Research Article: Effect of proton pump inhibitors on mortality of cirrhotic patients with pneumonia

Date Published: April 25, 2019

Publisher: Public Library of Science

Author(s): Tsung-Hsing Hung, Chih -Wei Tseng, Chih-Chun Tsai, Hsing-Feng Lee, Wisit Cheungpasitporn.

http://doi.org/10.1371/journal.pone.0216041

Abstract

Pneumonia is life-threatening in patients with liver cirrhosis. Proton pump inhibitors (PPIs) may increase the risk of these patients developing pneumonia. However, whether PPIs increase mortality in patients with cirrhosis and pneumonia remain unknown.

We used the Taiwan National Health Insurance Database to enroll 1,201 cirrhotic patients with pneumonia without active gastrointestinal bleeding who were receiving PPIs and were hospitalized between January 1, 2010 and December 31, 2013. A one-to-three propensity score match was performed to select a comparison group based on age, gender, and comorbid disorders.

The overall 30-day and 90-day all-cause mortality rates were 13.7% and 26.9% in the PPI group, and 14.3% and 25.1% in the non-PPI group, respectively. After Cox regression model adjusting for age, gender, and comorbid disorders, the hazard ratios of the effect of PPIs on 30-day and 30 to 90-day mortality were 0.94 (95% Confidence Interval [CI], 0.79–1.12, P = 0.468) and 1.26 (95% CI, 1.05–1.52; P = 0.013), respectively.

PPIs were not associated with 30-day mortality among cirrhotic patients with pneumonia but not active gastrointestinal bleeding. However, prolonged PPI therapy may be associated with higher mortality.

Partial Text

Patients with liver cirrhosis are prone to bacterial infections due to their impaired immune status, the increased use of invasive procedures, and alterations in the enteric flora [1, 2]. Bacterial infection is the main cause for hospitalization among patients with liver cirrhosis [3] and contributes to their fourfold greater likelihood of death over patients without cirrhosis [4]. In addition, bacterial infections can trigger and aggravate cirrhosis-related complications, such as hepatic encephalopathy, ascites, variceal bleeding, or hepatic renal syndrome [1–4].

The dataset contained 32,898 cases of cirrhotic patients with pneumonia. After excluding patients with active peptic ulcer bleeding or esophageal variceal bleeding; and those receiving panendoscopy, intravenous PPIs, or oral high-dose PPIs, we enrolled a total of 6,432 cirrhotic patients with pneumonia and without active gastrointestinal bleeding in this study. Of these patients, 1,201 cirrhotic patients with pneumonia were also receiving oral PPIs (PPI group). After 1:3 propensity score matching, we enrolled 3,603 cirrhotic patients with pneumonia but not receiving PPIs as the non-PPI group. Table 1 shows the demographic characteristics of the PPI and non-PPI groups. The overall 30-day and 90-day all-cause mortality rates were 13.7% and 26.9% in the PPI group, and 14.3% and 25.1% in the non-PPI group, respectively. After the Cox regression model was adjusted for age, gender, and other comorbid disorders, the HR of oral PPIs for 30-day mortality of cirrhotic patients with pneumonia and without active gastrointestinal bleeding was 0.94 (95% CI, 0.79–1.12, P = 0.468) compared to the non-PPI group. The adjusted HR of oral PPIs for 90-day mortality was 1.07 (95% CI, 0.95–1.22, P = 0.274) compared to the non-PPI group. The cumulative survival plot is shown in Fig 1. Other statistically significant prognostic factors are shown in Table 2.

This nationwide population-based study was designed to identify the association of PPIs with mortality among patients with cirrhosis and pneumonia, but no active gastrointestinal bleeding. Using the NHIRD, we enrolled a large population to provide reliable information that could be applied to clinical practice. Our study showed that PPIs were not associated with an increased risk of 30-day mortality among cirrhotic patients with pneumonia who did not have active gastrointestinal bleeding. However, the prolonged use of PPIs may correlate with a greater risk of mortality than that of non-PPI users.

 

Source:

http://doi.org/10.1371/journal.pone.0216041

 

Leave a Reply

Your email address will not be published.