1997 Volume 34 Issue 6 Pages 492-498
Clinical assessment and laboratory testing are important in predicting longevity and the outcomes of chronic diseases in the elderly. We therefore studied the prognostic utility of clinical and laboratory findings. A group of 168 elderly patients (70-97 years of age) with chronic diseases who were admitted to Kashiwazaki Kosei Hospital was studied. Data on 13 potentially prognostic factors were analyzed: dementia, being bedridden, anorexia, edema of the lower limbs, urinary incontinence, dyspnea, fever, hepatic dysfunction, renal dysfunction, anemia, hypoalbuminenmia, inflammation, and electrocardiographic abnormality. Kaplan-Meier survival curves were constructed, and either the log-rank or Wilcoxon method was used to look for significant differences in survival between patients with and without the factors listed above. Regression analysis was then done with the Coxproportional-hazardsmodel to study the factors that contributed to the shortest survival. Patients lived longer if they were not bedridden, anorexic, incontinent of urine, hypoalbuminemic, or if they had no inflammation (p<0.005). Analysis with the proportional-hazardsmodel revealed significant contributions from seven factors (p<0.05): being bedridden, anorexia, urinary incontinence, hepatic dysfunction, hypoalbuminemia, inflammation, and electrocardiographic abnormalities. The survival curve obtained by using these factors as independent variables in the proportional-hazardsmodel was similar to the Kaplan-Meier survival curve.
