Background and Purpose: We hypothesized that acute ischemic stroke (AIS) patients with hospital acquired infection (HAI) would have worse outcomes than patients with an infection present on admission (IPOA) or no infection.
Methods: We used the 2007-2009 California State Inpatient Database from the Health Care Utilization Project (HCUP). Patients who were over the age of 18 and hospitalized with AIS, defined by previously validated ICD-9 codes, were included. Validated ICD-9 codes for sepsis, pneumonia and urinary tract infections were used to categorize patients as having IPOA (infection diagnosed within 24 hours of admission), HAI (infection during admission), or no infection using federal guidelines. The outcomes of interest were length of hospital stay (LOS), and poor discharge outcome (discharge someplace other than home).
Results: A total of 133,678 patients with diagnosis of AIS were identified; 2.6% (N=3436) had HAI, 15.1% (N=20126) had IPOA, and 82.3% (N=110026) had no infection. Mean LOS was longer for patients with HAI compared to IPOA and no infection (18.5 vs. 8.5 vs. 5.1 days: p<0.0001). In adjusted models (Figure), LOS has longer for those with HAI compared to those with IPOA or no infection (8.2 vs. 6.4 vs. 5.5 days; p<0.0001). Adjusting for the same covariates plus LOS, HAI and IPOA remained associated with death (Figure). In adjusted models, patients with HAI or IPOA had higher odds of poor discharge outcome (HAI OR 1.82, 95% CI 1.51-2.21 and POA OR 1.49, 95% CI 1.38-1.61). In an adjusted model, patients with HAI are at increased odds for a poor discharge outcome when compared to those with an IPOA (OR 1.28, 95%CI 1.04-1.58). Using a mediation analysis, LOS explains 46% of the relationship between HAI and poor discharge outcomes, and 87% of the relationship between IPOA and poor discharge outcome.
Conclusion: IPOA and HAI were each associated with increased LOS. Even after accounting for LOS, patients with HAI had poor discharge outcomes and a higher risk of death.