Use of mathematically arterialised venous blood gas sampling: Comparison with arterial, capillary, and venous sampling

Abstract

<b>Background:</b> Arterial blood gas (ABG) sampling is essential for patients treated with non-invasive ventilation (NIV). Sampling is painful, though local anaesthesia is rarely used in UK ward-based practice. Capillary (CBG) and venous (VBG) are alternative methods, though limited by reliability and accuracy concerns. A newer Venous to Arterial Conversion method (v-TAC, Roche) is available, based on VBG measurements combined with SpO₂&nbsp;via pulse oximetry. In this study, we compare ABG sampling with v-TAC, CBG, and VBG. <b>Methods:</b> Time-matched serial ABG, v-TAC, CBG, and VBG samples in adult patients admitted to consider home NIV. The primary outcome was the agreement in PCO₂&nbsp;values and response. <b>Results:</b> 84 patients (45% female) recruited, of whom 52 started NIV. Mean age 63 (12) years. First-time sampling success was 88% for VBG/v-TAC, 67% for ABG, and 55% for CBG. ABG was more painful than v-TAC or CBG (p&lt;0.001). 6% of ABG samples were excluded due to venous contamination (pre-specified criteria). Bland-Altman analysis for PCO₂ and pH showed close agreement for ABG and v-TAC, but not CBG or VBG. Mean difference (SD) with ABG for PCO₂ was: v-TAC -0.01 (0.50) kPa, CBG -0.75 (0.69) kPa, VBG +1.00 (0.90) kPa. PaCO₂ responses to NIV were similar for ABG (0.53kPa), v-TAC (0.55kPa) and VBG (0.49kPa), but lower for CBG (0.16kPa). <b>Conclusions:</b> Compared to ABG, v-TAC sampling was easier and less painful: PaCO₂ and pH results were interchangeable. CBG and VBG showed poor agreement with ABG. These results challenge existing UK ward-based practice of repeated ABG sampling and the use of CBG as an ABG surrogate, plus support an easier and accurate approach for blood gas measurement.