To the Editor: Glycemic control in individuals with type 1 diabetes (T1D) is challenging and requires multidimensional evaluation and precise management. Glycated hemoglobin (HbA1c) is currently the gold standard for assessing glucose control and predicting diabetes prognosis. However, it exhibits limitations in terms of accuracy, which is influenced by clinical factors,[1] and its inability to reflect information about hypoglycemia and glycemic variability.[2] Continuous glucose monitoring (CGM) has revolutionized glycemic profile evaluation by introducing novel glycemic variables that surpass conventional parameters and offer insights into critical areas for improving glycemic control.

The glycemic risk index (GRI) is a novel glycemic metric derived from CGM that evaluates the risk of abnormal blood glucose levels by simultaneously reflecting both hyperglycemia and hypoglycemia risks.[3] Given the close association between hypoglycemic events and more severe clinical outcomes,[4,5] GRI particularly emphasizes the risk of hypoglycemia by assigning higher weight to its hypoglycemic components. In comparison with the risk of hyperglycemia, the GRI assigns nearly double the weight to hypoglycemia risk, with the risk of very low glucose levels (<3.0 mmol/L) being almost four times that of high glucose levels (>10.0–13.9 mmol/L). These notable features distinguish the GRI from HbA1c, enabling it to capture aspects that HbA1c cannot. However, whether GRI can address the limitations of HbA1c remains uncertain. Therefore, the primary objective of this study was to explore the implications of the GRI at different HbA1c levels and to investigate whether their combined use can offer enhanced guidance for clinical glucose management.

This study was approved by the Ethics Committee of the Second Xiangya Hospital of Central South University, and all patients provided signed informed consent (Approval No. 2019-198). The source cohort was registered at ClinicalTrials.gov (ID: 146 NCT03610984) before commencing the study. The inclusion and exclusion criteria and statistical analyses are shown in Supplementary Materials, https://links.lww.com/CM9/B863. Normally distributed variables (assessed by the Kolmogorov-Smirnov test) were presented as mean ± SD, skewed variables as median (first quartile [Q1], third quartile [Q3]), and categorical data were expressed as numbers and percentages.

A total of 281 patients aged >4 years with T1D duration exceeding 3 months were included in the study. The patients wore a flash glucose monitoring (FGM) system (Freestyle Libre; Abbott Diabetes Care, Rome, Italy) for a minimum of 14 days. It is worth noting that the analyzed dataset comprised 507 instances from CGM and clinical records, with 135 patients undergoing 1–5 follow-up visits [Supplementary Figure 1, https://links.lww.com/CM9/B863]. At each patient visit, we collected age, sex, diabetes duration, height, weight, body mass index (BMI), blood pressure, insulin administration method, daily insulin dosage (U·kg–1·day–1), and CGM data including time in range (TIR, 3.9–10.0 mmol/L), level 1 time below range (TBR) (TBR1, 3.0 mmol/L to <3.9 mmol/L), level 2 TBR (TBR2, <3.0 mmol/L), level 1 Time above range (TAR) (TAR1, >10.0 mmol/L to 13.9 mmol/L), and level 2 TAR (TAR2, >13.9 mmol/L). Venous blood samples were collected to measure HbA1c, fasting C-peptide (FCP), and 2 h C-peptide (2hCP) levels after a mixed-meal tolerance test (MMTT).

Of the 281 patients, 117 (41.6%) were male, and 201 (71.5%) experienced onset during childhood or adolescence, with a median age at onset of 10.9 (6.4, 20.0) years. The analysis of these 507 medical records revealed a median disease duration of 2.1 (1.0, 3.8) years. The patients’ ages ranged from 4 years to 72 years, and medical records of these patients comprising 176 (34.7%) adults and 331 (65.3%) children or adolescents visits. The mean values of HbA1c and TIR, indicative of glycemic control, were 7.3 ± 1.1% and 65.6% (54.4%, 76.8%), respectively. Less than half of T1D patients achieved the blood glucose target, with HbA1c <7% in 40.4% (205/507) visits and TIR >70% in 37.7% (191/507) visits. These results highlight the challenges in controlling blood glucose levels in patients with T1D and emphasize the need for more comprehensive glucose monitoring and management.

The GRI was calculated using the equation presented as following: GRI = 3.0 × hypoglycemic index (TBR2 + 0.8 × TBR1) + 1.6 × hyperglycemic index (TAR2 + 0.5 × TAR1). The mean GRI of all the T1D patients was 57.6 ± 20.6. More than half (53.3%, 270/507) of the records had a GRI score of 60 [Supplementary Table 1, https://links.lww.com/CM9/B863]. A lower GRI was observed in adults with T1D as well as in patients with a short disease duration, high C-peptide, and low HbA1c levels. Patients with HbA1c <7% exhibited significantly lower GRI values than those with HbA1c ≥7% (45.4 ± 20.2 vs. 65.9 ± 16.4, P <0.001). However, there was no statistically significant difference in the GRI between patients treated with continuous subcutaneous insulin infusion (CSII) and multiple daily insulin injection (MDI) [Supplementary Table 2, https://links.lww.com/CM9/B863].

To clarify the relationship between the GRI and glucose control parameters, such as HbA1c and CGM-derived metrics, we conducted a correlation analysis. The results revealed a positive correlation with HbA1c, TAR, TBR, and mean glucose (MG) and a negative correlation with TIR. The strongest correlation was found between the GRI and TIR (R = –0.896), with a moderate correlation between the GRI and HbA1c (R = 0.533) [Figure 1]. The correlation formula between GRI and TIR was Y = –1.16X + 132.59 (X = TIR, Y = GRI), indicating that at TIR = 70%, GRI = 51.4. The correlation formula between GRI and HbA1c was Y = 10.16X – 16.75 (X = HbA1c, Y = GRI), indicating that at HbA1c = 7%, GRI = 54.4. Based on these findings, we recommend a GRI threshold of 55.

Figure 1:

Relationship between GRI and other glycemic parameters by HbA1c levels, including HbA1c (A), TIR (B), TBR (C), hypoglycemic index (D), TAR (E), and hyperglycemic index (F). The red dot and blue dot represent the individuals with HbA1c <7% and HbA1c ≥7%, respectively. The purple line represents the correlation between GRI and other glycemic parameters for all participants in this study, while the red and blue lines represent the correlation for individuals with HbA1c <7% and HbA1c ≥7%, respectively. GRI: Glycemic risk index; HbA1c: Glycated hemoglobin A1c; TAR: Time above range; TBR: Time below range; TIR: Time in range.

Subgroup analysis based on an HbA1c threshold of 7% revealed varying correlation coefficients between the GRI and different glycemic parameters across patients with different HbA1c levels. In patients with HbA1c <7%, GRI showed no correlation with HbA1c or MG. The correlation coefficients between GRI and TIR, TBR, and the hypoglycemic index were higher in the HbA1c <7% group, while those between GRI and TAR and the hyperglycemic index were higher in the HbA1c ≥7% group [Figure 1 and Supplementary Table 3, https://links.lww.com/CM9/B863].

The HbA1c, GRI, and TIR among the analyzed patients with T1D were only partially aligned to achieve the control target ranges [Figure 1]. Although the GRI exhibited high agreement rates of 72.2% (148/205) with HbA1c and 93.7% (179/191) with TIR, the agreement rate decreased to 69.2%(132/191) when all three parameters were considered. Intriguingly, MG index did not differ in patients with HbA1c <7% regardless of GRI being ≥55 or <55. Moreover, the TIR in patients with GRI ≥55 was significantly lower than that in patients with GRI <55 (65.5% [59.7%, 68.9%] vs. 80.7% [75.4%, 86.6%], P <0.001) for those with HbA1c <7%. These results suggest that the interpretation of glycemic status via the GRI is not entirely equivalent to that of HbA1c or TIR.

To elucidate the implications of GRI at different HbA1c levels, the participants were categorized into four groups based on HbA1c and GRI status: (1) Dual-target group: HbA1c <7% and GRI <55; (2) Sole HbA1c-target group: HbA1c <7% and GRI ≥55; (3) Sole GRI-target group: HbA1c ≥7% and GRI <55; (4) Neither-target group: HbA1c ≥7% and GRI ≥55. The results are presented in Supplementary Table 4, https://links.lww.com/CM9/B863. The dual-target group exhibited the highest TIR (80.7% [75.4%, 86.6%]) and the lowest hyperglycemia index (6.3% [3.9%, 10.3%]), while the neither-target group had the lowest TIR (52.6% [43.0%, 60.2%]) and highest hyperglycemia index (26.7% [18.4%, 36.3%]). The group exclusively meeting the HbA1c target displayed the highest hypoglycemia index (11.0% [5.8%, 15.7%]), whereas the group exclusively meeting the GRI target had the lowest hypoglycemia index (3.3% [1.8%, 5.7%]). Interestingly, the sole GRI-target group, characterized by lower GRI levels, exhibited a higher TIR (71.5% [65.8%, 77.4%] vs. 65.5% [59.7%, 68.9%], P <0.001) despite having higher HbA1c levels compared with the sole HbA1c-target group. The GRI-target group also demonstrated a higher hyperglycemia index (13.9% [10.3%, 19.2%]) than the HbA1c-target group (11.0% [5.8%, 15.7%]), with significantly higher TAR1, while TAR2 remained similar in both groups, indicating that the group solely meeting the GRI target was predominantly affected by mild hyperglycemia. Based on these findings, the combination of HbA1c and the GRI can offer a more comprehensive evaluation of blood glucose levels, facilitating the management and regulation of blood glucose control.

In conclusion, the assessment of glycemic control has evolved to encompass a broader array of multi-dimensional parameters, and the GRI unquestionably enhances this perspective. This study explores the utility of the GRI as a new CGM metric in patients with T1D. Our findings showed a linear correlation between the GRI and both HbA1c and TIR, highlighting the multifaceted implications of the GRI across distinct HbA1c levels in T1D. Specifically, GRI exhibited a pronounced connection with hypoglycemic risk when HbA1c <7%, and a connection with hyperglycemic risk when HbA1c ≥7%. Combined with HbA1c, GRI provides a more comprehensive and accurate assessment of blood glucose management in patients with T1D.

Acknowledgements

The authors thank all of the patients, nurses, doctors and technicians involved at National Clinical Research Center for Metabolic Diseases for their efforts in data and sample collection.

Data Availability

The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Funding

This study was supported by the grants from the National Key R&D Program of China (No. 2022YFC2010102), the Natural Science Foundation of Hunan Province (No. 2021JC0003), and Sinocare Diabetes Foundation (No. 2020SD08).

Conflicts of interest

None.

References 1. Wright LA, Hirsch IB. Metrics beyond hemoglobin A1C in diabetes management: Time in range, hypoglycemia, and other parameters. Diabetes Technol Ther 2017;19:S16–S26. doi: 10.1089/dia.2017.0029. 2. Haynes A, Hermann JM, Miller KM, Hofer SE, Jones TW, Beck RW, et al. Severe hypoglycemia rates are not associated with HbA1c: A cross-sectional analysis of 3 contemporary pediatric diabetes registry databases. Pediatr Diabetes 2017;18:643–650. doi: 10.1111/pedi.12477. 3. Klonoff DC, Wang J, Rodbard D, Kohn MA, Li C, Liepmann D, et al. A glycemia risk index (GRI) of hypoglycemia and hyperglycemia for continuous glucose monitoring validated by clinician ratings. J Diabetes Sci Technol 2023;17:1226–1242. doi: 10.1177/19322968221085273. 4. Muneer M. Hypoglycaemia. Adv Exp Med Biol 2021;1307:43–69. doi: 10.1007/5584_2020_534. 5. International Hypoglycaemia Study Group. Hypoglycaemia, cardiovascular disease, and mortality in diabetes: Epidemiology, pathogenesis, and management. Lancet Diabetes Endocrinol 2019;7:385–396. doi: 10.1016/S2213-8587(18)30315-2.