• Izabela Kądziołka, Rafał Świstek, Karolina Borowska, Paweł Tyszecki, and Wojciech Serednicki.
• Department of Anaesthesiology and Intensive Therapy, University Hospital in Krakow, Poland.
• Anaesthesiol Intensive Ther. 2019 Jan 1; 51 (2): 107-111.

BackgroundDisease's severity classification systems are applied to measure the risk of death and to choose the best therapy for patients admitted to intensive care unit (ICU). The aim of the study was to verify risk of death calculated with APACHE II (Acute Physiology and Chronic Health Evaluation II), SAPS II (Simplified Acute Physiology Score II), SOFA (Sequential Organ Failure Assessment) and evaluate correlation between these scores. The usefulness of SOFA score as a sole scale also was assessed.MethodsThis was a retrospective study conducted in 30-beds ICU in Kraków, Poland. Every male and female patient over 18 years old who was admitted to the ICU between 18.04.2016 and 12.08.2016 was included in the analysis. Patients who were transferred from another ICU were excluded from the research. APACHE II, SAPS II, SOFA were calculated after admission using laboratory results and clinical examination. Discrimination and calibration were used to validate these scoring system.ResultsAnalysis included 86 patients. The outcomes were compared within survivors and non-survivors groups. The prediction of death was statistically significant only for APACHE II and SAPS II. The best AUROC was for APACHE II 0.737 and SAPS II 0.737; discrimination for SOFA was not statistically significant. There was high correlation only between SAPS II and APACHE II results (r ≥ 0.7, P < 0.01). The calibration was excellent for SAPS II, P = 0.991, and slightly worse for APACHE II, P = 0.685, and SOFA, P = 0.540. Patients who survived spent more days on ICU (P < 0.01), mean Length of Stay (LOS) in this group was 38.25 ± 16.80 days.ConclusionsAPACHE II and SAPS II scales have better discrimination, calibration and power to predict deaths on ICU than SOFA. Among these scales SOFA did not achieve expected results.

25 keywords...

hide…