Evaluation of a bedside blood ketone sensor: the effects of acidosis, hyperglycaemia and acetoacetate on sensor performance

To assess the performance of a handheld bedside ketone sensor in the face of likely metabolic disturbances in diabetic ketoacidosis, namely: pH, glucose and acetoacetate.

The effects of pH (7.44-6.83), glucose (5-50 mmol/l) and acetoacetate (0-5 mmol/l) were examined in venous blood to investigate the accuracy of betahydroxybutyrate measurement (0-5 mmol/l) by a handheld ketone sensor. Sensor results were compared with a reference method. Linear regression models were fitted to the difference between the methods with the concentration of metabolite as the explanatory factor.

Decreasing pH and increasing glucose had no effect on the accuracy of the handheld ketone sensor; the gradients of the fitted lines were -0.14 and -0.003, respectively. The 95% confidence intervals were -0.7-0.4 and -0.01-0.004, respectively (P = 0.59 and 0.4, respectively). In the acetoacetate study, a positive relationship between the sensor and reference method results was found, the gradient was 0.09. The 95% confidence interval was 0.05-0.14 (P < or = 0.001), indicating that high concentrations of acetoacetate interfere with the sensor performance.

Acidosis and hyperglycaemia have minimal effects on the sensor performance. However, high concentrations of acetoacetate result in some overestimation of betahydroxybutyrate. This bedside ketone sensor provides useful data over a broad range of conditions likely to be encountered during moderate to severe diabetic ketoacidosis.