Agreement between mathematically arterialised venous versus arterial blood gas values in patients undergoing non-invasive ventilation: a cohort study

Background Blood gas analysis is important for assessment of ventilatory function. Traditionally, arterial analysis has been used. A method for mathematically arterialising venous blood gas values has been developed. Our aim was to validate this method in patients undergoing non-invasive ventilation (NIV) in an emergency department (ED). Materials and methods This post hoc substudy of a prospective cohort study included adult patients undergoing NIV for acute respiratory compromise. When arterial blood gas analysis was required for clinical purposes, a venous sample was also drawn. Mathematically arterialised values were calculated independent of arterial values. Primary outcome of interest was agreement between mathematically arterialised venous and arterial values for pH and pCO(2). Bland-Altman agreement plot analysis was used. Results Eighty sample-pairs (58 patients) were studied. Mean difference for arterial pH (actual-calculated) was 0.01 pH units (95% limits of agreement: -0.04, 0.06). Mean difference for pCO(2) (actual-calculated) was -0.06 kPa (95% limits of agreement: -1.34, 1.22). Conclusions For patients undergoing NIV in an ED, agreement between mathematically arterialised venous values and arterial values was close for pH but only moderate for pCO(2). Depending on clinician tolerance for agreement, this method may be a clinically useful alternative to arterial blood gas analysis in the ED.