Trends in Time in Range–Related Publications and Clinical Trials: A Bibliometric Review

Published manuscripts (publications) and conference abstracts were searched from the following databases: MEDLINE, Embase, BIOSIS Previews, International Pharmaceutical Abstracts, Allied & Complementary Medicine, EMCare, and FSTA. Clinical trials were identified and sourced from clinical trial registries in Australia, China, Germany, India, Iran, the Netherlands, New Zealand, and the United States, as well as the University Hospital Medical Information Network and International Standard Randomized Controlled Trial Number registries.

Search terms were consistent across publications, conference abstracts, and clinical trial database searches. The search terms were “time in range,” “time-in-range,” or “TIR”; “time above range,” “time-above-range,” or “TAR”; and “time below range,” “time-below-range,” or “TBR.” For publications and abstracts, we applied a limit to obtain only those published between 2010 and 2021. Additionally, we included clinical trials with a start date during or after 2010. An end limit was not applied to clinical trials to assess the potential future of TIR as an outcome measure.

This review was not registered in databases such as PROSPERO because it was a bibliometric review describing trends in publications and clinical trials rather than a systematic review that follows a traditional PICO(T) format.

All identified publications, conference abstracts, and clinical trials were screened, and those not in English were excluded. Book chapters, nonhuman studies, and studies not reporting TIR as defined in the International Consensus for Time in Range recommendations (7) as an outcome were excluded.

Data extracted from publications and conference abstracts included manuscript/abstract category (clinical trial, real-world evidence, commentaries, systematic reviews, or meta-analysis), publication year, study region, and role of CGM (interventional or observational). Additionally, data on outcome measures were captured, including TIR reported as primary outcome (yes/no), A1C as any outcome (yes/no), TBR as any outcome (yes/no), and TAR as any outcome (yes/no). Studies that explicitly mentioned TIR as the primary outcome were marked as “yes” for TIR as a primary outcome. Clinical trial data extraction elements included the trial start and completion dates in addition to the above-listed data elements except for manuscript/abstract category. Additionally, the reasons for excluding any publications, conference abstracts, and clinical trials were documented. Also, the intervention was recorded when CGM was used as an observational tool.

CGM was classified as an interventional tool when it was used as an active intervention (alone or in combination with another intervention). For example, a study evaluating CGM to improve diabetes outcomes would be designated as interventional. The role of CGM was classified as observational when it was used to measure disease burden, effectiveness of other therapies or interventions, or relationships between TIR and outcomes. For example, a study evaluating the efficacy of sodium–glucose cotransporter 2 inhibitor therapy that included TIR as an outcome would be designated observational.

Data were extracted by investigators for all included publications and clinical trials. The majority of variables were categorized and defined a priori and operationalized as drop-down items in the Excel workbook to ensure consistency in the data-extraction process across multiple reviewers. Team discussions were conducted routinely to resolve any outstanding questions during the data-extraction process. Quality checks were conducted from a random sample of studies assigned to each data extractor to ensure consistency.

The literature and clinical trial search yielded 467 clinical trials, 665 publications, and 806 conference abstracts. Of these, 373 clinical trials (79.9%), 531 publications (79.8%), and 620 conference abstracts (76.9%) were included in the review (Table 1). The citations of the included trials, publications, and abstracts are available in the Supplementary Material. The top reason for exclusion was no relevant mention of TIR consistent with the consensus recommendations (7). TIR in many of these excluded studies was either captured from blood glucose monitoring using test strips or was related to time within the therapeutic range for an international normalized ratio. The majority of the publications (48%) were clinical trials, and the majority of conference abstracts (52%) were real-world analyses. The majority of the clinical trials, publications, and conference abstracts were conducted in North American or European settings. Approximately 54% of clinical trials, 47% of publications, and 47% of conference abstracts reported the role of CGM to be observational (i.e., CGM was not part of the intervention in these studies). There was wide variation in the interventions when the role of CGM was observational. Association of TIR/glycemic control with complications/A1C and insulin therapies were top categories when the role of CGM was observational. Thirty-two clinical trials, 89 publications, and 105 conference abstracts assessed the relationships between TIR and A1C and between TIR and diabetes complications in this review period.

A trend analysis for clinical trials was performed using the completion year. This was done to account for variability in study durations. There was a steady increase in clinical trials reporting TIR, especially between 2018 and 2022 (Figure 1). There was a decline observed in clinical trials reporting TIR that are scheduled to be completed after 2022. A trend analysis for publications revealed a dramatic increase in publications reporting or discussing TIR after 2018 (Figure 2). A similar trend was observed in conference abstracts, although the trend seemed to have stabilized in the 2020–2021 period (Figure 3). Notably, there was a 6-fold, 12-fold, and 4.5-fold increase in the number of clinical trials, publications, and conference abstracts, respectively, that reported TIR.

Forty-four percent of clinical trials, 35% of publications, and 35% of conference abstracts reported TIR as a primary outcome (Table 2). About 48%, 38%, and 64% of clinical trials, publications, and conference abstracts, respectively, reported TIR without including A1C. Finally, TBR was reported as an outcome in 68% of clinical trials, 73% of publications, and 65% of abstracts, whereas TAR was reported in 59% of clinical trials, 68% of publications, and 47% of abstracts. Thus, TBR was reported more often than TAR.

Findings from this review highlight the marked increase in number of clinical trials, publications, and conference abstracts that report or discuss TIR. Together, these findings highlight the growing body of TIR-related evidence and the increasing significance of TIR as an outcome measure. A decline was observed in current clinical trials that plan to report TIR when they are completed in years after 2022 (Figure 1). This decline may be attributable to coronavirus disease 2019 (COVID-19) pandemic–related disruptions in clinical trial enrollments and potentially the fact that many trials may not have been registered yet.

There was a lag observed in the trend of reporting TIR between conference abstracts and publications, with abstracts showing an earlier increase in TIR reporting (Figures 1 and 2). This difference in reporting rates may be attributable to the time lag between the submission of abstracts and the usual delays associated with the scientific journal publication process. It is possible that the number of TIR-related publications may stabilize in 2022 compared with 2021.

The growing body of TIR-related evidence highlights the increasing acceptance of CGM-derived metrics by providers, researchers, and manufacturers. Although TIR may be a relatively new metric, several clinical trials, publications, and abstracts explored relationships between TIR and A1C and between TIR with diabetes complications during this investigation’s review period. This evidence provides the clinical and scientific rationale for the use of TIR as an outcome measure. A recent international consensus statement on CGM metrics for clinical trials recommended standardization of CGM use and CGM metrics as study-specific end points or supportive complementary glucose metrics (16).

Given the growing body of evidence, quality accrediting bodies and regulatory agencies should explore inclusion of CGM-derived metrics in their frameworks, which currently rely on A1C outcomes. A recent white paper from the National Committee for Quality Assurance lists recommendations that incorporate CGM-derived metrics that complement A1C (15). First, it highlights the use of the glucose management indicator (GMI) when a recent A1C value is not available. GMI is an estimated A1C derived from CGM data. Next, it highlights the use of a TIR-related threshold (i.e., the number of people achieving a 5% improvement in TIR) as a quality metric. Finally, it describes the use of TIR- and TBR-related metrics to optimize outcomes and minimize the risk of hypoglycemic events. From a regulatory standpoint, recent discussions at the U.S. Food and Drug Administration have explored the value of metrics beyond A1C that could be included in a more patient-centered regulatory framework (17). Outcomes such as TBR (a measure of hypoglycemia), weight loss, and patient-reported outcomes have been discussed. Additionally, this discussion included one patient’s perspective that “TIR really defined the daily experience of living with diabetes” (18). These developments could explain some of the findings of this review related to the observational use of CGM.

A significant proportion of studies reported the use of CGM as an observational tool for other therapies or programs (Table 1). These therapies included insulin, noninsulin agents, digital health interventions, and telehealth programs. Furthermore, a significant proportion of studies reported TIR as a primary outcome, and some reported TIR without complementary A1C (Table 2). These findings highlight the growing use of TIR as an outcome measure, either as a complement to or as a replacement for A1C.

Our findings also emphasize that TIR can be a useful end point beyond A1C for value-based contracts. Value-based contracts have traditionally relied on metrics such as A1C, health care costs, health resource utilization, and medication adherence. However, A1C values may be difficult to capture and operationalize as part of a value-based contract. Risk-sharing agreements that use CGM-derived metrics can be easier to implement and operationalize with the digital ecosystems that CGMs enable. Furthermore, these outcomes are more patient-centered and account for the daily experiences of people with diabetes. The Diabetes Leadership Council supports leveraging TIR in value-based contracts, as reflected in its 2020 consensus statement, which stated that “value-based insurance design in diabetes will fall short if payers and providers emphasize A1C but neglect TIR, reducing hypoglycemia, cardiovascular and renal protection, behavioral health, improved quality of life, and other measures that people with diabetes value” (19). Although the details of such arrangements are unknown, payers have partnered with diabetes technology manufacturers in value-based contracts that leverage TIR data (20).

Although this review highlights the growing significance of TIR, there are several challenges that limit its widespread application, especially from a quality metric standpoint. First, not everyone has access to CGM, even those for whom CGM is considered a standard of care (i.e., people with diabetes who are on intensive insulin therapy regimens) (21). This accessibility issue limits the utility and population-level application of TIR-related quality metrics. However, quality metrics such as the percentage of people on an intensive insulin regimen who use CGM could be considered as a process measure that aligns with professional guideline recommendations. In addition, incorporating disparities and health equity considerations could help to alleviate existing disparities in access to diabetes technology that prevail across racial/ethnic lines and insurance types (22–24). Doing so will increase the user base and pave the way for adoption of TIR-related quality metrics, especially among people who have an intensive insulin therapy regimen.

Although data are available supporting associations of TIR with A1C and complications, there have been limited long-term randomized controlled trials to help establish these relationships. This review also highlights a gap in evidence that explores the direct three-way relationships among TIR, A1C, and diabetes complications.

Additionally, CGM accuracy and performance can vary among available devices, and such differences can potentially affect the utility of TIR-related quality metrics. However, the diabetes technology landscape continues to evolve to produce more accurate devices. Similar limitations exist with available A1C testing methods, which cannot only be confounded by a variety of factors, but also yield different values depending on whether a laboratory-based test or a rapid test is performed (2–4,25).

To the best of our knowledge, this is the first bibliometric review of TIR-related publications, conference abstracts, and clinical trials. Although trends in publications and conference abstracts describe the past and present use of TIR-related data, clinical trials were included to also describe future trends in TIR-related metrics in ongoing or future studies.

This comprehensive review has some limitations. First, our inclusion of clinical trials, conference abstracts, and publications can duplicate and amplify some of the trends observed. For example, results from a clinical trial may be presented as an abstract and culminate as a publication. To reduce the impact of this duplication on the trends, results are presented separately for clinical trials, publications, and conference abstracts. Next, abstracts may have appeared as encore presentations at multiple conferences, which could have led to multiple counts of the same abstract. Finally, this review only included studies that aligned with the TIR definition used in the international consensus report and did not differentiate for the recommended pregnancy and gestational diabetes mellitus target range of 63–140 mg/dL (7,8). There were several instances in which terms such as “time in target” or “time between 70 and 100 mg/dL” were derived from CGM data; however, they were excluded because they did not align with the consensus report definition. Nonetheless, this limitation would only result in underestimating the number of studies reporting CGM-derived data.

Further analyses could address the limitations listed above. Additionally, non-English publications and abstracts, a delineation between interventional use and observational use of CGM, and subgroup analysis by region or country could be included. Understanding the application of CGM metrics for different populations grouped by age, diabetes type, and other baseline or demographic factors may also be of interest.

The findings of this review reveal a marked increase in the number of clinical trials, publications, and conference abstracts reporting CGM-derived metrics such as TIR. This trend underscores the increasing acceptance and significance of TIR as an outcome measure in diabetes management. Stakeholders, especially in quality accrediting and regulatory bodies, should explore incorporating CGM-derived measures into their decision frameworks; however, they may be constrained by the current limited use of diabetes technology among people with diabetes.

The authors thank Dr. Bree Hamman, a research information scientist at Abbott Diabetes Care, for assistance with the search strategy and drafting of the supplementary material. They also acknowledge Dr. William Perlman, a contractor with Writing Assistance, Inc., for his medical writing services.

This study was funded by Abbott Diabetes Care.

P.M.P. was an employee of Abbott Diabetes Care during the study. R.M.A., N.D., C.B.L., M.A.F., and N.S.V. are employees of Abbott Diabetes Care. No other potential conflicts of interest relevant to this article were reported.

P.M.P. contributed to the concept and design; acquisition, analysis, and interpretation of data; drafting and critical revision of the manuscript for important intellectual content; administrative, technical, and logistic support; and supervision. R.M.A., N.D., C.B.L., and M.A.F. contributed to the acquisition of data, analysis, and interpretation of data and critical revision of the manuscript for important intellectual content. N.S.V. contributed to the concept and design, analysis and interpretation of data, and critical revision of the manuscript for important intellectual content. P.M.P. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Preliminary findings of this study were presented at the American Diabetes Association’s 82nd Scientific Sessions in New Orleans, LA, 3–7 June 2022.