ABSTRACT NUMBER: ESOC2026A1465 PREVENTIVE STRATEGIES FOR STROKE-ASSOCIATED PNEUMONIA: A SYSTEMATIC REVIEW WITH NETWORK AND COMPONENT NETWORK META-ANALYSES

Abstract Background and aims Pneumonia is a common and serious complication in acute stroke patients, and various preventive interventions have been studied. This review aims to estimate the relative efficacy of individual and combined preventive strategies on pneumonia incidence within one month after stroke. Methods MEDLINE, Embase, WoS, Scopus, CINAHL, CENTRAL, and ClinicalTrials.gov were searched up to October 2025. Network and component network meta-analyses of randomized controlled trials (RCTs) were performed in two populations: acute stroke patients and patients with post-stroke dysphagia (PSD). Results For acute stroke patients, 31 studies were included in the meta-analyses. Dysphagia screening/management [odds ratio (OR) 0.48, 95% confidence interval (CI) 0.32–0.72] and respiratory muscle training [OR 0.35, 95% CI 0.23–0.56] reduced pneumonia incidence compared with usual care. Combining these interventions with oral hygiene care (OHC) and early mobilization improved efficacy (Figure 1). For PSD patients, 39 studies were analyzed. Instrumental swallowing rehabilitation (ISR) [OR 0.31, 95% CI 0.12–0.78], OHC [OR 0.29, 95% CI 0.12–0.73], and pharmacotherapy [OR 0.27, 95% CI 0.14–0.53] were effective in preventing pneumonia. Conventional swallowing rehabilitation (CSR) [OR 0.50, 95% CI 0.22–1.17] did not significantly reduce pneumonia risk. However, CSR demonstrated significant preventive effects when combined with one or more additional interventions. These combinations involved ISR, OHC, pharmacotherapy, acupuncture, oral sensory stimulation, breathing therapy, or intermittent tube feeding (Figure 2). Conclusions Effective pneumonia prevention in stroke patients should prioritize the management of dysphagia and respiratory function. In PSD patients, CSR may be insufficient and should be supplemented with additional preventive interventions. Conflict of interest Name of author: nothing to disclose Figure 1 - belongs to Results Figure 2 - belongs to Results