Proton Pump Inhibitor Use in Patients With Cirrhosis and Its Association With Spontaneous Bacterial Peritonitis

Abstract

Background: Cirrhosis is a major cause of morbidity and mortality in the United States, with spontaneous bacterial peritonitis (SBP) being a serious complication. Established SBP risk factors include gastrointestinal bleeding and low ascitic protein, but the role of proton pump inhibitors (PPIs) remains unclear. Objective: This study evaluated the impact of PPI use on primary SBP development in hospitalized patients with cirrhosis. Additional objectives included reviewing PPI prescribing patterns and associated clinical outcomes. Methods: An institutional review board–approved, retrospective chart review was conducted on adults (≥18 years) with cirrhosis admitted for presumed SBP between June 1, 2022, and June 30, 2024. Exclusion criteria included pregnancy, incarceration, and recent or current upper gastrointestinal bleeding. Patients were grouped by PPI exposure, defined as PPI use prior to admission. The primary outcome was SBP incidence; secondary outcomes included mortality and hepatic decompensation events. Results: Eighty-one patients were included: 42 reported home PPI therapy, and 39 did not. SBP incidence was 33.3% in the PPI group versus 20.5% in the no PPI group (χ² = 0.249, P = 0.618). Worsening ascites occurred in 99%, encephalopathy in 42%, varices in 11%, and suspected hepatorenal syndrome in 21%. In-hospital mortality was 9.9%. PPI indications were often undocumented. Conclusion and Relevance: Although no significant association was found between home PPI use and SBP, frequent undocumented use and potential overuse of PPIs underscore the need for targeted intervention. Pharmacists are well-positioned to lead stewardship efforts by reviewing indications and minimizing unnecessary therapy to enhance safety and outcomes.

Introduction

Liver cirrhosis is a significant source of morbidity and mortality in the United States, with a recent National Health Interview Survey estimating that approximately 4.5 million adults aged 18 years and older have been diagnosed with liver disease. ¹ Cirrhosis is the end stage of chronic liver disease that is characterized by an asymptomatic “compensated” phase followed by a progressive phase marked by liver function decline termed “decompensated cirrhosis.^”2 Changes in portal pressure and liver function drive the transition to the decompensated phase of cirrhosis resulting in ascites, gastrointestinal bleeding, encephalopathy, and jaundice; this transition greatly impacts patient’s morbidity and mortality. ² In addition, advanced liver dysfunction in cirrhosis causes an alteration of the immune system and promotes bacterial translocation which increases patients’ susceptibility to spontaneous bacterial peritonitis (SBP).^{3 -8}

Spontaneous bacterial peritonitis is one of the most common and life-threatening bacterial infections in hospitalized patients with cirrhosis. ⁹ SBP refers to the infection of the ascitic fluid in the absence of another potential source for intra-abdominal infection. One SBP episode confers substantial medical costs to patients and hospital systems and often prolongs hospitalization. Specifically, a study by Saleem et al ¹⁰ reported that SBP was associated with a 59% increase in hospital charges and a 50% longer length of stay compared with cirrhotic patients without SBP (P < 0.0001). Similarly, in patients with decompensated cirrhosis, SBP was linked to an 18.4% increase in costs and a 14.4% longer hospitalization. ¹¹ Risk factors for SBP include upper gastrointestinal bleeds, previous SBP episodes, low ascitic protein, and elevated model for end-stage liver disease (MELD) scores.^{12 -14} In addition to these proven risk factors, it is hypothesized that proton pump inhibitors (PPIs) may contribute to SBP.

Proton pump inhibitors are among the most prescribed medications in the United States; however, their overuse contributes to undue economic costs and has been attributed to an increasing number of adverse effects. ¹⁵ A national analysis reported that outpatient PPI use more than doubled between 2002 and 2009, with 62.9% of users having no documented gastrointestinal complaint, diagnosis, or other appropriate indication for therapy. ¹⁶ PPIs raise gastric pH by decreasing gastric acid, effectively treating diseases such as gastroesophageal reflux disease (GERD), peptic ulcer disease (PUD), and Barrett esophagus. While PPIs have an established role, their acid suppressive action promotes intestinal bacterial overgrowth. Furthermore, the half-life of PPIs is increased in patients with cirrhosis. A gastric environment primed for bacterial overgrowth coupled with an increased PPI half-life may increase SBP risk in patients with cirrhosis and ascites.^3,4,6,17 Additional long-term risks of PPI therapy include osteoporosis, hypomagnesemia, iron deficiency, and Clostridioides difficile colitis. ¹⁸ To minimize harm, the American Gastroenterological Association recommends routine reassessment of PPI indications, dose, and duration to guide deprescribing efforts, ¹⁵ while the American Association for the Study of Liver Diseases (AASLD) recommends avoiding them in acute-on-chronic liver failure unless strongly indicated. ¹⁹

Observational and retrospective studies have sought to describe the impact of PPI therapy on hepatic decompensation events, specifically the first SBP occurrence^{4,6,20 -26} as well as on long-term mortality in patients with cirrhosis,^6,24,26,27 however the results lack consistency and generalizability. The variability in findings may be attributed to varying definitions of PPI use based on dose and duration of acid reducing therapy, along with the inclusion of cirrhotic patients with varying levels of decompensation. While awaiting results from prospective randomized trials, ²⁸ further retrospective studies are warranted to clarify the role of PPI therapy in cirrhosis and its link to SBP. The goal of this study was to evaluate the impact of PPIs on SBP development in patients with cirrhosis. In addition, this study sought to assess the appropriateness of PPI therapy, including medication choice, dosing, duration, and documented indications.

Methods

This was an Institutional Review Board approved retrospective chart review of patients with cirrhosis. Patients were eligible for inclusion if they were 18 years of age or older, admitted between June 1, 2022 and June 30, 2024 for evaluation of SBP, and had a documented diagnosis of cirrhosis as identified by International Classification of Diseases, Tenth Revision (ICD-10) codes. Patients were excluded if they had a recent (within 2 weeks prior to hospitalization) or current upper gastrointestinal bleed, a recent or current variceal bleed, were pregnant, or were incarcerated at the time of admission. These exclusions were identified based on documentation in the electronic health record, including admission records and history and physical notes.

The primary outcome of this study was the incidence of SBP, defined as a polymorphonuclear leukocyte (PMN) count greater than 250 cells/mL in ascitic fluid, with or without a positive ascitic fluid culture. Based on the hypothesis that PPI therapy increases the risk of SBP (Figure 1), patients were categorized into PPI and non-PPI groups for comparison. PPI therapy was defined as documented use of a PPI prior to hospital admission based on the home medication list or admission history and physical.

Figure 1.

Primary outcome: incidence of SBP.

Abbreviations: PPI, proton pump inhibitor; SBP, spontaneous bacterial peritonitis.

Secondary outcomes included use of PPI therapy prior to hospital admission, with documentation of drug name, dosage, and indication, when available. Additional clinical endpoints included in-hospital mortality, receipt of hemodialysis, and admission to the intensive care unit (ICU). Hepatic decompensation events were also evaluated and defined as follows:

• New or worsening ascites, as documented by providers, confirmed by radiographic imaging, or indicated by the need for repeat paracentesis

• New or worsening hepatic encephalopathy, based on clinical documentation, escalation of lactulose therapy, or initiation of rifaximin

• Newly diagnosed esophageal varices, identified through imaging or endoscopic reports

• Suspected new-onset hepatorenal syndrome (HRS), determined by ICD-10 diagnosis codes or administration of albumin for HRS, terlipressin, midodrine and octreotide on the general medical floor, or norepinephrine and octreotide in the ICU.

Demographic, clinical, and medication data were collected via retrospective chart review. Variables collected included age, sex, body mass index, race, smoking status, alcohol use, comorbidities, liver disease etiology, and hepatic decompensation events. Hospital course data included length of stay, ICU admission, dialysis, and temperature. Home and inpatient medication use was recorded with agent, dose, frequency, and indication when available. Paracentesis details and ascitic fluid analysis were documented, along with admission laboratory values including blood counts, liver function tests, coagulation profile, and renal markers.

Study data were extracted from the electronic medical record. The primary end point and other dichotomous measures are reported as percentages and were analyzed using χ². All continuous data are reported as either a mean with standard deviation for parametric data sets or a median with interquartile range if a nonparametric data set. Parametric data sets were analyzed using Student T test and nonparametric using Mann-Whitney U.

Results

A total of 176 patients were identified and screened for inclusion. The most common reasons for exclusion were absence of a documented cirrhosis ICD-10 code (n = 35), lack of an eligible hospital admission associated with a cirrhosis code (n = 33), and presence of a concurrent upper gastrointestinal bleed (n = 12). A complete list of exclusion criteria is provided in Figure 2. The overall cohort had a median age of 54 years, 57% were male, and alcoholic liver disease was the most common etiology of cirrhosis (45.7%). Among patients in the PPI group, 67% had a documented diagnosis of GERD or PUD, compared with 10% in the non-PPI group. Documented home medications prior to admission included lactulose in 38% and rifaximin in 12.8% of patients in the PPI group, compared with 12.8% and 0%, respectively, in the non-PPI group. Detailed demographic and baseline characteristics are presented in Table 1.

Table 1.

Baseline Demographics.

	Overall (N = 81)	PPI (n = 42)	No PPI (n = 39)
Age, years, median (IQR)	54 (43, 64.5)	58.5 (49.8, 65.5)	51 (40, 59)
Gender, male, n (%)	46 (56.8)	21 (50.0)	25 (64.1)
BMI, kg/m2, median (IQR)	29.1 (24.5, 34.8)	27.2 (23.4, 35)	31 (24.9, 34.6)
Comorbidities, n (%)
Atrial Fibrillation	2 (2.5)	2 (4.8)	0 (0)
Cardiovascular Disease	50 (61.7)	29 (69)	21 (53.8)
CKD	9 (11.1)	7 (16.7)	2 (5.1)
GERD	28 (34.6)	24 (57.1)	4 (10.3)
PUD	4 (4.9)	4 (9.5)	0 (0)
Current Alcohol Use, n (%)	33 (40.7)	11 (26.1)	22 (56.4)
Etiology of Liver Disease a , n (%)
Alcoholic Liver Disease	37 (45.7)	19 (45.2)	18 (46.2)
Chronic Viral Hepatitis	4 (4.9)	0 (0)	4 (10.3)
Nonalcoholic Steatohepatitis	13 (16)	9 (21.4)	4 (10.3)
Unknown	29 (35.8)	14 (33.3)	15 (38.5)
MELD Score, median (IQR)	17 (13, 24)	16 (13, 21)	21 (15, 25.25)
Laboratory Values at Admission, median (IQR)
WBC count, K/μL	9 (5.6, 12.2)	8.7 (4.6, 11.1)	9.8 (6.5, 13)
Platelets, K/μL	152 (103, 208)	139 (101, 196)	156 (109, 225)
Serum Creatinine, K/μL	1 (0.7, 1.5)	1.1 (0.8, 1.7)	0.8 (0.6, 1.4)
AST, U/L	75 (43, 110)	62 (37, 100)	98 (58, 126)
ALT, U/L	29 (22.5, 50)	26 (22, 37)	43 (27, 55)
Total bilirubin, mg/dL	2.9 (1.5, 5.7)	2.7 (1.1, 4.7)	3.8 (1.8, 6.3)
Home Medication Prior to Admission, n (%)
Beta-Blocker	32 (39.5)	19 (45.2)	13 (33.3)
Proton Pump Inhibitor	42 (51.9)	41 (100)	0 (0)
Lactulose	21 (25.9)	16 (38.1)	5 (12.8)
Rifaximin	7 (8.6)	7 (16.7)	0 (0)
Anticoagulant Agent	7 (8.6)	6 (14.3)	1 (2.6)
Antiplatelet Agent	8 (9.9)	7 (16.7)	1 (2.6)

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; CKD, chronic kidney disease; GERD, gastroesophageal reflux disease; IQR, interquartile range; MELD, model for end-stage liver disease; PUD, peptic ulcer disease.

^aPatient could have more than one etiology.

Figure 2.

Inclusion and exclusion.

Abbreviations: ICD-10, International Classification of Diseases, Tenth Revision code.

The overall incidence of SBP in the study population was 23.4%. Among patients receiving PPI therapy prior to admission, 33.3% developed SBP, compared with 20.5% of those not on PPI therapy; this difference was not statistically significant (χ² = 0.249; P = 0.618). Ascitic fluid analysis was performed on samples obtained during the first paracentesis of hospital admission. The median volume of ascitic fluid removed was approximately 5 L (IQR, 2.1-6.2 L), with similar volumes between groups. The median total nucleated cell count was 310 K/μL (IQR, 134-1098 K/μL). Positive ascitic fluid cultures were identified in 4.9% of patients, more frequently in the PPI group (7.1%) than the non-PPI group (2.6%). Additional fluid analysis details are shown in Table 2.

Table 2.

Ascitic Fluid Analysis From the First Paracentesis Upon Admission.

	Overall (N = 81)	PPI (n = 42)	No PPI (n = 39)
Volume of Fluid Removal, L, median (IQR)	5 (2.1, 6.2)	5.1 (2.2, 7.5)	4.9 (1.6, 6.1)
Ascitic Fluid a , median (IQR)
Total Nucleated Count, K/μL	310 (134, 1098)	257 (104.5, 1124)	340 (143.8, 925.8)
Fluid Neutrophils, %	24 (10, 67)	24 (10.5, 73.5)	22.5 (9, 59.25)
Albumin, g/dL	1.1 (0.6, 1.5)	1.1 (0.55, 1.3)	1.3 (0.68, 2.78)
Positive Ascitic Fluid Culture, n (%)	4 (4.9)	3 (7.1)	1 (2.6)
PMN, cells/μL, median (IQR)	46 (10.7, 338.2)	34.6 (7.2, 496.5)	55.1 (19, 298.3)

Abbreviations: IQR, interquartile range; PMN, polymorphonuclear leukocytes.

^aA total of 4 paracentesis had no cell counts performed.

The median hospital length of stay was 6 days (IQR, 3-11), with a slightly longer stay in the PPI group (6.5 days; IQR, 3.8-12) than the non-PPI group (5 days; IQR, 3-11). Intensive care unit admission occurred in 29.6% of patients overall, including 23.8% of the PPI group and 35.9% of the non-PPI group. Dialysis was required in 4.9% of patients, with similar rates between groups. In-hospital mortality was 9.9% overall, lower in the PPI group (4.8%) compared with the non-PPI group (15.4%). Hepatic decompensation events in the overall cohort included new or worsening ascites in 99% of patients, new or worsening hepatic encephalopathy in 42%, newly diagnosed esophageal varices in 11%, and hepatorenal syndrome in 5%. In the overall cohort, 99% of patients had new or worsening ascites, 42% had hepatic encephalopathy, 11% had new esophageal varices, and 21% developed suspected hepatorenal syndrome. Additional secondary outcomes are summarized in Table 3.

Table 3.

Secondary Outcomes.

	Overall (N = 81)	PPI (n = 42)	No PPI (n = 39)
Hospital length of stay, days, median (IQR)	6 (3, 11)	6.5 (3.8, 12)	5 (3, 11)
ICU admission, n (%)	24 (29.6)	10 (23.8)	14 (35.9)
Received dialysis, n (%)	4 (4.9)	2 (4.8)	2 (5.1)
In-hospital mortality, n (%)	8 (9.9)	2 (4.8)	6 (15.4)

Abbreviations: ICU, intensive care unit; IQR, interquartile range.

Among the 81 patients enrolled, most received antibiotics (89%) and had an order for albumin (74%). About half were treated for hepatic encephalopathy with lactulose and rifaximin (each 56%). Octreotide and norepinephrine were ordered in 20% of patients, while midodrine was ordered in 10%. The primary indication for antibiotic therapy was intra-abdominal infections, including SBP treatment and prophylaxis, accounting for 50.8% of cases. The most commonly used antibiotics were ceftriaxone, cefepime, piperacillin-tazobactam, vancomycin, and metronidazole.

Discussion

This retrospective study assessed the association between outpatient PPI use prior to hospital admission and the incidence of primary SBP in patients with cirrhosis, finding no statistically significant difference in SBP rates between patients receiving home PPI therapy and those not treated with PPIs. This study cohort exhibited advanced liver disease, with a median MELD score of 17. Among these patients, 99% experienced new or worsening ascites, 42% had hepatic encephalopathy, 11% developed new esophageal varices, and 21% developed suspected hepatorenal syndrome. In addition, many patients were already receiving lactulose and rifaximin prior to admission, further underscoring the severity of their condition. These findings highlight the high-risk nature of this population and their increased likelihood of decompensation events, which in turn elevate the risk of mortality.

Prior studies assessing the association between PPI use and the risk of SBP in patients with cirrhosis have reported inconsistent findings.^{4,6,20 -26} A major contributor to this variability is the heterogeneity in inclusion and exclusion criteria across studies. For instance, some investigations focused exclusively on hospitalized patients with decompensated cirrhosis, ²¹ while others employed broader definitions based on clinical criteria, administrative coding such as ICD diagnoses,^6,20,25,27 or surrogate indicators like receipt of diuretics for ascites control.^4,29 Exclusion criteria also differed considerably—some studies omitted patients with a prior history of SBP or concurrent infections, ²⁹ receipt of intravenous PPI therapy, ⁶ history of liver transplantation,^{4,21,23,29,30} recent variceal or gastrointestinal bleeding,^4,6,21,25,29 or baseline decompensated cirrhosis.^23,27 These methodological inconsistencies complicate direct comparison of outcomes and likely contribute to the conflicting evidence surrounding the potential risk of SBP associated with PPI use in cirrhotic patients. While the exclusion of patients with recent gastrointestinal bleeding or prior liver transplantation is justifiable, given their independent associations with SBP risk, excluding patients with decompensated cirrhosis is more questionable, as they represent the population most susceptible to developing SBP.

In addition to differences in study design, definitions of PPI exposure were often imprecise, with limited information on duration of use. In this study, we selected outpatient PPI use for comparison based on the hypothesis that prolonged exposure may pose a greater risk of SBP than short-term inpatient therapy. However, limitations inherent to retrospective home medication data may compromise the accuracy of assessing chronic PPI exposure. Definitions of PPI use vary substantially across studies, reflecting differences in both timing and dosage criteria. One study identified an increased risk associated with PPI use exceeding 180 days. ²¹ Others have evaluated cumulative dosing, ²⁷ classified patients as baseline users if a prescription was filled within 90 days of index date, ²³ or included short-term use, such as within 1 week prior to SBP recurrence ²⁹ or within two weeks of hospital admission. ²⁵ Nevertheless, several studies, including ours, are constrained by reliance on single point-in-time documentation, which provides only a binary indication of PPI use without information on therapy duration, initiation date, or patient adherence.^4,7,20,26

This study identified a high incidence of potentially inappropriate PPI use in hospitalized cirrhotic patients. Of the 81 patients included, 42 reported home PPI therapy prior to admission, while 53 patients received PPI therapy during hospitalization, including 16 who were newly started on PPIs. Indications for PPI therapy were poorly documented, with only 37.7% of all inpatient recipients and roughly half of the newly initiated patients having a clear clinical rationale noted. These findings raise concerns about PPI overuse and insufficient documentation, highlighting a critical need for regular reassessment of PPI use regarding indication, dose, and duration. Pharmacists are uniquely positioned to lead medication stewardship efforts by reviewing therapy appropriateness, ensuring adherence to evidence-based guidelines, and facilitating deprescribing when warranted. Such pharmacist-driven interventions could reduce unnecessary PPI exposure, decrease complications, and improve patient safety and outcomes in this vulnerable population.

In addition to PPI stewardship, pharmacists also play a key role in the identification, prompt treatment, and appropriate secondary prophylaxis of SBP. The AASLD advises empiric antibiotic therapy and intravenous albumin for the treatment of SBP with albumin being of highest benefit for patients who present with or are at risk of acute kidney injury.^13,31 Antibiotic use was well documented in this study, with approximately 50.8% of antibiotic orders indicated for intra-abdominal infections, including SBP and prophylaxis. However, albumin dosing appropriateness could not be evaluated due to data collection limitations, as only medication orders without dosing details were captured. Despite these limitations, adherence to guidelines recommending timely antibiotic and albumin therapy remains essential to improving patient outcomes. Following the acute management of SBP, lifelong secondary prophylaxis is recommended for all patients who survive an episode of SBP, typically with agents such as ciprofloxacin or sulfamethoxazole/trimethoprim. Ensuring appropriate use and adherence is critical to reducing SBP recurrence. While it is often assumed that PPI use may increase the risk of both primary and secondary SBP, the extent to which PPI therapy influences recurrence of SBP remains unclear. Two cohort studies with similar findings have examined the relationship between PPI use and secondary SBP risk. Kim et al ²⁹ found that PPI use was not associated with secondary or recurrent SBP; however, their study excluded high-risk patients, including those with gastrointestinal bleeding, immunosuppressive therapy, liver transplant, concurrent gastrointestinal infections, or recent SBP within 30 days. Similarly, Janka et al ²² reported that chronic PPI use increased the risk of primary SBP in patients without prior SBP but did not elevate risk in those with a history of the infection, suggesting that antibiotic prophylaxis may mitigate the risk in patients with previous SBP. The authors hypothesized that PPI-induced bacterial overgrowth and translocation contribute to SBP, while prophylactic antibiotics may counterbalance this effect. These findings underscore the importance of maintaining secondary prophylaxis in patients with prior SBP, particularly when ongoing PPI therapy is required for compelling indications.

Chronic liver disease and cirrhosis remain major contributors to mortality in the United States, ranked as the 10th leading underlying cause of death from 2018 to 2022, according to a 2024 Centers for Disease Control and Prevention report. ³² Mortality remains markedly high among patients with decompensated cirrhosis, particularly in those who develop SBP, which significantly worsens prognosis. In this cohort, in-hospital mortality was 4.8% in patients on home PPI therapy compared with 15.1% in those not receiving PPIs. However, no post-discharge follow-up data were collected, limiting the assessment to in-hospital outcomes only. Although these findings suggest a lower in-hospital mortality rate in the PPI group, the study was not adequately powered to draw definitive conclusions regarding the impact of PPI use on mortality. Prior studies have reported mixed results on the relationship between PPI use and mortality in cirrhotic patients with SBP, with some indicating increased risk ²⁷ while others show no significant association.^7,24 A meta-analysis by Yu et al ³⁰ found that PPI therapy was not associated with in-hospital or 30-day mortality in patients with SBP, suggesting that short-term use may not significantly impact early mortality outcomes. Similarly, Hung et al ⁶ reported no association between PPI use and short-term mortality in patients with cirrhosis; however, long-term use was significantly associated with increased mortality. More recently, Wong et al ²⁶ conducted a meta-analysis evaluating both short-term and long-term mortality, as well as hepatic decompensation events, in cirrhotic patients receiving PPI therapy compared with those not on PPIs. While no increase in short-term mortality was observed, PPI use was associated with increased long-term mortality and higher risks of infections, including SBP and hepatic encephalopathy. ²⁶ Collectively, these findings reinforce the need for careful consideration of the risks and benefits of PPI.

This study has several limitations. First, its retrospective design limits the ability to establish causal relationships and is subject to potential confounding and documentation bias. Second, the duration of PPI therapy prior to admission was not consistently available, making it difficult to assess the impact of chronic versus short-term PPI exposure on clinical outcomes. Third, although there was a numerical trend toward more SBP in the PPI group, the small sample size likely limited our power to detect a statistically significant difference. In addition, the initial screening of patients for inclusion was conducted by the health system’s information technology team using ICD-10 codes. While this approach facilitated case identification, it may have introduced classification errors or excluded eligible patients due to coding inconsistencies, thereby contributing to the smaller final sample size. These limitations highlight the need for larger, prospective studies with standardized data collection methods to more accurately assess the relationship between PPI use and clinical outcomes in patients with cirrhosis.

Conclusion and Relevance

In conclusion, this study found no statistically significant increase in the incidence of spontaneous bacterial peritonitis among cirrhotic patients receiving proton pump inhibitor therapy. Pharmacists are well-positioned to lead stewardship efforts by rigorously reviewing PPI indications, optimizing dosing and duration, minimizing unnecessary therapy, and collaborating with multidisciplinary teams to enhance medication safety and improve clinical outcomes in this vulnerable population.