RESUMEN
BACKGROUND: During acute on chronic hypercarbic respiratory failure (AHRF), arterial pH is associated with non-invasive ventilation (NIV) failure and mortality. Venous blood gas (VBG) has been proposed as a substitute for arterial blood gas, based on a good agreement between venous and arterial values. We assessed the predictive value of admission VBG on intubation rate, NIV failure and mortality during AHRF. METHODS: Retrospective chart review of inpatients admitted between 2009 and 2015 with AHRF who had VBG performed on admission. Demographic, clinical and biological data were collected throughout the hospital course. RESULTS: 196 patients were included and hospital survival was not significantly associated with initial venous pH, PCO2 or HCO3-. Patients requiring intubation had significantly lower venous pH [7.29 (7.24-7.33) vs 7.31 (7.28-7.36), P = .04] while venous PCO2 and HCO3- did not differ as compared to non-intubated patients. Intubation within 48 h of admission was associated with significantly lower venous pH [7.28 (7.24-7.30) vs 7.32 (7.28-7.37), P = .002] and higher PCO2 [72 (63-92) mm Hg vs 62 (52-75) mm Hg, P = .04]. Among 69 patients receiving NIV, there were no differences in venous pH [7.29 (7.25-7.31) vs 7.30 (7.27-7.35), P = .3] or PCO2 [68 (44-74) mm Hg vs 70 (55-97) mm Hg, P = .23] associated with subsequent intubation. Using c statistics, we observed poor performances of venous pH, PCO2 or HCO3- for prediction of NIV failure, intubation or hospital mortality. CONCLUSIONS: Our results do not support the use of VBG on admission as a predictor for NIV failure, intubation and mortality during AHRF.