Maajid Mohi Ud Din Malik
The bicentric retrospective analysis by Kitchen and colleagues [1] reports that proton pump inhibitor (PPI) treatment is independently associated with gram-positive spontaneous bacterial peritonitis (SBP) in cirrhotic patients (adjusted OR 2.72, 95% CI 1.34–5.51). Their clinical message—that empiric antibiotic choice should account for an enriched gram-positive spectrum in PPI users—is plausible and pragmatically useful. How strongly that inference holds, however, depends on residual confounding the present design does not fully address. First, prior hospitalisation within 90 days was the most strongly imbalanced baseline variable (82.5% PPI vs. 62.0% non-PPI, p = 0.002, table 1). This covariate was not retained in the multivariable model; only the binary nosocomial-SBP indicator was. Recent healthcare exposure is among the strongest documented drivers of gram-positive—particularly Enterococcus and Staphylococcus—colonisation in cirrhotic patients [2]. The nosocomial-SBP variable captures only the index admission, not cumulative 90-day exposure to inpatient antibiotics, indwelling devices or contact-precaution organisms. A model including both nosocomial status and recent hospitalisation would more honestly partition the variance currently attributed to PPI. Second, the 18-year inclusion window (January 2007–July 2025) spans an era of substantial change in both PPI prescribing intensity and SBP microbial epidemiology. Independent of PPI use, large series have documented a secular drift from gram-negative toward gram-positive SBP [3, 4]. If PPI prescription increased over the study period while gram-positive SBP increased for unrelated reasons, the cross-sectional association is partly an era effect. The supplementary stratified analysis (figure S1) acknowledges this but does not enter calendar year or, more appropriately, a pre/post midpoint indicator as a covariate in the principal regression. Third, the non-PPI group is unusual. Only 50 of 227 patients (22%) a proportion well below the published prevalence of PPI use among hospitalised cirrhotics [5] comprise it, and they had numerically worse liver function (Child-Pugh 11 (11–12) vs. 11 (10–11), bilirubin 5.3 vs. 3.0 mg/dL) yet less prior healthcare contact. This pattern is compatible with channelling bias: PPI discontinuation in end-stage disease (palliation, polypharmacy review) may select patients who differ from PPI users in unobserved ways relevant to the gut microbiome. Fourth, the external validation cohort (n = 24 PPI, n = 6 non-PPI) carries insufficient information to validate the primary finding. The single non-PPI patient with gram-positive SBP (1/6 = 16.7%) sits within an exact binomial 95% confidence interval of approximately 0.4% to 64.1%. Presenting this as supporting evidence inflates the apparent reproducibility of the result. The clinical recommendation heightened gram-positive coverage when starting empiric therapy in PPI-treated cirrhotic patients has plausible microbiological grounds and may prove correct. Confirming it as causal, however, will require prospective registry data with cumulative healthcare-exposure metrics, era stratification and a non-PPI denominator larger than 50. Yours sincerely, Maajid Mohi Ud Din Malik Maajid Mohi Ud Din Malik: supervision, formal analysis, software, writing – review and editing, writing – original draft, investigation, conceptualization, methodology, validation, visualization. The author has nothing to report. The author declares no conflicts of interest. This article is linked to Kitchen et al. papers. To view this article, visit https://doi.org/10.1111/apt.70593. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.