Background: Interpretation of glucose sensor results requires clarification of the relationship between interstitial (IG) and blood (BG) glucose. We examined the delay of IG upon BG change and reinvestigated the push-pull phenomenon in type 1 diabetes patients. The push-pull phenomenon postulates that IG shows a delayed increase but earlier decrease compared to BG. If so, postprandial sensor curves should have narrower peak widths than BG curves.
Methods: For both sensors a two-point calibration procedure was used. Delay was assessed by shifting combined fitted postprandial glucose sensor curves horizontally. The sensor and BG peak widths of the separately fitted curves were assessed and compared. Peak width was re-assessed for the microdialysis sensor using raw current values to rule out any calibration effect on the shape of the curve. The contribution of instrumental delay to the earlier reported 7.1-min delay of the microdialysis sensor was calculated.
Results: No delay [-2.2 +/- 6.2 (SD) min] was seen for the needle-type sensor. Instrumental delay was >6.2 min for the microdialysis sensor, accounting for more than 87% of the total reported delay of 7.1 +/- 5.5 min. Mean peak width for the BG curves was 100.8 +/- 25.0 min, for the needle-type sensor curves 110.0 +/- 20.5 min, and for the microdialysis sensor curves 104.6 +/- 21.7 min (P = 0.052 and P = 0.11 vs. BG, respectively). Mean peak width for the uncalibrated microdialysis current values was 105.0 +/- 23.1 min, which was not different from the peak width of the BG curves (P = 0.347).
Conclusions: IG-BG delay may be smaller than previously postulated. The sensor curves tended to have broader peaks than the BG curves, in contrast to the expected narrower peaks predicted by the push-pull phenomenon. This argues against the existence of the push-pull phenomenon.