We agree with Dr. Radermecker’s (1) concerns regarding the applicability/accuracy of real-time continuous glucose monitoring (CGM) for adjusting insulin therapy in the intensive care unit (ICU). This was, however, not our aim, which was to document the duration/intensity of unacceptable glycemia in ICU patients (2). For this purpose, we used the best available method—a posteriori calibration. Since we observed that insulin therapy based on discontinuous glucose measurements failed to maintain normoglycemia, we suggested that it might be improved using online CGM.

Regarding the lag time between blood and interstitial fluid glucose, we acknowledge that both physiological parameters (equilibration of glucose across the capillary endothelial barrier) and device specifics (sampling frequency, microdialysis perfusion rate) should be considered (3). In a recent study, the delay between blood and glucose sensed by the GlucoDay was 7 min and mainly explained by instrument delay (6 min) (4). The lag time is consistent, irrespective of increments/decrements in glycemia and insulin levels (3). The hemodynamic alterations we encountered (hypotension, shock, vasopressor/inotropic need) did not worsen error grid analysis results (2). Such variables would rather affect subcutaneous glucose recovery/microdialysis, resulting in a calibration issue, rather than in a sensor performance issue. A lag time of <10 min is clinically acceptable, since online adjustment of insulin dose should be based on immediate detection by CGM of unacceptable rates of change (>25 mg · dl−1 · h−1).

Criteria for evaluation of CGM performance and calibration are urgently needed. CGM accuracy improved after 6- instead of 2-point calibration (2). Calibration should be performed in times of glucose stability (<10% change over 9 min), a paradigm used in the GlucoDay. From our preliminary results, it seems that for real-time CGM, a single calibration after 2 h is not satisfactory and that verification using conventional glucose measurements is still mandatory. Whether progressive adjustment using later calibrations improves accuracy needs further investigation.

The use of continuous glucose–error grid analysis to evaluate clinical accuracy of CGM systems for ICU patients use is open for discussion (4) and might need adjustment. Insulin therapy in the ICU aims to titrate glycemia in a tight range (80–110 mg/dl), and changes of >25 mg · dl−1 · h−1 (0.4 mg · dl−1 · min−1) already require a change in insulin dose (5). The currently used boundaries of 1 mg · dl−1 · min−1 are therefore not rigorous enough.

Hopefully these observations will stimulate a debate on current glycemic monitoring, the choice of dynamic boundaries for rate–error grid analysis, the need for standard methods to assess accuracy, and the definition of valid requirements for CGM systems in the ICU.

1.
Radermecker RP: Intensive insulin therapy in the intensive care unit: assessment by continuous glucose monitoring (Letter).
Diabetes Care
29
:
2762
–2763,
2006
2.
De Block C, Manuel-y-Keenoy B, Van Gaal L, Rogiers P: Intensive insulin therapy in the intensive care unit: assessment by continuous glucose monitoring.
Diabetes Care
29
:
1750
–1756,
2006
3.
Steil GM, Rebrin K, Hariri F, Jinagonda S, Tadros S, Darwin C, Saad MF: Interstitial fluid glucose dynamics during insulin-induced hypoglycaemia.
Diabetologia
48
:
1833
–1840,
2005
4.
Wentholt IM, Hoekstra JB, DeVries JH: A critical appraisal of the continuous glucose-error grid analysis.
Diabetes Care
29
:
1805
–1811,
2006
5.
Goldberg PA, Siegel MD, Sherwin RS, Halickman JI, Lee M, Bailey VA, Lee SL, Dziura JD, Inzucchi SE: Implementation of a safe and effective insulin infusion protocol in a medical intensive care unit.
Diabetes Care
27
:
461
–467,
2004