By Richard R. Watkins, MD, MS, FACP, FIDSA, FISAC

Professor of Internal Medicine, Northeast Ohio Medical University, Rootstown, OH

SYNOPSIS: A nationwide cohort study of adults in Denmark found that proton-pump inhibitor (PPI) use was associated with a moderately increased risk of community-associated Clostridioides difficile infection, and the risk remained elevated up to one year after PPI treatment had ended.

SOURCE: Inghammar M, Svanström H, Voldstedlund M, et al. Proton-pump inhibitor use and the risk of community-associated Clostridium difficile infection. Clin Infect Dis 2021;72:e1084-e1089.

The authors of several observational studies over the last 15 years have reported an association between Clostridioides difficile infection (CDI) and proton-pump inhibitor (PPI) use. However, many studies have been criticized for their observational design and lack of sufficient adjustment for confounding variables. One way to adjust for confounders is a relatively new type of study design called the self-controlled case-series (SCCS), which compares periods of exposure and non-exposure within individuals, thereby controlling for all confounders that remain constant over the observation window. Inghammar et al conducted a large nationwide study to determine the risk of community-associated CDI (CA-CDI) in adults prescribed PPIs using an SCCS design.

The study included Danish adults age 20 years and older in a nationwide database diagnosed with CA-CDI between February 2010 and December 2013. A case of CA-CDI was defined based on guidelines from the Infectious Diseases Society of America as a first positive test for C. difficile based on culture, molecular assay, or toxin test among individuals who reported symptoms in an outpatient setting or two or fewer days after hospitalization, who produced no other positive CDI tests within the preceding eight weeks, and had not been hospitalized in the preceding 12 weeks. PPI exposure was defined based on four periods: new use, defined as a PPI prescription among individuals without PPI use in the prior 365 days; current use, defined as ongoing treatment with one tablet per day of a PPI from the first day of treatment until treatment cessation; intermediate use, defined as the period zero to six months (0-179 days) after treatment cessation; and past use, defined as six to 12 months (180-364 days) after treatment cessation.

There were 3,583 cases of CA-CDI in 3,338 individuals during the study period. The median age was 65 years (interquartile range, 44 to 80 years), and 38% were male. CA-CDI occurred in 964 individuals who currently were using PPIs, 324 cases occurred after intermediate use, 123 occurred after past use, and 2,172 occurred during periods without PPI use. The adjusted incident rate ratio (IRR) was 2.03 (95% CI, 1.74-2.36), comparing PPI use with nonuse. The increased risk continued to be elevated in later periods: 1.54 (95% CI, 1.31-1.80) for zero to six months and 1.24 (95% CI, 1.00-1.53) for six to 12 months after current use.

After comparing the incidence of CA-CDI during current use of PPIs with periods of nonuse, the unadjusted IRR was found to be 2.78 (95% CI, 2.40-3.22). Adjusting for hospitalization, antibiotic use, and corticosteroid use resulted in an adjusted IRR of 2.03 (95% CI, 1.74-2.36). The increased risk was lower but still elevated in later periods (adjusted IRR, 1.54 [95% CI, 1.31-1.80] for zero to six months and adjusted IRR, 1.24 [1.00-1.53] for six to 12 months after current use of PPIs). Estimates for the association between current use of PPIs and CA-CDI were similar regarding sex and age.

COMMENTARY

Most previous studies of the association between PPIs and CDI were conducted with hospitalized patients. These individuals tend to be older, sicker, and have more antibiotic exposures than patients who develop CDI in community settings. Therefore, the analysis by Inghammar et al is interesting because it focused on this latter, less studied group. The key finding was the use of PPIs was associated with approximately two times the risk for acquiring CA-CDI. The risk was lower after PPI treatment stopped, but it remained significantly higher up to 12 months afterward.

Clinicians should realize there is mounting evidence suggesting an association between PPI use and CDI. They should carefully assess the potential benefits of PPI use compared to the risk of CA-CDI, especially in patients with a history of CDI. In such cases, perhaps an alternative agent would be a better choice, or the PPI could be used but for a limited time. The mechanism by which PPI use increases the risk for CDI has not yet been fully elucidated. It is believed PPIs suppress the capacity of the normal microbiome in limiting the proliferation of C. difficile. Further experimental evidence is needed to test this hypothesis.

Despite the robust study design, a few limitations are worth mentioning. First, some PPIs can be purchased over the counter in Denmark, which could have led to misclassification of drug exposure. Second, although the SCCS design enables patients to be used as their own controls, which minimizes the effect of time-fixed confounders, residual confounding remains theoretically possible since there was no randomization here. Finally, the results of the study might not be generalizable to other populations in different geographic areas.

Definitive proof that PPIs cause CDI likely will require a randomized clinical trial. Considering the relatively low incidence of CDI in the community, this would require many individuals to be enrolled. Until then, clinicians must be aware of the risks associated with PPIs and prescribe them with the best available evidence in mind.