ISSN 1470-3947 (print) | ISSN 1479-6848 (online)

Endocrine Abstracts (2019) 64 013 | DOI: 10.1530/endoabs.64.013

The interpretative value of CGM-derived parameters in type 1 diabetes depends on the level of glycaemic control

S Helleputte1, P Calders1, B Pauwels2, S Shadid2, T De Backer3 & B Lapauw2

1Ghent University, Gent, Belgium; 2Department of Endocrinology, University Hospital, Gent, Belgium; 3Department of Cardiology, University Hospital, Gent, Belgium.

Background and aims: HbA1c as a gold-standard measure for glycaemic control has been criticized for several years: lacking the ability to capture hypoglycaemia, time spent in different glucose ranges or to reflect glycaemic variability (GV) are frequently reported shortcomings. Continuous glucose monitoring (CGM) can overcome this, and several new CGM-derived parameters have been proposed to provide additional insights, with the concept of time in range (TIR) and other parameters reflecting glycaemic control and variability being put forward. In this study, we aimed to examine the (inter)relation between these new promising parameters and other indices of glycaemic control in a group of T1DM patients.

Methods: In this observational study, ninety-eight T1DM patients with a minimum disease duration of 10 years and without known macrovascular complications were enrolled in a screening program. Patients were equipped with a blinded Dexcom G4 CGM device for seven days. TIR was defined as time spent in glucose range of 70–180 mg/dl, time in hypoglycaemia was subdivided in total (<70 mg/dl) and level 2 (<55 mg/dl); and time in hyperglycaemia in total (>180 mg/dl) and level 2 (>250 mg/dl). GV was determined by coefficient of variation (COV: mean blood glucose (MBG) divided by S.D. of glucose values). Pearson correlations were used to examine associations between these parameters.

Results: 95 patients (age: 45±10 years; HbAc1: 7.67±0.75%) were included in CGM data analysis (MBG: 159±31; TIR 56.1±14.9%; COV: 43.4±7.8%). Nineteen patients showed good glycaemic control with HbA1c values <7%, 46 patients moderate (7–8%) and 30 patients poor glycaemic control (HbA1c >8%). As expected, HbA1c was significantly associated with MBG (r=0.483; P<0.001) and time spent in hyperglycaemia (total: r=0.509; level 2: r=0.470; P<0.001), but not with time in hypoglycaemia and COV, even after analysis in HbA1c subgroups. In the entire cohort, TIR was negatively associated with HbA1c (r=−0.508; P<0.001), MBG (r=−0.851; P<0.001) and time spent in hyperglycaemia (total: r=−0.924; level 2: r=−0.855; P<0.001), but not with time spent in hypoglycaemia. However, subgroup analysis showed that TIR did associate with shorter time in level 2 hypoglycaemia in patients with good (r=−0.596; P=0.007) and moderate (r=−0.252; P=0.047) glycaemic control. In contrast, COV was strongly positively associated with time in hypoglycaemia (total: r=0.750; level 2: r=0.740; P<0.001), but not with time in hyperglycaemia. Once more, subgroup analysis showed that COV did correlate with time in hyperglycaemia in the lowest HbA1c group (total: r: 0.588; level 2: r=0.662; P<0.01) and with time in level 2 hyperglycaemia in the moderate group (r=0.285; P=0.024). The relationship between TIR and COV was modulated by HbA1c levels as well. TIR did not correlate with COV in the whole group. However, TIR was negatively associated with COV again in patients with good (r=−0.832, P<0.001) and moderate (r=−0.469, P<0.001) glycaemic control.

Conclusion: This study provides arguments for the added value of using CGM-derived parameters as TIR and COV in evaluating glycaemic control in T1DM patients, as they relate with clinical important situations such as level 2 hyper- and hypoglycaemia respectively. It should be noted however, that the interpretation and interrelation of these parameters depends on the level of glycaemic control of the individual patient, adding less in those with poor glycaemic control as it seems not to reflect hypoglycaemia or GV.

Article tools

My recent searches

No recent searches.