Objective. An accurate identification of patients at the need for prioritized diagnostics and care are crucial in the emergency department (ED). Blood gas (BG) analysis is a widely available laboratory test, which allows to measure vital parameters, including markers of ventilation and perfusion. The aim of our analysis was to assess whether blood gas parameters in patients with dyspnea at an increased risk of respiratory failure admitted to the ED can predict short-term outcomes. Methods. The study group eventually consisted of 108 patients, with available BG analysis. The clinical and laboratory parameters were retrospectively evaluated, and three groups were distinguishedarterial blood gas (ABG), venous blood gas (VBG), and mixed blood gas. The primary endpoint was short-term, all-cause mortality during the follow-up of median (quartile 1quartile 3) 2 (14) months. The independent risk factors for mortality that could be obtained from blood gas sampling were evaluated. Results. The short-term mortality was 35.2 (38/108). Patients who died were more frequently initially assigned to the red triage risk group, more burdened with comorbidities, and the median SpO₂ on admission was significantly lower than in patients who survived the follow-up period. In the multivariable analysis, lactate was the strongest independent predictor of death, with 1mmol/L increasing all-cause mortality by 58 in ABG (95 CI: 1.012.47), by 80 in VBG (95 CI: 1.132.88), and by 68 in the mixed blood gas analysis (95 CI: 1.222.31), what remained significant in VBG and mixed group after correction for base excess. In each group, pH, pO₂, and pCO₂ did not predict short-term mortality. Conclusions. In patients admitted to the ED due to dyspnea, at risk of respiratory failure, lactate levels in arterial, venous, and mixed blood samples are independent predictors of short-term mortality.

(1). Background: Therapeutic plasma exchange (TPE) is a method of extracorporeal plasma filtration designed to remove immunoglobulins and pro-inflammatory factors as pathogenesis of numerous diseases. Procedures are performed in intensive care units (ICUs); however, the complications and treatment outcomes remain unclear. The aim of the study was to evaluate clinical outcomes and identify risk factors of complications associated with TPE. (2). Methods: In this multi-center, retrospective, 5-year cohort study, we analyzed a database of 56 adult ICU patients who had undergone TPE at two academic hospitals and one regional hospital. (3). Results: In our study, the median APACHE II was 7.5 IQR 12.5 (4–16.5) and SOFA score was 2 IQR 4 (1–5). In the multivariate analysis, the APACHE II (p = 0.043) and SOFA score (p = 0.046) were the predictors of prolonged length of stay. The SOFA score was a predictor of hospital-acquired infection (HAI) (p = 0.011). Multivariate logistic regression revealed age (p = 0.011), SOFA (p = 0.011), and APACHE II score (p = 0.032) as independent predictors of the development of shock as a TPE complication. (4). Conclusions: Our results suggest that the SOFA and APACHE II scores are significant predictors of the length of hospitalization and the occurrence of shock. In addition, the SOFA score was a predictor of HAI in patients treated TPE in ICU.