Introduction The risk of failure of non-invasive ventilation (NIV) in AHRF patients is greater with higher degrees of acidosis. The BTS guidelines suggest that patients with pH <7.26 have a higher risk of NIV failure which implies a higher risk of endotracheal intubation.
Aims and objectives We set out to verify the strength of this known association of initial arterial pH (pre-commencement of NIV) and the risk of endotracheal intubation in a large dataset.
Methods A retrospective analysis of the initial ABG values on 1003 episodes of NIV at a dedicated respiratory NIV unit from 1 August 2004 to 31 December 2009. We predicted the probability of endotracheal intubation (T) using pH, pCO2 and pO2 variables, using a logistic regression model. The fitted equation was log [T/(1−T)]=53.0704–7.7064 [pH]–0.0752 [CO2]–0.005[O2] which was applied to find the strength of association and expressed as p-values.
Results In 1003 recorded episodes of acute NIV, the data entry was complete in 998 episodes on 712 unique patients. The p-values for the significance in predicting the risk of endotracheal intubation from the initial (pre-NIV) pH, CO2 and O2 are 0.0036, 0.3551, and 0.8929 respectively. The summary statistic for initial pH is stated in Abstract P158 Table 1.
Conclusion Our survey confirms that out of initial pH, pCO2 and pO2, pH is the significant factor (p=0.0036) in predicting the risk of endotracheal intubation of acute hypercapnic respiratory failure patients treated with NIV.