442-P: Excess Burden of Adverse In-Hospital Outcomes in Patients with Diabetes Hospitalized for Stroke

Abstract

Cerebrovascular accidents (CVA) represent a major complication in diabetes (DM). Real-life evidence as to whether modern management of CVA and DM have softened this relationship is currently limited. We assessed the prevalence of DM, previously known or newly-diagnosed, in all patients (N=542, males 51.5%) admitted for CVA at the Stroke Unit (SU) of Verona University Hospital from 1/1/2015 to 12/31/2016. DM was ascertained by previous diagnosis, glucose-lowering therapy at admission/discharge or admittance plasma glucose (PG) ≥11.1 mmol/L. Prevalence of DM was 21.03% (78.1% known-DM; 21.9% new-DM). In-hospital death rate was 10.5%. Compared to non-DM, patients with DM showed an increased risk of death (15.8 vs. 9.1%; OR 1.87 95% CI, 1.03-3.41) and complications (48.2 vs. 33.3%; 1.87, 1.22-2.86) with similar duration of hospitalization (mean±SEM, 9.14±1.06 vs. 8.80±0.39 days, p=0.72). All deaths occurred in the SU. The “survivors” with DM transferred to non-intensive ward had an out-of-SU hospital stay twice as longer than non-DM (2.32±1.02 vs. 0.98±0.19 days, p=0.038). After multivariable adjustment, DM remained predictor of complications (adj-OR 1.87, 1.22-2.86) but not mortality (1.79, 0.92-3.46). Patients in the highest PG quartile (≥6.55 mmol/L) displayed the highest risk of death (7.89, 2.73-22.79; PG<4.77 mmol/L, ref.) and complications (2.30, 1.24-4.23), independent of diabetes status, glucose-lowering medications, CVA treatment and established CVA risk factors. Individuals in the older age tertile (≥81 years) were at increased risk of death (3.25, 1.40-7.54; age <72 years, ref.), while risk of infectious/cardiorespiratory complications was higher at younger age (3.13, 1.42-6.9). No difference was observed in new-DM vs. known-DM. In conclusion, these data highlight that DM frequently occurs in patients admitted for CVA and adds an excess burden of adverse clinical outcomes that urgently calls for strategies to anticipate DM diagnosis in high-risk individuals. Disclosure M. Dauriz: None. E. Tregnaghi: None. L. Santi: None. T. Lucianer: None. A. Altomari: None. E. Rinaldi: None. S. Tardivo: None. C. Bovo: None. E. Bonora: None.