Agreement between two plasma bicarbonate assays in critically ill patients.

Previous studies have suggested that measurement of plasma bicarbonate concentration using the Henderson-Hasselbalch equation may be unreliable, particularly in critically ill patients. We examined the agreement between two plasma bicarbonate concentration assays in critically ill patients. Data were collected from records of routine daily blood samples. Paired samples were taken at the same time from arterial lines. A Bland-Altman analysis was used to compare two bicarbonate assays in clinical use. The first used the Henderson-Hasselbalch equation for blood-gas machine calculations. The second used a spectrophotometric enzymatic technique. Comparing the enzymatic method to the calculated method (enzymatic minus calculated) the bias was -1.6 mmol/l (95% CI: -1.2 to -2.0 mmol/l). The limits of agreement were -5.85 mmol/l to 2.65 mmol/l. This study found poor agreement between the two bicarbonate assays. This poor agreement is clinically important but the causes are unclear. We suggest further investigation of the reliability of bicarbonate assays.