Assessment of the prognosis of the outcome of severe community-acquired pneumonia caused by Klebsiella pneumoniae

Abstract

Introduction. The problem of an adequate assessment of the prognosis of the outcome of severe community-acquired pneumonia (CAP) is particularly difficult if it is caused by an unusual pathogen for it, in particular Klebsiella pneumoniae.The objective was to develop the approach for predicting the survival of a heterogeneous population of patients with CAP caused by Klebsiella pneumoniae using statistical approaches based on artificial neural networks.Materials and methods. The design is a retrospective, multicenter, controlled, non-randomized study. Inclusion criteria: clinical, laboratory and radiological diagnosis of CAP associated with Klebsiella pneumoniae with a SOFA score of 2 or more points. The development included 100 patients. 50 died. The prognostic significance of the SOFA, APACHE II, PSI/PORT, Glasgow and Charlson comorbidity index, procalciotonin, C-reactive protein scales was evaluated. The data obtained were evaluated in the StatPlus 7 and Pycharm GPT programs.Results. None of the stated scales has shown its significance. There were no statistically significant differences between the surviving and deceased patients in terms of the level of biomarkers studied. In this regard, we have compiled a logistic regression equation for assessing the prognosis based on a combination of the SOFA score, the Charlson index and the procalcitonin level.Conclusion. In assessing the prognosis of outcome in patients with CAP caused by Klebsiella pneumoniae, it is advisable to use a combination of data from the SOFA score, Charlson comorbidity index and procalciotonin levels. Threshold critical values are SOFA score of more than 4 points, Charlson comorbidity index of more than 7 points, procalciotonin level of more than 2 ng/ml.