P3514Gender differences in prognosis for total mortality among patients with ADHF: results of four-year follow-up

Abstract

Abstract Aim To obtain information on the effect of various predictors on the prognosis of a patient with acute decompensated heart failure (ADHF). Methods We investigated a sample of patients (n=719) hospitalized due to ADHF for 4 years (average observation duration was 2.3 years). 27 predictors were included in the model (clinical conditions and clinical signs of ADHF). We used regression tree method with CHAID algorithm with 10-fold cross-validation to determine the most important predictors of the prognosis. Results The most significant predictor for total mortality at the first branching was the gender of the patient. Mortality rate was 65.0% among males vs 47.9% among females (p<0.001). In the subgroup of males, the next most significant predictor for the prognosis was the level of systolic blood pressure (SBP): mortality rates was 79.6% with SBP ≤125 mm Hg vs. 57.2% with SBP >125 mm Hg (p=0.001). In the subgroup of males with SBP >125 mm Hg mortality rates differed depending on the presence of X-ray signs of congestion; when present, the total mortality was 65.7% vs. 47.8 in the absence (p=0.01). In the subgroup of females, the first most significant predictor of prognosis was diabetes mellitus (DM): mortality rates was 60.1% with DM vs. 41.2% without DM (p<0.001). The worst prognosis was found among those who are over 78 years old (55.8%) compared with younger (32.4%, p=0.001) for females with DM. The model has high sensitivity (74.9%) but low specificity (50.3%). In the ROC analysis, AUC was 0.673 (95% CI: 0.634–0.712). Regression tree for total mortality rate Conclusion The 4-year survival rate for patients with ADHF varies by gender. Among men, the most important predictors of prognosis were the level of SBP and the presence of congestion in the lungs at the hospital admission. Among women, DM and age had the greatest influence on the prognosis.