis, aspergillosis, and coccidiosis have stayed the same. A comparison of readmissions for SBP and SFP is shown below (Figure 1) (Figure 2).

Table 1.

Descriptive Statistics by Readmission Status at 30 days.

Characteristic	Overall			Readmission with SBP			Readmission with Fungal Infection
	Not-readmitted N = 290,690a	Readmitted N = 102,505a	P-valueb	Not-readmitted N = 384,717a	Readmitted N = 8,478a	P-valueb	Not-readmitted N = 390,089a	Readmitted N = 3,106a	P-valueb
Age	58.1 (0.05)	57.3 (0.07)	<0.001	57.9 (0.05)	55.1 (0.19)	<0.001	57.9 (0.0)	56.0 (0.3)	<0.001
Female	108,204 (37%)	39,320 (38%)	<0.001	144,446 (38%)	3078 (36%)	0.101	145,863 (37%)	1661 (53%)	<0.001
Insurance			<0.001			<0.001			0.097
Medicare	109,998 (38%)	40,288 (39%)		147,498 (38%)	2788 (33%)		149,091 (38%)	1195 (39%)
Medicaid	77,367 (27%)	30,725 (30%)		105,197 (27%)	2894 (34%)		107,211 (28%)	880 (28%)
Private	67,989 (23%)	21,277 (21%)		87,336 (23%)	1930 (23%)		88,524 (23%)	742 (24%)
Self-pay	21,218 (7.3%)	5936 (5.8%)		26,632 (6.9%)	522 (6.2%)		26,983 (6.9%)	171 (5.5%)
Other	13,459 (4.6%)	4104 (4.0%)		17,231 (4.5%)	333 (3.9%)		17,452 (4.5%)	112 (3.6%)
Unknown	659	175		822	12		828	6
Charlson index category			<0.001			0.276			<0.001
1	39,200 (13%)	12,665 (12%)		50,677 (13%)	1188 (14%)		51,549 (13%)	316 (10%)
2	22,541 (7.8%)	7511 (7.3%)		29,422 (7.6%)	630 (7.4%)		29,851 (7.7%)	200 (6.5%)
≥3	228,950 (79%)	82,329 (80%)		304,619 (79%)	6660 (79%)		308,689 (79%)	2590 (83%)
Hospital bedsize			0.006			0.747			0.072
Small	44,097 (15%)	14,861 (14%)		57,657 (15%)	1301 (15%)		58,517 (15%)	440 (14%)
Medium	75,919 (26%)	26,834 (26%)		100,521 (26%)	2232 (26%)		102,009 (26%)	744 (24%)
Large	170,674 (59%)	60,810 (59%)		226,539 (59%)	4945 (58%)		229,562 (59%)	1922 (62%)
Hospital location/teaching status			<0.001			0.064			0.827
Metropolitan teaching	213,129 (73%)	76,226 (74%)		283,028 (74%)	6327 (75%)		287,075 (74%)	2279 (73%)
Metropolitan non-teaching	60,144 (21%)	20,905 (20%)		79,319 (21%)	1731 (20%)		80,398 (21%)	652 (21%)
Non-metropolitan	17,418 (6.0%)	5373 (5.2%)		22,371 (5.8%)	420 (5.0%)		22,616 (5.8%)	175 (5.6%)

SBP, Spontaneous bacterial peritonitis.

^aMean (SE); n (%).

^bWilcoxon rank-sum test for complex survey samples for continous variables; chi-squared test with Rao & Scott's second-order correction for categorical variables.

Table 2.

Multivariable Logistic Regression Analysis Showing Predictors of 30-Day Readmission.

Characteristic	Overall			Readmission with SBP			Readmission with Fungal Infection
	ORa	95% CIa	P-value	ORa	95% CIa	P-value	ORa	95% CIa	P-value
Age	0.99	0.99, 0.99	<0.001	0.98	0.98, 0.98	<0.001	0.98	0.98, 0.99	<0.001
Female	1.05	1.02, 1.07	<0.001	0.95	0.89, 1.01	0.12	1.93	1.74, 2.14	<0.001
Insurance
Medicare	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
Medicaid	0.94	0.91, 0.97	<0.001	1.08	0.99, 1.18	0.074	0.8	0.69, 0.93	0.003
Private	0.76	0.74, 0.79	<0.001	0.92	0.84, 1.01	0.091	0.86	0.74, 1.00	0.047
Self-pay	0.67	0.64, 0.71	<0.001	0.77	0.66, 0.89	<0.001	0.64	0.49, 0.84	0.001
Other	0.76	0.71, 0.80	<0.001	0.81	0.68, 0.96	0.017	0.71	0.53, 0.94	0.019
Charlson index category
1	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
2	1.04	1.00, 1.09	0.068	1	0.87, 1.15	>0.9	1.16	0.89, 1.50	0.3
≥3	1.11	1.08, 1.15	<0.001	1	0.92, 1.10	>0.9	1.45	1.23, 1.72	<0.001
Hospital bedsize
Small	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
Medium	1.05	1.01, 1.09	0.015	0.98	0.88, 1.08	0.6	0.96	0.81, 1.15	0.7
Large	1.05	1.02, 1.09	0.003	0.95	0.87, 1.04	0.3	1.09	0.94, 1.27	0.3
Hospital location/teaching status
Metropolitan teaching	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
Metropolitan non-teaching	0.98	0.96, 1.01	0.2	1	0.93, 1.08	>0.9	1.07	0.95, 1.21	0.3
Non-metropolitan	0.88	0.83, 0.92	<0.001	0.89	0.76, 1.05	0.2	1.04	0.81, 1.33	0.8

SBP, Spontaneous bacterial peritonitis.

^aOR = Odds Ratio, CI = Confidence Interval.

Table 3.

Trend Analyses for 30 day Readmission.

Year	Overall, N = 393,195	2016, N = 69,666	2017, N = 74,146	2018, N = 78,969	2019, N = 82,989	2020, N = 87,425	P valuea for trend	OR/RR	95% CI
Index admission
Mean length of stay (days)	6.2	6.3	6.1	6.2	6.2	6.4	0.636	1	0.99, 1.01
Mean total hospital cost ($, 2020)b	20,183	20,273	19,582	19,481	20,120	21,305	0.684	1	0.99, 1.02
Readmission, any type
Anytype, readmission	102,505 (26%)	18,350 (26%)	19,047 (26%)	20,551 (26%)	21,811 (26%)	22,746 (26%)	0.86	1	0.99, 1.01
In-hospital mortality - all readmissions in 30 days (%)	10,983 (8.7%)	1922 (8.5%)	2015 (8.6%)	2098 (8.3%)	2282 (8.6%)	2667 (9.6%)	0.066	1.02	1.00, 1.04
Mean length of stay (days)	6.6	6.5	6.4	6.5	6.6	6.7	<0.001	1.01	1.00, 1.01
Mean total hospital cost ($, 2020)b	20,770	19,749	20,093	20,041	20,854	22,694	<0.001	1.02	1.01, 1.03
Readmission with SBP
SBP, readmission	8478 (2.2%)	1421 (2.0%)	1525 (2.1%)	1756 (2.2%)	1816 (2.2%)	1959 (2.2%)	0.042	1.02	1.00, 1.05
In-hospital mortality - all readmissions in 30 days (%)	1551 (18%)	247 (17%)	277 (18%)	347 (19%)	326 (18%)	354 (17%)	0.599	0.99	0.93, 1.04
Mean length of stay (days)	9.4	9.4	9.5	9.4	9.3	9.3	0.001	0.99	0.98, 1.00
Mean total hospital cost ($, 2020)b	32,839	32,458	34,876	31,533	31,531	33,876	0.437	0.99	0.96, 1.02
Readmission with candidiasis
Candidiasis, readmisssion	2979 (0.8%)	560 (0.8%)	525 (0.7%)	607 (0.8%)	638 (0.8%)	649 (0.7%)	0.602	0.99	0.95, 1.03
In-hospital mortality - all readmissions in 30 days (%)	458 (15%)	80 (14%)	95 (18%)	78 (12%)	89 (14%)	116 (17%)	0.605	0.98	0.89, 1.07
Mean length of stay (days)	14.9	14.4	14.7	15.3	14.7	15.4	0.316	1	1.00, 1.01
Mean total hospital cost ($, 2020)b	56,210	52,045	55,830	60,010	54,300	58,238	0.788	0.99	0.93, 1.05
Readmission with cryptococcosis
Cryptococcosis, readmission	53 (<0.1%)	11 (<0.1%)	7 (<0.1%)	10 (<0.1%)	20 (<0.1%)	4 (<0.1%)	0.416	0.9	0.70, 1.16
In-hospital mortality - all readmissions in 30 days (%)	22 (40%)	3 (24%)	3 (46%)	4 (44%)	7 (33%)	4 (100%)	0.112	1.76	0.93, 3.97
Mean length of stay (days)	18.5	9.7	39	25.1	10	15.3	0.47	0.97	0.91, 1.04
Mean total hospital cost ($, 2020)b	90,030	30,346	172,728	177,186	28,894	71,502	0.722	0.94	0.60, 1.47
Readmission with aspergillosis
Aspergillosis (with pneumonia), readmission	60 (<0.1%)	10 (<0.1%)	14 (<0.1%)	13 (<0.1%)	6 (<0.1%)	18 (<0.1%)	0.996	1	0.75, 1.34
In-hospital mortality - all readmissions in 30 days (%)	22 (35%)	6 (53%)	6 (41%)	5 (36%)	0 (0%)	6 (32%)	0.203	0.62	0.30, 1.29
Mean length of stay (days)	32.3	17	35.5	46	34.3	27.1	<0.001	1.19	1.13, 1.26
Mean total hospital cost ($, 2020)b	220,850	132,015	295,012	354,552	153,821	110,428	0.323	1.15	0.82, 1.61
Readmission with coccidiosis
Coccidiosis, readmisssion	32 (<0.1%)	5 (<0.1%)	9 (<0.1%)	3 (<0.1%)	4 (<0.1%)	10 (<0.1%)	0.901	1.02	0.72, 1.46
In-hospital mortality - all readmissions in 30 days (%)	7 (20%)	1 (20%)	1 (14%)	2 (56%)	0 (0%)	3 (21%)	1	NA (due to low numbers)	NA (due to low numbers)
Mean length of stay (days)	14.6	7.2	31.6	50.5	5	3.3	<0.001	0.72	0.62, 0.82
Mean total hospital cost ($, 2020)b	104,437	35,961	241,479	422,702	17,323	14,684	0.132	0.61	0.22, 1.62

OR, odds ration for index admission and readmission; RR, risk ratio for LOS and total cost; SBP, Spontaneous bacterial peritonitis.

^aP value adjusted for age, sex, grouped Charlson comorbidity index, insurance type, and hospital characteristics; Logistic regression was used for index admission, readmission, and in-hospital mortality; Poisson regression for length of stay; generalized linear model with gamma distribution for total hospital cost.

^bAdjusted for inflation using Consumer Price Index (CPI) for Medical care services in U.S. city average, all urban consumers, chained, with 2020 as the reference point.

Figure 1.

30-day readmissions for SBP and SFP with reference to all cirrhosis readmission [2016 (n)–18,350, 2017 (n)–19,047, 2018 (n)– 20,551, 2019 (n)–21,811, 2020 (n)– 22,746].

Figure 2.

Comparing 30-day readmissions for SBP vs SFP vs all cirrhosis readmissions.

The overall mortality for index admission and 30-day readmission has increased. The mortality for 30-day readmissions for SBP has stayed the same with a mean of 18% and for SFP the mean is 16%, but the trend has increased over the years from 15% to 18% (Figure 3). The mean LOS for index admission has been 6.2 days. The mean LOS for 30-day readmission for SBP has been 9.4 days (Figure 4). The 30-day index admission total charges for SBP are $42,258 and SFP are $51,739. The 30-day readmission total charges for SBP are $64,266 and for SFP $89,913 (Figure 5). THC were calculated using HCUP Cost-to-Charge Ratio files and adjusted for inflation based on the CPI for medical care services in the U.S., with 2020 as the reference year.

Figure 3.

Comparison of 30-day readmission, mortality, and LOS of various fungal infections.

Figure 4.

Mean LOS for fungal infection – 30-day readmissions after index admission.

Figure 5.

Total Hospital Costs for 30 days for various fungal infections with year 2020 as a reference and adjusted for inflation based on the Consumer Price Index for medical care services in the U.S.

Trend Analysis for 90-day Readmissions

For 90-day readmission, there was an increasing but not significant trend with 3.9% in 2016 to 4.1% in 2020 (P = 0.157) (sTable 1) (sTable 2). Trends for 90-day readmissions for candidiasis have decreased from 1.5% to 1.3% (P = 0.093), mortality increased from 14% to 15% (P = 0.152), LOS decreased from 14.8 days to 14.3, hospital charges decreased from $54,189 to $53,184. For cryptococcosis readmissions have stayed the same, mortality increased from 29% to 31% (P = 0.853), LOS decreased from 17.9 days to 15.3 days, and hospital charges increased from $50,175 to $84,123. For aspergillosis mortality decreased from 50% to 35% (P = 0.137), LOS increased from 17.9 days to 33 days, and hospital charges from $121,522 to $159,333. For coccidiosis, mortality decreased from 14% to 12% (P = 0.214), LOS decreased from 8 days to 3.4 days, and hospital charges decreased from $37,917 to $13,150. The 90-day readmissions for cryptococcosis, aspergillosis, and coccidiosis have stayed the same (Figure 6).

Figure 6.

90-day readmissions for SBP and SFP.

The mortality for 90 days for SBP is the same with a mean of 17%, and for SFP is 16%. The mean LOS for 90-day readmission for SBP was 9.4. The 90-day index admission total charges for SBP are $40, 622 and SFP are $48,518. The 90-day readmission total charges for SBP are $65,472 and for SFP $88,093 (sTable 3).

Discussion

Important findings observed from this analysis of hospitalizations for fungal infections include increased mortality, LOS, and hospital costs. Early recognition of SFP in diagnostic workup, especially in patients unresponsive to antibiotic treatment is crucial for improved outcomes in SFP.

In patients with liver cirrhosis, SBP and SFP are some of the serious complications encountered in hospital admission. They significantly impact mortality, morbidity, and utilization of healthcare resources. This study focused on readmission trends among liver cirrhosis patients with SBP and SBF and the importance of effectively recognizing and managing the infection to mitigate the risk of readmissions and improve patient outcomes.

Bacterial infections are present in approximately one-third of patients with cirrhosis who are hospitalized, a much higher prevalence than in those without cirrhosis.⁸ Bacterial translocation, the passage of bacteria from the gut to the bloodstream and other extraintestinal sites, and decreased host defenses have been implicated in the pathogenesis of these spontaneous infections^9,10. Among spontaneous infections, the most common (∼60%) are Gram-negative bacteria and fungi representing less than 5% of infections.¹¹ More recently, there has been a shift toward Gram-positive and multidrug-resistant organisms (MDRO), particularly in nosocomial and healthcare-associated SBP.^{11, 12, 13} Prompt diagnosis and treatment of SBP are crucial to prevent disease progression, systemic inflammation, and multiorgan failure, which may necessitate hospital readmissions.^{14, 15, 16}

The mechanisms underlying the association between SBP and SFP in readmission are complex and multifactorial. SBP may lead to hepatic decompensation, and renal impairment requiring dialysis, hospital-based interventions, and close monitoring.^17,18 Similarly, fungal infections may exacerbate liver dysfunction, precipitate sepsis, and trigger hemodynamic instability, necessitating aggressive management and hospital readmissions.^19,20 The presence of SBP may predispose patients to fungal superinfections, further complicating clinical management and increasing the likelihood of readmissions.²¹ SFP is defined as fungal infection of ascitic fluid in the presence of ascitic neutrophil count of greater than 250 cells/mL and culture of ascitic fluid is positive for fungi.¹⁴

From our analysis, the 30-day and 90-day readmission mortality rates for SBP and SFP were very similar, ranging between 16% and 18%. However, SFP had a significantly longer LOS and higher readmission costs than SBP. Potential explanations for this discrepancy include diminished provider suspicion towards SFP, prolonged fungal culture times leading to delayed treatment initiation, and an elevated severity of fungal peritonitis relative to bacterial peritonitis.²² Fungi are much larger compared to bacteria, so fungal translocation across the gut requires higher intestinal permeability which is associated with increased severity of the cirrhosis.²³

Among various fungal infections, 30-day and 90-day readmissions are highest for candidiasis. For the 30-day readmission data, the in-hospital mortality is highest for cryptococcus followed by aspergillosis, Coccidioides, and candidiasis. The LOS is highest for aspergillosis followed by cryptococcus, candidiasis, and Coccidioides. The THC are highest for aspergillosis followed by Coccidioides, cryptococcus, and Candidiasis. For the 90-day readmission data, the in-hospital mortality is highest for aspergillosis followed by cryptococcosis, Coccidioides, and candidiasis. The LOS is highest for aspergillosis followed by cryptococcus, candidiasis, and Coccidioides. The THC are highest for aspergillosis followed by Coccidioides, cryptococcus, and Candidiasis.

For the significant trends among the years, in 30- and 90-day readmissions, there are increasing trends for in-hospital mortality observed for candidiasis and cryptococcus, and a decreasing trend for aspergillosis while the trend for Coccidioides increased for 30 days and decreased with 90-day admissions. The P-value for the trends is not significant due to the small sample size from the NRD. Consequently, patients with more severe cirrhosis are more likely to develop SFP, requiring longer hospitalization courses. Although SFP is rare, high clinical suspicion and earlier treatment can significantly reduce the costs associated with readmissions.²³ A lack of improvement within 48 h after admission is linked to an increased risk of SFP. Hence, fungi should be sought as potential pathogens in ceftriaxone or cefotaxime-resistant peritonitis which occurs in approximately 7%–17% of cirrhotic patients.^24,25 Candidal infections overall are the costliest fungal infections in the United States, with a cost of over $1.2 billion.²⁶ Candida PCR of peritoneal fluid is an inexpensive test that takes an average of 2 days and has a sensitivity between 80% and 90%.^{27, 28, 29}

SFP is a severe and often underdiagnosed complication in patients with decompensated liver cirrhosis. The clinical relevance of SFP is underscored by its high mortality rates, with studies reporting in-hospital mortality ranging from 33.3% to 100% and one-month mortality between 50% and 73.3%.³⁰ This high mortality emphasizes the need for early recognition and prompt management of SFP to improve patient outcomes.

Ascitic fluid analysis should include both PMN count and cultures for both bacteria and fungi. In cases where patients are not responding to antibiotics or meet criteria for severe sepsis, higher Model for End-Stage Liver Disease (MELD) or Child-Pugh scores, long-term antibiotic use, or recent invasive procedures testing early for fungal pathogens can prevent progression to systemic infections.³¹ Standard ascitic fluid culture methods can delay diagnosis, as fungal growth often requires extended culture time. Clinicians should consider using rapid diagnostic tools, such as polymerase chain reaction assays, 1,3-beta-D-glucan testing, or antigen-based tests for fungi, which may enhance early detection compared to traditional cultures for critically ill cirrhotic patients^30,32.

In high-risk patients, with prior hospitalization/episodes of SBP/recent use of antibiotics, starting antifungal prophylaxis could reduce the risk of SFP. Echinocandins, such as caspofungin, are recommended as initial therapy due to their broad-spectrum efficacy and lower toxicity compared to azoles.^33,34 De-escalation to fluconazole may be appropriate once culture results confirm susceptibility and the patient's condition stabilizes.³⁴ This should be carefully weighed, considering potential resistance and adverse effects of antifungals.³⁵ The question of whether to use antifungal prophylaxis in high-risk cirrhotic patients remains an area for further research. Until more definitive guidelines are established, prophylaxis should be considered on a case-by-case basis. This measure can potentially reduce hospital costs by reducing LOS and secondary complications.

Drawbacks of NRD: The NRD is an annual dataset containing inpatient records for patients discharged within a given calendar year. It includes records for patients who were admitted in the previous year but discharged in the current year, but it excludes patients who were admitted in the current year and discharged in the following year. It is not possible to merge data across different years to create a multi-year database. The patient linkage numbers (NRD_VisitLink) do not allow tracking of the same individual across years, and the hospital identifiers in HOSP_NRD do not remain consistent over time. Additionally, the NRD cannot track patients who were admitted to a hospital in one state and later readmitted or transferred to a hospital in another state.

Strengths and limitations

The strengths of the study include the utilization of a comprehensive hospitalization dataset, the NRD, which is the largest all-payer inpatient hospitalization database in the United States, which improves the generalizability of our findings. In addition, we utilized a systematic and validated algorithm to identify cases of SBP and SFP among hospitalizations with alcoholic and non-alcoholic cirrhosis.

However, our study has certain limitations. The database primarily records hospitalizations rather than individual patient data, resulting in separate entries for each inpatient stay and potentially duplicating records for patients transferred between hospitals within the same state. NRD lacks information on medication adherence, access to subspecialty care, severity or control of cirrhosis, intensive care unit stay at the time of admission, or readmissions. Further, given our focus on inpatient disease burden, might have overlooked SBP and SFP instances in outpatient settings, particularly among undiagnosed or non-hospitalized patients were not accurately captured in our analyses. Another limitation stems from the utilization of ICD-10 codes to document hospitalizations, which could introduce coding errors. Despite the limitations, the large sample size, scientific inquiries, long study duration, and analysis techniques enable this study to emphasize the importance of peritonitis associated with cirrhosis readmissions.

Patients with SFP had significantly higher mortality than patients with SBP. SFP is a potentially fatal complication and was found to be significantly associated with higher rates of readmissions. Fungal infections, particularly those caused by Candida species, have high morbidity and mortality. Among the readmissions, Candida was by far the most common fungal pathogen, though Cryptococcus and Aspergillus infections were much rarer, they had a higher cost and longer stay. Clinicians should prioritize fungal infections in their differential diagnoses and consider the most prevalent fungal infections in their region for better patient outcomes.

Credit authorship contribution statement

SC as a first author analyzed and interpreted the data, generated tables, and graphs, wrote most of the sections, and was a major contributor to writing the manuscript. TA as a second and third author analyzed the data and was a contributor to the discussion section of the manuscript. RW as a senior author reviewed the article and made necessary changes to the manuscript. All authors read and approved the final manuscript.

Funding

Not applicable.

Consent for publication

Not applicable.

Availability of data and material

Data has been collected from the Nationwide HCUP database https://hcup-us.ahrq.gov/databases.jsp.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supplementary data

The following is the Supplementary data to this article.