MedSafer preprint reports PPI deprescription shows no 30-day CDI benefit

Conventional wisdom dictates that removing proton pump inhibitors (PPIs) from a patient’s regimen directly lowers their short-term risk of C. difficile infection (CDI). Yet, new analysis from the MedSafer cluster randomized trial suggests this belief is not fully supported by the data. Clinical risk models, it seems, misattribute an outsized protective benefit to such deprescription efforts.

What the paper actually did and what it reports The preprint analyzes deprescription events within the MedSafer cluster randomized trial population and measures CDI incidence in the 30-day window after PPI deprescription. The authors report that deprescribing PPIs did not produce a statistically significant decline in 30-day CDI risk in this cohort of 4,990 hospitalized older adults, directly challenging the assumption that removing PPIs meaningfully lowers short-term CDI incidence.

The result is presented as a secondary analysis of an existing randomized trial, not as a primary endpoint trial designed specifically to test CDI outcomes.

How the finding maps to the evidence layer and its limits Because this is a preprint and a secondary analysis, the claim should be seen as provisional. The MedSafer dataset offers randomized cluster assignment for deprescribing prompts, but the CDI outcome here was not the trial's primary endpoint; the paper's own framing therefore leaves open questions about statistical power to detect relatively rare infectious events within a 30-day period after medication change.

The preprint does not alter the higher-level observational literature on PPI-associated CDI risk, but it does show that, at least in this constrained hospitalized older-adult population and timeframe, removing PPIs did not translate into an immediate measurable protective effect.

Why the consensus read — that deprescribing reduces CDI — is incomplete The prevailing clinical consensus has leaned on observational associations between PPI exposure and CDI incidence. But observational associations can reflect confounding by indication (sicker patients both receive PPIs and have higher CDI risk), exposure misclassification, or dose-duration effects not captured by short-term deprescribing efforts.

The MedSafer secondary analysis suggests that when you intervene to deprescribe at hospital discharge or during admission, the anticipated downstream reduction in CDI may be smaller or slower than models assume. If CDI risk models priced PPI deprescription as a near-term lever to cut hospital CDI rates, those models are likely overstating the immediate benefit.

The skeptical counter-read the preprint doesn't resolve

A reasonable counter is that the null result reflects insufficient event counts, an excessively short 30-day follow-up, incomplete adherence to deprescribing recommendations, or residual confounding despite randomization at the cluster level; any of these would mask a true effect. Because the analysis is secondary, skeptics can argue it was never powered to settle this question definitively.

The preprint itself does not put the issue to rest for programs that target CDI reduction as a primary objective of deprescribing.

What this changes for hospitals and guideline-makers in the next 12–18 months If other datasets replicate this null short-term effect, infection-control teams and formulary committees should stop treating PPI deprescribing as a reliable, immediate CDI-reduction intervention when allocating limited intervention bandwidth and infection-prevention dollars. Instead, deprescribing may still be justified for other harms (overuse, fracture risk, medication burden), but its CDI-prevention benefit should be repriced toward uncertainty in short-term models.

This reframing shifts the decision metric from "deprescribe to reduce CDI now" to "deprescribe for broader medication-safety reasons, with CDI benefits uncertain and possibly delayed."

Who benefits, who is exposed, and the overlooked middle Pharmacists and clinicians focused on medication safety keep the upside of deprescribing (reduced polypharmacy harms), while hospital epidemiologists relying on modelled CDI savings are exposed to mispriced expectations. The under-noticed middle is infection-control procurement and quality teams that allocate nurse time and programmatic funding based on projected CDI reductions from deprescribing pathways — those budgets and staffing choices may be misallocated if the CDI dividend is smaller than assumed.

Signals that would falsify this reinterpretation

This thesis would be disproved if a meta-analysis of randomized trials by Q3 2025 found a statistically significant CDI reduction after deprescription, if a major infectious disease society updated guidelines by Q4 2024 to recommend PPI deprescription primarily for CDI prevention, or if a large prospective study by Q2 2025 demonstrated a clear dose-response between PPI exposure and CDI after rigorous confounder control; absent one of these signals, the MedSafer secondary analysis stands as an important corrective to overconfident risk models.

No one in the reported packet is on the record about these findings; until independent confirmatory analyses arrive, hospitals should treat the medRxiv result as preliminary but plausible evidence that short-term CDI risk models overvalue the protective effect of deprescribing PPIs.