Correlation and agreement between arterial and central venous blood pH, P02) PC02 and HC03- values of mechanically ventilated patients in intensive care unit: A prospective observational study

Abstract

Background: The procedure for arterial blood sampling can be technically difficult with limitations and complications.

Aims: To evaluate the correlation and agreement between arterial and central venous blood pH, PO₂, PCO₂ and HCO₃^- values and infer whether central venous blood gas (CVBG) values could replace arterial blood gas (ABG) values. Design: Prospective observational study.

Methods and Material: A total of 100 samples were collected from 50 adult normotensive and normothermic patients requiring mechanical ventilation. Arterial blood was collected from radial artery and within 2 minutes central venous blood was withdrawn from the same patient. Correlation and agreement was tested using Pearson's Correlation and Bland Altman Analysis.

Results: The pH, PO₂, PCO₂ and HCO₃^- of CVBG correlated significantly with arterial values (r_pH = 0.88, p < 0.001; rpo₂ = 0.358, p <0.05; rpco₂ = 0.470, p < 0.001 and Thco₃ =0.714, p < 0.001). Regression equations were derived to predict AVG values from CVBG values as follows: Arterial pH = 0.879 x central venous pH + 0.9422 (constant), arterial PO₂ = 0.421 x central venous PO₂ +114.4 (constant), R² = 0.128, arterial PCO₂ = 0.429 x central venous PCO₂ + 24.627 (constant), R² = 0.2205 and arterial HCO₃ = 1.045 x central venous HCO₃ + 3.402 (constant), R² = 0.5101.The mean arterial minus venous difference for_pH, PO₂, PCO₂, and bicarbonate was 0.053 ±0.014, 56.04 ±15.74, 2.20 ± 4.4 and 4.30 ±1.64 respectively. Bland-Altman plots for agreement of pH, PO₂, PCO₂, and bicarbonate showed 95% limits of agreement of -0.04 to 0.146, -52.51 to 164.59, -26.61 to 31.01 and -7.0 to 15.6, respectively.

Conclusions: The arterial pH, PO₂, PCO₂ and HCO₃^- values correlated well with central venous values. However, only the arterial pH value can replace the central venous pH value.