In a retrospective study to evaluate the impact of proton pump inhibitors (PPIs) on all-cause mortality in patients with cirrhosis, researchers found reduced mortality only in those hospitalized for gastrointestinal bleeding. They reported increased liver-related mortality associated with PPIs in all other patients with cirrhosis.
Patients on PPIs had an 18% reduction in all-cause mortality versus other patients if they had gastrointestinal bleeding. But in those without bleeding, PPIs were associated with a 23% increase in liver-related mortality.
Further analysis suggested that the mortality increase could be related to a 21% increased risk for severe infection with PPI exposure in patients with cirrhosis, as well as a 64% increased risk for decompensation.
“My takeaway from this study is that there should be a nuanced understanding of PPIs and cirrhosis,” corresponding author Nadim Mahmud, MD, MS, University of Pennsylvania, Philadelphia, said in an interview, adding that, if they are to be used in this setting, there should be “a very compelling indication.”
Based on the new analysis, Dr. Mahmud explained, in a patient with cirrhosis hospitalized with a potentially ulcer-related upper gastrointestinal bleed, “we shouldn’t be afraid” to use PPIs “out of fear of potential infection or decompensation because our data demonstrate pretty strongly that that sort of patient may have a mortality benefit.”
In contrast, patients with cirrhosis and “vague abdominal discomfort” are often started on a PPI “just to see if that helps,” Dr. Mahmud said, and they may stay on the medication “in perpetuity, just because they’re so ubiquitously prescribed.”
“In that patient, we should recognize that there is a potential risk of increased infection and decompensation,” he said. There “should be an active effort to deprescribe the PPI or at the very least reduce it to the minimum dose needed for efficacy, if it’s treating a symptom.”
The research was published online in Gastroenterology.
Looking at the big picture of PPIs in people with cirrhosis
The authors noted that the half-life of PPIs is “prolonged in patients with cirrhosis” and that alterations in the gastrointestinal microbiota as a result of gastric acid suppression “may allow for bacterial overgrowth and translocation,” thus increasing the risk for infections.
However, studies of the impact of PPIs on adverse outcomes in patients with cirrhosis have often been hampered by numerous limitations, such as small sample sizes, a “limited ability to control for complex confounding,” or a “narrow focus” on hospitalized patients.
To overcome these problems, the team retrospectively examined data from the Veterans Outcomes and Costs Associated with Liver Diseases cohort, including all adults with incident cirrhosis between January 2008 and June 2021.
They excluded patients with Fibrosis-4 scores less than 1.45 at baseline, as well as those with prior liver transplantation, decompensated cirrhosis at baseline, a diagnosis of hepatocellular carcinoma within 6 months of the index date, and less than 6 months of follow-up.
In all, 76,251 patients with incident cirrhosis met the inclusion criteria, 21% of whom were on a PPI at baseline. The most commonly used PPIs were omeprazole (76.7%), followed by pantoprazole (22.2%) and lansoprazole (0.1%).
Those taking the drugs were more likely than other patients to be White, have metabolic and cardiovascular comorbidities, have a higher median body mass index, and were more likely to have cirrhosis because of alcohol-related liver disease or metabolic-associated fatty liver disease.
Over 49 months of follow-up, all-cause mortality was recorded for 37.5% of patients, of whom 59% experienced non–liver-related death and 41% liver-related mortality.
Multivariate analysis revealed that PPI exposure was not associated with all-cause mortality overall but was significantly associated with reduced all-cause mortality in patients with hospitalization for gastrointestinal bleeding, at a hazard ratio of 0.88.
However, PPI exposure in patients without gastrointestinal bleeding was associated with an increased risk for liver-related mortality (HR, 1.23), but a reduced risk for non–liver-related mortality (HR, 0.88).
Dr. Mahmud and colleagues found that PPI exposure was significantly associated with severe infection (HR, 1.21) and cirrhosis decompensation (HR, 1.64).
The authors suggested that these increased risks “may mediate the observed increased in liver-related mortality.”