Comparative clinical outcomes of ampicillin/sulbactam versus antipseudomonal broad-spectrum antibiotics for community-onset aspiration pneumonia: a propensity score-matched retrospective cohort study based on guideline-recommended strategies in a secondary care hospital in Japan.

Objective

Aspiration pneumonia is highly prevalent among older adults and poses a substantial clinical burden, particularly in aging and rural societies such as Japan. Although the 2024 Japanese Respiratory Society guidelines recommend ampicillin/sulbactam (ABPC/SBT) as first-line therapy, empirical use of antipseudomonal broad-spectrum antibiotics remains common in routine clinical practice. However, evidence supporting the guideline-recommended escalation and de-escalation strategies is limited. Patients and

Methods

We conducted a single-center retrospective cohort study of adult patients hospitalized with community-onset aspiration pneumonia between January 2020 and December 2023. Patients received initial empirical therapy with either ABPC/SBT or antipseudomonal broad-spectrum antibiotics (piperacillin/tazobactam or meropenem). The primary outcome was 30-day all-cause mortality. Secondary outcomes included Clostridioides difficile infection (CDI), pneumonia recurrence, length of hospital stay, and duration of antibiotic therapy. Propensity score matching (PSM; up to 1:2) was performed using guideline-based variables, including quick Sequential Organ Failure Assessment score, A-DROP score, and resistant-pathogen risk factors. Subgroup analyses were conducted according to guideline-defined escalation and de-escalation recommendations.

Results

After matching, 97 patients were included in the analysis (ABPC/SBT, n=63; broad-spectrum, n=34). In the matched cohort, the broad-spectrum group had significantly higher 30-day mortality (23.5% vs. 6.3%, P=0.022) and CDI incidence (11.8% vs. 1.6%, P=0.049). In subgroup analyses, the most pronounced between-group difference was observed for 30-day mortality in the post-PSM escalation-recommended subgroup (broad 25.0% vs. narrow 4.0%; exact odds ratio 7.77, 95% confidence interval 1.33-82.77; P=0.009), whereas several other subgroup estimates were imprecise because of the small number of events.

Conclusion

Empirical treatment with antipseudomonal broad-spectrum antibiotics did not improve clinical outcomes and was associated with increased mortality and CDI. These findings support guideline-recommended narrow-spectrum-first strategies and highlight the importance of systematic assessment of disease severity and risk factors for resistant pathogens to guide empirical antibiotic selection and strengthen antimicrobial stewardship, particularly in rural and resource-limited healthcare settings.