In the present study, we aimed to investigate the effects of continuous glucose monitoring (CGM) on blood glucose levels, body weight, blood pressure, and hypoglycaemia in patients with type 2 diabetes mellitus using a meta-analysis of randomized controlled trials (RCTs). A literature search was performed using MEDLINE, Cochrane Controlled Trials Registry, and ClinicalTrials.gov. RCTs using CGM in patients with type 2 diabetes mellitus were then selected. Statistical analysis included calculation of the standardized mean difference (SMD) or risk ratio and 95% confidence intervals (CIs) using a random effects model. After literature search, seven RCTs (669 patients) satisfied the eligibility criteria established herein and were included into the meta-analysis. Compared with the self-monitoring blood glucose group, the CGM group exhibited significantly lower HbA1c levels (SMD, −0.35; 95% CI, −0.59–−0.10;
The number of patients suffering from type 2 diabetes mellitus is increasing worldwide, with estimates suggesting that approximately 300 million individuals could develop the disease by 2050 [
Self-monitoring blood glucose (SMBG) has been proven to be useful for long-term glycaemic control in patients with type 2 diabetes mellitus [
Continuous glucose monitoring (CGM) allows for continuous measurement of interstitial glucose levels in subcutaneous tissues and evaluation of the detailed blood glucose profile of the patient. CGM includes retrospective CGM (r-CGM), which is used for retrospective examination of lifestyle problems and pharmacotherapy adjustment after understanding the blood glucose profile over several days, and real-time CGM (RT-CGM), which confirms the blood glucose profile in real-time. Studies have shown that utilization of such CGM approaches promotes favourable blood glucose control by changing patient behaviours or pharmacotherapy adjustment [
A 2013 meta-analysis that examined the influence of CGM on blood glucose levels in patients with type 2 diabetes mellitus indicated significant improvements in HbA1c levels [
A literature search was performed on 1st February 2018 using MEDLINE (from 1960), Cochrane Controlled Trials Registry (from 1960), and ClinicalTrials.gov. The search strategy was “(type 2 diabet
We created a data extraction form listing the characteristics of studies included in the present study (i.e., key author’s name, publication year, study location, sample size, patient’s baseline information, basic treatment, and treatment duration). Continuous variables were expressed as mean values, standard deviations, standard errors, or 95% confidence intervals (CIs), whereas binary variables were expressed as percentages (%). Studies comparing one SMBG group with two or more intervention groups were treated as two or more studies sharing an SMBG group. Two authors (SI and RK) independently evaluated the quality of research included in the present study. Accordingly, Cochrane’s risk of bias tool was used for evaluating quality [
Given that continuous variables in each study appeared to be expressed using different units, analysis was performed using standardized mean difference (SMD) and 95% CIs. Binary variables were analyzed using the risk ratio (RR) and 95% CIs. When only the standard error or
A total of 1126 papers were extracted from the literature search, among which seven RCTs (669 patients) satisfied the eligibility criteria and were included in the meta-analysis (Figure
Study flow diagram.
Characteristics of CGM interventions included in the present meta-analysis compared with SMBG interventions.
Reference | Year | Region | No. of patients | Age (years) | % women | BMI (kg/m2) | Body weight (kg) | Duration of DM (years) | HbA1c (%) | Hypertension (%) | Dyslipidaemia (%) | Prior CVD (%) | Diabetes treatment | Study duration (weeks) | Type of RT-CGM | Frequency of sensor usage (%) |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
RT-CGM vs. SMBG | ||||||||||||||||
Yoo et al. [ |
2008 | Korea | 65 | 57.5 | 50 | 25.7 | 65.7 | 13.3 | 8.7 | NR | NR | NR | Insulin alone or OADs alone or insulin+OADs | 12 | Guardian-RT MiniMed (Medtronic) | NR |
Ehrhardt et al. [ |
2011 | US | 100 | 60 | 56 | 32.7 | 197 | NR | 8.2 | NR | NR | NR | Diet+exercise or OADs alone or OADs+GLP-1 or basal insulin, alone or in combination | 12 | Dexcom SEVEN (Dexcom) | 68 |
Beck et al. [ |
2017 | US | 158 | 60 | 43.9 | 37 | 105 | 18 | 8.5 | 82 | 63 | 4 | Insulin alone or insulin+OADs | 24 | Dexcom G4 Platinum (Dexcom) | 92 |
r-CGM vs. SMBG | ||||||||||||||||
Allen et al. [ |
2008 | US | 52 | 57 | 48 | 33.8 | NR | 8.5 | 8.4 | NR | NR | NR | Diet+exercise | 8 | Medtronic MiniMed (Medtronic) | NR |
Cosson et al. [ |
2009 | France | 25 | 57 | 27.2 | 30 | NR | 10.5 | 9.2 | NR | NR | NR | Insulin alone or insulin+OADs | 12 | The GlucoDay system (Menarini Diagnostics) | NR |
Ajjan et al. [ |
2016 | UK | 45 | 55.5 | 26.7 | 33.2 | 93.9 | 15.8 | 9.2 | NR | NR | NR | Insulin | 100 (days) | FreeStyle Navigator (Abbotts) | NR |
Haak et al. [ |
2017 | Germany | 224 | 59.5 | 25 | 33.3 | 99 | 18 | 7.5 | NR | NR | NR | Insulin or CSII | 24 | FreeStyle Libre Pro (Abbotts) | NR |
Unless otherwise indicated, data are shown as mean values. CGM: continuous glucose monitoring; SMBG: self-monitoring blood glucose; RT-CGM: real-time continuous glucose monitoring; r-CGM: retrospective continuous glucose monitoring; DM: diabetes mellitus; BMI: body mass index; CVD: cardiovascular diseases; OADs: oral antidiabetic drugs; CSII: continuous subcutaneous insulin infusion; NR: not reported.
Among RCTs included herein, proportions of appropriate assessments for each domain were as follows: random sequence generation, 85.7% (6/7); allocation concealment, 85.7% (6/7); blinding of participants and personnel, 0% (0/7); blinding of outcome assessors, 14.2% (1/7); incomplete data, 71.4% (5/7); and selective reporting, 100% (7/7). The quality of the included RCTs varied considerably, with none of the included studies having a low risk of bias. Generally, the overall risk of bias was high, with most of the bias originating from blinding of participants, personnel, and outcome assessors. As there were <10 RCTs, a funnel plot was not created.
Seven trials regarding HbA1c were included in the meta-analysis [
Forest plot presenting the meta-analysis based on standardized mean differences (SMDs) for the effect of CGM versus SMBG on HbA1c levels. SMDs in the individual studies are presented as squares with 95% confidence intervals (CIs) presented as extending lines. Pooled SMD with its 95% CI is presented as a diamond. CGM: continuous glucose monitoring; SMBG: self-monitoring blood glucose; RT-CGM: real-time continuous glucose monitoring; r-CGM: retrospective continuous glucose monitoring.
Four trials regarding body weight were included in the meta-analysis [
Forest plot presenting the meta-analysis based on standardized mean differences (SMDs) for the effect of CGM versus SMBG on body weight. SMDs in the individual studies are presented as squares with 95% confidence intervals (CIs) presented as extending lines. Pooled SMD with its 95% CI is presented as a diamond. CGM: continuous glucose monitoring; SMBG: self-monitoring blood glucose; RT-CGM: real-time continuous glucose monitoring; r-CGM: retrospective continuous glucose monitoring.
Three trials regarding time spent with hypoglycaemia were included in the meta-analysis [
Forest plot presenting the meta-analysis based on standardized mean differences (SMDs) for the effect of CGM versus SMBG on time spent with hypoglycaemia (<70 mg/dL). SMDs in the individual studies are presented as squares with 95% confidence intervals (CIs) presented as extending lines. Pooled SMD with its 95% CI is presented as a diamond. CGM: continuous glucose monitoring; SMBG: self-monitoring blood glucose; RT-CGM: real-time continuous glucose monitoring; r-CGM: retrospective continuous glucose monitoring.
Forest plot presenting the meta-analysis based on standardized mean differences (SMDs) for the effect of CGM versus SMBG on time spent with hyperglycaemia (>180 mg/dL). SMDs in the individual studies are presented as squares with 95% confidence intervals (CIs) presented as extending lines. Pooled SMD with its 95% CI is presented as a diamond. CGM: continuous glucose monitoring; SMBG: self-monitoring blood glucose; RT-CGM: real-time continuous glucose monitoring; r-CGM: retrospective continuous glucose monitoring.
Two trials regarding systolic blood pressure were included in the meta-analysis [
Forest plot presenting the meta-analysis based on standardized mean differences (SMDs) for the effect of CGM versus SMBG on systolic blood pressure. SMDs in the individual studies are presented as squares with 95% confidence intervals (CIs) presented as extending lines. Pooled SMD with its 95% CI is presented as a diamond. CGM: continuous glucose monitoring; SMBG: self-monitoring blood glucose; RT-CGM: real-time continuous glucose monitoring; r-CGM: retrospective continuous glucose monitoring.
Forest plot presenting the meta-analysis based on standardized mean differences (SMDs) for the effect of CGM versus SMBG on diastolic blood pressure. SMDs in the individual studies are presented as squares with 95% confidence intervals (CIs) presented as extending lines. Pooled SMD with its 95% CI is presented as a diamond. CGM: continuous glucose monitoring; SMBG: self-monitoring blood glucose; RT-CGM: real-time continuous glucose monitoring; r-CGM: retrospective continuous glucose monitoring.
Diabetes-specific scales used in the included trials were the Diabetes Treatment Satisfaction Questionnaire (DTSQ), Diabetes Quality of Life (DQoL), Diabetes Distress Scale (DDS), CGM Satisfaction Scale, etc. (Table
Changes in various patient-reported outcome scores in the CGM and SMBG groups.
Within-group change, mean (SD) | Between-group change, mean (SD) | ||||||
---|---|---|---|---|---|---|---|
CGM group | SMBG group | CGM group | SMBG group | ||||
Baseline | End of study | Baseline | End of study | ||||
DTSQ | |||||||
Ajjan et al. [ |
— | 13.39 | — | 13.52 | — | — | 0.936 |
Haak et al. [ |
— | 13.1 (0.5) | — | 9.0 (0.7) | — | — | <0.001 |
DQoL | |||||||
Haak et al. [ |
— | — | — | — | −0.2 (0.0) | 0.0 (0.0) | 0.025 |
DDS | |||||||
Beck et al. [ |
1.9 (0.8) | 1.8 (0.9) | 2.0 (0.8) | 1.8 (0.6) | — | — | — |
CGM Satisfaction Scale | |||||||
Beck et al. [ |
— | 4.3 (0.4) | — | — | — | — | — |
CGM: continuous glucose monitoring; DTSQ: Diabetes Treatment Satisfaction Questionnaire; DQoL: Diabetes Quality of Life; DDS: Diabetes Distress Scale.
In this study, we examined the influence of CGM on blood glucose levels, weight, blood pressure, and frequency of hypoglycaemia in patients with type 2 diabetes mellitus using a meta-analysis of RCTs. Accordingly, our results revealed that HbA1c levels and time spent with hypoglycaemia were significantly lower in the CGM group than in the SMBG group. Conversely, no difference in body weight and blood pressure was observed between the CGM and SMBG groups.
One 2013 meta-analysis involving four RCTs that collectively examined the effects of RT-CGM and r-CGM in patients with type 2 diabetes mellitus indicated that the CGM treatment group had significantly lower HbA1c levels than the SMBG group [
We showed no difference in body weight change between the CGM and SMBG groups. However, although the study by Beck et al. [
With regard to influence on hypoglycaemia, we showed that the RT-CGM group spent less time with hypoglycaemia than the SMBG group. A previous study examining the utility of CGM for type 1 diabetes observed a shortening in the time spent with hypoglycaemia because of CGM intervention. In general, CGM intervention exhibits greater hypoglycaemic effect among patients with high hypoglycaemic frequency at baseline, such as those with type 1 diabetes [
One study on the effect on blood pressure included herein showed that the CGM group had no reduction in systolic and diastolic blood pressure compared with the SMBG group. In another study included herein, Allen et al. found that the r-CGM group exhibited lower blood pressure during the collection period than the SMBG group. However, as indicated in a previous study [
Although a meta-analysis regarding treatment satisfaction after CGM intervention had not been conducted, the present study included one trial [
Large-scale clinical trials have shown that strict blood glucose management contributes to the reduction of the risk for vascular complications in patients with type 2 diabetes mellitus [
The present study had several limitations. First, given the few number of RCTs included, the present study might have had insufficient power to detect differences between groups. Second, although previous studies on RT-CGM interventions had indicated that the frequency of CGM sensor use influences its effects on HbA1c levels [
The present study examined the effects of CGM on blood glucose levels, body weight, blood pressure, and hypoglycaemia in patients with type 2 diabetes mellitus using a meta-analysis of RCTs. The results revealed that the CGM group had significantly lower HbA1c levels and shorter time spent with hypoglycaemia than the SMBG group. On the other hand, no difference in body weight and blood pressure had been observed between the CGM and SMBG groups. As previously mentioned, given the few RCTs included as well as the presence of heterogeneity, care may be needed when interpreting the results of the present study. Accordingly, further studies addressing the limitations presented herein may be necessary.
The authors declare that they have no conflicts of interest.
The authors would like to thank the staff members of the Department of Metabolic Diseases at Ise Red Cross Hospital for their cooperation in this study.