Dissertation is devoted to the management optimization of active native valve infective endocarditis (ANVIE). The study is based on management of 136 patients with definite ANVIE as defined by the Duke criteria and 3 patients in whom ANVIE diagnosed after death. Using the backward stepwise binary logistic regression analysis the following factors were identified as independent predictors of unfavorable course of ANVIE: blood urea level, mmol/L (regression coefficient B=0,222; odds ratio (OR): 1,25; 95% confidence interval (CI): 1,09-1,43; p=0,001), large (>=10mm) vegetations (B=1,749; OR: 5,75; CI: 1,79-18,4; p=0,003) and congestive heart failure (B=1,608; OR: 4,99; CI: 1,73-14,4; p=0,003); the constant in the model = - 4,499). Overall 85% cases could be predicted correctly by this statistical model. Surgery in active phase (B= -1,559; OR=0,210; p=0,004), blood urea level, mmol/L (B=0,212; OR=1,24; p<0,001) and NYHA heart failure class (for NYHA class III: B=2,054; OR=7,8; p=0,013; for NYHA class IV: B=2,186; OR=8,9; p=0,019; reference category NYHA class II) were identified as independent predictors of 6-month mortality in ANVIE. Use early surgery instead of surgery in this analysis revealed early surgery as an independent predictor of 6-month mortality with B= -1,752 and OR=0,173 (p=0,023). The key point of the optimization, with the aim of reducing the overall mortality, was to compare the risks of the operation with the risk of continuing the isolated medical treatment in patients without indications for emergent or urgent surgery. Results of the application of the new approach are estimated by comparison of mortality and survival in the main group of patients with ANVIE (68 patients treated in 2007-2013 years) and a comparison group (42 patients treated in 2004-2006 years), formed in chronological order. The new approach has led to reduction in in-hospital (from 26% to 15%), 30-day (from 29% to 15%) and 6-month (from 31% to 21%) mortality and better survival in the main group of patients with ANVIE compared to the comparison group with the log-rank test (khi square =5,098 і p=0,024).
