Use of continuous glucose monitoring (CGM) in people with diabetes may provide a more complete picture of glycemic control than glycated hemoglobin (HbA1c) measurements, which do not capture day-to-day fluctuations in blood glucose levels. The randomized, crossover, phase IV SWITCH PRO study assessed time in range (TIR), derived from CGM, following treatment with insulin degludec or insulin glargine U100 in patients with type 2 diabetes at risk for hypoglycemia. This post hoc analysis evaluated the relationship between TIR and HbA1c, following treatment intensification during the SWITCH PRO study.
Correlation between absolute values for TIR (assessed over 2-week intervals) and HbA1c, at baseline and at the end of maintenance period 1 (M1; week 18) or maintenance period 2 (M2; week 36), were assessed by linear regression and using the Spearman correlation coefficient (rs). These methods were also used to assess correlation between change in TIR and change in HbA1c from baseline to the end of M1, both in the full cohort and in subgroups stratified by baseline median HbA1c (≥ 7.5% [≥ 58.5 mmol/mol] or < 7.5% [< 58.5 mmol/mol]).
A total of 419 participants were included in the analysis. A moderate inverse linear correlation was observed between TIR and HbA1c at baseline (rs −0.54), becoming stronger following treatment intensification during maintenance periods M1 (weeks 17–18: rs −0.59) and M2 (weeks 35–36: rs −0.60). Changes in TIR and HbA1c from baseline to end of M1 were also linearly inversely correlated in the full cohort (rs –0.40) and the subgroup with baseline HbA1c ≥ 7.5% (rs −0.43). This was less apparent in the subgroup with baseline HbA1c < 7.5% (rs −0.17) (p-interaction = 0.07).
Results from this post hoc analysis of data from SWITCH PRO, one of the first large interventional clinical studies to use TIR as the primary outcome, further support TIR as a valid clinical indicator of glycemic control.
Trial registration: ClinicalTrials.gov identifier, NCT03687827.
Why carry out this study?
The use of continuous glucose monitoring (CGM) in people with diabetes, enabling measurement of time in the target glycemic range (TIR), may provide a more complete picture of glycemic control than glycated hemoglobin (HbA1c) measurements alone.
There are limited data on the relationship between TIR and HbA1c during treatment, and on the value of TIR as an additional metric for the assessment of glycemic control.
The current post hoc analysis evaluated the correlation between TIR and HbA1c, and also between change in TIR and change in HbA1c, during the randomized, crossover, phase IV SWITCH PRO study, following treatment with insulin degludec or insulin glargine U100 in basal insulin-treated patients with type 2 diabetes at risk for hypoglycemia.
What was learned from the study?
Inverse linear correlations were observed between TIR and HbA1c (greater TIR being associated with lower HbA1c), with the strongest correlation following treatment intensification. Changes from baseline in TIR and HbA1c, following treatment intensification, were also inversely linearly correlated (increases in TIR associated with reductions in HbA1c) in the full cohort and in participants with HbA1c ≥ 7.5% at baseline, although a correlation was less apparent in participants with HbA1c < 7.5% at baseline.
Despite these clear correlations, there was a wide scatter of data, indicating that TIR (and other metrics) provides information about glycemic control that cannot be discerned from HbA1c alone, and which at least complements it. These results, therefore, support the role of TIR as a metric of glycemic control that may benefit people with diabetes (and their healthcare providers) in disease management.
Clinical trials in diabetes have traditionally assessed glycated hemoglobin (HbA1c) as the primary efficacy outcome to determine treatment-related changes in glycemic control. HbA1c readings provide an indication of glycemic control over the preceding 3 months, but do not capture day-to-day fluctuations of blood glucose levels in people with diabetes . For this reason, two people with the same HbA1c measurements could have quite different glucose profiles . The use of continuous glucose monitoring (CGM) to continuously measure glucose levels in the interstitial fluid has made it possible to determine time in range (TIR), time above range (TAR), and time below range (TBR), all expressed as the percentage of time spent within the respective ranges . CGM metrics may, therefore, provide a more comprehensive picture of glycemic control than HbA1c alone, reflecting daily changes in glucose levels [1, 3, 4].
Evidence suggests that, like HbA1c, TIR is also associated with diabetes-related complications, with a greater TIR correlating with fewer microvascular and macrovascular complications [3,4,5]. To date, however, there have been limited diabetes therapeutic trials utilizing TIR as the primary outcome, and thus there are limited data on how closely TIR tracks and compares with HbA1c changes during treatment, and whether this metric offers additional discrimination or insight in assessing pharmacologic treatments compared with HbA1c.
The randomized, crossover, phase IV SWITCH PRO trial compared the effect of insulin degludec (degludec) versus insulin glargine U100 (IGlar U100), in patients with type 2 diabetes (T2D), on glycemic control as assessed by TIR as the primary outcome. SWITCH PRO was a follow-up study to the SWITCH 2 trial, which also compared these two insulins using a similar design . With the addition of blinded professional CGM, the aim of SWITCH PRO was to specifically characterize time spent in different glycemic ranges. The results of SWITCH PRO demonstrated superiority of degludec over glargine U100 for TIR (70–180 mg/dl [3.9–10.0 mmol/l]) . More time spent in a narrower glycemic range (70–140 mg/dl [3.9–7.8 mmol/l]) and less nocturnal time below range were also seen with degludec versus IGlar U100 . To better understand how well the emerging TIR metric compares with the traditional metric of HbA1c in assessing therapeutic interventions, we undertook this post hoc analysis to evaluate the correlation between TIR and HbA1c, and also between change in TIR and change in HbA1c, during the SWITCH PRO trial.
The design of the SWITCH PRO trial (ClinicalTrials.gov identifier: NCT03687827) has been reported previously . Briefly, basal insulin-treated patients with T2D and at least one risk factor for hypoglycemia were treated with degludec or glargine U100 during a 16-week titration and 2-week maintenance phase, then crossed over to the other treatment for the same time periods (Fig. 1). During the titration period, insulin dose was adjusted once weekly based on pre-breakfast self-monitored blood glucose and investigator discretion, to a blood glucose target of 70–90 mg/dl (3.9–5.0 mmol/l).
Glucose was evaluated using blinded professional CGM, using the Abbott Freestyle Libre Pro system (Abbott Laboratories, Chicago, IL, USA), which consisted of a sensor applied to the participant’s arm and a reader (set for date, time, and target glucose range) for data upload . Glucose levels were collected every 15 min from interstitial fluid during the 2-week run-in period and maintenance periods, and were not visible to the participants or investigators until after each CGM period. TIR was defined as the percentage of time spent in the glycemic range of 70–180 mg/dl (3.9–10.0 mmol/l), in line with the International Consensus on Time in Range , during the 2-week run-in period (weeks −2 and −1) and the 2-week maintenance periods (M1 [weeks 17 and 18] and M2 [weeks 35 and 36]).
All pre-specified endpoint analyses in SWITCH PRO were performed on a final analysis set (n = 448) comprising participants remaining on the assigned treatment and completing ≥ 70% of 2 weeks of CGM measurements (as recommended)  in each maintenance period (M1 and M2). Of these, 419 participants also had ≥ 70% of 2 weeks of CGM measurements during the 2-week run-in (baseline) period. SWITCH PRO was conducted in accordance with ethical principles derived from international guidelines across 67 sites in five countries (USA, Canada, Poland, South Africa, and Slovakia) between 2 October 2018 and 27 December 2019 . All participants gave their written informed consent prior to inclusion in the study.
Prior to initiation of the SWITCH PRO study, the protocol, consent form, and subject information sheet were reviewed and approved according to local regulations by both the appropriate health authorities and by an independent ethics committee (IEC) and/or institutional review board (IRB). A list of the IECs/IRBs is provided in Electronic Supplementary Material (ESM) Table S1. The SWITCH PRO study was conducted in accordance with ethical principles derived from international guidelines, including the Declaration of Helsinki  and the International Conference on Harmonisation (ICH) Good Clinical Practice .
In the current post hoc analysis, the correlation between absolute values for TIR and HbA1c at baseline and at the end of M1 (week 18) and M2 (week 36) was assessed through linear regression models and using the Spearman correlation coefficient (rs). Regression models and rs were also used to assess the correlation between change in TIR and change in HbA1c from baseline to the end of M1 (week 18) in the full cohort and in subgroups stratified by baseline median HbA1c (≥ 7.5% [≥ 58.5 mmol/mol] or < 7.5% [< 58.5 mmol/mol]).
Baseline characteristics were generally similar between the full cohort (n = 419) and baseline HbA1c subgroups (≥ 7.5% [≥ 58.5 mmol/mol; n = 212] or < 7.5% [< 58.5 mmol/mol; n = 207]) (Table 1).
Correlation Between TIR and HbA1c
Scatter plots showed a moderate inverse linear correlation between TIR and HbA1c at baseline (rs −0.54), and this became stronger following treatment intensification at the end of M1 (rs −0.59) and M2 (rs −0.60) (Fig. 2a). Improved glycemic control following treatment intensification was signified by both greater TIR and reduced HbA1c. This was evidenced by the distribution of scatter points becoming more centered around the lower right quadrant at the end of M1 and M2 compared with the plots at baseline, with less variability in the data and a steeper regression line at M1 and M2 compared with baseline (Fig. 2a). HbA1c values were also lower for any TIR value after treatment intensification: for example, 70% TIR corresponded to an estimated HbA1c of 7.5% (58.5 mmol/mol) at baseline, 7.2% (55.2 mmol/mol) in M1, and 7.2% (55.2 mmol/mol) in M2 (Fig. 2a).
Correlation Between Change in TIR and Change in HbA1c
An inverse linear correlation was demonstrated between change in TIR and change in HbA1c, from baseline to the end of M1, in the full cohort (rs −0.40) and in the baseline HbA1c ≥ 7.5% (≥ 58.5 mmol/mol) subgroup (rs −0.43), but this was less apparent in the baseline HbA1c < 7.5% (< 58.5 mmol/mol) subgroup (rs −0.17) (p-interaction = 0.07) (Fig. 2b). A 5% increase in TIR corresponded to estimated decreases in HbA1c of −0.4% (−4.4 mmol/mol), −0.7% (−7.7 mmol/mol), and −0.1% (−1.1 mmol/mol) in the full cohort, baseline HbA1c ≥ 7.5% (≥ 58.5 mmol/mol) subgroup, and baseline HbA1c < 7.5% (< 58.5 mmol/mol) subgroup, respectively.
This post hoc analysis of data from participants in the SWITCH PRO study with T2D and increased risk of hypoglycemia, treated with basal insulin, showed an inverse linear correlation between TIR and HbA1c, with greater TIR following treatment intensification corresponding to lower HbA1c values. An inverse linear correlation was also seen between change in TIR and change in HbA1c from baseline to the end of the first maintenance period, with increasing TIR again corresponding to a lowering of HbA1c levels. Notably, there was little concordance between change in TIR and change in HbA1c in those with a baseline HbA1c < 7.5% (< 58.5 mmol/mol).
Our data are not the first to show such correlations between TIR and HbA1c. An evaluation of selected paired HbA1c and TIR data from 18 previously published articles, by linear regression analysis and Pearson's correlation coefficient, previously found a good correlation between the two metrics (R = −0.84; R2 = 0.71) ; for every absolute 10% change in %TIR, there was a 0.8% (9 mmol/mol) change in HbA1c. These data are therefore broadly in agreement with the results shown here.
As shown in Fig. 2a and b, however, the data are widely scattered, supporting the premise that HbA1c and TIR can be relatively crude surrogates of each other when it comes to individual patients. Where patients have both low HbA1c and low TIR values, this might indicate frequent episodes of hypoglycemia. There were also a few individual patients in our study for whom TIR exceeded 70% but HbA1c approached 9% (74.9 mmol/mol). This apparent disparity might be explained if, during their time outside of the target range, these patients had extremely high blood glucose levels. Although they will display a good overall TIR, they will also have a high mean HbA1c. It is also possible that HbA1c may not be an accurate reflection of average blood glucose in certain individuals because of variation in red blood cell physiology.
An observation of potential clinical importance was that the association between change in TIR and change in HbA1c was more apparent in patients with baseline HbA1c ≥ 7.5% (≥ 58.5 mmol/mol) than in those with HbA1c < 7.5% (< 58.5 mmol/mol). A 5% increase in TIR corresponded to a decrease in HbA1c of −0.4% (−4.4 mmol/mol) overall in the full cohort and −0.7% (−7.7 mmol/mol) in the subgroup with baseline HbA1c ≥ 7.5% (≥ 58.5 mmol/mol). However, in patients with HbA1c < 7.5% (< 58.5 mmol/mol), a 5% increase in TIR was associated with only a −0.1% (−1.1 mmol/mol) change in HbA1c. While a 5% increase in TIR has previously been shown to be associated with clinically significant benefits for individuals with type 1 diabetes or T2D , the −0.1% (−1.1 mmol/mol) correlative estimated change in HbA1c seen in this lower HbA1c range would be less compelling. Our observation therefore suggests that, for a given change in TIR at high baseline HbA1c levels, there is a greater change in HbA1c than for the same change in TIR at lower baseline HbA1c levels. Beneficial changes in TIR may therefore become less detectable in terms of a further HbA1c decrease as the baseline HbA1c values decrease. Similar findings were reported by Beck and colleagues, who stated that an increase in TIR of 10% was associated with a change in HbA1c of approximately −1% (−10.9 mmol/mol) for those with a high baseline HbA1c (≥ 8.0% [≥ 63.9 mmol/mol]), and with only a change of −0.4% (4.4 mmol/mol) for those with a lower baseline HbA1c (7.0–7.9% [53.0–62.9 mmol/mol]) . Poorer correlation between TIR and HbA1c in patients with lower baseline HbA1c may also be due to increased TBR for the target glycemic range. In this treat-to-target trial, patients experiencing hypoglycemic episodes or pre-breakfast self-monitored plasma glucose below range would have had their insulin dose down-titrated and further decreases in HbA1c would not have been seen.
Research shows that over one-third of people with T2D are not achieving the internationally recommended HbA1c targets (HbA1c < 7.0–8.5% [< 53.0–69.4 mmol/mol], depending on age and complications)  and that many people with diabetes currently spend less than half of their day in the recommended glycemic target range [3, 13, 14]. When used in conjunction with HbA1c, CGM data, such as TIR, TBR, and TAR, provide a more complete picture of glucose levels throughout the day and night. This may help empower people with diabetes to better manage their condition, giving them practical insights into the factors driving daily fluctuations in glucose levels, such as diet, exercise, insulin dosage, and insulin timing. These metrics may also be used to inform treatment decisions by healthcare professionals [3, 15, 16]. Ultimately, it is hoped that the use of these new metrics should lead to an improved quality of glycemic control and, in turn, to a reduction in the number of diabetes-related complications.
Our findings should be interpreted in the light of a few study limitations. This was a post hoc analysis of a crossover study intended to compare TIR outcomes between two basal insulins. The study was not designed to assess HbA1c as a primary outcome, hence the on-drug treatment periods of 18 weeks were relatively short. HbA1c values reflect what has occurred up to 90 days back in time, so they might not have fully stabilized when the real-time CGM data were collected in the maintenance periods. This study evaluated the correlation between TIR and HbA1c, two metrics used to describe the quality of glycemia. As noted previously, HbA1c does not capture day-to-day fluctuations of blood glucose levels in people with diabetes. The same is true for TIR, as both metrics reflect overall glycemia and do not account for variability in blood glucose levels. Also, as noted previously, the treat-to-target approach to insulin dose titration might also have limited the spread of the TIR, TAR and TBR data. Nevertheless, clear correlations were observed, consistent with previous studies, and the spread of the data in the scatterplots highlights the complementary nature of TIR and HbA1c in evaluating glycemic control.
This post hoc analysis of the SWITCH PRO trial expands the evidence base in support of the role of TIR, assessed using CGM, as a valid clinical indicator of glycemic control during therapeutic intervention. SWITCH PRO is one of the first large clinical studies to compare therapeutics using TIR as the primary outcome rather than HbA1c. The results presented here are in agreement with the increasing momentum behind the use of TIR as a key endpoint for assessment of glycemic control and prediction of the risk of diabetes-related complications in clinical trials [17,18,19], and that TIR may possibly emerge as a preferred metric to HbA1c for clinical management.
Suwa T, Ohta A, Matsui T, et al. Relationship between clinical markers of glycemia and glucose excursion evaluated by continuous glucose monitoring (CGM). Endocr J. 2010;57:135–40.
Dunn TC, Hayter GA, Doniger KJ, Wolpert HA. Development of the likelihood of low glucose (LLG) algorithm for evaluating risk of hypoglycemia: a new approach for using continuous glucose data to guide therapeutic decision making. J Diabetes Sci Technol. 2014;8:720–30.
Battelino T, Danne T, Bergenstal RM, et al. Clinical targets for continuous glucose monitoring data interpretation: recommendations from the International Consensus on Time in Range. Diabetes Care. 2019;42:1593–603.
Vigersky RA, McMahon C. The relationship of hemoglobin A1C to time-in-range in patients with diabetes. Diabetes Technol Ther. 2019;21:81–5.
Beck RW, Bergenstal RM, Riddlesworth TD, et al. Validation of time in range as an outcome measure for diabetes clinical trials. Diabetes Care. 2019;42:400–5.
Wysham C, Bhargava A, Chaykin L, et al. Effect of Insulin degludec vs insulin glargine U100 on hypoglycemia in patients with type 2 diabetes: the SWITCH 2 randomized clinical trial. JAMA. 2017;318:45–56.
Goldenberg RM, Aroda VR, Billings LK, et al. Effect of insulin degludec versus insulin glargine U100 on time in range: SWITCH PRO, a crossover study of basal insulin-treated adults with type 2 diabetes and risk factors for hypoglycaemia. Diabetes Obes Metab. 2021;23:2572–81.
Abbott Laboratories. FreeStyle Libre Pro System. 2021. https://provider.myfreestyle.com/freestyle-libre-pro-product.html. Accessed 17 May 2021.
World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA. 2013;310:2191–4.
The International Council for Harmonisation. ICH Harmonised Tripartite Guideline. Guideline for Good Clinical Practice E6 (R2). Current step 5 version. 2016. https://www.ema.europa.eu/en/ich-e6-r2-good-clinical-practice-scientific-guideline. Accessed 03 Jan 2023.
Beck RW, Bergenstal RM, Cheng P, et al. The relationships between time in range, hyperglycemia metrics, and HbA1c. J Diabetes Sci Technol. 2019;13:614–26.
Kazemian P, Shebl FM, McCann N, Walensky RP, Wexler DJ. Evaluation of the cascade of diabetes care in the United States, 2005–2016. JAMA Intern Med. 2019;179:1376–85.
DiMeglio LA, Kanapka LG, DeSalvo DJ, et al. Time spent outside of target glucose range for young children with type 1 diabetes: a continuous glucose monitor study. Diabet Med. 2020;37:1308–15.
Sandig D, Grimsmann J, Reinauer C, et al. Continuous glucose monitoring in adults with type 1 diabetes: real-world data from the German/Austrian Prospective Diabetes Follow-Up Registry. Diabetes Technol Ther. 2020;22:602–12.
Chehregosha H, Khamseh ME, Malek M, Hosseinpanah F, Ismail-Beigi F. A view beyond HbA1c: role of continuous glucose monitoring. Diabetes Ther. 2019;10:853–63.
Runge AS, Kennedy L, Brown AS, et al. Does time-in-range matter? Perspectives from people with diabetes on the success of current therapies and the drivers of improved outcomes. Clin Diabetes. 2018;36:112–9.
Bajaj HS, Bergenstal RM, Christoffersen A, et al. Switching to once-weekly insulin icodec versus once-daily insulin glargine U100 in type 2 diabetes inadequately controlled on daily basal insulin: a phase 2 randomized controlled trial. Diabetes Care. 2021;44:1586–94.
Battelino T, Bosnyak Z, Danne T, et al. InRange: comparison of the second-generation basal insulin analogues glargine 300 U/mL and degludec 100 U/mL in persons with type 1 diabetes using continuous glucose monitoring-study design. Diabetes Ther. 2020;11:1017–27.
Advani A. Positioning time in range in diabetes management. Diabetologia. 2020;63:242–52.
We thank Rizi E. Parvaresh, previously of Novo Nordisk, for his contribution to the analyses of data.
This study was sponsored by Novo Nordisk A/S. The Rapid Service Fee is also funded by Novo Nordisk.
Medical Writing and Editorial Assistance
Medical writing and editorial support for the development of this manuscript, under the direction of the authors, were provided by Jane Blackburn and Helen Marshall of Ashfield MedComms, an Ashfield Health company, and funded by Novo Nordisk A/S.
All authors confirm that they meet the International Committee of Medical Journal Editors (ICJME) uniform requirements for authorship and that they have contributed to the roles detailed below. Novo Nordisk was involved in the trial design and protocol development, provided logistical support, and obtained the data, which were evaluated jointly by the authors and the sponsor. Richard M. Bergenstal 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. All authors made substantial contributions to conception or design of the work, and to the acquisition, analysis, or interpretation of data. All authors were involved in the drafting and critical revision of the work for important intellectual content, and all approved the final version to be published.
This manuscript is based on work that has been previously presented in abstract form at the American Diabetes Association 2021 virtual meeting, 25–29 June, 2021 (abstract numbers 617-P and 618-P).
Ronald M. Goldenberg has received research support from Abbott, AstraZeneca, Boehringer Ingelheim, Eli Lilly, GlaxoSmithKline, Janssen, Medtronic, Merck, Novartis, Novo Nordisk, Roche, Sanofi, and Takeda; has served on advisory panels for AstraZeneca, Boehringer Ingelheim, Eli Lilly, HLS Therapeutics, Janssen, Merck, Novo Nordisk, Roche, Sanofi, and Takeda; has served on speaker bureaus for Abbott, AstraZeneca, Boehringer Ingelheim, Eli Lilly, HLS Therapeutics, Janssen, Merck, Novo Nordisk, Sanofi, and Servier; and reports consulting for AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen, Novo Nordisk, and Takeda. Vanita R. Aroda has in the past 3 years acted as a consultant for Applied Therapeutics, Fractyl, Novo Nordisk, Pfizer, and Sanofi; has had a spouse who is/was an employee of Merck Research Laboratories and Janssen; and has received research support (institutional contracts) from Applied Therapeutics/Medpace, Eli Lilly; Fractyl/Premier, Novo Nordisk, and Sanofi/Medpace. Liana K. Billings has served on advisory panels or consulting for Novo Nordisk, Eli Lilly, Sanofi, Bayer, and Xeris Pharmaceuticals. Anders Meller Donatsky, Marie Frederiksen, and Balamurali Kalyanam are employees of Novo Nordisk A/S. David C. Klonoff is a consultant to AI Health, Eoflow, Integrity, Lifecare, Medtronic, Novo Nordisk, Roche Diagnostics, and Thirdwayv. Richard M. Bergenstal has received research support, consulted for, or has been on a scientific advisory board for Abbott Diabetes Care, Ascensia, Bigfoot Biomedical, CeCur, DexCom, Lilly, Hygieia, Novo Nordisk, Onduo, Roche Diabetes Care, Sanofi, Medtronic, Zealand, and Insulet. His employer, non-profit HealthPartners Institute, contracts for his services and no personal income goes to Dr Bergenstal.
Compliance with Ethics Guidelines
Prior to initiation of the SWITCH PRO study, the protocol, consent form, and subject information sheet were reviewed and approved according to local regulations by appropriate health authorities, and by an independent ethics committee (IEC) and/or institutional review board (IRB). A list of the IECs/IRBs is provided in ESM Table S1. The SWITCH PRO study was conducted in accordance with ethical principles derived from international guidelines including the Declaration of Helsinki  and the International Conference on Harmonisation (ICH) Good Clinical Practice . All participants gave their informed consent prior to inclusion in the study.
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.
Below is the link to the electronic supplementary material.
About this article
Cite this article
Goldenberg, R.M., Aroda, V.R., Billings, L.K. et al. Correlation Between Time in Range and HbA1c in People with Type 2 Diabetes on Basal Insulin: Post Hoc Analysis of the SWITCH PRO Study. Diabetes Ther 14, 915–924 (2023). https://doi.org/10.1007/s13300-023-01389-2
- Professional CGM
- Long-acting basal insulin
- Type 2 diabetes
- Basal insulins
- Insulin analogs
- Insulin treatment