Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Utility of Real-Time and Retrospective Continuous Glucose Monitoring in Patients with Type 2 Diabetes Mellitus: A Meta-Analysis of Randomized Controlled Trials

Utility of Real-Time and Retrospective Continuous Glucose Monitoring in Patients with Type 2... Hindawi Journal of Diabetes Research Volume 2019, Article ID 4684815, 10 pages https://doi.org/10.1155/2019/4684815 Review Article Utility of Real-Time and Retrospective Continuous Glucose Monitoring in Patients with Type 2 Diabetes Mellitus: A Meta-Analysis of Randomized Controlled Trials Satoshi Ida , Ryutaro Kaneko, and Kazuya Murata Department of Diabetes and Metabolism, Ise Red Cross Hospital, Mie, Japan Correspondence should be addressed to Satoshi Ida; bboy98762006@yahoo.co.jp Received 2 August 2018; Revised 23 October 2018; Accepted 28 October 2018; Published 15 January 2019 Academic Editor: Andrea Scaramuzza Copyright © 2019 Satoshi Ida et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 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; P =0 006) and shorter time spent with hypoglycaemia (SMD, −0.42; 95% CI, −0.70–−0.13; P =0 004). Conversely, no differences in body weight and blood pressure were observed between the groups. CGM in patients with type 2 diabetes mellitus could reduce HbA1c levels and time spent with hypoglycaemia. However, because few RCTs were included in this present study and heterogeneity was also noted, care should be taken when interpreting the results. 1. Introduction finger pricking several times per day is not only troublesome but also painful [8]. Furthermore, understanding detailed blood sugar fluctuations, such as elevated blood glucose after The number of patients suffering from type 2 diabetes melli- meals or asymptomatic hypoglycaemia, may be difficult [9]. tus is increasing worldwide, with estimates suggesting that approximately 300 million individuals could develop the Continuous glucose monitoring (CGM) allows for con- tinuous measurement of interstitial glucose levels in subcuta- disease by 2050 [1, 2]. Previous studies have revealed that neous tissues and evaluation of the detailed blood glucose strict blood glucose control is extremely important for profile of the patient. CGM includes retrospective CGM preventing microangiopathy and macrovascular disorders (r-CGM), which is used for retrospective examination of life- [3, 4]. Primary treatment for type 2 diabetes mellitus includes diet/exercise therapy, whereas pharmacotherapy is adminis- style problems and pharmacotherapy adjustment after understanding the blood glucose profile over several days, tered only when diet therapy/exercise therapy is insufficient. However, in many cases, favourable blood glucose control and real-time CGM (RT-CGM), which confirms the blood glucose profile in real-time. Studies have shown that utiliza- cannot be achieved through the aforementioned therapeutic interventions alone [5, 6]. tion of such CGM approaches promotes favourable blood Self-monitoring blood glucose (SMBG) has been proven glucose control by changing patient behaviours or pharma- to be useful for long-term glycaemic control in patients with cotherapy adjustment [10, 11]. type 2 diabetes mellitus [7]. However, this method places A 2013 meta-analysis that examined the influence of considerable burden on the patient given that performing CGM on blood glucose levels in patients with type 2 diabetes 2 Journal of Diabetes Research 2.3. Statistical Analysis. Given that continuous variables in mellitus indicated significant improvements in HbA1c levels [12]. However, the aforementioned study included only a few each study appeared to be expressed using different units, randomized controlled trials (RCTs) and did not examine analysis was performed using standardized mean difference whether CGM intervention had a direct hypoglycaemic (SMD) and 95% CIs. Binary variables were analyzed using reduction effect or an influence on weight. In the present the risk ratio (RR) and 95% CIs. When only the standard study, therefore, we aimed to investigate the effects of CGM error or p values were described, standard deviation was cal- on blood glucose levels, body weight, blood pressure, and culated with reference to the method used by Altman and hypoglycaemia in patients with type 2 diabetes mellitus using Bland [13]. When no description for the standard deviation was present, it was calculated from 95% CIs, t values, or a meta-analysis of RCTs. p values [14]. A random effects model was used for anal- ysis; I was used for evaluating statistical heterogeneity (I ≥ 50% was regarded as heterogeneous) [15]. When the 2. Materials and Methods number of RCTs included in the analysis was ≥10, a funnel plot was created for evaluating publication bias [14]. Further- 2.1. Study Selection. A literature search was performed on 1st more, previous studies have reported that baseline HbA1c February 2018 using MEDLINE (from 1960), Cochrane Controlled Trials Registry (from 1960), and ClinicalTrials.- levels and age may affect the influence of CGM on HbA1c levels [16, 17]. Therefore, when heterogeneity was noted, a gov. The search strategy was “(type 2 diabet or T2DM or NIDDM or non-insulin dependent diabet ) AND [continu- metaregression analysis was conducted on whether baseline ∗ ∗ HbA1c levels, age, and frequency of CGM sensor use affected ous glucose and (monitor or sensing or sensor )] or the impact of CGM on HbA1c levels. Moreover, RevMan [continuous subcutaneous glucose and (monitor or sensing or sensor )] or CGM or CGMS or real-time CGM or version 5.3 (Cochrane Collaboration, https://tech.cochrane. org/revman/download, July/2017) and STATA version 12.1 RT-CGM or flash glucose monitor or FGM or (Stata Corporation LP, College Station, TX) were used for sensor-augmented insulin pump or SAP AND (randomized the analysis. controlled trial or controlled clinical trial or randomized or randomised or placebo or randomly).” The present study included RCTs that evaluated the effect of CGM on blood 3. Results glucose levels, body weight, hypoglycaemic frequency, 3.1. Description of Included Studies. A total of 1126 papers and other parameters in type 2 diabetes. Moreover, we were extracted from the literature search, among which seven included RCTs that compared CGM and SMBG regardless RCTs (669 patients) satisfied the eligibility criteria and were of diet/exercise therapy, oral hypoglycaemic agent use, and included in the meta-analysis (Figure 1) [18–24]. The charac- injectable formulation administration. The exclusion criteria teristics of the seven RCTs are summarized in Table 1. were as follows: non-RCT studies, those involving animal Accordingly, three RCTs used RT-CGM [18–20], whereas experiments, those that targeted patients with gestational four used r-CGM [21–24]. The mean age of the subjects diabetes, those with insufficient data for analysis, and dupli- was 58.0 years. Moreover, women comprised 39.5% of the cate literature. Two authors (SI and RK) independently subjects, the duration of diabetes was 14.0 years, and the test assessed whether each document satisfied the eligibility period was 15.1 weeks. criteria established herein. In case of disagreements between interpretations by the two authors, a third reviewer (KM) 3.2. Assessment of Potential Bias. Among RCTs included was consulted. herein, proportions of appropriate assessments for each domain were as follows: random sequence generation, 2.2. Data Extraction and Quality Assessment. We created a 85.7% (6/7); allocation concealment, 85.7% (6/7); blinding data extraction form listing the characteristics of studies of participants and personnel, 0% (0/7); blinding of outcome included in the present study (i.e., key author’s name, publi- assessors, 14.2% (1/7); incomplete data, 71.4% (5/7); and cation year, study location, sample size, patient’s baseline selective reporting, 100% (7/7). The quality of the included information, basic treatment, and treatment duration). Con- RCTs varied considerably, with none of the included studies tinuous variables were expressed as mean values, standard having a low risk of bias. Generally, the overall risk of bias deviations, standard errors, or 95% confidence intervals was high, with most of the bias originating from blinding of (CIs), whereas binary variables were expressed as percentages participants, personnel, and outcome assessors. As there (%). Studies comparing one SMBG group with two or more were <10 RCTs, a funnel plot was not created. intervention groups were treated as two or more studies shar- 3.3. HbA1c. Seven trials regarding HbA1c were included in ing an SMBG group. Two authors (SI and RK) independently evaluated the quality of research included in the present the meta-analysis [18–24], with 369 and 291 pooled subjects study. Accordingly, Cochrane’s risk of bias tool was used belonging to the CGM and SMBG groups, respectively. An I for evaluating quality [12]. Six domains (random sequence value of 64% (P =0 01) confirmed the presence of heteroge- generation, allocation concealment, blinding of personnel neity. The CGM group had significantly lower HbA1c levels and participants, blinding of outcome assessors, incomplete than the SMBG group (SMD, −0.42; 95% CI, −0.70–−0.13; data, and selective reporting) were evaluated using low, P =0 004; Figure 2). When RT-CGM and r-CGM were moderate, and high risk of bias. viewed separately, the comparison between the RT-CGM Journal of Diabetes Research 3 Studies identified through the database search aer duplicates were removed (n = 1126) Non-relevant studies excluded (n = 998) Full-text articles assessed for eligibility (n =128) Studies excluded (n =121): (i) Non-randomized trial: 12 (ii) Does not evaluate CGM: 37 (iii) Necessary data not provided: (iv) Protocol paper: 7 (v) Patients do not have type 2 diabetes: 49 Studies included in meta-analysis (n = 7) Figure 1: Study flow diagram. and SMBG groups resulted in an SMD of −0.45 (95% CI, the RT-CGM and SMBG groups were not included in −0.67–−0.23; P <0 001), whereas the comparison between these analyses. the r-CGM and SMBG groups resulted in an SMD of −0.43 (95% CI, −0.99–0.13; P =0 13). In addition, despite perform- 3.6. Blood Pressure. Two trials regarding systolic blood pres- sure were included in the meta-analysis [19, 21], with 77 and ing metaregression analysis because of heterogeneity, base- line HbA1c levels (P =0 244) and age (P =0 068) did not 75 pooled subjects in the CGM and SMBG groups, respec- affect the impact of CGM on HbA1c. tively. An I value of 75% (P =0 05) confirmed heterogene- ity. No difference in the systolic blood pressure was observed between the CGM and SMBG groups (SMD, 3.4. Body Weight. Four trials regarding body weight were −0.26; 95% CI, −0.94–0.42; P =0 46; Figure 6). When included in the meta-analysis [18–20, 23], with 191 and 177 RT-CGM and r-CGM were viewed separately, the compari- pooled subjects belonging to the CGM and SMBG groups, son between the RT-CGM and SMBG groups resulted in an respectively. An I value of 47% (P =0 13) suggested no SMD of 0.06 (95% CI, −0.33–0.45; P =0 76), whereas a com- heterogeneity. No difference in body weight change was parison between the r-CGM and control or SMBG group noted between the CGM and SMBG groups (SMD, 0.04; resulted in an SMD of −0.63 (95% CI, −1.19–−0.08; 95% CI, −0.26–0.34; P =0 78; Figure 3). When RT-CGM P =0 03). The same two trials were used for studying the and r-CGM were viewed separately, the comparison diastolic blood pressure in the meta-analysis [19, 21]. No dif- between the RT-CGM and SMBG groups resulted in an ference in the diastolic blood pressure was observed between SMD of 0.12 (95% CI, −0.19–0.42; P =0 45), whereas the the CGM and SMBG groups (SMD, −0.03; 95% CI, −0.35– comparison between the r-CGM and SMBG groups resulted 0.29; P =0 87; Figure 7). When RT-CGM and r-CGM were in an SMD of −0.33 (95% CI, −0.95–0.29; P =0 30). viewed separately, the comparison between the RT-CGM and SMBG groups resulted in an SMD of 0.01 (95% CI, 3.5. Time Spent with Hypoglycaemia (<70 mg/dL) and −0.38–0.40; P =0 96), whereas a comparison between the Hyperglycaemia (>180 mg/dL). Three trials regarding r-CGM and SMBG groups resulted in an SMD of −0.10 time spent with hypoglycaemia were included in the (95% CI, −0.64–0.45; P =0 730). meta-analysis [21, 22, 24], with 181 and 104 pooled subjects in the CGM and SMBG groups, respectively. An I value of 3.7. CGM Satisfaction and Quality of Life. Diabetes-specific 0% (P =0 86) suggested no heterogeneity. The CGM group scales used in the included trials were the Diabetes Treatment exhibited significantly shorter time spent with hypoglycaemia Satisfaction Questionnaire (DTSQ), Diabetes Quality of Life than the SMBG group (SMD, −0.35; 95% CI, −0.59–−0.10; (DQoL), Diabetes Distress Scale (DDS), CGM Satisfaction P =0 006; Figure 4). Moreover, two trials regarding time Scale, etc. (Table 2). Accordingly, although three trials spent with hyperglycaemia were included in the [20, 23, 24] evaluated the aforementioned scales, a meta-analysis [21, 24], with 170 and 90 pooled subjects in meta-analysis was not performed because of the different the CGM and SMBG groups, respectively. An I value of scales used for each study. Two trials utilizing the DTSQ, 0% (P =0 53) indicated no heterogeneity. No difference DQoL, and CGM Satisfaction Scale revealed that treatment in time spent with hyperglycaemia was observed between satisfaction was higher in the CGM group than in the SMBG the CGM and SMBG groups (SMD, 0.07; 95% CI, group [20, 24]. However, in the remaining trial utilizing the −0.19–0.32; P =0 60; Figure 5). Moreover, tests comparing DTSQ [23], no difference in the degree of treatment 4 Journal of Diabetes Research Table 1: Characteristics of CGM interventions included in the present meta-analysis compared with SMBG interventions. Body Study Frequency No. of Age % BMI Duration of HbA1c Hypertension Dyslipidaemia Prior Diabetes Type of Reference Year Region weight duration of sensor patients (years) women (kg/m ) DM (years) (%) (%) (%) CVD (%) treatment RT-CGM (kg) (weeks) usage (%) RT-CGM vs. SMBG Insulin alone or Guardian-RT Yoo et al. 2008 Korea 65 57.5 50 25.7 65.7 13.3 8.7 NR NR NR OADs alone or 12 MiniMed NR [18] insulin+OADs (Medtronic) Diet+exercise or OADs alone or OADs Dexcom Ehrhardt 2011 US 100 60 56 32.7 197 NR 8.2 NR NR NR +GLP-1 or 12 SEVEN 68 et al. [19] basal insulin, (Dexcom) alone or in combination Dexcom G4 Beck Insulin alone or 2017 US 158 60 43.9 37 105 18 8.5 82 63 4 24 Platinum 92 et al. [20] insulin+OADs (Dexcom) r-CGM vs. SMBG Medtronic Allen 2008 US 52 57 48 33.8 NR 8.5 8.4 NR NR NR Diet+exercise 8 MiniMed NR et al. [21] (Medtronic) The GlucoDay Cosson Insulin alone or system 2009 France 25 57 27.2 30 NR 10.5 9.2 NR NR NR 12 NR et al. [22] insulin+OADs (Menarini Diagnostics) FreeStyle Ajjan 100 2016 UK 45 55.5 26.7 33.2 93.9 15.8 9.2 NR NR NR Insulin Navigator NR et al. [23] (days) (Abbotts) Haak FreeStyle Libre 2017 Germany 224 59.5 25 33.3 99 18 7.5 NR NR NR Insulin or CSII 24 NR et al. [24] Pro (Abbotts) 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. Journal of Diabetes Research 5 CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 1.1.1 RT-CGM vs. SMBG Yoo et al. -1.1 1.21 32 -0.4 1.04 33 13.9% -0.61 [-1.11, -0.12] 2008 Ehrhardt et al. -1 1.28 50 -0.5 1.26 50 16.5% -0.39 [-0.79, 0.01] 2011 Beck et al. -0.8 0.44 79 -0.5 0.89 79 18.7% -0.43 [-0.74, -0.11] 2017 Subtotal (95% CI) 161 162 49.1% -0.45 [-0.67, -0.23] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.52, df = 2 (P = 0.77); I = 0% Test for overall effect: Z = 4.00 (P < 0.0001) 1.1.2 r-CGM vs. SMBG Allen et al. -1.16 1.04 27 -0.32 1.02 25 12.3% -0.80 [-1.37, -0.24] 2008 Cosson et al. -0.63 0.34 11 -0.31 0.29 14 7.7% -0.99 [-1.83, -0.15] 2009 Ajjan et al. 68 11.9 30 72 11.9 15 11.2% -0.33 [-0.95, 0.29] 2016 Haak et al. -0.28 1.01 140 -0.41 1.16 75 19.6% 0.12 [-0.16, 0.40] 2017 Subtotal (95% CI) 208 129 50.9% -0.43 [-0.99, 0.13] 2 2 2 Heterogeneity: tau = 0.24; chi = 12.80, df = 3 (P = 0.005); I = 77% Test for overall effect: Z = 1.52 (P = 0.13) Total (95% CI) 369 291 100.0% -0.42 [-0.70, -0.13] 2 2 2 Heterogeneity: tau = 0.09; chi = 16.58, df = 6 (P = 0.01); I = 64% -1 -0.5 0 0.5 1 Test for overall effect: Z = 2.88 (P = 0.004) Favors CGM Favors SMBG 2 2 Test for subgroup differences: chi = 0.00, df = 1 (P = 0.96), I = 0% Figure 2: 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. CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 2.2.1 RT-CGM vs. SMBG Yoo et al. -2.2 13.47 32 -1.4 13.58 33 22.2% -0.06 [-0.54, 0.43] 2008 Ehrhardt et al. -3.4 37.35 50 -0.8 49.09 50 28.0% -0.06 [-0.45, 0.33] 2011 Beck et al. 1.3 3.6 79 -0.2 4.5 79 33.9% 0.37 [0.05, 0.68] 2017 Subtotal (95% CI) 161 162 84.0% 0.12 [-0.19, 0.42] 2 2 2 Heterogeneity: tau = 0.03; chi = 3.62, df = 2 (P = 0.16); I = 45% Test for overall effect: Z = 0.75 (P = 0.45) 2.2.2 r-CGM vs. SMBG Ajjan et al. 94.7 3.87 30 96 3.87 15 16.0% -0.33 [-0.95, 0.29] 2016 Subtotal (95% CI) 30 15 16.0% -0.33 [-0.95, 0.29] Heterogeneity: Not applicable Test for overall effect: Z = 1.04 (P = 0.30) Total (95% CI) 191 177 100.0% 0.04 [-0.26, 0.34] 2 2 2 Heterogeneity: tau = 0.04; chi = 5.62, df = 3 (P = 0.13); I = 47% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.28 (P = 0.78) Favors CGM Favors SMBG 2 2 Test for subgroup differences: Chi = 1.59, df = 1 (P = 0.21), I = 37.3% Figure 3: 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. satisfaction was observed between the CGM and SMBG hypoglycaemia in patients with type 2 diabetes mellitus using groups. Two trials utilizing DDS found no significant differ- a meta-analysis of RCTs. Accordingly, our results revealed ences in scores between the CGM and SMBG groups [20]. 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 4. Discussion pressure was observed between the CGM and SMBG groups. In this study, we examined the influence of CGM on blood One 2013 meta-analysis involving four RCTs that collec- glucose levels, weight, blood pressure, and frequency of tively examined the effects of RT-CGM and r-CGM in 6 Journal of Diabetes Research CGM SMBG Std. mean difference Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 4.1.1 r-CGM vs. SMBG Cosson et al. -2 6.41 11 7 35.41 14 9.5% -0.32 [-1.12, 0.47] 2009 Aljan et al. 0.57 0.67 30 0.7 0.67 15 15.5% -0.19 [-0.81, 0.43] 2016 Haak et al. -0.71 1.63 140 -0.09 1.59 75 75.0% -0.38 [-0.67, -0.10] 2017 Subtotal (95% CI) 181 104 100.0% -0.35 [-0.59, -0.10] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.31, df = 2 (P = 0.86); I = 0% Test for overall effect: Z = 2.78 (P = 0.006) Total (95% CI) 181 104 100.0% -0.35 [-0.59, -0.10] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.31, df = 2 (P = 0.86); I = 0% Test for overall effect: Z = 2.78 (P = 0.006) -1 -0.5 0 0.5 1 Test for subgroup differences: Not applicable Favors CGM Favors SMBG Figure 4: 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. CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 6.1.1 r-CGM vs. SMBG Ajjan et al. 9.13 3.66 30 9.55 3.66 15 17.0% -0.11 [-0.73, 0.51] 2016 Haak et al. 1 5.37 140 0.4 6.14 75 83.0% 0.11 [-0.17, 0.39] 2017 Subtotal (95% CI) 170 90 100.0% 0.07 [-0.19, 0.32] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.40, df = 1 (P = 0.53); I = 0% Test for overall effect: Z = 0.53 (P = 0.60) Total (95% CI) 170 90 100.0% 0.07 [-0.19, 0.32] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.40, df = 1 (P = 0.53); I = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.53 (P= 0.60) Favors CGM Favors SMBG Test for subgroup differences: Not applicable Figure 5: 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. Std. mean difference CGM SMBG Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 7.1.1 RT-CGM vs. SMBG Ehrhardt et al. -1.8 18.02 50 -3 20.45 50 54.3% 0.06 [-0.33, 0.45] 2011 Subtotal (95% CI) 50 50 54.3% 0.06 [-0.33, 0.45] Heterogeneity: Not applicable Test for overall effect: Z = 0.31 ( P= 0.76) 7.1.2 r-CGM vs. SMBG Allen et al. -7 16 27 3 15 25 45.7% -0.63 [-1.19, -0.08] 2008 Subtotal (95% CI) 27 25 45.7% -0.63 [-1.19, -0.08] Heterogeneity: Not applicable Test for overall effect: Z = 2.23 (P = 0.03) Total (95% CI) 77 75 100.0% -0.26 [-0.94, 0.42] 2 2 2 Heterogeneity: tau = 0.18; chi = 4.00, df = 1 (P = 0.05); I = 75% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.74 (P = 0.46) Favors CGM Favors SMBG 2 2 Test for subgroup differences: chi = 4.00, df = 1 (P = 0.05), I = 75.0% Figure 6: 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. patients with type 2 diabetes mellitus indicated that the CGM that the CGM group had significantly lower HbA1c levels treatment group had significantly lower HbA1c levels than than the SMBG group. However, when RT-CGM and the SMBG group [11]. Similarly, the present study revealed r-CGM were viewed separately, we found that although the Journal of Diabetes Research 7 CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 8.1.1 RT-CGM vs. SMBG Ehrhardt et al. -1.3 11.24 50 -1.4 9.99 50 65.9% 0.01 [-0.38, 0.40] 2011 Subtotal (95% CI) 50 50 65.9% 0.01 [-0.38, 0.40] Heterogeneity: Not applicable Test for overall effect: Z = 0.05 (P = 0.96) 8.1.2 r-CGM vs. SMBG Allen et al. -3 11 27 -2 9 25 34.1% -0.10 [-0.64, 0.45] 2008 Subtotal (95% CI) 27 25 34.1% -0.10 [-0.64, 0.45] Heterogeneity: Not applicable Test for overall effect: Z = 0.35 (P = 0.73) Total (95% CI) 77 75 100.0% -0.03 [-0.35, 0.29] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.10, df = 1 (P = 0.75); I = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.17 (P = 0.87) Favors CGM Favors SMBG 2 2 Test for subgroup differences: chi = 0.10, df = 1 (P = 0.75), I = 0% Figure 7: 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. Table 2: 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 P value Baseline End of study Baseline End of study DTSQ Ajjan et al. [23] — 13.39 — 13.52 —— 0.936 Haak et al. [24] — 13.1 (0.5) — 9.0 (0.7) —— <0.001 DQoL 0.025 Haak et al. [24] ——— — −0.2 (0.0) 0.0 (0.0) DDS Beck et al. [20] 1.9 (0.8) 1.8 (0.9) 2.0 (0.8) 1.8 (0.6) —— — CGM Satisfaction Scale Beck et al. [20] — 4.3 (0.4) —— — — — CGM: continuous glucose monitoring; DTSQ: Diabetes Treatment Satisfaction Questionnaire; DQoL: Diabetes Quality of Life; DDS: Diabetes Distress Scale. P <0 05. RT-CGM group had predominantly lower HbA1c levels than Beck et al. [20] showed that the RT-CGM group tended to the SMBG group, no significant difference in HbA1c levels have greater body weight than the SMBG group, the other had been found between the r-CGM and SMBG groups. three trials [18, 19, 23] showed no change or even a decrease in body weight. The daily amount of insulin administered in According to a systematic review of patients with type 1 dia- betes, RT-CGM has a greater blood glucose-ameliorating Beck et al.’s study increased compared with the baseline. effect than r-CGM [25]. The use of RT-CGM helps patients However, this remained unchanged or decreased in the other not only adjust diabetes medication dosage but also under- three trials. Moreover, Beck et al.’s study revealed that stand changes in blood glucose levels on a monitor and be although patients in the RT-CGM group had improved blood conscious of lifestyle factors, such as meals and exercise, glucose levels because of an increase in snacking as a result of thereby ameliorating blood glucose levels [18, 21, 26]. Con- hypoglycaemia or an increase in insulin levels to correct versely, r-CGM increases physical activity and blood glucose blood glucose levels, an increase in body weight could have amelioration and inhibits the onset of complications [21]. been present. Accordingly, blood glucose management using Nevertheless, further studies are needed to determine CGM in patients with type 2 diabetes mellitus necessitates whether RT-CGM improves HbA1c in patients with type 2 paying close attention to the insulin dose and changes in diabetes mellitus to a greater extent than r-CGM. weight [26]. We showed no difference in body weight change between With regard to influence on hypoglycaemia, we showed the CGM and SMBG groups. However, although the study by that the RT-CGM group spent less time with hypoglycaemia 8 Journal of Diabetes Research levels or a ≥10% improvement from baseline values contrib- than the SMBG group. A previous study examining the utility of CGM for type 1 diabetes observed a shortening in the time utes to the inhibition of future cardiovascular events and has spent with hypoglycaemia because of CGM intervention. In been indicated as clinically significant amelioration [28–30]. Given that hypoglycaemia and blood glucose fluctuations, general, CGM intervention exhibits greater hypoglycaemic effect among patients with high hypoglycaemic frequency at which are believed to be related to various poor outcomes, baseline, such as those with type 1 diabetes [17]. Among could be underestimated in patients in type 2 diabetes melli- the studies included in the present meta-analysis, the time tus [31], understanding detailed blood glucose profiles spent with hypoglycaemia per day at patient baseline ranged through CGM may be useful. In recent years, the increase in healthcare costs has been noted as a problem. Reportedly, from 3 to 60 min, which may be considered relatively short [22–24]. Nevertheless, CGM intervention shortened the time CGM intervention is useful in terms of cost effectiveness in spent with hypoglycaemia, suggesting its practicality for patients with type 1 diabetes [32] and in those with type 2 shortening time spent with hypoglycaemia in patients with diabetes [33, 34], although the number of reports is limited type 2 diabetes mellitus. However, given that RCTs compar- for the latter type of patients. Further investigations are needed on effects of CGM intervention in patients with type ing the RT-CGM and SMBG groups had not been included in the present analysis, further investigation is necessary. 2 diabetes to alleviate complications, to reduce the inci- One study on the effect on blood pressure included dence of cardiovascular disease, and to improve QOL and herein showed that the CGM group had no reduction in sys- cost effectiveness. tolic and diastolic blood pressure compared with the SMBG The present study had several limitations. First, given the few number of RCTs included, the present study might have group. In another study included herein, Allen et al. found that the r-CGM group exhibited lower blood pressure during had insufficient power to detect differences between groups. the collection period than the SMBG group. However, as Second, although previous studies on RT-CGM interventions indicated in a previous study [11], given the inclusion of had indicated that the frequency of CGM sensor use influ- counselling on exercise therapy based on r-CGM data, the ences its effects on HbA1c levels [35], this had not been examined because of a lack of sufficient data. Third, we can- independent impact of r-CGM might have not been observed. However, most of the patients in trials included not deny the possibility that some literature could have been herein had been administered hypotensive medication for missed while searching the databases, which could have influenced the results of the present study. Fourth, the obser- blood pressure management. Accordingly, baseline blood pressure management appeared to be the reason why inter- vation period and evaluation items of each RCT included herein varied greatly. Therefore, it appeared necessary to vention effects of CGM had not been observed. Moreover, assessing the influence of CGM on blood pressure had been pay close attention to the interpretation of the results and generally difficult given the few studies included. generalization. Finally, the quality of RCTs included in the present study was generally low. Moreover, given the pres- Although a meta-analysis regarding treatment satisfac- tion after CGM intervention had not been conducted, the ence of heterogeneity, there could be concern regarding the validity of the results derived from the present study. present study included one trial [20, 24] that indicated increased treatment satisfaction and another [23] in which The present study examined the effects of CGM on blood no change was noted. Accordingly, the shortening of time glucose levels, body weight, blood pressure, and hypoglycae- mia in patients with type 2 diabetes mellitus using a spent with hypoglycaemia has been speculated to be the rea- son for such differences. In a previous study on patients with meta-analysis of RCTs. The results revealed that the CGM group had significantly lower HbA1c levels and shorter time type 1 diabetes, the decrease in hypoglycaemic frequency had been indicated to be closely related to patient satisfaction spent with hypoglycaemia than the SMBG group. On the [27]. In our study, there are similar observations wherein a other hand, no difference in body weight and blood pressure had been observed between the CGM and SMBG groups. As shortening of time spent with hypoglycaemia because of CGM in two trials resulted in increased treatment satisfac- previously mentioned, given the few RCTs included as well as tion, but limited shortening of time spent with hypoglycae- the presence of heterogeneity, care may be needed when mia in one study resulted in unchanged satisfaction. Hence, interpreting the results of the present study. Accordingly, based on the trials involving patients with type 2 diabetes further studies addressing the limitations presented herein may be necessary. mellitus included herein, the shortening of time spent with hypoglycaemia because of CGM intervention may perhaps lead to increased treatment satisfaction. Large-scale clinical trials have shown that strict blood Conflicts of Interest glucose management contributes to the reduction of the risk for vascular complications in patients with type 2 diabetes The authors declare that they have no conflicts of interest. mellitus [3, 4]. However, avoiding the risk of hypoglycaemia and maintaining patient QOL are also extremely important for glucose management. The present meta-analysis showed Acknowledgments that the CGM group exhibited a significantly greater degree of HbA1c reduction (a decrease of approximately 1% from The authors would like to thank the staff members of the the baseline value) and shorter time spent with hypoglycae- Department of Metabolic Diseases at Ise Red Cross Hospital mia than the SMBG group. A ≥0.5% improvement in HbA1c for their cooperation in this study. Journal of Diabetes Research 9 system on glycemic control in type 1 diabetic patients: system- References atic review and meta-analysis of randomized trials,” European [1] World Health Organization (WHO), Global Report on Diabe- Journal of Endocrinology, vol. 166, no. 4, pp. 567–574, 2012. tes, World Health Organisation, Geneva, Switzerland, 2016, [17] J. C. Pickup, “The evidence base for diabetes technology: May 2018, https://www.who.int. appropriate and inappropriate meta-analysis,” Journal of Dia- [2] H. King, R. E. Aubert, and W. H. Herman, “Global burden of betes Science and Technology, vol. 7, no. 6, pp. 1567–1574, diabetes, 1995–2025: prevalence, numerical estimates, and 2013. projections,” Diabetes Care, vol. 21, no. 9, pp. 1414–1431, [18] H. J. Yoo, H. G. An, S. Y. Park et al., “Use of a real time contin- uous glucose monitoring system as a motivational device for poorly controlled type 2 diabetes,” Diabetes Research and [3] UK Prospective Diabetes Study (UKPDS) Group, “Intensive blood-glucose control with sulphonylureas or insulin com- Clinical Practice, vol. 82, no. 1, pp. 73–79, 2008. pared with conventional treatment and risk of complications [19] N. M. Ehrhardt, M. Chellappa, M. S. Walker, S. J. Fonda, and in patients with type 2 diabetes (UKPDS 33),” The Lancet, R. A. Vigersky, “The effect of real-time continuous glucose vol. 352, no. 9131, pp. 837–853, 1998. monitoring on glycemic control in patients with type 2 diabe- tes mellitus,” Journal of Diabetes Science and Technology, [4] F. M. Turnbull, C. Abraira, R. J. Anderson et al., “Intensive glu- vol. 5, no. 3, pp. 668–675, 2011. cose control and macrovascular outcomes in type 2 diabetes,” Diabetologia, vol. 52, no. 11, pp. 2288–2298, 2009. [20] R. W. Beck, T. D. Riddlesworth, K. Ruedy et al., “Continuous glucose monitoring versus usual care in patients with type 2 [5] W. H. Herman and P. Zimmet, “Type 2 diabetes: an epidemic diabetes receiving multiple daily insulin injections: a random- requiring global attention and urgent action,” Diabetes Care, ized trial,” Annals of Internal Medicine, vol. 167, no. 6, vol. 35, no. 5, pp. 943-944, 2012. pp. 365–374, 2017. [6] A. Nanditha, R. C. W. Ma, A. Ramachandran et al., “Diabetes [21] N. A. Allen, J. A. Fain, B. Braun, and S. R. Chipkin, “Continu- in Asia and the pacific: implications for the global epidemic,” ous glucose monitoring counseling improves physical activity Diabetes Care, vol. 39, no. 3, pp. 472–485, 2016. behaviors of individuals with type 2 diabetes: a randomized [7] O. Schnell, H. Alawi, T. Battelino et al., “Self-monitoring of clinical trial,” Diabetes Research and Clinical Practice, vol. 80, blood glucose in type 2 diabetes: recent studies,” Journal of no. 3, pp. 371–379, 2008. Diabetes Science and Technology, vol. 7, no. 2, pp. 478–488, [22] E. Cosson, E. Hamo-Tchatchouang, L. Dufaitre-Patouraux, J. R. Attali, J. Pariès, and P. Schaepelynck-Bélicar, “Multicen- [8] H. P. Chase, L. M. Kim, S. L. Owen et al., “Continuous subcu- tre, randomised, controlled study of the impact of continuous taneous glucose monitoring in children with type 1 diabetes,” sub-cutaneous glucose monitoring (GlucoDay ) on glycaemic Pediatrics, vol. 107, no. 2, pp. 222–226, 2001. control in type 1 and type 2 diabetes patients,” Diabetes & [9] E. Boland, T. Monsod, M. Delucia, C. A. Brandt, S. Fernando, Metabolism, vol. 35, no. 4, pp. 312–318, 2009. and W. V. Tamborlane, “Limitations of conventional methods [23] R. A. Ajjan, K. Abougila, S. Bellary et al., “Sensor and software of self-monitoring of blood glucose: lessons learned from 3 use for the glycaemic management of insulin-treated type 1 days of continuous glucose sensing in pediatric patients with and type 2 diabetes patients,” Diabetes and Vascular Disease type 1 diabetes,” Diabetes Care, vol. 24, no. 11, pp. 1858– Research, vol. 13, no. 3, pp. 211–219, 2016. 1862, 2001. [24] T. Haak, H. Hanaire, R. Ajjan, N. Hermanns, J. P. Riveline, and [10] N. A. Allen, J. A. Fain, B. Braun, and S. R. Chipkin, “Continu- G. Rayman, “Flash glucose-sensing technology as a replace- ous glucose monitoring in non–insulin-using individuals with ment for blood glucose monitoring for the management of type 2 diabetes: acceptability, feasibility, and teaching opportu- insulin-treated type 2 diabetes: a multicenter, open-label ran- nities,” Diabetes Technology & Therapeutics, vol. 11, no. 3, domized controlled trial,” Diabetes Therapy, vol. 8, no. 1, pp. 151–158, 2009. pp. 55–73, 2017. [11] N. Poolsup, N. Suksomboon, and A. Kyaw, “Systematic review [25] M. Langendam, Y. M. Luijf, L. Hooft, J. H. DeVries, A. H. and meta-analysis of the effectiveness of continuous glucose Mudde, and R. J. P. M. Scholten, “Continuous glucose monitoring (CGM) on glucose control in diabetes,” Diabetol- monitoring systems for type 1 diabetes mellitus,” Cochrane ogy & Metabolic Syndrome, vol. 5, no. 1, p. 39, 2013. Database of Systematic Reviews, no. 1, article CD008101, 2012. [12] M. C. Simmonds, J. P. T. Higginsa, L. A. Stewartb, J. F. [26] R. A. Vigersky, S. J. Fonda, M. Chellappa, M. S. Walker, and Tierneyb, M. J. Clarke, and S. G. Thompson, “Meta-analysis N. M. Ehrhardt, “Short- and long-term effects of real-time of individual patient data from randomized trials: a review continuous glucose monitoring in patients with type 2 diabe- of methods used in practice,” Clinical Trials, vol. 2, no. 3, tes,” Diabetes Care, vol. 35, no. 1, pp. 32–38, 2012. pp. 209–217, 2005. [27] R. W. Beck, T. Riddlesworth, K. Ruedy et al., “Effect of contin- [13] D. G. Altman and J. M. Bland, “Detecting skewness from sum- uous glucose monitoring on glycemic control in adults with mary information,” BMJ, vol. 313, no. 7066, p. 1200, 1996. type 1 diabetes using insulin injections: the DIAMOND ran- [14] J. P. T. Higgins and S. Green, Eds., Cochrane Handbook for domized clinical trial,” JAMA, vol. 317, no. 4, pp. 371–378, Systematic Reviews of Interventions, John Wiley & Sons, Ltd, 2011, The Cochrane Collaboration, version 5.1.0. March [28] E. Cummins, P. Royle, A. Snaith et al., “Clinical effectiveness 2018, https://training.cochrane.org/handbook. and cost-effectiveness of continuous subcutaneous insulin [15] J. P. Higgins, S. G. Thompson, J. J. Deeks, and D. G. Altman, infusion for diabetes: systematic review and economic evalua- “Measuring inconsistency in meta-analyses,” BMJ, vol. 327, tion,” Health Technology Assessment, vol. 14, no. 11, 2010. no. 7414, pp. 557–560, 2003. [29] Writing Team for the Diabetes Control and Complications [16] A. Szypowska, A. Ramotowska, K. Dzygalo, and D. Golicki, Trial/Epidemiology of Diabetes Interventions and Complica- “Beneficial effect of real-time continuous glucose monitoring tions, “Effect of intensive therapy on the microvascular 10 Journal of Diabetes Research complications of type 1 diabetes mellitus,” JAMA, vol. 287, no. 19, pp. 2563–2569, 2002. [30] I. M. Stratton, A. I. Adler, H. A. W. Neil et al., “Association of glycaemia with macrovascular and microvascular complica- tions of type 2 diabetes (UKPDS 35): prospective observational study,” BMJ, vol. 321, no. 7258, pp. 405–412, 2000. [31] S. F. E. Praet, R. J. F. Manders, R. C. R. Meex et al., “Glycaemic instability is an underestimated problem in type II diabetes,” Clinical Science, vol. 111, no. 2, pp. 119–126, 2006. [32] W. Wan, M. R. Skandari, A. Minc et al., “Cost-effectiveness of continuous glucose monitoring for adults with type 1 diabetes compared with self-monitoring of blood glucose: the DIA- MOND randomized trial,” Diabetes Care, vol. 41, no. 6, pp. 1227–1234, 2018. [33] J. A. Sierra, M. Shah, M. S. Gill et al., “Clinical and economic benefits of professional CGM among people with type 2 diabetes in the United States: analysis of claims and lab data,” Journal of Medical Economics, vol. 21, no. 3, pp. 225–230, [34] S. J. Fonda, C. Graham, J. Munakata, J. M. Powers, D. Price, and R. A. Vigersky, “The cost-effectiveness of real-time continuous glucose monitoring (RT-CGM) in type 2 diabetes,” Journal of Diabetes Science and Technology, vol. 10, no. 4, pp. 898–904, 2016. [35] T. Battelino, S. Liabat, H. J. Veeze, J. Castañeda, A. Arrieta, and O. Cohen, “Routine use of continuous glucose monitoring in 10 501 people with diabetes mellitus,” Diabetic Medicine, vol. 32, no. 12, pp. 1568–1574, 2015. MEDIATORS of INFLAMMATION The Scientific Gastroenterology Journal of World Journal Research and Practice Diabetes Research Disease Markers Hindawi Hindawi Publishing Corporation Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 http://www www.hindawi.com .hindawi.com V Volume 2018 olume 2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 International Journal of Journal of Immunology Research Endocrinology Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com BioMed PPAR Research Research International Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Journal of Obesity Evidence-Based Journal of Journal of Stem Cells Complementary and Ophthalmology International Alternative Medicine Oncology Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2013 Parkinson’s Disease Computational and Behavioural Mathematical Methods AIDS Oxidative Medicine and in Medicine Neurology Research and Treatment Cellular Longevity Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Diabetes Research Hindawi Publishing Corporation

Utility of Real-Time and Retrospective Continuous Glucose Monitoring in Patients with Type 2 Diabetes Mellitus: A Meta-Analysis of Randomized Controlled Trials

Loading next page...
 
/lp/hindawi-publishing-corporation/utility-of-real-time-and-retrospective-continuous-glucose-monitoring-HHkhWnJDKz

References (38)

Publisher
Hindawi Publishing Corporation
Copyright
Copyright © 2019 Satoshi Ida et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ISSN
2314-6745
eISSN
2314-6753
DOI
10.1155/2019/4684815
Publisher site
See Article on Publisher Site

Abstract

Hindawi Journal of Diabetes Research Volume 2019, Article ID 4684815, 10 pages https://doi.org/10.1155/2019/4684815 Review Article Utility of Real-Time and Retrospective Continuous Glucose Monitoring in Patients with Type 2 Diabetes Mellitus: A Meta-Analysis of Randomized Controlled Trials Satoshi Ida , Ryutaro Kaneko, and Kazuya Murata Department of Diabetes and Metabolism, Ise Red Cross Hospital, Mie, Japan Correspondence should be addressed to Satoshi Ida; bboy98762006@yahoo.co.jp Received 2 August 2018; Revised 23 October 2018; Accepted 28 October 2018; Published 15 January 2019 Academic Editor: Andrea Scaramuzza Copyright © 2019 Satoshi Ida et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 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; P =0 006) and shorter time spent with hypoglycaemia (SMD, −0.42; 95% CI, −0.70–−0.13; P =0 004). Conversely, no differences in body weight and blood pressure were observed between the groups. CGM in patients with type 2 diabetes mellitus could reduce HbA1c levels and time spent with hypoglycaemia. However, because few RCTs were included in this present study and heterogeneity was also noted, care should be taken when interpreting the results. 1. Introduction finger pricking several times per day is not only troublesome but also painful [8]. Furthermore, understanding detailed blood sugar fluctuations, such as elevated blood glucose after The number of patients suffering from type 2 diabetes melli- meals or asymptomatic hypoglycaemia, may be difficult [9]. tus is increasing worldwide, with estimates suggesting that approximately 300 million individuals could develop the Continuous glucose monitoring (CGM) allows for con- tinuous measurement of interstitial glucose levels in subcuta- disease by 2050 [1, 2]. Previous studies have revealed that neous tissues and evaluation of the detailed blood glucose strict blood glucose control is extremely important for profile of the patient. CGM includes retrospective CGM preventing microangiopathy and macrovascular disorders (r-CGM), which is used for retrospective examination of life- [3, 4]. Primary treatment for type 2 diabetes mellitus includes diet/exercise therapy, whereas pharmacotherapy is adminis- style problems and pharmacotherapy adjustment after understanding the blood glucose profile over several days, tered only when diet therapy/exercise therapy is insufficient. However, in many cases, favourable blood glucose control and real-time CGM (RT-CGM), which confirms the blood glucose profile in real-time. Studies have shown that utiliza- cannot be achieved through the aforementioned therapeutic interventions alone [5, 6]. tion of such CGM approaches promotes favourable blood Self-monitoring blood glucose (SMBG) has been proven glucose control by changing patient behaviours or pharma- to be useful for long-term glycaemic control in patients with cotherapy adjustment [10, 11]. type 2 diabetes mellitus [7]. However, this method places A 2013 meta-analysis that examined the influence of considerable burden on the patient given that performing CGM on blood glucose levels in patients with type 2 diabetes 2 Journal of Diabetes Research 2.3. Statistical Analysis. Given that continuous variables in mellitus indicated significant improvements in HbA1c levels [12]. However, the aforementioned study included only a few each study appeared to be expressed using different units, randomized controlled trials (RCTs) and did not examine analysis was performed using standardized mean difference whether CGM intervention had a direct hypoglycaemic (SMD) and 95% CIs. Binary variables were analyzed using reduction effect or an influence on weight. In the present the risk ratio (RR) and 95% CIs. When only the standard study, therefore, we aimed to investigate the effects of CGM error or p values were described, standard deviation was cal- on blood glucose levels, body weight, blood pressure, and culated with reference to the method used by Altman and hypoglycaemia in patients with type 2 diabetes mellitus using Bland [13]. When no description for the standard deviation was present, it was calculated from 95% CIs, t values, or a meta-analysis of RCTs. p values [14]. A random effects model was used for anal- ysis; I was used for evaluating statistical heterogeneity (I ≥ 50% was regarded as heterogeneous) [15]. When the 2. Materials and Methods number of RCTs included in the analysis was ≥10, a funnel plot was created for evaluating publication bias [14]. Further- 2.1. Study Selection. A literature search was performed on 1st more, previous studies have reported that baseline HbA1c February 2018 using MEDLINE (from 1960), Cochrane Controlled Trials Registry (from 1960), and ClinicalTrials.- levels and age may affect the influence of CGM on HbA1c levels [16, 17]. Therefore, when heterogeneity was noted, a gov. The search strategy was “(type 2 diabet or T2DM or NIDDM or non-insulin dependent diabet ) AND [continu- metaregression analysis was conducted on whether baseline ∗ ∗ HbA1c levels, age, and frequency of CGM sensor use affected ous glucose and (monitor or sensing or sensor )] or the impact of CGM on HbA1c levels. Moreover, RevMan [continuous subcutaneous glucose and (monitor or sensing or sensor )] or CGM or CGMS or real-time CGM or version 5.3 (Cochrane Collaboration, https://tech.cochrane. org/revman/download, July/2017) and STATA version 12.1 RT-CGM or flash glucose monitor or FGM or (Stata Corporation LP, College Station, TX) were used for sensor-augmented insulin pump or SAP AND (randomized the analysis. controlled trial or controlled clinical trial or randomized or randomised or placebo or randomly).” The present study included RCTs that evaluated the effect of CGM on blood 3. Results glucose levels, body weight, hypoglycaemic frequency, 3.1. Description of Included Studies. A total of 1126 papers and other parameters in type 2 diabetes. Moreover, we were extracted from the literature search, among which seven included RCTs that compared CGM and SMBG regardless RCTs (669 patients) satisfied the eligibility criteria and were of diet/exercise therapy, oral hypoglycaemic agent use, and included in the meta-analysis (Figure 1) [18–24]. The charac- injectable formulation administration. The exclusion criteria teristics of the seven RCTs are summarized in Table 1. were as follows: non-RCT studies, those involving animal Accordingly, three RCTs used RT-CGM [18–20], whereas experiments, those that targeted patients with gestational four used r-CGM [21–24]. The mean age of the subjects diabetes, those with insufficient data for analysis, and dupli- was 58.0 years. Moreover, women comprised 39.5% of the cate literature. Two authors (SI and RK) independently subjects, the duration of diabetes was 14.0 years, and the test assessed whether each document satisfied the eligibility period was 15.1 weeks. criteria established herein. In case of disagreements between interpretations by the two authors, a third reviewer (KM) 3.2. Assessment of Potential Bias. Among RCTs included was consulted. herein, proportions of appropriate assessments for each domain were as follows: random sequence generation, 2.2. Data Extraction and Quality Assessment. We created a 85.7% (6/7); allocation concealment, 85.7% (6/7); blinding data extraction form listing the characteristics of studies of participants and personnel, 0% (0/7); blinding of outcome included in the present study (i.e., key author’s name, publi- assessors, 14.2% (1/7); incomplete data, 71.4% (5/7); and cation year, study location, sample size, patient’s baseline selective reporting, 100% (7/7). The quality of the included information, basic treatment, and treatment duration). Con- RCTs varied considerably, with none of the included studies tinuous variables were expressed as mean values, standard having a low risk of bias. Generally, the overall risk of bias deviations, standard errors, or 95% confidence intervals was high, with most of the bias originating from blinding of (CIs), whereas binary variables were expressed as percentages participants, personnel, and outcome assessors. As there (%). Studies comparing one SMBG group with two or more were <10 RCTs, a funnel plot was not created. intervention groups were treated as two or more studies shar- 3.3. HbA1c. Seven trials regarding HbA1c were included in ing an SMBG group. Two authors (SI and RK) independently evaluated the quality of research included in the present the meta-analysis [18–24], with 369 and 291 pooled subjects study. Accordingly, Cochrane’s risk of bias tool was used belonging to the CGM and SMBG groups, respectively. An I for evaluating quality [12]. Six domains (random sequence value of 64% (P =0 01) confirmed the presence of heteroge- generation, allocation concealment, blinding of personnel neity. The CGM group had significantly lower HbA1c levels and participants, blinding of outcome assessors, incomplete than the SMBG group (SMD, −0.42; 95% CI, −0.70–−0.13; data, and selective reporting) were evaluated using low, P =0 004; Figure 2). When RT-CGM and r-CGM were moderate, and high risk of bias. viewed separately, the comparison between the RT-CGM Journal of Diabetes Research 3 Studies identified through the database search aer duplicates were removed (n = 1126) Non-relevant studies excluded (n = 998) Full-text articles assessed for eligibility (n =128) Studies excluded (n =121): (i) Non-randomized trial: 12 (ii) Does not evaluate CGM: 37 (iii) Necessary data not provided: (iv) Protocol paper: 7 (v) Patients do not have type 2 diabetes: 49 Studies included in meta-analysis (n = 7) Figure 1: Study flow diagram. and SMBG groups resulted in an SMD of −0.45 (95% CI, the RT-CGM and SMBG groups were not included in −0.67–−0.23; P <0 001), whereas the comparison between these analyses. the r-CGM and SMBG groups resulted in an SMD of −0.43 (95% CI, −0.99–0.13; P =0 13). In addition, despite perform- 3.6. Blood Pressure. Two trials regarding systolic blood pres- sure were included in the meta-analysis [19, 21], with 77 and ing metaregression analysis because of heterogeneity, base- line HbA1c levels (P =0 244) and age (P =0 068) did not 75 pooled subjects in the CGM and SMBG groups, respec- affect the impact of CGM on HbA1c. tively. An I value of 75% (P =0 05) confirmed heterogene- ity. No difference in the systolic blood pressure was observed between the CGM and SMBG groups (SMD, 3.4. Body Weight. Four trials regarding body weight were −0.26; 95% CI, −0.94–0.42; P =0 46; Figure 6). When included in the meta-analysis [18–20, 23], with 191 and 177 RT-CGM and r-CGM were viewed separately, the compari- pooled subjects belonging to the CGM and SMBG groups, son between the RT-CGM and SMBG groups resulted in an respectively. An I value of 47% (P =0 13) suggested no SMD of 0.06 (95% CI, −0.33–0.45; P =0 76), whereas a com- heterogeneity. No difference in body weight change was parison between the r-CGM and control or SMBG group noted between the CGM and SMBG groups (SMD, 0.04; resulted in an SMD of −0.63 (95% CI, −1.19–−0.08; 95% CI, −0.26–0.34; P =0 78; Figure 3). When RT-CGM P =0 03). The same two trials were used for studying the and r-CGM were viewed separately, the comparison diastolic blood pressure in the meta-analysis [19, 21]. No dif- between the RT-CGM and SMBG groups resulted in an ference in the diastolic blood pressure was observed between SMD of 0.12 (95% CI, −0.19–0.42; P =0 45), whereas the the CGM and SMBG groups (SMD, −0.03; 95% CI, −0.35– comparison between the r-CGM and SMBG groups resulted 0.29; P =0 87; Figure 7). When RT-CGM and r-CGM were in an SMD of −0.33 (95% CI, −0.95–0.29; P =0 30). viewed separately, the comparison between the RT-CGM and SMBG groups resulted in an SMD of 0.01 (95% CI, 3.5. Time Spent with Hypoglycaemia (<70 mg/dL) and −0.38–0.40; P =0 96), whereas a comparison between the Hyperglycaemia (>180 mg/dL). Three trials regarding r-CGM and SMBG groups resulted in an SMD of −0.10 time spent with hypoglycaemia were included in the (95% CI, −0.64–0.45; P =0 730). meta-analysis [21, 22, 24], with 181 and 104 pooled subjects in the CGM and SMBG groups, respectively. An I value of 3.7. CGM Satisfaction and Quality of Life. Diabetes-specific 0% (P =0 86) suggested no heterogeneity. The CGM group scales used in the included trials were the Diabetes Treatment exhibited significantly shorter time spent with hypoglycaemia Satisfaction Questionnaire (DTSQ), Diabetes Quality of Life than the SMBG group (SMD, −0.35; 95% CI, −0.59–−0.10; (DQoL), Diabetes Distress Scale (DDS), CGM Satisfaction P =0 006; Figure 4). Moreover, two trials regarding time Scale, etc. (Table 2). Accordingly, although three trials spent with hyperglycaemia were included in the [20, 23, 24] evaluated the aforementioned scales, a meta-analysis [21, 24], with 170 and 90 pooled subjects in meta-analysis was not performed because of the different the CGM and SMBG groups, respectively. An I value of scales used for each study. Two trials utilizing the DTSQ, 0% (P =0 53) indicated no heterogeneity. No difference DQoL, and CGM Satisfaction Scale revealed that treatment in time spent with hyperglycaemia was observed between satisfaction was higher in the CGM group than in the SMBG the CGM and SMBG groups (SMD, 0.07; 95% CI, group [20, 24]. However, in the remaining trial utilizing the −0.19–0.32; P =0 60; Figure 5). Moreover, tests comparing DTSQ [23], no difference in the degree of treatment 4 Journal of Diabetes Research Table 1: Characteristics of CGM interventions included in the present meta-analysis compared with SMBG interventions. Body Study Frequency No. of Age % BMI Duration of HbA1c Hypertension Dyslipidaemia Prior Diabetes Type of Reference Year Region weight duration of sensor patients (years) women (kg/m ) DM (years) (%) (%) (%) CVD (%) treatment RT-CGM (kg) (weeks) usage (%) RT-CGM vs. SMBG Insulin alone or Guardian-RT Yoo et al. 2008 Korea 65 57.5 50 25.7 65.7 13.3 8.7 NR NR NR OADs alone or 12 MiniMed NR [18] insulin+OADs (Medtronic) Diet+exercise or OADs alone or OADs Dexcom Ehrhardt 2011 US 100 60 56 32.7 197 NR 8.2 NR NR NR +GLP-1 or 12 SEVEN 68 et al. [19] basal insulin, (Dexcom) alone or in combination Dexcom G4 Beck Insulin alone or 2017 US 158 60 43.9 37 105 18 8.5 82 63 4 24 Platinum 92 et al. [20] insulin+OADs (Dexcom) r-CGM vs. SMBG Medtronic Allen 2008 US 52 57 48 33.8 NR 8.5 8.4 NR NR NR Diet+exercise 8 MiniMed NR et al. [21] (Medtronic) The GlucoDay Cosson Insulin alone or system 2009 France 25 57 27.2 30 NR 10.5 9.2 NR NR NR 12 NR et al. [22] insulin+OADs (Menarini Diagnostics) FreeStyle Ajjan 100 2016 UK 45 55.5 26.7 33.2 93.9 15.8 9.2 NR NR NR Insulin Navigator NR et al. [23] (days) (Abbotts) Haak FreeStyle Libre 2017 Germany 224 59.5 25 33.3 99 18 7.5 NR NR NR Insulin or CSII 24 NR et al. [24] Pro (Abbotts) 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. Journal of Diabetes Research 5 CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 1.1.1 RT-CGM vs. SMBG Yoo et al. -1.1 1.21 32 -0.4 1.04 33 13.9% -0.61 [-1.11, -0.12] 2008 Ehrhardt et al. -1 1.28 50 -0.5 1.26 50 16.5% -0.39 [-0.79, 0.01] 2011 Beck et al. -0.8 0.44 79 -0.5 0.89 79 18.7% -0.43 [-0.74, -0.11] 2017 Subtotal (95% CI) 161 162 49.1% -0.45 [-0.67, -0.23] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.52, df = 2 (P = 0.77); I = 0% Test for overall effect: Z = 4.00 (P < 0.0001) 1.1.2 r-CGM vs. SMBG Allen et al. -1.16 1.04 27 -0.32 1.02 25 12.3% -0.80 [-1.37, -0.24] 2008 Cosson et al. -0.63 0.34 11 -0.31 0.29 14 7.7% -0.99 [-1.83, -0.15] 2009 Ajjan et al. 68 11.9 30 72 11.9 15 11.2% -0.33 [-0.95, 0.29] 2016 Haak et al. -0.28 1.01 140 -0.41 1.16 75 19.6% 0.12 [-0.16, 0.40] 2017 Subtotal (95% CI) 208 129 50.9% -0.43 [-0.99, 0.13] 2 2 2 Heterogeneity: tau = 0.24; chi = 12.80, df = 3 (P = 0.005); I = 77% Test for overall effect: Z = 1.52 (P = 0.13) Total (95% CI) 369 291 100.0% -0.42 [-0.70, -0.13] 2 2 2 Heterogeneity: tau = 0.09; chi = 16.58, df = 6 (P = 0.01); I = 64% -1 -0.5 0 0.5 1 Test for overall effect: Z = 2.88 (P = 0.004) Favors CGM Favors SMBG 2 2 Test for subgroup differences: chi = 0.00, df = 1 (P = 0.96), I = 0% Figure 2: 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. CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 2.2.1 RT-CGM vs. SMBG Yoo et al. -2.2 13.47 32 -1.4 13.58 33 22.2% -0.06 [-0.54, 0.43] 2008 Ehrhardt et al. -3.4 37.35 50 -0.8 49.09 50 28.0% -0.06 [-0.45, 0.33] 2011 Beck et al. 1.3 3.6 79 -0.2 4.5 79 33.9% 0.37 [0.05, 0.68] 2017 Subtotal (95% CI) 161 162 84.0% 0.12 [-0.19, 0.42] 2 2 2 Heterogeneity: tau = 0.03; chi = 3.62, df = 2 (P = 0.16); I = 45% Test for overall effect: Z = 0.75 (P = 0.45) 2.2.2 r-CGM vs. SMBG Ajjan et al. 94.7 3.87 30 96 3.87 15 16.0% -0.33 [-0.95, 0.29] 2016 Subtotal (95% CI) 30 15 16.0% -0.33 [-0.95, 0.29] Heterogeneity: Not applicable Test for overall effect: Z = 1.04 (P = 0.30) Total (95% CI) 191 177 100.0% 0.04 [-0.26, 0.34] 2 2 2 Heterogeneity: tau = 0.04; chi = 5.62, df = 3 (P = 0.13); I = 47% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.28 (P = 0.78) Favors CGM Favors SMBG 2 2 Test for subgroup differences: Chi = 1.59, df = 1 (P = 0.21), I = 37.3% Figure 3: 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. satisfaction was observed between the CGM and SMBG hypoglycaemia in patients with type 2 diabetes mellitus using groups. Two trials utilizing DDS found no significant differ- a meta-analysis of RCTs. Accordingly, our results revealed ences in scores between the CGM and SMBG groups [20]. 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 4. Discussion pressure was observed between the CGM and SMBG groups. In this study, we examined the influence of CGM on blood One 2013 meta-analysis involving four RCTs that collec- glucose levels, weight, blood pressure, and frequency of tively examined the effects of RT-CGM and r-CGM in 6 Journal of Diabetes Research CGM SMBG Std. mean difference Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 4.1.1 r-CGM vs. SMBG Cosson et al. -2 6.41 11 7 35.41 14 9.5% -0.32 [-1.12, 0.47] 2009 Aljan et al. 0.57 0.67 30 0.7 0.67 15 15.5% -0.19 [-0.81, 0.43] 2016 Haak et al. -0.71 1.63 140 -0.09 1.59 75 75.0% -0.38 [-0.67, -0.10] 2017 Subtotal (95% CI) 181 104 100.0% -0.35 [-0.59, -0.10] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.31, df = 2 (P = 0.86); I = 0% Test for overall effect: Z = 2.78 (P = 0.006) Total (95% CI) 181 104 100.0% -0.35 [-0.59, -0.10] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.31, df = 2 (P = 0.86); I = 0% Test for overall effect: Z = 2.78 (P = 0.006) -1 -0.5 0 0.5 1 Test for subgroup differences: Not applicable Favors CGM Favors SMBG Figure 4: 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. CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 6.1.1 r-CGM vs. SMBG Ajjan et al. 9.13 3.66 30 9.55 3.66 15 17.0% -0.11 [-0.73, 0.51] 2016 Haak et al. 1 5.37 140 0.4 6.14 75 83.0% 0.11 [-0.17, 0.39] 2017 Subtotal (95% CI) 170 90 100.0% 0.07 [-0.19, 0.32] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.40, df = 1 (P = 0.53); I = 0% Test for overall effect: Z = 0.53 (P = 0.60) Total (95% CI) 170 90 100.0% 0.07 [-0.19, 0.32] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.40, df = 1 (P = 0.53); I = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.53 (P= 0.60) Favors CGM Favors SMBG Test for subgroup differences: Not applicable Figure 5: 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. Std. mean difference CGM SMBG Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 7.1.1 RT-CGM vs. SMBG Ehrhardt et al. -1.8 18.02 50 -3 20.45 50 54.3% 0.06 [-0.33, 0.45] 2011 Subtotal (95% CI) 50 50 54.3% 0.06 [-0.33, 0.45] Heterogeneity: Not applicable Test for overall effect: Z = 0.31 ( P= 0.76) 7.1.2 r-CGM vs. SMBG Allen et al. -7 16 27 3 15 25 45.7% -0.63 [-1.19, -0.08] 2008 Subtotal (95% CI) 27 25 45.7% -0.63 [-1.19, -0.08] Heterogeneity: Not applicable Test for overall effect: Z = 2.23 (P = 0.03) Total (95% CI) 77 75 100.0% -0.26 [-0.94, 0.42] 2 2 2 Heterogeneity: tau = 0.18; chi = 4.00, df = 1 (P = 0.05); I = 75% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.74 (P = 0.46) Favors CGM Favors SMBG 2 2 Test for subgroup differences: chi = 4.00, df = 1 (P = 0.05), I = 75.0% Figure 6: 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. patients with type 2 diabetes mellitus indicated that the CGM that the CGM group had significantly lower HbA1c levels treatment group had significantly lower HbA1c levels than than the SMBG group. However, when RT-CGM and the SMBG group [11]. Similarly, the present study revealed r-CGM were viewed separately, we found that although the Journal of Diabetes Research 7 CGM SMBG Std. mean differe nce Std. mean difference Study or subgroup Mean SD Total Mean SD Total Weight IV, random, 95% CI Year IV, random, 95% CI 8.1.1 RT-CGM vs. SMBG Ehrhardt et al. -1.3 11.24 50 -1.4 9.99 50 65.9% 0.01 [-0.38, 0.40] 2011 Subtotal (95% CI) 50 50 65.9% 0.01 [-0.38, 0.40] Heterogeneity: Not applicable Test for overall effect: Z = 0.05 (P = 0.96) 8.1.2 r-CGM vs. SMBG Allen et al. -3 11 27 -2 9 25 34.1% -0.10 [-0.64, 0.45] 2008 Subtotal (95% CI) 27 25 34.1% -0.10 [-0.64, 0.45] Heterogeneity: Not applicable Test for overall effect: Z = 0.35 (P = 0.73) Total (95% CI) 77 75 100.0% -0.03 [-0.35, 0.29] 2 2 2 Heterogeneity: tau = 0.00; chi = 0.10, df = 1 (P = 0.75); I = 0% -1 -0.5 0 0.5 1 Test for overall effect: Z = 0.17 (P = 0.87) Favors CGM Favors SMBG 2 2 Test for subgroup differences: chi = 0.10, df = 1 (P = 0.75), I = 0% Figure 7: 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. Table 2: 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 P value Baseline End of study Baseline End of study DTSQ Ajjan et al. [23] — 13.39 — 13.52 —— 0.936 Haak et al. [24] — 13.1 (0.5) — 9.0 (0.7) —— <0.001 DQoL 0.025 Haak et al. [24] ——— — −0.2 (0.0) 0.0 (0.0) DDS Beck et al. [20] 1.9 (0.8) 1.8 (0.9) 2.0 (0.8) 1.8 (0.6) —— — CGM Satisfaction Scale Beck et al. [20] — 4.3 (0.4) —— — — — CGM: continuous glucose monitoring; DTSQ: Diabetes Treatment Satisfaction Questionnaire; DQoL: Diabetes Quality of Life; DDS: Diabetes Distress Scale. P <0 05. RT-CGM group had predominantly lower HbA1c levels than Beck et al. [20] showed that the RT-CGM group tended to the SMBG group, no significant difference in HbA1c levels have greater body weight than the SMBG group, the other had been found between the r-CGM and SMBG groups. three trials [18, 19, 23] showed no change or even a decrease in body weight. The daily amount of insulin administered in According to a systematic review of patients with type 1 dia- betes, RT-CGM has a greater blood glucose-ameliorating Beck et al.’s study increased compared with the baseline. effect than r-CGM [25]. The use of RT-CGM helps patients However, this remained unchanged or decreased in the other not only adjust diabetes medication dosage but also under- three trials. Moreover, Beck et al.’s study revealed that stand changes in blood glucose levels on a monitor and be although patients in the RT-CGM group had improved blood conscious of lifestyle factors, such as meals and exercise, glucose levels because of an increase in snacking as a result of thereby ameliorating blood glucose levels [18, 21, 26]. Con- hypoglycaemia or an increase in insulin levels to correct versely, r-CGM increases physical activity and blood glucose blood glucose levels, an increase in body weight could have amelioration and inhibits the onset of complications [21]. been present. Accordingly, blood glucose management using Nevertheless, further studies are needed to determine CGM in patients with type 2 diabetes mellitus necessitates whether RT-CGM improves HbA1c in patients with type 2 paying close attention to the insulin dose and changes in diabetes mellitus to a greater extent than r-CGM. weight [26]. We showed no difference in body weight change between With regard to influence on hypoglycaemia, we showed the CGM and SMBG groups. However, although the study by that the RT-CGM group spent less time with hypoglycaemia 8 Journal of Diabetes Research levels or a ≥10% improvement from baseline values contrib- than the SMBG group. A previous study examining the utility of CGM for type 1 diabetes observed a shortening in the time utes to the inhibition of future cardiovascular events and has spent with hypoglycaemia because of CGM intervention. In been indicated as clinically significant amelioration [28–30]. Given that hypoglycaemia and blood glucose fluctuations, general, CGM intervention exhibits greater hypoglycaemic effect among patients with high hypoglycaemic frequency at which are believed to be related to various poor outcomes, baseline, such as those with type 1 diabetes [17]. Among could be underestimated in patients in type 2 diabetes melli- the studies included in the present meta-analysis, the time tus [31], understanding detailed blood glucose profiles spent with hypoglycaemia per day at patient baseline ranged through CGM may be useful. In recent years, the increase in healthcare costs has been noted as a problem. Reportedly, from 3 to 60 min, which may be considered relatively short [22–24]. Nevertheless, CGM intervention shortened the time CGM intervention is useful in terms of cost effectiveness in spent with hypoglycaemia, suggesting its practicality for patients with type 1 diabetes [32] and in those with type 2 shortening time spent with hypoglycaemia in patients with diabetes [33, 34], although the number of reports is limited type 2 diabetes mellitus. However, given that RCTs compar- for the latter type of patients. Further investigations are needed on effects of CGM intervention in patients with type ing the RT-CGM and SMBG groups had not been included in the present analysis, further investigation is necessary. 2 diabetes to alleviate complications, to reduce the inci- One study on the effect on blood pressure included dence of cardiovascular disease, and to improve QOL and herein showed that the CGM group had no reduction in sys- cost effectiveness. tolic and diastolic blood pressure compared with the SMBG The present study had several limitations. First, given the few number of RCTs included, the present study might have group. In another study included herein, Allen et al. found that the r-CGM group exhibited lower blood pressure during had insufficient power to detect differences between groups. the collection period than the SMBG group. However, as Second, although previous studies on RT-CGM interventions indicated in a previous study [11], given the inclusion of had indicated that the frequency of CGM sensor use influ- counselling on exercise therapy based on r-CGM data, the ences its effects on HbA1c levels [35], this had not been examined because of a lack of sufficient data. Third, we can- independent impact of r-CGM might have not been observed. However, most of the patients in trials included not deny the possibility that some literature could have been herein had been administered hypotensive medication for missed while searching the databases, which could have influenced the results of the present study. Fourth, the obser- blood pressure management. Accordingly, baseline blood pressure management appeared to be the reason why inter- vation period and evaluation items of each RCT included herein varied greatly. Therefore, it appeared necessary to vention effects of CGM had not been observed. Moreover, assessing the influence of CGM on blood pressure had been pay close attention to the interpretation of the results and generally difficult given the few studies included. generalization. Finally, the quality of RCTs included in the present study was generally low. Moreover, given the pres- Although a meta-analysis regarding treatment satisfac- tion after CGM intervention had not been conducted, the ence of heterogeneity, there could be concern regarding the validity of the results derived from the present study. present study included one trial [20, 24] that indicated increased treatment satisfaction and another [23] in which The present study examined the effects of CGM on blood no change was noted. Accordingly, the shortening of time glucose levels, body weight, blood pressure, and hypoglycae- mia in patients with type 2 diabetes mellitus using a spent with hypoglycaemia has been speculated to be the rea- son for such differences. In a previous study on patients with meta-analysis of RCTs. The results revealed that the CGM group had significantly lower HbA1c levels and shorter time type 1 diabetes, the decrease in hypoglycaemic frequency had been indicated to be closely related to patient satisfaction spent with hypoglycaemia than the SMBG group. On the [27]. In our study, there are similar observations wherein a other hand, no difference in body weight and blood pressure had been observed between the CGM and SMBG groups. As shortening of time spent with hypoglycaemia because of CGM in two trials resulted in increased treatment satisfac- previously mentioned, given the few RCTs included as well as tion, but limited shortening of time spent with hypoglycae- the presence of heterogeneity, care may be needed when mia in one study resulted in unchanged satisfaction. Hence, interpreting the results of the present study. Accordingly, based on the trials involving patients with type 2 diabetes further studies addressing the limitations presented herein may be necessary. mellitus included herein, the shortening of time spent with hypoglycaemia because of CGM intervention may perhaps lead to increased treatment satisfaction. Large-scale clinical trials have shown that strict blood Conflicts of Interest glucose management contributes to the reduction of the risk for vascular complications in patients with type 2 diabetes The authors declare that they have no conflicts of interest. mellitus [3, 4]. However, avoiding the risk of hypoglycaemia and maintaining patient QOL are also extremely important for glucose management. The present meta-analysis showed Acknowledgments that the CGM group exhibited a significantly greater degree of HbA1c reduction (a decrease of approximately 1% from The authors would like to thank the staff members of the the baseline value) and shorter time spent with hypoglycae- Department of Metabolic Diseases at Ise Red Cross Hospital mia than the SMBG group. A ≥0.5% improvement in HbA1c for their cooperation in this study. Journal of Diabetes Research 9 system on glycemic control in type 1 diabetic patients: system- References atic review and meta-analysis of randomized trials,” European [1] World Health Organization (WHO), Global Report on Diabe- Journal of Endocrinology, vol. 166, no. 4, pp. 567–574, 2012. tes, World Health Organisation, Geneva, Switzerland, 2016, [17] J. C. Pickup, “The evidence base for diabetes technology: May 2018, https://www.who.int. appropriate and inappropriate meta-analysis,” Journal of Dia- [2] H. King, R. E. Aubert, and W. H. Herman, “Global burden of betes Science and Technology, vol. 7, no. 6, pp. 1567–1574, diabetes, 1995–2025: prevalence, numerical estimates, and 2013. projections,” Diabetes Care, vol. 21, no. 9, pp. 1414–1431, [18] H. J. Yoo, H. G. An, S. Y. Park et al., “Use of a real time contin- uous glucose monitoring system as a motivational device for poorly controlled type 2 diabetes,” Diabetes Research and [3] UK Prospective Diabetes Study (UKPDS) Group, “Intensive blood-glucose control with sulphonylureas or insulin com- Clinical Practice, vol. 82, no. 1, pp. 73–79, 2008. pared with conventional treatment and risk of complications [19] N. M. Ehrhardt, M. Chellappa, M. S. Walker, S. J. Fonda, and in patients with type 2 diabetes (UKPDS 33),” The Lancet, R. A. Vigersky, “The effect of real-time continuous glucose vol. 352, no. 9131, pp. 837–853, 1998. monitoring on glycemic control in patients with type 2 diabe- tes mellitus,” Journal of Diabetes Science and Technology, [4] F. M. Turnbull, C. Abraira, R. J. Anderson et al., “Intensive glu- vol. 5, no. 3, pp. 668–675, 2011. cose control and macrovascular outcomes in type 2 diabetes,” Diabetologia, vol. 52, no. 11, pp. 2288–2298, 2009. [20] R. W. Beck, T. D. Riddlesworth, K. Ruedy et al., “Continuous glucose monitoring versus usual care in patients with type 2 [5] W. H. Herman and P. Zimmet, “Type 2 diabetes: an epidemic diabetes receiving multiple daily insulin injections: a random- requiring global attention and urgent action,” Diabetes Care, ized trial,” Annals of Internal Medicine, vol. 167, no. 6, vol. 35, no. 5, pp. 943-944, 2012. pp. 365–374, 2017. [6] A. Nanditha, R. C. W. Ma, A. Ramachandran et al., “Diabetes [21] N. A. Allen, J. A. Fain, B. Braun, and S. R. Chipkin, “Continu- in Asia and the pacific: implications for the global epidemic,” ous glucose monitoring counseling improves physical activity Diabetes Care, vol. 39, no. 3, pp. 472–485, 2016. behaviors of individuals with type 2 diabetes: a randomized [7] O. Schnell, H. Alawi, T. Battelino et al., “Self-monitoring of clinical trial,” Diabetes Research and Clinical Practice, vol. 80, blood glucose in type 2 diabetes: recent studies,” Journal of no. 3, pp. 371–379, 2008. Diabetes Science and Technology, vol. 7, no. 2, pp. 478–488, [22] E. Cosson, E. Hamo-Tchatchouang, L. Dufaitre-Patouraux, J. R. Attali, J. Pariès, and P. Schaepelynck-Bélicar, “Multicen- [8] H. P. Chase, L. M. Kim, S. L. Owen et al., “Continuous subcu- tre, randomised, controlled study of the impact of continuous taneous glucose monitoring in children with type 1 diabetes,” sub-cutaneous glucose monitoring (GlucoDay ) on glycaemic Pediatrics, vol. 107, no. 2, pp. 222–226, 2001. control in type 1 and type 2 diabetes patients,” Diabetes & [9] E. Boland, T. Monsod, M. Delucia, C. A. Brandt, S. Fernando, Metabolism, vol. 35, no. 4, pp. 312–318, 2009. and W. V. Tamborlane, “Limitations of conventional methods [23] R. A. Ajjan, K. Abougila, S. Bellary et al., “Sensor and software of self-monitoring of blood glucose: lessons learned from 3 use for the glycaemic management of insulin-treated type 1 days of continuous glucose sensing in pediatric patients with and type 2 diabetes patients,” Diabetes and Vascular Disease type 1 diabetes,” Diabetes Care, vol. 24, no. 11, pp. 1858– Research, vol. 13, no. 3, pp. 211–219, 2016. 1862, 2001. [24] T. Haak, H. Hanaire, R. Ajjan, N. Hermanns, J. P. Riveline, and [10] N. A. Allen, J. A. Fain, B. Braun, and S. R. Chipkin, “Continu- G. Rayman, “Flash glucose-sensing technology as a replace- ous glucose monitoring in non–insulin-using individuals with ment for blood glucose monitoring for the management of type 2 diabetes: acceptability, feasibility, and teaching opportu- insulin-treated type 2 diabetes: a multicenter, open-label ran- nities,” Diabetes Technology & Therapeutics, vol. 11, no. 3, domized controlled trial,” Diabetes Therapy, vol. 8, no. 1, pp. 151–158, 2009. pp. 55–73, 2017. [11] N. Poolsup, N. Suksomboon, and A. Kyaw, “Systematic review [25] M. Langendam, Y. M. Luijf, L. Hooft, J. H. DeVries, A. H. and meta-analysis of the effectiveness of continuous glucose Mudde, and R. J. P. M. Scholten, “Continuous glucose monitoring (CGM) on glucose control in diabetes,” Diabetol- monitoring systems for type 1 diabetes mellitus,” Cochrane ogy & Metabolic Syndrome, vol. 5, no. 1, p. 39, 2013. Database of Systematic Reviews, no. 1, article CD008101, 2012. [12] M. C. Simmonds, J. P. T. Higginsa, L. A. Stewartb, J. F. [26] R. A. Vigersky, S. J. Fonda, M. Chellappa, M. S. Walker, and Tierneyb, M. J. Clarke, and S. G. Thompson, “Meta-analysis N. M. Ehrhardt, “Short- and long-term effects of real-time of individual patient data from randomized trials: a review continuous glucose monitoring in patients with type 2 diabe- of methods used in practice,” Clinical Trials, vol. 2, no. 3, tes,” Diabetes Care, vol. 35, no. 1, pp. 32–38, 2012. pp. 209–217, 2005. [27] R. W. Beck, T. Riddlesworth, K. Ruedy et al., “Effect of contin- [13] D. G. Altman and J. M. Bland, “Detecting skewness from sum- uous glucose monitoring on glycemic control in adults with mary information,” BMJ, vol. 313, no. 7066, p. 1200, 1996. type 1 diabetes using insulin injections: the DIAMOND ran- [14] J. P. T. Higgins and S. Green, Eds., Cochrane Handbook for domized clinical trial,” JAMA, vol. 317, no. 4, pp. 371–378, Systematic Reviews of Interventions, John Wiley & Sons, Ltd, 2011, The Cochrane Collaboration, version 5.1.0. March [28] E. Cummins, P. Royle, A. Snaith et al., “Clinical effectiveness 2018, https://training.cochrane.org/handbook. and cost-effectiveness of continuous subcutaneous insulin [15] J. P. Higgins, S. G. Thompson, J. J. Deeks, and D. G. Altman, infusion for diabetes: systematic review and economic evalua- “Measuring inconsistency in meta-analyses,” BMJ, vol. 327, tion,” Health Technology Assessment, vol. 14, no. 11, 2010. no. 7414, pp. 557–560, 2003. [29] Writing Team for the Diabetes Control and Complications [16] A. Szypowska, A. Ramotowska, K. Dzygalo, and D. Golicki, Trial/Epidemiology of Diabetes Interventions and Complica- “Beneficial effect of real-time continuous glucose monitoring tions, “Effect of intensive therapy on the microvascular 10 Journal of Diabetes Research complications of type 1 diabetes mellitus,” JAMA, vol. 287, no. 19, pp. 2563–2569, 2002. [30] I. M. Stratton, A. I. Adler, H. A. W. Neil et al., “Association of glycaemia with macrovascular and microvascular complica- tions of type 2 diabetes (UKPDS 35): prospective observational study,” BMJ, vol. 321, no. 7258, pp. 405–412, 2000. [31] S. F. E. Praet, R. J. F. Manders, R. C. R. Meex et al., “Glycaemic instability is an underestimated problem in type II diabetes,” Clinical Science, vol. 111, no. 2, pp. 119–126, 2006. [32] W. Wan, M. R. Skandari, A. Minc et al., “Cost-effectiveness of continuous glucose monitoring for adults with type 1 diabetes compared with self-monitoring of blood glucose: the DIA- MOND randomized trial,” Diabetes Care, vol. 41, no. 6, pp. 1227–1234, 2018. [33] J. A. Sierra, M. Shah, M. S. Gill et al., “Clinical and economic benefits of professional CGM among people with type 2 diabetes in the United States: analysis of claims and lab data,” Journal of Medical Economics, vol. 21, no. 3, pp. 225–230, [34] S. J. Fonda, C. Graham, J. Munakata, J. M. Powers, D. Price, and R. A. Vigersky, “The cost-effectiveness of real-time continuous glucose monitoring (RT-CGM) in type 2 diabetes,” Journal of Diabetes Science and Technology, vol. 10, no. 4, pp. 898–904, 2016. [35] T. Battelino, S. Liabat, H. J. Veeze, J. Castañeda, A. Arrieta, and O. Cohen, “Routine use of continuous glucose monitoring in 10 501 people with diabetes mellitus,” Diabetic Medicine, vol. 32, no. 12, pp. 1568–1574, 2015. MEDIATORS of INFLAMMATION The Scientific Gastroenterology Journal of World Journal Research and Practice Diabetes Research Disease Markers Hindawi Hindawi Publishing Corporation Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 http://www www.hindawi.com .hindawi.com V Volume 2018 olume 2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 International Journal of Journal of Immunology Research Endocrinology Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com BioMed PPAR Research Research International Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Journal of Obesity Evidence-Based Journal of Journal of Stem Cells Complementary and Ophthalmology International Alternative Medicine Oncology Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2013 Parkinson’s Disease Computational and Behavioural Mathematical Methods AIDS Oxidative Medicine and in Medicine Neurology Research and Treatment Cellular Longevity Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal

Journal of Diabetes ResearchHindawi Publishing Corporation

Published: Jan 15, 2019

There are no references for this article.