Differences in Continuous Glucose Monitoring Metrics Between Prediabetes and Normoglycemia: A Systematic Review and Meta-Analysis

Abstract

Background:

Prediabetes often remains undiagnosed until it progresses to type 2 diabetes mellitus (T2DM), particularly among individuals with obesity or a family history of diabetes. Conventional tests, including fasting plasma glucose, hemoglobin A1c, and the oral glucose tolerance test, provide only static snapshots of glycemia and may fail to capture variability and postprandial excursions. Continuous glucose monitoring (CGM) offers dynamic insight into glucose regulation and may complement traditional diagnostic tools.

Methods:

This systematic review and meta-analysis evaluated differences in CGM-derived metrics between individuals with prediabetes and those with normoglycemia. PubMed, EMBASE, and CENTRAL were searched from inception to September 3, 2025 (PROSPERO: CRD42024608658). The primary outcome was the mean amplitude of glycemic excursions (MAGE); secondary outcomes included time above range, 24-hour mean glucose, coefficient of variation, and time in range. Sixteen studies met the inclusion criteria, and ten were included in the quantitative synthesis (n = 1657).

Results:

Prediabetes was consistently associated with higher CGM values compared with normoglycemia: MAGE (mean difference [MD] = 9.41 mg/dL, 95% confidence interval [CI]: 4.31, 15.31), TAR% (MD = 5.68%, 95% CI: 1.04, 10.32), and 24-hour mean glucose (MD = 7.91 mg/dL, CI: 6.27, 9.55).

Conclusions:

These results provide the first quantitative evidence that CGM can discriminate between prediabetes and normoglycemia, supporting its potential as a complementary tool for refined metabolic risk assessment. Further prospective studies are needed to determine its predictive value for progression to T2DM.

Keywords

CGM HbA1c normoglycemia prediabetes

Introduction

Type 2 diabetes mellitus (T2DM) is a global health problem, affecting more than 462 million people worldwide, and is projected to increase to 643 million.¹ Prediabetes, an intermediate stage between normoglycemia and T2DM, is defined as elevated blood glucose levels below the diagnostic threshold for diabetes and represents a critical opportunity for intervention to prevent disease progression.² However, it frequently remains undiagnosed, as approximately one in three adults in the United States has prediabetes, and more than 80% are unaware of their condition.³ This is mainly due to the asymptomatic nature of prediabetes and the limitations of traditional glucose measurements.

Standard methods, such as fasting plasma glucose (FPG), hemoglobin A1c (HbA1c), and the oral glucose tolerance test (OGTT) do not adequately capture glucose fluctuations. HbA1c levels can be affected by factors, such as red blood cell turnover, anemia, and pregnancy,⁴ while FPG does not account for postprandial glucose changes. In addition, the OGTT can be stressful and time-consuming, and may miss peak glucose levels.⁵ These limitations may contribute to the underdiagnosis of prediabetes and delay intervention, thereby increasing the risk of developing diabetes and its associated complications, including kidney failure, cardiovascular disease, neuropathy, and vision loss.²

Continuous glucose monitoring (CGM) offers a more comprehensive approach by providing real-time, dynamic insight into glucose variability. Since 2012, the American Diabetes Association (ADA) has recommended CGM for individuals with type 1 diabetes mellitus (T1DM), and more recently, its use has been extended to people with T2DM, including those not treated with insulin.⁶ These recommendations currently apply to individuals with diabetes, in whom CGM has demonstrated benefits in reducing hypoglycemic events and improving glycemic control.⁷

Emerging evidence suggests that CGM metrics, such as 24-hour mean glucose and mean amplitude of glycemic excursions (MAGE), also differ between individuals with prediabetes and those with normoglycemia.⁸ By continuously tracking glucose variability and postprandial responses, CGM provides detailed information on glycemic regulation that complements conventional diagnostic methods.⁹ Despite growing interest in CGM beyond diabetes management, no previous meta-analysis has quantitatively synthesized evidence comparing CGM metrics between prediabetes and normoglycemia. Understanding these differences could help bridge the gap between physiological glucose dynamics and conventional clinical screening.

This study aimed to systematically evaluate whether CGM metrics differ between individuals with prediabetes and those with normoglycemia, thereby providing evidence to support the potential role of CGM in individuals without diabetes.

Methods

Protocol

This systematic review and meta-analysis adhered to the Cochrane Handbook¹⁰ and the 2020 PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines¹¹ (Supplementary Table S1). The protocol was registered in PROSPERO (CRD42024608658). Initially designed to assess dysglycemia, the scope was refined before data extraction to focus specifically on comparisons between prediabetes and normoglycemia, thereby reducing heterogeneity. Predefined outcomes included MAGE, time above range (TAR), 24-hour mean glucose, time in range (TIR), and coefficient of variation (CV). In this study, TIR refers to the percentage of time spent between 70 and 140 mg/dL, a range commonly applied in populations without diabetes, and often referred to as time in tight range (TITR) in the literature. The study was conducted within the Systems Education model, coordinated by the Centre for Translational Medicine at Semmelweis University.¹²

Eligibility Criteria

Observational studies of individuals without diabetes were eligible if CGM metrics were reported separately for prediabetes (HbA1c = 5.7%-6.5%, fasting glucose 100-125 mg/dL, or 2-hour OGTT = 140-199 mg/dL) and normoglycemia. Studies, including diabetes, severe comorbidities, pregnancy, or glucose-lowering therapy were excluded. Only full-text original articles were included.

Search Strategy, Data Extraction, and Management

PubMed, EMBASE, and CENTRAL were searched from inception to September 3, 2025, without language restrictions. Full strategies are provided in Supplementary Document S1. Additional studies were identified through reference screening and citation chasing.¹³ After duplicate removal, four reviewers screened titles, abstracts, and full texts; discrepancies were resolved by consensus, with a κ value greater than 0.75 at all stages.

Records were managed in EndNote 21 and Microsoft Excel.^14,15 Extracted data included study characteristics, demographics, and CGM metrics (MAGE, TAR, TIR, CV, and mean glucose).

When impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) were reported separately, they were combined as “prediabetes.” Missing standard deviations (SDs) were calculated from SEs, CIs, or P values; medians were converted to mean using the methods described by Luo et al¹⁶ and Shi et al.¹⁷ Abnormal CGM thresholds were defined from prior studies.^18-20

Risk of Bias Assessment

The risk of bias was assessed using the ROBINS-I²¹ across seven domains; any disagreements were resolved through discussion.

Statistical Analysis

Effect sizes were expressed as mean differences (MDs) with 95% confidence intervals (CIs). For each study, sample size, mean, and SD were extracted or derived separately for prediabetes and normoglycemia groups, and MDs were calculated as the difference between group means. Pooled estimates were generated using inverse-variance random-effects models when at least two studies contributed data. Between-study variance (τ²) was estimated using the restricted maximum-likelihood method, and heterogeneity was quantified with I². Confidence intervals for τ² were obtained using the Q-profile approach. Subgroup analyses were performed by device type, ethnicity, and monitoring duration, and sensitivity analyses excluded high-risk studies. Additional subset analyses were conducted, including only studies published from 2019 onward. All analyses were performed in R (v4.3) using the meta and dmetar packages.^22-25

Results

Search and Selection

We identified a total of 21 764 studies. After duplicate removal and eligibility screening, 12 studies remained. An additional four studies were included after citation tracking. In total, 16 studies were selected for the systematic review^26,27 (Table S4), with ten of these studies included in the meta-analysis. The study selection process is presented in Figure 1.

Figure 1.

PRISMA flowchart of the article selection process.

Study Characteristics

Sixteen studies (2010-2024) were included, totaling 1657 participants in the quantitative synthesis. Most used prospective cross-sectional designs with CGM over defined monitoring periods were conducted in the United States, Japan, Spain, China, Greece, and Sri Lanka. All studies used ADA/European Association for the Study of Diabetes-approved devices (Abbott Freestyle Libre, Dexcom G6/G7, Medtronic Guardian) (Table 1).

Table 1.

Characteristics of the Included Studies and Participants.

Author	Study site	Total number of patients (female %)	Age in years (SD or min-max)^a	Prediabetes group number of patients (female %)	Prediabetes group BMI (SD)	Normoglycemia group number of patients (female %)	Normoglycemia group BMI	Calibration method (F/N)^b	MARD value range^c	High-risk cohort (Y/N)^d	CGM mode (blinded/ unblinded)^e
Rizos et al²⁶	Greece	36 (61.1)	51 (31.3-76.3)	18 (61.1)	27.34 (3.8)	18 (61.1)	21.13 (3.2)	N	9.1-16.3	N	U
Spartano et al²⁷	USA	1023 (64.5)	58.5 (NA)	463 (45.8)	29.3 (5.1)	560 (64.1)	26.5 (4.8)	F	8.2-10.1	N	B
Kashiwagi et al²⁸	Japan	40 (50)	40.3 (NA)	14 (43)	21.5 (2.1)	26 (54)	21.1 (2.8)	F	7.8-9.2	N	NA
Dorcely et al²⁹	USA	15 (20)	50 (14)	8 (0)	NA	7 (43)	31.14 (5.7)	F	7.8-9.2	Y	U
Rodriguez-Segade et al³⁰	Spain	322 (57.5)	44 (13)	68 (57.4)	NA	254 (57.5)	26.3 (4.2)	N	9.1-16.3	N	U
Hanefeld et al³¹	USA	80 (34)	NA	42 (21.42)	28.87 (4.12)	28 (53)	28.35 (3.47)	N	9.1-16.3	N	NA
Madhu et al³²	Sri Lanka	60 (65)	46.3 (8.8)	20 (65)	28.92 (2.42)	20 (65)	29.01 (4.2)	N	9.1-16.3	Y	U
Xu et al¹⁹	China	40 (55)	38 (8)	29 (55)	NA	11 (NA)	NA	N	9.1-16.3	N	NA
Nomura et al³³	Japan	29 (14)	38 (6.9)	7 (14)	NA	22 (14)	25 (2.3)	N	9.1-16.3	N	NA
Chakarova et al³⁴	Bulgaria	42 (NA)	NA (NA)	NA	30,3 (5.3)	18 (77.7)	24,8 (6.9)	F	7.8-9.2	Y	B
DuBose et al³⁵	USA	39 (51)	4,4 (1.4)	NA	NA	NA	NA	F	8.2-10.1	N	NA
Sundberg et al²⁰	Sweden	15 (66.6)	5.4 (1.6)	NA	NA	NA	NA	F	8.2-10.1	Y	NA
Fox et al³⁶	USA	74 (69)	(9-65)	NA	NA	NA	NA	F, N	8.2-16.3	Y	B
Shah et al³⁷	USA	153 (101)	31 (21)	NA	NA	NA	NA	F	8.2-10.1	Y	B
Shilo et al³⁸	Israel	8315 (54.4)	51.3 (7.58)	NA	NA	NA	NA	F	7.8-9.2	N	NA
Borg et al¹⁸	USA, Europe, Africa	80 (69)	41 (13.8)	NA	NA	NA	NA	N	9.1-16.3	N	NA

Abbreviations: BMI: body mass index; CGM: continuous glucose monitor; MARD: mean absolute relative difference.

Parameters are presented as mean ± standard deviation, or median with minimum-maximum values.

F = factory-calibrated CGM, N = manually calibrated CGM.

The MARD values were not reported in the listed studies but were added based on validation and pivotal trials of the respective CGM devices.^39-41

High-risk (risk-enriched) cohort = intentional recruitment based on baseline risk: family history of diabetes (≥1 first-degree relative) and/or obesity (BMI ≥30 kg/m²).

B = blinded (professional CGM; participants cannot view readings), U = unblinded (real-time CGM), NA = not available.

Mage

Eight studies (n = 602) showed significantly higher MAGE in prediabetes compared with normoglycemia, with a pooled MD of 9.81 mg/dL (95% CI: 4.31, 15.31) (Figure 2a). Although heterogeneity was high (I² = 77%), all studies reported elevated MAGE in prediabetes.

Figure 2.

Mean difference in (a) MAGE (mg/dL, n = 602), (b) TAR (%, n = 1484), and (c) 24-hour CGM mean (mg/dL, n = 1606) between individuals with prediabetes and normoglycemia.

TAR (>140 mg/dL)

Five studies (n= 1484) demonstrated higher TAR in prediabetes (MD = 5.68%, 95% CI: 1.04, 10.32; I² = 90%) (Figure 2b), indicating longer exposure to hyperglycemia.

Twenty-Four-Hour CGM Mean

Eight studies (n = 1606) showed significantly higher 24-hour mean glucose levels in prediabetes (MD = 7.91 mg/dL, 95% CI: 6.27-9.55; I² = 10%) (Figure 2c).

CV%

Six studies (n = 557) revealed higher glycemic variability in prediabetes, with a pooled MD in CV of 1.54% (95% CI: 0.67, 2.41; I² = 0%) (Figure 3a).

Figure 3.

Mean difference in (a) CV (%, n = 557) and (b) TIR (%, n = 1421) between individuals with prediabetes and normoglycemia.

TIR (70-140 mg/dL)

Four studies (n = 1421) indicated a non-significant trend toward lower TIR in prediabetes (MD = −4.75%, 95% CI: –10.47, 0.97; I² = 89%) (Figure 3b).

Risk of Bias Assessment

Using ROBINS-I,²² three studies were rated as low risk, six as moderate, and one as serious, primarily due to participant selection and intervention classification. Excluding this single high-risk study²⁹ in sensitivity analysis did not meaningfully alter the pooled estimates, indicating robustness of the results. Variability in CGM duration (3-14 days) may also have contributed to between-study differences (Table S2).

Publication Bias and Heterogeneity

Publication bias could not be assessed because fewer than ten studies were included. Heterogeneity was moderate to high in several studies. To explore potential sources, we conducted subgroup analyses (Table S5, Figures S6-S10).

A significant difference between subgroups accompanied by a substantial reduction in heterogeneity would indicate a true subgroup effect; however, our analyses did not confirm such differences. For MAGE, neither device type (P = .816; residual heterogeneity I² = 79%, τ² = 37.75; Figure S7) nor race (P = .660; residual I² = 74%, τ² = 34.34; Figure S8) explained the observed heterogeneity. In contrast, monitoring duration showed a significant subgroup effect: studies with <7 days of wear had a significantly larger pooled MD (14.01 mg/dL) than those with ≥7 days (5.80 mg/dL: P = .034; residual I² = 51%, τ² = 9.35; Figure S6).

For 24-hour mean glucose, neither monitoring duration (P = .628; residual I² = 22%, τ² = 1.77) nor device type (P= .456; residual I² = 16%, τ² = 3.01) materially altered pooled effects (factory-calibrated MD = 6.64 [95% CI: 2.35, 10.93], manually calibrated MD = 8.66 [3.49 13.83], other MD = 8.60 [6.99, 10.21]; Figures S9 and S10).

Leave-one-out analyses (Figures S1-S5) showed that blinded CGM studies contributed most to heterogeneity. Excluding Spartano et al²⁷ reduced heterogeneity in TIR without affecting the direction or magnitude of the pooled effect, suggesting that CGM mode (blinded vs unblinded) influences variability through behavioral responses to real-time feedback (Figure S4). However, this pattern was not replicated in another blinded study³⁴ (Figures S11-S13).

Because standalone CGM systems were approved after 2018, additional subset analyses were performed, restricting the meta-analyses to studies published from 2019 onward (Figures S14-S17). Pooled analyses were feasible for MAGE, 24-hour mean glucose, CV, and TIR. The pooled effect for 24-hour mean glucose remained statistically significant and directionally consistent with the primary analysis (MD = 7.84 mg/dL; 95% CI: 5.51, 10.17), with minimal heterogeneity. For MAGE, CV, and TIR, effect directions were consistent with the main analyses, although statistical significance was attenuated, likely due to reduced sample size and consequent loss of statistical power in the restricted data set.

Sensitivity analyses excluding high-risk studies resulted in minor changes in pooled estimates and heterogeneity (Figures S11-S13). The certainty of evidence, assessed using the GRADE approach, is summarized in Table S3.

Discussion

Key Findings and Interpretation

This meta-analysis identified consistent differences in CGM metrics between individuals with prediabetes and those with normoglycemia. It provides the first quantitative synthesis, showing that dynamic CGM measures (MAGE, TAR, CV, and mean glucose) differ significantly between groups, offering additional insights into glucose regulation across the prediabetes spectrum. These findings support the integration of CGM metrics into refined metabolic risk assessment frameworks and motivate further prospective validation in populations without diabetes.

Clinical Relevance and Risk Stratification

Prediabetes often remains unrecognized,⁴² yet timely identification is crucial since lifestyle modification can prevent up to 58% of diabetes cases.⁴³ Earlier recognition of glycemic instability reduces healthcare burden and improves long-term outcomes.^44,45 Our findings confirm that individuals with prediabetes have significantly higher glycemic variability and mean glucose than those with normoglycemia, supporting CGM as a complementary tool for refined risk stratification. Madhu et al³² similarly reported prolonged hyperglycemic periods in individuals with IGT, reinforcing these associations across diverse cohorts.

Continuous glucose monitoring complements FPG, HbA1c, and OGTT, which, although diagnostic, provide only static measurements. Its continuous nature enables detection of postprandial fluctuations and variability not visible in single-point tests, making it particularly informative for individuals with borderline or discordant laboratory results.

Barriers and Gaps in Glucose Screening

Despite the growing burden of diabetes, screening remains suboptimal. In Europe, nearly one-third of diabetes cases are undiagnosed,⁴⁶ and only 55.6% of individuals aged ≥15 years report blood glucose testing in the past year.⁴⁷ Conventional tests, such as the OGTT, are time-consuming and may miss transient glycemic excursions. Studies show that even normoglycemic individuals experience brief postprandial spikes above 140 mg/dL.^16,29 Dorcely et al²⁹ observed that more than half of participants with HbA1c < 5.7% had elevated 1-hour plasma glucose and higher CGM variability. These findings highlight that static tests may underestimate subtle metabolic disturbances that CGM can detect over several days of wear.

Technological and Physiological Insight

Advances in CGM technology, including factory calibration, improved sensor accuracy, and smartphone integration, have enhanced usability and accessibility.⁴⁸ Factory-calibrated devices, such as the Freestyle Libre and Dexcom G6, eliminate the need for fingerstick calibration, thereby increasing adherence.⁴⁹ Continuous glucose monitoring is now integral to the care of T1DM and T2DM⁵⁰ and is increasingly used in individuals without diabetes, where variability reflects early metabolic risk.^51-53 Although T2DM was once uncommon in children, its prevalence has increased.⁵⁴ Continuous glucose monitoring studies in healthy children revealed brief glucose excursions above 140 mg/dL and below 70 mg/dL despite normal averages, emphasizing physiological.^18,35 In older adults without diabetes, more than 90% of glucose values fall between 71 and 120 mg/dL, and excursions above 140 mg/dL occur only 0.4% of the time.³⁵ A multicenter study reported a median TIR of 96% (70-140 mg/dL), providing reference data across various age groups.³⁷

In a cohort of 8315 individuals without diabetes, fasting glucose fluctuated daily by 7.5 mg/dL, and 40% were reclassified when multiple sequential CGM readings were analyzed,³⁸ underscoring the limits of single fasting measurements.

Ethnic and regional factors may influence glucose metabolism and the interpretation of CGM results.⁵⁵ Subgroup analyses revealed no significant differences in pooled effect estimates between racial groups, though data were limited. Notably, CGM use remains lower among Black, Hispanic, and other marginalized populations despite socioeconomic adjustment,⁵⁶ warranting further equity-focused research.

Limitations of Oral Glucose Tolerance Test and Single-Time Measurements

During a 75-g OGTT, glucose may exceed 140 mg/dL within two hours, even in individuals with normal tolerance, while baseline and two-hour readings may miss transient peaks.⁵⁷ Dorcely et al²⁹ found that over half of participants with HbA1c < 5.7% had elevated 1-hour glucose and higher CGM means, indicating early dysglycemia. Similarly, the A1C-derived average glucose study reported that 93% of normoglycemic individuals exceeded 140 mg/dL for a median of 26 min/day.¹⁶ These data illustrate that OGTT and other single-time tests may underestimate short-term fluctuations, whereas CGM provides continuous characterization of variability.

Strengths, Limitations, and Sources of Bias

This meta-analysis, including 1657 participants across diverse settings, applied standardized CGM definitions and showed consistent differences in mean glucose, MAGE, CV, and TAR between prediabetes and normoglycemia. The reproducibility and alignment of direction and magnitude strengthen confidence in its validity.

However, CGM systems (Abbott, Medtronic, Dexcom) varied in calibration, lag time, and accuracy, introducing heterogeneity.⁵⁸ Monitoring duration (3-14 days) and dietary standardization also differed across studies. Importantly, there is currently no established consensus on the optimal CGM wear duration required to reliably estimate glycemic variability and related CGM metrics in individuals with prediabetes. Our subgroup analyses indicated that monitoring duration influenced effect estimates for certain outcomes, suggesting that variability in sampling length may introduce uncertainty and limit direct comparability across studies. Some studies used unblinded CGM, allowing real-time feedback that may have prompted behavioral changes (Hawthorne effect) and reduced TAR.^59-62 Future trials should employ blinded designs or short wash-in periods to minimize such bias.

Prediabetes definitions varied across the included studies, with some relying solely on HbA1c and FPG, whereas others incorporated OGTT-based criteria. Individuals with isolated IFG and isolated IGT represent distinct metabolic phenotypes with different underlying pathophysiology. The absence of systematic OGTT testing in some cohorts may therefore have led to misclassification, whereby a proportion of participants categorized as normoglycemic may in fact have had undetected prediabetes or early dysglycemia. This heterogeneity in diagnostic criteria could have attenuated observed effect sizes and contributed to between-study variability.

Formal publication bias testing was not feasible due to the limited number of studies per outcome (<10). Funnel plot asymmetry suggested possible small-study effects and selective reporting. Moreover, limited access and cost constrain broader CGM use in many regions.⁶³

Importantly, subset analyses restricted to studies published from 2019 onward yielded directionally consistent results across the main CGM metrics, indicating that the observed differences between prediabetes and normoglycemia are not driven by earlier CGM technologies. Despite these factors, all analyses consistently demonstrated higher CGM metrics in individuals with prediabetes, supporting the robustness of these findings.

Implications for Practice and Future Research

Translating these findings into clinical and public health practice is essential for maximizing CGM’s potential.^64,65 CGM captures distinct glycemic patterns differentiating prediabetes from normoglycemia, offering additional context to borderline or discordant HbA1c, FPG, or OGTT results. Such insight may support personalized risk assessment and targeted lifestyle modification. Before integration into screening frameworks, standardized metrics and diagnostic thresholds should be established in prospective cohorts. Available evidence suggests that in many settings, a 10- to 14-day monitoring period with ≥70% data completeness provides more stable estimates of key CGM metrics, whereas a shorter wear period may increase imprecision.^49,66,67 Age and body mass index (BMI) modify postprandial glycemia; thus, stratified reporting is warranted.^68,69,28

Prediabetes represents a progressive metabolic continuum rather than a single homogeneous state, spanning early dysglycemia to stages approaching overt T2DM. Emerging staging frameworks for T2DM further emphasize this concept.⁷⁰ CGM metrics may therefore differ between early and more advanced prediabetes phenotypes, reflecting differences in insulin sensitivity and β-cell function. Although the present meta-analysis could not stratify participants by prediabetes stage due to limitations in the available literature, future studies integrating CGM metrics with clinical staging may enable refined risk stratification and earlier identification of individuals at highest progression risk.

Beyond conventional TAR-based metrics, future CGM research should increasingly focus on real-life glucose kinetics, including the duration and temporal patterns of postprandial hyperglycemia. Quantifying how long glucose values remain above clinically relevant thresholds (eg, >140, >180, and >200 mg/dL) after meals may provide complementary insights into metabolic flexibility, insulin sensitivity, and early dysglycemia in both healthy individuals and those with prediabetes. Such dynamic metrics may improve the physiological interpretation of CGM data and enhance early risk stratification beyond static summary measures.

Expanding the use of CGM in populations without diabetes will require education and interdisciplinary collaboration. Training programs for patients and healthcare providers should emphasize interpretation and behavioral guidance based on variability data.⁷¹ Policymakers and funding agencies should support further research to establish the role of CGM in early metabolic risk assessment and precision prevention.

Conclusion

This meta-analysis provides the first quantitative synthesis showing that CGM metrics consistently differ between individuals with prediabetes and those with normoglycemia. These results indicate that CGM captures dynamic aspects of glucose regulation that complement conventional static measures, supporting its potential use in refined metabolic risk assessment. Further prospective studies are warranted to define standardized CGM thresholds and evaluate their contribution to future precision-screening and preventive care strategies.

Supplemental Material

sj-docx-1-dst-10.1177_19322968261426384 – Supplemental material for Differences in Continuous Glucose Monitoring Metrics Between Prediabetes and Normoglycemia: A Systematic Review and Meta-Analysis

Supplemental material, sj-docx-1-dst-10.1177_19322968261426384 for Differences in Continuous Glucose Monitoring Metrics Between Prediabetes and Normoglycemia: A Systematic Review and Meta-Analysis by Gabriella Rákóczi, Judit Nagy, Shaghayegh Jozaee, Shirin Jozaee, Boglárka Lilla Szentes, Jimin Lee, Anett Rancz, Péter Hegyi, Gergely Agócs and Emese Sipter in Journal of Diabetes Science and Technology

Footnotes

Acknowledgements

None to declare

Abbreviations

ADA, American Diabetes Association; BMI, body mass index; CGM, continuous glucose monitoring; CI, confidence interval; CV, coefficient of variation; EASD, European Association for the Study of Diabetes; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; MAGE, mean amplitude of glycemic excursions; MARD, mean absolute relative difference; MD, mean difference; OGTT, oral glucose tolerance test; SD, standard deviation; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus; TAR, time above range; TITR, time in tight range; TIR, time in range.

Author Contributions

G.R. contributed to conceptualization, data curation, investigation, project administration, and writing—original draft; J.N. contributed to conceptualization, data curation, investigation, and writing—review and editing; Sha.J. contributed to data curation, investigation, and writing—review and editing; Shi.J. contributed to data curation, investigation, and writing—review and editing; G.A. contributed to formal analysis, investigation, validation, visualization, and writing—review and editing; B.L.Sz. contributed to formal analysis, investigation, validation, and visualization J.L. contributed to conceptualization, methodology, supervision, and writing—review and editing; A.R. contributed to conceptualization, supervision, writing—review and editing; P.H. contributed to conceptualization, methodology, supervision, and writing—review and editing; E.S. contributed to conceptualization, supervision, and writing—review and editing.

All authors certify that they have participated sufficiently in the work to take public responsibility for the content, including participation in the concept, design, analysis, writing, or revision of the manuscript.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding was provided by the Centre for Translational Medicine, Semmelweis University, Hungarian Ministry of Innovation and Technology, National Research, Development, and Innovation Fund (TKP2021-EGA-23 to Péter Hegyi), and NKFIH project grants K131996 and K147265 (to Péter Hegyi).

Ethical Approval

No ethical approval was required for this systematic review with meta-analysis, as all data were already published in peer-reviewed journals. No patients were involved in the design, conducting or interpretation of our study.

ORCID iD

Gabriella Rákóczi

Data Availability Statement

The data sets used in this study are available in the full-text articles included in the systematic review and meta-analysis.

Supplemental Material

Supplemental material for this article is available online.

References

Khan

MAB

Hashim

King

Govender

Mustafa

Al Kaabi

. Epidemiology of type 2 diabetes—global burden of disease and forecasted trends. J Epidemiol Glob Health. 2019;10(1):107. doi:10.2991/jegh.k.191028.001.

Zhang

Sun

Zhai

Early detection of type 2 diabetes risk: limitations of current diagnostic criteria. Front Endocrinol (Lausanne). 2023;14:1260623. doi:10.3389/fendo.2023.1260623.

Wang

Weng

Kraschnewski

Pichardo-Lowden

AR.

Prediabetes awareness barriers and enablers: national health and nutrition examination survey 2011–2020. AJPM Focus. 2025;4(5):100378. doi:10.1016/j.focus.2025.100378.

Abass

A-E

Musa

Rayis

Adam

Gasim

GI.

Glycated hemoglobin and red blood cell indices in non-diabetic pregnant women. Clin Pract. 2017;7(4):999. doi:10.4081/cp.2017.999.

Kuzuya

Nakagawa

Satoh

, et al. Report of the Committee on the classification and diagnostic criteria of diabetes mellitus. Diabetes Res Clin Pract. 2002;55(1):65-85. doi:10.1016/S0168-8227(01)00365-5.

American Diabetes Association Professional Practice Committee; ElSayed

McCoy

Aleppo

, et al. Introduction and methodology: standards of care in diabetes—2025. Diabetes Care. 2025;48(Suppl 1):S1-S5. doi:10.2337/dc25-SINT.

Didyuk

Econom

Guardia

Livingston

Klueh

Continuous glucose monitoring devices: past, present, and future focus on the history and evolution of technological innovation. J Diabetes Sci Technol. 2021;15(3):676-683. doi:10.1177/1932296819899394.

Wang

Yang

, et al. Glucose fluctuations in subjects with normal glucose tolerance, impaired glucose regulation and newly diagnosed type 2 diabetes mellitus. Clin Endocrinol (Oxf). 2012;76(6):810-815. doi:10.1111/j.1365-2265.2011.04205.x.

Berkovic

Strollo

Ceriello

Is now the time to review the diabetes diagnostic criteria. Diabetes Metab Res Rev. 2025;41(4):e70046. doi:10.1002/dmrr.70046.

10.

Chandler

Hopewell

Cochrane methods—twenty years experience in developing systematic review methods. Syst Rev. 2013;2(1):76. doi:10.1186/2046-4053-2-76.

11.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71.

12.

Hegyi

Varró

Systems education can train the next generation of scientists and clinicians. Nat Med. 2024;30(12):3399-3400. doi:10.1038/s41591-024-03315-w.

13.

Citation Chaser. Citation Chaser tool. https://estech.shinyapps.io/citationchaser/. Published 2025. Accessed September 7, 2025.

14.

Clarivate. EndNote download page. https://support.clarivate.com/Endnote/s/article/Download-EndNote?language=en_US. Published 2025. Accessed September 5, 2025.

15.

Microsoft Corporation. Microsoft Office/Office 365; Microsoft Global Headquarters in Redmond. Washington, DC: Microsoft Office Locations Page. https://www.microsoft.com/en-us/about/office-locations. Published 2025. Accessed September 7, 2025.

16.

Luo

Wan

Liu

Tong

Optimally estimating the sample mean from the sample size, median, mid-range, and/or mid-quartile range. Stat Methods Med Res. 2018;27(6):1785-1805. doi:10.1177/0962280216669183.

17.

Shi

Luo

Weng

, et al. Optimally estimating the sample standard deviation from the five-number summary. Res Synth Methods. 2020;11(5):641-654. doi:10.1002/jrsm.1429.

18.

Borg

Kuenen

Carstensen

, et al.; ADAG Study Group. Real-life glycaemic profiles in nondiabetic individuals with low fasting glucose and normal HbA1c: the A1C-Derived Average Glucose (ADAG) study. Diabetologia. 2010;53(8):1608-1611. doi:10.1007/s00125-010-1741-9.

19.

Zhu

Yan

, et al. Characteristics of real-life glucose profiles monitored by continuous glucose monitoring system in persons with normal glucose tolerance. Zhonghua Yi Xue Za Zhi. 2012;92(26):1820-1823. doi: 10.3760/cma.j.issn.0376-2491.2012.26.007

20.

Sundberg

Forsander

Continuous glucose monitoring in healthy children aged 2-8 years. Diabetes Technol Ther. 2018;20(2):113-116. doi:10.1089/dia.2017.0270.

21.

Sterne

Hernán

Reeves

, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919. doi:10.1136/bmj.i4919.

22.

R, Core Team. R: A Language and Environment for Statistical Computing (Version 4.4.1). Vienna, Austria: R Foundation for Statistical Computing; 2024. https://www.R-project.org/. Accessed February 16, 2026.

23.

Harrer

Cuijpers

Furukawa

Ebert

DD.

Doing Meta-Analysis with R: A Hands-On Guide. 1st ed. New York: Chapman and Hall/CRC; 2021. doi:10.1201/9781003107347.

24.

Schwarzer

. meta: general package for meta-analysis (Version 7.0.0) [Computer Software]. Github. https://github.com/guido-s/meta. Accessed September 9, 2025.

25.

Cuijpers

Furukawa

Ebert

. dmetar: companion R package for the guide doing meta-analysis in R (Version 0.1.0) [Computer software]. https://dmetar.protectlab.org. Published 2024. Accessed September 9, 2025.

26.

Rizos

Kanellopoulou

Filis

, et al. Difference on glucose profile from continuous glucose monitoring in people with prediabetes vs normoglycemic individuals: a matched-pair analysis. J Diabetes Sci Technol. 2024;18(2):414-422. doi:10.1177/19322968221123530.

27.

Spartano

Sultana

Lin

, et al. Defining continuous glucose monitor time in range in a large, community-based cohort without diabetes. J Clin Endocrinol Metab. 2025;110(4):1128-1134. doi:10.1210/clinem/dgae626.

28.

Kashiwagi

Inaishi

Kinoshita

, et al. Assessment of glycemic variability and lifestyle behaviors in healthy nondiabetic individuals according to the categories of body mass index. PLoS ONE. 2023;18(10):e0291923. doi:10.1371/journal.pone.0291923.

29.

Dorcely

Sifonte

Popp

, et al. Continuous glucose monitoring and 1-h plasma glucose identifies glycemic variability and dysglycemia in high-risk individuals with HbA1c <5.7%: a pilot study. Endocrine. 2022;77(2):403-407. doi:10.1007/s12020-022-03109-5.

30.

Rodriguez-Segade

Rodriguez

Camiña

, et al. Continuous glucose monitoring is more sensitive than HbA1c and fasting glucose in detecting dysglycaemia in a Spanish population without diabetes. Diabetes Res Clin Pract. 2018;142:100-109. doi:10.1016/j.diabres.2018.05.026.

31.

Hanefeld

Sulk

Helbig

Thomas

Köhler

Differences in glycemic variability between normoglycemic and prediabetic subjects. J Diabetes Sci Technol. 2014;8(2):286-290. doi:10.1177/1932296814522739.

32.

Madhu

Muduli

Avasthi

Abnormal glycemic profiles by CGMS in obese first-degree relatives of type 2 diabetes mellitus patients. Diabetes Technol Ther. 2013;15(6):461-465. doi:10.1089/dia.2012.0333.

33.

Nomura

Saitoh

Kim

Yamanouchi

Glycemic profiles of healthy individuals with low fasting plasma glucose and HbA1c. ISRN Endocrinol. 2011;2011:435047. doi:10.5402/2011/435047.

34.

Chakarova

Dimova

Grozeva

Tankova

Assessment of glucose variability in subjects with prediabetes. Diabetes Res Clin Pract. 2019;151:56-64. doi:10.1016/j.diabres.2019.03.038.

35.

DuBose

Kanapka

Bradfield

Sooy

Beck

Steck

AK.

Continuous glucose monitoring profiles in healthy, nondiabetic young children. J Endocr Soc. 2022;6(6):bvac060. doi:10.1210/jendso/bvac060.

36.

Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group; Fox

Beck

Xing

Variation of interstitial glucose measurements assessed by continuous glucose monitors in healthy, nondiabetic individuals. Diabetes Care. 2010;33(6):1297-1299. doi:10.2337/dc09-1971.

37.

Shah

DuBose

, et al. Continuous glucose monitoring profiles in healthy nondiabetic participants: a multicenter prospective study. J Clin Endocrinol Metab. 2019;104(10):4356-4364. doi:10.1210/jc.2018-02763.

38.

Shilo

Keshet

Rossman

, et al. Continuous glucose monitoring and intrapersonal variability in fasting glucose. Nat Med. 2024;30(5):1424-1431. doi:10.1038/s41591-024-02908-9.

39.

Bailey

Ahmann

Brazg

, et al. Accuracy and acceptability of the 6-day Enlite continuous subcutaneous glucose sensor. Diabetes Technol Ther. 2014;16(5):277-283. doi:10.1089/dia.2013.0222.

40.

Welsh

Gao

Derdzinski

, et al. Accuracy, utilization, and effectiveness comparisons of different continuous glucose monitoring systems. Diabetes Technol Ther. 2019;21(3):128-132. doi:10.1089/dia.2018.0374.

41.

Alva

Brazg

Castorino

Kipnes

Liljenquist

Liu

Accuracy of the third generation of a 14-day continuous glucose monitoring system. Diabetes Ther. 2023;14(4):767-776. doi:10.1007/s13300-023-01385-6.

42.

Tabák

Herder

Rathmann

Brunner

Kivimäki

Prediabetes: a high-risk state for diabetes development. Lancet. 2012;379(9833):2279-2290. doi:10.1016/S0140-6736(12)60283-9.

43.

Knowler

Barrett-Connor

Fowler

, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002;346(6):393-403.

44.

Zhang

Barker

Chowdhury

Zhang

Cost-effectiveness of interventions to prevent and control diabetes mellitus: a systematic review. Diabetes Care. 2010;33(8):1872-1894. doi:10.2337/dc10-0843.

45.

Barry

Roberts

Oke

Vijayaraghavan

Normansell

Greenhalgh

Efficacy and effectiveness of screen and treat policies in prevention of type 2 diabetes: systematic review and meta-analysis. BMJ. 2017;356:i6538. doi:10.1136/bmj.i6538.

46.

World Health Organization. Diabetes [Fact sheet]. WHO Regional Office for Europe. https://www.who.int/europe/news-room/fact-sheets/item/diabetes. Published 2024. Accessed September 15, 2025.

47.

Eurostat. European Health Interview Survey (EHIS)—Physical activity statistics. https://ec.europa.eu/eurostat/databrowser/product/page/HLTH_EHIS_PA2E. Published 2025. Accessed June 4, 2025.

48.

Pettus

Edelman

SV.

Recommendations for using real-time continuous glucose monitoring data for insulin adjustments in type 1 diabetes. J Diabetes Sci Technol. 2017;11(1):138-147. doi:10.1177/1932296816663747.

49.

Battelino

Danne

Bergenstal

, et al. Clinical targets for continuous glucose monitoring data interpretation: recommendations from the international consensus on time in range. Diabetes Care. 2019;42(8):1593-1603. doi:10.2337/dci19-0028.

50.

Agarwal

Schechter

Gonzalez

Long

JA.

Racial–ethnic disparities in diabetes technology use among young adults with type 1 diabetes. Diabetes Technol Ther. 2021;23(4):306-313. doi:10.1089/dia.2020.0338.

51.

Montaser

Farhy

Rich

SS.

Enhancing type 1 diabetes immunological risk prediction with continuous glucose monitoring and genetic profiling. Diabetes Technol Ther. 2025;27(4):292-300. doi:10.1089/dia.2024.0496.

52.

Mader

Wong

Freckmann

, et al. The use of continuous glucose monitoring to diagnose stage 2 type 1 diabetes. J Diabetes Sci Technol. 2025;19(4):1109-1127. doi:10.1177/19322968251333441.

53.

Montaser

Farhy

Kovatchev

BP.

Novel detection and progression markers for diabetes based on continuous glucose monitoring data dynamics. J Clin Endocrinol Metab. 2024;110(1):254-262. doi:10.1210/clinem/dgae379.

54.

Patterson

Zhang

, et al. Worldwide estimates of incidence of type 2 diabetes in children and adolescents in 2021. Diabetes Res Clin Pract. 2022;185:109785. doi:10.1016/j.diabres.2022.109785.

55.

Alkabbani

Cromer

Kim

, et al. Overall uptake and racial, ethnic, and socioeconomic disparities in the use of continuous glucose monitoring devices among insulin-treated older adults with type 2 diabetes. Diabetes Care. 2025;48(8):1377-1385. doi:10.2337/dca25-0006.

56.

Vrany

Hill-Briggs

Ephraim

Myers

Garnica

Fitzpatrick

SL.

Continuous glucose monitors and virtual care in high-risk, racial and ethnic minority populations: toward promoting health equity. Front Endocrinol (Lausanne). 2023;14:1083145. doi:10.3389/fendo.2023.1083145.

57.

Hijikata

Higa

Ichijo

Hirose

A comparison of meal tolerance test and oral glucose tolerance test for predicting insulin therapy in patients with gestational diabetes. Food Nutr Res. 2021;2021:65. doi:10.29219/fnr.v65.5490.

58.

Freckmann

Eichenlaub

Waldenmaier

, et al. Clinical performance evaluation of continuous glucose monitoring systems: a scoping review and recommendations for reporting. J Diabetes Sci Technol. 2023;17(6):1506-1526. doi:10.1177/19322968231190941.

59.

McCambridge

Witton

Elbourne

DR.

Systematic review of the Hawthorne effect: new concepts are needed to study research participation effects. J Clin Epidemiol. 2014;67(3):267-277. doi:10.1016/j.jclinepi.2013.08.015.

60.

Muchmore

Sharp

Vaughn

Benefits of blinded continuous glucose monitoring during a randomized clinical trial. J Diabetes Sci Technol. 2011;5(3):676-680. doi:10.1177/193229681100500321.

61.

Longo

Sperling

Personal versus professional continuous glucose monitoring: when to use which on whom. Diabetes Spectr. 2019;32(3):183-193. doi:10.2337/ds18-0093.

62.

Ehrhardt

Montour

Berberian

Vasconcelos

Comstock

Wright

LA-C

. A randomized clinical trial of a culturally tailored diabetes education curriculum with and without real-time continuous glucose monitoring in a Latino population with type 2 diabetes: the CUT-DM with continuous glucose monitoring study [published online ahead of print April 10, 2025]. J Diabetes Sci Technol. doi:10.1177/19322968251331526.

63.

Jospe

Richardson

Saleh

, et al. Leveraging continuous glucose monitoring as a catalyst for behaviour change: a scoping review. Int J Behav Nutr Phys Act. 2024;21(1):74. doi:10.1186/s12966-024-01622-6.

64.

Hegyi

Erőss

Izbéki

Párniczky

Szentesi

Accelerating the translational medicine cycle: the Academia Europaea pilot. Nat Med. 2021;27(8):1317-1319. doi:10.1038/s41591-021-01458-8.

65.

Hegyi

Petersen

Holgate

, et al. Academia Europaea position paper on translational medicine: the cycle model for translating scientific results into community benefits. J Clin Med. 2020;9(5):1532. doi:10.3390/jcm9051532.

66.

Bergenstal

RM.

Understanding continuous glucose monitoring data. ADA Clin Compend. 2018;2018:20-23. doi:10.2337/db20181-20.

67.

Battelino

Alexander

Amiel

, et al. Continuous glucose monitoring and metrics for clinical trials: an international consensus statement. Lancet Diabetes Endocrinol. 2023;11(1):42-57. doi:10.1016/S2213-8587(22)00319-9.

68.

Jarvis

PRE

Cardin

Nisevich-Bede

McCarter

. Continuous glucose monitoring in a healthy population: understanding the post-prandial glycemic response in individuals without diabetes mellitus. Metabolism. 2023;146:155640. doi:10.1016/j.metabol.2023.155640.

69.

Köhlmoos

Dittmar

Glycemic variability and control by CGM in healthy older and young adults and their relationship with diet. J Endocr Soc. 2025;9(7):bvaf081. doi:10.1210/jendso/bvaf081.

70.

Shah

Battelino

Bergenstal

Phillip

Staging prediabetes and type 2 diabetes: the time to start is now. Lancet Diabetes Endocrinol. 2026;14(1):8-10. doi:10.1016/S2213-8587(25)00333-X.

71.

Gehr

Holder

Kulzer

, et al. Spectrum: a training and treatment program for continuous glucose monitoring for all age groups. J Diabetes Sci Technol. 2017;11(2):284-289. doi:10.1177/1932296816661735.