Sequential respiratory support in septic patients undergoing continuous renal replacement therapy: A study based on MIMIC-III database.

Objective: Oxygen and hemodynamic management are important for providing a sufficient adequate oxygen-containing blood to the organs for septic patients. In present study, we aimed to explore the application of sequential respiratory support (SRS) and the association of SRS with the outcome of septic patients who needed continuous renal replacement therapy (CRRT).

Methods: We extracted the medical information of septic patients who received CRRT within 24 h of intensive care unit (ICU) admission from the MIMIC-III v1.4. SRS was defined as receiving firstly oxygen therapy followed by mechanical ventilation (MV) within 24 h of admission to ICU. The was performed to compare the differences in clinical characteristics and outcomes of patients with or without SRS. Finally, we developed regression models to analyze the effects of SRS on hospital mortality.

Results: A total of 181 patients entered in this study, and there were 80 patients undergoing MV including SRS group (n = 61) and non-SRS group (n = 19). In the multivariate regression, the value of SRS was associated with the lower risk of hospital mortality adjusted by minimum systolic BP (SBP), maximum lactate, vasopressor use, and sequential organ failure assessment (SOFA) score or Logistic Organ Dysfunction System (LODS) scores within the first 24 h of ICU stay. After adjusted by SBP, maximum lactate, vasopressor use, SOFA, and LODS, there were 31 patients in SRS group with a and 18 cases in non-SRS group, displaying a significantly lower hospital mortality in SRS group than that in patients without S (19.4 % 83.3 %,  < 0.001). In addition, age, qSOFA, necessitating the administration of vasopressor, and duration of vasopressor were significantly correlated with the hospital mortality in septic patients undergoing CRRT and SRS.

Conclusions: Receiving SRS within the first 24 h upon admission to the ICU was independently associated with the hospital mortality in patient with sepsis undergoing CRRT, and patients who were directly received MV had a high risk of death.