1151 UNDER-RECOGNITION OF SLEEP APNEA IN PATIENTS HOSPITALIZED FOR ACUTE ISCHEMIC STROKE

Abstract Introduction: There is a bidirectional association between sleep apnea (SA) and acute ischemic stroke (AIS). Previously estimated prevalence of SA in patients with AIS ranged 60%-80%. We hypothesized that the routine clinical recognition of SA in patients with AIS is below reported range. Methods: We retrospectively analyzed all patients with AIS admitted to Mayo Clinic hospital in Jacksonville, FL between 2008 and 2014. We abstracted the demographic data, pertinent clinical variables, known or suspected diagnosis of SA and compliance with the treatment. We assessed the hospital mortality and the need for mechanical ventilation (MV). Results: There were 994 individual patients admitted with AIS within the study period. Majority were white (87%), female (52%), median BMI was 26 (23, 31) and median age was 75 years (IQR 64, 84). Median NIH Stroke Scale was 5 (IQR 2, 10), and median Glasgow Coma Scale was 15 (IQR 13, 15). Only 190 (19%) patients were considered of having SA (known diagnosis or documented clinical suspicion), of which only 42 (23%) received any treatment for SA in the hospital. Based on the limited and frequently extrapolated records of compliance, 47 patients were previously compliant with the home therapy for SA. The hospital mortality was 5%. In univariate analysis, only NIHSS, GCS, SA and congestive heart failure were significantly associated with mortality. More patients without known SA died compared to those with SA (5.6% vs. 1%, p=0.004). After adjusting in multivariate logistic regression, only NIHSS and GCS retained independent statistical significance. Compared to patients without SA, those with SA had lower NIHSS (p=0.009) and higher GCS (p=0.008). The patients compliant with home therapy had lower NIHSS than the noncompliant ones (p=0.011). More patients without SA required MV (10.9%) compared to those with SA (4.7%, p=0.009). Conclusion: The low prevalence of SA in patients with AIS is likely related to under-recognition. This is supported by the fact that the patients with recognized SA had better clinical outcomes, at least in part because of administered therapy for SA. However, only independent predictors of mortality were well established severity scales NIHSS and GCS. Support (If Any): Mayo Foundation