Background: Post-stroke aspiration pneumonia is associated with higher mortality and disability. Patients are often kept Nil By Mouth (NBM) until after dysphagia screen and declared safe to resume oral intake. Stroke is associated with immunosuppression. An attractive hypothesis is that pneumonia post stroke is related to immunosuppression and stroke severity. In this study the aim is to assess the proportional contribution of the variables and pneumonia. 
Method: This is a retrospective study of admissions to the stroke unit at Monash Health in 2015. We collected data on demographic variables, admission diagnoses, time to triage, imaging and stroke severity. Variables found to be associated with pneumonia were entered into hierarchical partition analysis to determine the proportion contribution of each variable. This method permits assessment of proportional contribution of each variable to the model.
Results: Within 2015, there were 627 patients with ischemic and hemorrhagic stroke (age 71.2±15.4, National of Health Institute Stroke Score/NIHSS 7.6±7.8, male=55.4%). The frequency of pneumonia was 5.4%, week day admission was 43.7% and NBM status was 72.8%. The regression analyses were restricted to 452 patients with reliable time of dysphagia screen. Univariable analyses show stroke severity, time to dysphagia screen, Charlson comorbidity score and age to be significantly associated with pneumonia. Week day versus weekend admission was not associated with pneumonia. The hierarchical partition analysis showed that stroke severity accounted for (76.2%), Charlson (11.6%), age (7.9%) and time to dysphagia screen (4.2%) 
Conclusion: Stroke severity status play key role in determining risk for post stroke pneumonia in this retrospective study.

Stroke severity versus time to dysphagia screen as risk factor for post-stroke pneumonia