Unadjusted and adjusted analyses of association between persistent use of proton pump inhibitors and hospital readmission*
| Outcome | Crude SHR (95% CI) | Minimally adjusted SHR (95% CI) | Fully adjusted SHR (95% CI) |
|---|---|---|---|
| All-cause readmission (n = 767) | 1.44 (1.25–1.67) | 1.40 (1.21–1.63) | 1.31 (1.12–1.53) |
| Potentially PPI-related readmission (n = 62) | 1.39 (0.83–2.32) | 1.33 (0.80–2.22) | 1.24 (0.74–2.08) |
| Pneumonia-related readmission (n = 34) | 1.70 (0.83–3.49) | 1.56 (0.76–3.20) | 1.40 (0.68–2.87) |
| Fracture-related readmission (n = 25) | 1.03 (0.47–2.28) | 1.04 (0.48–2.30) | 0.96 (0.42–2.23) |
Note: CI = confidence interval, PPI = proton pump inhibitor, SHR = subdistribution hazard ratio.
↵* Competing-risk regression based on the method by Fine and Gray, (17) with death as competing event. Minimal adjustment model adjusted for age, sex and Charlson Comorbidity Index. Full adjustment model adjusted for age, sex, Charlson Comorbidity Index, medication count, anticoagulant use, study site, admission ward, number of previous hospital admissions, intervention arm and discharge destination. Potentially PPI-related readmissions included readmissions with pneumonia, fracture, nephritis or bacterial intestinal infection.