Continuous glucose monitoring umbrella review
Archives of Public Health
Summary
This umbrella review synthesizes evidence from 31 systematic reviews (SRs), of which 20 included meta-analyses, comparing continuous glucose monitoring (CGM) with self-monitoring of blood glucose or usual care (SMBG/UC) in adults with type 2 diabetes mellitus (T2DM). Searches were conducted across eight databases—including MEDLINE, Embase, Cochrane, CINAHL, SCOPUS, Web of Science and Epistemonikos—up to June 28, 2024. SRs were screened and data extracted independently by two investigators, with conflicts resolved by a third. Outcomes of interest included changes in hemoglobin A₁c (HbA₁c), glycemic variability metrics (time-in-range [TIR], time-above-range [TAR], time-below-range [TBR]) and patient-reported outcome measures (PROMs). Risk of bias was assessed using AMSTAR-2 for SRs and RoB-2, NOS or ROBINS-I for primary studies. Overlap in primary studies was quantified using GROOVE, revealing very high overlap (corrected covered area >20%), which prompted primary-study-level meta-analyses. A random-effects model was fitted for mean differences in pre-post changes and SMDs for PROMs, with subgroup and sensitivity analyses by CGM modality, funding source, insulin use and risk of bias. Evidence certainty was graded with GRADE. Meta-analysis of 34 primary studies (n = 11 494) demonstrated that CGM led to a significantly greater reduction in HbA₁c compared to SMBG/UC (mean difference [MD] = –0.40% [95% CI: –0.54 to –0.25], I² = 81%), invariant across CGM modality, funding and insulin use. Fourteen studies (n = 1452) showed a significant TIR increase (MD = 6.00% [95% CI: 3.13 to 8.88], I² = 75%), and 11 studies (n = 1113) showed a significant TAR decrease (MD = –4.33% [95% CI: –8.37 to –0.28], I² = 78%). TBR change was non-significant (MD = –0.33% [95% CI: –0.75 to 0.09], I² = 67%). PROMs (seven domains, 21 instruments, n = 485-938) showed no significant differences (SMD range 0.10–1.32, all p>0.05). Adverse events were limited to self-limited cutaneous reactions (1 event per 8 weeks of wear). GRADE rated evidence as moderate for HbA₁c, TIR and TAR; low for TBR; very low for PROMs. The findings support CGM adoption alongside SMBG in standard T2DM care.
Study Design
Interventions
Study Type
Outcomes
Duration and Size
Study Population
Age Range
Sex
Geography
Other Criteria
Methodology
We performed an umbrella review of systematic reviews (SRs) and meta-analyses (MAs) evaluating continuous glucose monitoring (CGM) versus self-monitoring of blood glucose or usual care (SMBG/UC) in adults with type 2 diabetes mellitus. Eight databases (MEDLINE, Embase, Cochrane, CINAHL, Epistemonikos, SCOPUS, Web of Science, CINAHL) were searched to June 28, 2024. After deduplication, titles/abstracts were screened and full texts assessed independently by two reviewers, with a third adjudicator resolving conflicts. Data were extracted on study characteristics, risk of bias (AMSTAR-2 for SRs; RoB-2, NOS, ROBINS-I for primary studies), and outcomes: HbA₁c, time-in-range (TIR), time-above-range (TAR), time-below-range (TBR), and patient-reported outcome measures (PROMs). Overlap of primary studies across MAs was quantified via GROOVE (corrected covered area). High overlap (>20%) triggered primary-study-level meta-analyses. Random-effects models pooled mean differences for pre-post changes and standardized mean differences (SMDs) for PROMs, with subgroup and sensitivity analyses by CGM modality, funding source, insulin use, and risk of bias. Publication bias was assessed by funnel plots and Egger’s test. Evidence certainty was graded using GRADE.
Interventions
The intervention of interest was continuous glucose monitoring (CGM), encompassing real-time CGM, flash/intermittently scanned CGM, retrospective CGM and professional CGM devices worn subcutaneously to sample interstitial glucose continuously over days to weeks. CGM was compared against self-monitoring of blood glucose (SMBG), which involves finger-prick capillary glucose measurements at discrete time points, or usual care without continuous monitoring. CGM enables detailed glycemic profiling—time-in-range, glycemic variability metrics—and supports behavioral adjustments, whereas SMBG provides intermittent glucose snapshots with potential barriers related to pain and inconvenience.
Key Findings
Meta-analysis of 34 primary studies (n = 11 494) found CGM use was associated with a mean HbA₁c reduction of 0.40% (95% CI: –0.54 to –0.25; I² = 81%) compared to SMBG/UC, invariant across CGM modality, funding and insulin therapy. Fourteen studies (n = 1452) demonstrated a 6.00 percentage-point increase in time-in-range (TIR; 95% CI: 3.13 to 8.88; I² = 75%) and eleven studies (n = 1113) demonstrated a 4.33 percentage-point decrease in time-above-range (TAR; 95% CI: –8.37 to –0.28; I² = 78%). Time-below-range (TBR) change was non-significant (–0.33%; 95% CI: –0.75 to 0.09). PROMs across seven domains showed no significant differences (SMD range 0.10–1.32). Adverse events were limited to mild cutaneous reactions. GRADE rated evidence moderate for HbA₁c, TIR and TAR; low for TBR, very low for PROMs.
Comparison with other Studies
While numerous systematic reviews and meta-analyses have previously examined CGM’s impact on glycemic control in type 2 diabetes, heterogeneity in inclusion criteria, CGM modality (e.g., flash vs. real-time CGM), and outcome metrics has limited consensus. Earlier SRs focusing solely on flash glucose monitoring reported mixed results in HbA₁c reduction and insufficient data on patient-reported outcomes, whereas those targeting real-time CGM demonstrated more consistent glycemic benefits. Our umbrella review, the most contemporaneous and comprehensive to date, integrates findings from 31 SRs up to mid-2024 and addresses overlap bias through GROOVE analysis. Unlike prior overviews that often pooled endpoint HbA₁c differences only, we conducted rigorous pre-post meta-analyses of primary trials covering a broad range of CGM technologies and patient subgroups, including both insulin-treated and non–insulin-treated cohorts. Our effect estimates for HbA₁c (MD –0.40%) and TIR (+6.00%) exceed minimal clinically important differences, reinforcing earlier RCT-level evidence while providing stronger certainty through GRADE. Furthermore, by systematically quantifying adverse events and PROMs, we clarify that CGM’s glycemic benefits do not extend to measurable improvements in patient-reported quality of life or treatment satisfaction—an area requiring further qualitative and mixed-methods research. This contrasts with smaller cohort studies suggesting usability gains, highlighting the need for standardized PROM instruments in future CGM trials.
Journal Reference
Tan YY, Suan E, Koh GCH, Suhairi SB, Tyagi S. Effectiveness of continuous glucose monitoring in patient management of type 2 diabetes mellitus: an umbrella review of systematic reviews from 2011 to 2024. Arch Public Health. 2024;82:231. doi:10.1186/s13690-024-01459-2
© 2025 deDiabetes. Licensed under CC BY-NC (Attribution-NonCommercial)
Stay informed. Stay ahead.
Subscribe now for the latest breakthroughs, expert insights, and cutting-edge updates in diabetes care—delivered straight to your inbox.