Sensor glucose levels (SGs) after the capillary blood glucose level used to calibrate (BGc) are not influenced by the future BGc in personal continuous glucose monitoring (CGM) (psCGM), but by the future BGc in professional CGM (pfCGM). This leads to a stronger influence of the BGc on SG just after the BGc (SGja) in psCGM than that in pfCGM. Inaccurate SGs or BGc causes dissociation. We investigated the accuracy of psCGM and pfCGM. We cross-sectionally investigated 192 sets including the BGc and SGs before and after the BGc in 12 patients (pts) with type 2 diabetes (96 sets in 6 pts with psCGM (GUARDIAN CONNECT) and pfCGM each). Preprandial and bedtime glucose levels on days 2-5 (of wearing CGM devices) were evaluated in all pts. We calculated the degree of dissociation between the BGc and nearest SGs (DBS) using the following formula: |[(SG just before BGc (SGjb) - BGc) + (SGja - BGc)]/2| = calibration index (CaI) A. We calculated the effect of the BGc on post-SGs (EBpS) using the following formulae: |(SGja - SGjb) - [(SGjb - SG 15 min before SGjb)/3]| = a, |(SGjb - BGc)|/|(SGja - BGc) (if 0, 1)| = b, √(a × b) = CaI B. CaI C was defined as CaI A + CaI B. We defined cases wherein the BGc had dissociation cause potential (CBGc) as both I and II in psCGM group (psCGMg) (I: CaI B > mean in each of two consecutive sets, II: magnitude relation between SGjb and BGc was different between two consecutive sets). CaI B was significantly higher in psCGMg than in pfCGM group (pfCGMg) (p < 0.001), but CaI A and CaI C were not different. CaI A and CaI B were significantly correlated in psCGMg (r = 0.54, p < 0.001), but did not correlate in pfCGMg. High CaI B was significantly associated with high CBGc (p = 0.002). The optimal cutoff value for CBGc in CaI B was 5.5 (sensitivity: 88%, specificity: 85%, area under the curve: 0.87 p < 0.001) in psCGMg. The EBpS was stronger in psCGMg than in pfCGMg, while the DBS was not different. The result of CaI C shows equal accuracy of psCGM to pfCGM in terms of retrospective data. At CaI B > 5.5, recalibration is necessary.


S. Takeishi: None. H. Tsuboi: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at