The effect of ambient temperature on the analytical and clinical performance of a glucose meter was examined. A total of 114 venous whole blood samples were analysed for glucose by a reference method, and by a glucose meter at 21-22 degrees C, room temperatures, 26-27 degrees C and 33-34 degrees C. Glucose meter readings at each temperature were compared with the reference values and evaluated by analysis of variance, Spearman's correlation, the percentage of glucose meter readings within +/- 10% of the reference value and error grid analysis. Analysis of covariance was used to determine the effect of temperature on glucose meter readings. There were no significant differences in the glucose meter readings and in accuracy of the meter readings between different temperatures. Temperature was not a significant independent determinant of the glucose meter readings. For each glucose concentration, the precision of the meter and clinical performance were comparable between the different temperatures. In conclusion, ambient temperature does not affect the accuracy, precision and clinical performance of the Omnitest Sensor.