We read with interest the comments of Ward (1) on our microdialysis-based continuous glucose monitoring system. We especially appreciate that he shares our delight on the stability (i.e., lack of time-dependent signal decline) of this system, which has been verified by standardized regression analysis.
Furthermore, Ward raises the important general question of the optimal way to evaluate a continuous glucose monitoring device—an open question so far. Our suggestion is to perform an evaluation of a continuous glucose monitoring device in a stepwise approach consisting of a technical evaluation followed by a clinical evaluation.
First, there is certainly a need for an accuracy and precision analysis (2) to evaluate the technical performance of a device. Obviously, for this evaluation, glucose values generated by test and reference method are matched according to the moment at which the samples are taken and not according to the moment at which they are measured. Consequently, for the technical evaluation presented, it was not only justified but also mandatory to correct the data for the time lag between sample collection and analysis.
We agree with Ward that there is also a need for a clinical evaluation to predict the clinical efficiency of continuous glucose monitoring devices. But how should an appropriate clinical evaluation of real-time continuous glucose monitoring devices be done? Ward proposed an evaluation by performing a simple accuracy and precision analysis of the data without correcting for the time lag between sample collection and measurement display. This approach totally ignores the additional information on the trend of glucose change, which is a major advantage of real-time continuous glucose monitoring. Therefore, we reckon that it is not appropriate to only use the proposed method for clinical evaluation of a real-time continuous glucose monitoring device. After finalization of the development of algorithms 1) providing trend information and 2) integrating hypo- and hyperglycemia alarms, a more sophisticated clinical evaluation of our device should follow.
There is no doubt that minimizing the system’s characteristic time lag between sample collection within the body and the subsequent display of the measurement result will increase the system’s effectiveness. Decreasing the inner volume and the length of the tubing as well as increasing the perfusion velocity of our microdialysis-based system are realizable tasks for future developments.
References
Address correspondence to Karsten Jungheim, MD, German Diabetes Research Institute, Auf’m Hennekamp 65, D-40225 Duesseldorf, Germany. E-mail: [email protected].
K.J., K.J.W., and A.J.S. received unrestricted research grant support from Roche Diagnostics. L.H. and T.K. received unresticted research grant support from and are members of the advisory board on subcutaneous continuous glucose-monitoring systems for Roche Diagnostics.
