Background: Arterial base deficit (ABD) measurement is a standard test for assessment of the trauma patient's metabolic response to shock. Venous blood is readily available earlier during the trauma resuscitation. The aim of this study is to analyze the difference (correlation, agreement, clinical significance) between the first peripheral venous base deficit (pVBD) and the first ABD during trauma resuscitation.
Methods: Consecutive trauma patients >18 years presenting to John Hunter Hospital (JHH), Newcastle, Australia, from January 2007 until July 2007 requiring arterial blood gas sampling had a peripheral venous blood gas performed simultaneously. A survey of JHH trauma clinicians and members of the American Association for the Surgery of Trauma was performed to determine a clinically relevant difference between two serial base deficit measurements. Pearson correlation and Bland-Altman tests were performed.
Results: During the 7-month period, 127 patients (79% men, mean age, 46.3 [±18.4 years] and median injury severity score of 15 [interquartile range, 8-23; range, 1-75]) were included into the study. The average peripheral ABD (pABD) and pVBD were -2.2 mmol/L±3.8 mmol/L and -1.3 mmol/L±3.8 mmol/L, respectively. The average difference between measurements was 0.9 (range, -1.7 to +3.5; 95% confidence interval, 0.7-1.0) with pVBD>pABD. The Pearson test showed highly significant correlation (r=0.97, p<0.0001). The survey of 11 JHH and 56 American Association for the Surgery of Trauma clinicians determined 2 mmol/L as clinically relevant difference between two base deficit measurements. All individual paired sample's difference sat within the clinically relevant limits and>95% (121 of 127) of samples sat within the 1.96 standard deviation acceptable by the Bland-Altman plot.
Conclusion: There is near perfect correlation and clinically acceptable agreement between pABD and pVBD values on simultaneous testing. pVBD is an acceptable test to assess trauma patients' initial metabolic status when occult blood loss suspected.