Early effects of sodium-glucose co-transporter 2 inhibitors in type 2 diabetes: study based on continuous glucose monitoring

Among the T2DM patients admitted to the University of Occupational and Environmental Health, Japan between May 2014 and June 2016, who were being treated with an SGLT2 inhibitor, we selected 24 subjects who had undergone 1-week evaluation of blood glucose dynamics using a CGM system (CGMS^® Gold™ or iPro2, Medtronic MinMed, Inc.). Other inclusion criteria, used in this study, are the following: (1) age >20 years; (2) blood glucose level at admission <300 mg/dL, (3) absence of diabetic ketosis or non-ketotic hyperosmolar coma, and (4) absence of cardiac arrhythmias. Patients with aspartate aminotransferase (AST) ≥100 IU/L or alanine aminotransferase (ALT) ≥100 IU/L, patients with serum creatinine level ≥2.0 mg/dL or estimated glomerular filtration rate (eGFR) <30 mL/min/1.73 m², infectious diseases, acute coronary syndrome, anemia and/or using erythropoiesis stimulating agents were also excluded from the study.

Blood glucose dynamics were evaluated on day −1, 0 (the day of commencement of treatment), 3, and 7. The primary endpoint was the time to onset of action of SGLT2 inhibitors.

The study protocol was approved by the ethics review committees of the University of Occupational and Environmental Health and the participating medical centers. Informed consent was obtained from all subjects.

A previous study indicated that interstitial glucose concentrations measured by CGM correlate with venous blood glucose levels [6]. CGM measurements represent glucose concentrations in the interstitial fluid, but since the introduction of the self-monitoring blood glucose (SMBG) technique, the measured value is considered to represent blood glucose level. Because the iPro2 has insufficient stability within 24 h after its placement, the data after 24 h were used in order to avoid bias related to CGM placement or insufficient stability of the monitoring system. All patients received optimal meals (25 kcal/kg of ideal body weight; 60% carbohydrate, 15–20% protein and 20–25% fat) during CGM. The data recorded by the CGM using a SMBG device were retrieved and the MBG, standard deviation (SD), coefficient of variation (CV), percentage time at glucose level <70, 70–140, and ≥140 mg/dL, mean amplitude of glycemic excursion (MAGE), largest amplitude of glycemic excursion (LAGE), M value, mean postprandial glucose excursion (MPPGE), maximum (Max) and minimum blood glucose level (Min) before each meal, and blood glucose at 1 and 2 h after each meal were computed. MAGE, which was proposed by Service et al. [7], represents fluctuations in blood glucose levels over a 24-h period and was calculated from the daily variations in blood glucose level. The M value is calculated as published previously [8]: \( {{\Sigma \left| { 10 \times { \log }\left( {{\text{GR}}t/ 100} \right)} \right|^{ 3} } \mathord{\left/ {\vphantom {{\varSigma \left| { 10 \times { \log }\left( {{\text{GR}}t/ 100} \right)} \right|^{ 3} } n}} \right. \kern-0pt} n} \), where GRt is the glucose reading at time t, and n is the number of observations. The M value index is used to quantify the change in postprandial blood glucose, as proposed by Schlichtkrull et al. [9]. To assess postprandial glucose excursions from CGM data, MPPGE was calculated as the arithmetic mean of the differences between the postprandial peak glucose values and the corresponding preprandial glucose values for meals.

All numerical values are expressed as mean ± SD, and their normal distribution was tested using the Shapiro–Wilk test. The change in each CGM parameter over time was analyzed using repeated measures analysis of variance, and any noted significant difference was subjected to multiple comparisons using Tukey’s test. The correlation between the change in urinary glucose and that in each CGM parameter was analyzed using the Pearson’s correlation if it was normally distributed; otherwise Spearman’s correlation was used. A difference with a p value of <0.05 was considered significant. The statistical package for the social sciences software version 22.0 was used for the data analysis (SPSS Inc., Chicago, IL).

The clinical characteristics of the 24 subjects (14 men and 10 women) are shown in Table 1. The mean age, duration of morbidity, and HbA1c level were 56.2 ± 8.9 years (range 40–71 years), 5.3 ± 6.8 years (range 0–25 years), and 10.1 ± 2.4% (range 7.3–17.0%), respectively.

All subjects were treated with one SGLT2 inhibitor in addition to dietary therapy once hospitalized. The SGLT2 inhibitors ipragliflozin, dapagliflozin, and tofogliflozin were used in 17, 4, and 2 subjects, respectively and canagliflozin in 1 subject. This study was conducted in clinical settings. The antidiabetic therapy was modified in 4 patients when adding SGLT2 inhibitors, considering the background and treatment goal for each patient: SU drug treatment was discontinued in 1 patient; metformin treatment was added in 1 patient; and treatment with a DPP4 inhibitor and insulin were discontinued in 1 patient each.

Table 2 and Fig. 1 summarize the effects of treatment with SGLT2 inhibitors on various CGM parameters. The MBG and M value significantly decreased on day 0 and further decreased by day 7. The SD, MAGE, and LAGE improved significantly by day 7. The percentage time at ≥140 mg/dL, Max, and Min significantly decreased on day 3 and further improved by day 7 while the percentage time at <70 mg/dL and MPPGE remained unchanged.

We also examined the blood glucose profile on the day treatment was initiated. Analysis of the change in blood glucose from day −1 to 0 showed that the levels before and 1 h after breakfast remained the same (176.5–170.5 and 226.7–207.4 mg/dL, respectively) while those at 2 h after breakfast and before lunch significantly decreased (249.8–218.9 and 206.5–174.2 mg/dL, respectively).

Urinary glucose level increased significantly from day 3 (p < 0.001), with values at 18.8 ± 26.6, 84.3 ± 31.9, and 71.3 ± 31.7 g/day before the start of treatment, at day 3, and at day 7, respectively. The relationship between changes in CGM parameters and urinary glucose before treatment and by day 7 is shown in Table 3. The change in urinary glucose (Δurinary glucose) correlated with ΔSD (r = −0.468, p = 0.024), ΔCV (r = −0.654, p = 0.001), ΔMAGE (r = −0.520, p = 0.011), ΔLAGE (r = −0.465, p = 0.025), ΔMPPGE (r = −0.436, p = 0.038), and Min (r = 0.418, p = 0.047), but not with Δpercent time at ≥140, <70, and 70–140 mg/dL or the ΔM value.

Table 4 describes the relationship between CGM parameters and urinary glucose on day 7. Urinary glucose correlated with MBG (r = 0.414, p = 0.050), percent time at ≥140 (r = 0.468, p = 0.024) and 70–140 mg/dL (r = −0.470, p = 0.023), and Min (r = 0.474, p = 0.022), but not with SD, CV, percent time at <70 mg/dL, MAGE, LAGE, M value, MPPGE, or Max.