RESUMO
BACKGROUND: Frailty, demographic and clinical variables linked to incident diseases (e.g., dehydration, inflammation) contribute to poor outcomes in older patients acutely hospitalized. Their predictivity on short-, intermediate- and long-term mortality in a comprehensive model has been scarcely investigated. AIMS: To test the performance of a predictive tool considering frailty and inflammation as well as age, sex and impaired hydration status on 1-year mortality in acutely admitted older patients. METHODS: Retrospective observational study including 529 medical patients (age 84.6 ± 7.3 years). At hospital admission, frailty was assessed by the Multidimensional Prognostic Index (MPI). The Glasgow Prognostic Score (GPS) was used to grade systemic inflammation. Serum osmolarity was calculated to assess hydration. RESULTS: After adjusting for age, sex, GPS and osmolarity, the severe-risk MPI was a strong predictor for 1-year mortality (OR 4.133; 95% CI 2.273-7.516; p < 0.001). Age > 85 years, male sex, GPS-2 and serum osmolarity > 300 mOsm/L were independent predictors of mortality in the same multivariable model. The MPI alone showed a moderate discrimination power (AUC 0.678; 95% CI 0.628-0.729; p < 0.001) on 1-year mortality, which increased by 12.5% after the addition of the above predictors in the fully adjusted regression model (AUC 0.763; 95% CI 0.719-0.807; p < 0.001). The severe-risk MPI adjusted for the same factors was also an independent predictor of mortality after 60 and 180 days since hospital admission. DISCUSSION: Inflammation and impaired hydration are potentially modifiable risk factors for severe outcomes in older acutely hospitalized patients. A model combining GPS, age, gender, and plasma osmolarity improved the accuracy of MPI at admission in predicting long-term mortality.