A Comparison of Daily Glucose Fluctuation Between GCK-MODY and Type 2 Diabetes Using Continuous Glucose Monitoring Technology Free

Glucokinase variant-induced maturity-onset diabetes of the young (GCK-MODY), discovered nearly three decades ago (1–4), is one of the most common subtypes of MODY (5). Patients with GCK-MODY exhibit mild, stable hyperglycemia and a very low risk of chronic diabetic complications (6,7). A suggested explanation for the low risk of complications is mild hyperglycemia resulting from a defect in the glucose sensor of glucokinase (8–10). In type 2 diabetes, the time in range (TIR) generated from continuous glucose monitoring (CGM) is negatively associated with diabetic retinopathy (11). However, it is unclear whether patients with GCK-MODY have a higher TIR than type 2 diabetes mellitus (T2DM) patients with equivalent A1C.

The 2-h glucose increments (2-h GI) during oral glucose tolerance test (OGTT) are small in GCK-MODY (12,13). However, the 2-h GI cannot completely and precisely reflect glucose fluctuations. To date, the quantitative description of glucose excursion in the daily life of GCK-MODY patients has not been investigated. Consequently, whether GCK-MODY patients have a lower glucose excursion than T2DM patients with equivalent A1C remains unclear. In 2019, the Advanced Technologies & Treatments for Diabetes consensus recommendations standardized the CGM data utilization and report in diabetes (14). In this study, we aimed to describe natural glucose features in GCK-MODY by using professional-mode flash glucose monitoring (FGM) data.

We conducted a cross-sectional case-control study at Peking University People’s Hospital in China between June 2020 and October 2022. Twenty-five patients with GCK-MODY were enrolled in the study. Patients were diagnosed by DNA sequencing between 2016 and 2021 and were followed up in our clinic after diagnosis. The mutations are shown in Supplementary Table 1, and the pathogenicity of these variants was in line with the American College of Medical Genetics and Genomics guidelines (15) as described in Supplementary Tables 1 and 2 and Supplementary Fig. 1.

Twenty-five A1C-matched subjects with T2DM were selected from our clinic (1:1). If the difference in A1C between GCK-MODY and T2DM was within ±0.2%, it was considered to be matched. All of the matched T2DM patients were newly diagnosed and drug naive. GCK gene sequencing was performed to exclude possible GCK-MODY. The criteria for the diagnosis of diabetes were in accordance with the American Diabetes Association guidelines for 2021 (16).

In addition, we collected data from 25 age-, BMI-, and sex-matched subjects with normal glucose tolerance (NGT) for GCK-MODY (1:1). The inclusion and exclusion criteria are described in the Supplementary Material.

This study was approved by the Ethics Committee of Peking University People’s Hospital (Beijing, People’s Republic of China), and informed consent was obtained from all participants.

To reveal the glucose profile in GCK-MODY, we chose a professional-mode FGM system (Freestyle Libre H; Abbott Diabetes Care, Witney, U.K.). All glucose readings were masked to avoid glucose monitoring–driven behavior modification during the 2-week monitoring period. All participants performed normal daily activities during FGM. A1C was tested using a Premier Hb9210 Automatic glycosylated hemoglobin analyzer (Trinity Biotech, Wicklow, Ireland) using high-performance liquid chromatography.

Metrics of average glucose and glucose excursions were generated from FGM data, including mean blood glucose (MBG), glucose management indicator (GMI), the maximum and minimum of glucose (MAX and MIN), coefficient of variation (CV), mean amplitude of glycemic excursions (MAGE), standard deviation of blood glucose (SDBG), mean of daily difference (MODD), average daily risk range (ADRR), largest amplitude of glycemic excursions (LAGE), high blood glucose index (HBGI), low blood glucose index (LBGI), TIR (glucose within 3.9–10 mmol/L), time above range (TAR) (glucose ≥ 10 mmol/L), and time below range (TBR) (glucose <3.9 mmol/L). Daytime and nighttime were defined as the time intervals from 6:00 a.m. to 11:59 p.m. and 12:00 a.m. to 5:59 a.m., respectively.

Continuous variables were expressed as mean ± SD with normal distributions and otherwise as medians and interquartile ranges. Nonnormally distributed variables were logarithmically transformed prior to analysis. The Student’s t test was used to compare variables with a normal distribution between the groups. The χ² test was used to analyze categorical variables. Logistic regression analysis was used to calculate odds ratios (ORs), 95% CI, and P values, with the diagnosis of GCK-MODY as the dependent variable after adjusting for age, sex, and BMI. All statistical analyses were performed using SPSS version 23 for Windows (17). Statistical significance was set at P < 0.05.

Receiver operating characteristic (ROC) curve analysis was used to evaluate the capability of glucose excursion metrics to discriminate between GCK-MODY and T2DM. The optimal cutoff values of the glucose excursion metrics were determined according to the Youden index. ROC curves were compared between the glucose excursion parameters. MedCalc Statistical Software version 19.0 (18) was used for the ROC curve analysis.

Data and resources are available upon request.

We enrolled 25 GCK-MODY patients, 25 A1C-matched patients with T2DM, and 25 age-, BMI- and sex-matched NGT subjects. The clinical characteristics are shown in Table 1. GCK-MODY patients were followed up in our clinic for 2.1 ± 1.9 years. None of them had diabetic retinopathy or diabetic kidney disease. In the T2DM group, 23 patients (92%) agreed to undergo genetic testing, and did not carry GCK mutations. The main clinical features of the two patients without genetic testing were not compatible with GCK-MODY. One patient had hyperinsulinemia and responded well to metformin treatment. The other patient had a high OGTT 2-h glucose (19.24 mmol/L) and macrovascular complication as shown in Supplementary Material.

FGM was successfully conducted, with an average effective data ratio of 99.9 ± 0.59%. The ambulatory glucose profile (AGP) is shown in Fig. 1A–C. The FGM metrics showed that the CV was 22.26 ± 4.74% (<36% reference target) (14) and TIR was 93.00% (90.00–96.50%) in GCK-MODY. We also merged the AGP of GCK-MODY, T2DM, and NGT to enhance visual comparison (Fig. 1D).

Overall, MBG and GMI were similar between GCK-MODY and T2DM, and the glucose excursion was significantly lower in the GCK-MODY group than that in the T2DM group (Table 1 and Supplementary Fig. 3). Next, we evaluated the FGM glucose distribution during the day and night. ADRR, CV, SDBG, MAGE, MODD, and LAGE were significantly lower in GCK-MODY patients than those in T2DM patients during the daytime, but all the difference disappeared during the nighttime (Table 2). SDBG, MAGE, MODD, and LAGE were significantly higher in the GCK-MODY patients than those in the NGT patients, regardless of daytime or nighttime. However, CV was an exception, as it was similar between GCK-MODY and NGT throughout the day (Table 2 and Supplementary Fig. 3).

Logistic regression analysis revealed that CV (OR 0.727, 95% CI 0.593–0.893, P = 0.002), SDBG (OR 0.077, 95% CI 0.010–0.624, P = 0.016), MAGE (OR 0.343, 95% CI 0.150–0.784, P = 0.011), MODD (OR 0.065, 95% CI 0.006–0.752, P = 0.029), ADRR (OR 0.537, 95% CI 0.301–0.957, P = 0.035), and TIR (OR 1.098, 95% CI 1.001–1.205, P = 0.048) were independent predictive factors for GCK-MODY after adjusting for age, sex, and BMI. Multivariable logistic regression showed that CV was the most important flash glucose parameter (OR 0.604, 95% CI 0.398–0.916, P = 0.018).

Next, we evaluated six FGM-based fluctuation parameters (CV, SDBG, MAGE, MODD, ADRR, and TIR) to differentiate GCK-MODY from T2DM. These parameters performed well with an area under the curve from 0.758 to 0.875 (Table 3 and Supplementary Fig. 4). Considering both sensitivity and specificity, we combined the cutoff values of the six fluctuation parameters and found that fulfilling any four of them was the best way to distinguish GCK-MODY from T2DM, yielding a sensitivity of 80% and specificity of 92%.

In our study, the 25 GCK-MODY patients came from 12 families. In order to avoid the influences of relatedness, an additional subgroup analysis was done in GCK-MODY probands (n = 12). Similarly, glucose excursion was also significantly lower in the GCK-MODY probands than that in T2DM (Supplementary Table 3).

In this study, we adopted a case-control design and used FGM for 14 days to comprehensively examine the glucose profile of GCK-MODY patients during their daily lives. Previously, the glucose features of GCK-MODY were mainly evaluated using OGTT (19), intravenous glucose tolerance test (IVGTT) (20), or glucose clamps combined with a strict diet (20). A1C was used during the long-term follow-up but provided no information on glycemic variability (7,21). Only two recent case reports have been published on GCK-MODY using FGM (22). To our knowledge, this is the first case-control study to evaluate the AGP of GCK-MODY.

GCK plays a key role as a “glucose sensor” in maintaining the glucose homeostasis. During physiological processes, an increase in glucose activates GCK, which promotes the conversion of glucose to glucose-6-phosphate. ATP-sensitive K⁺ channels are closed, voltage-gated Ca²⁺ channels are open, and insulin secretion is stimulated (23). Heterozygous inactivating mutations in GCK alter its kinetic parameters, resulting in an upward shift in fasting blood glucose (FBG) (24).

Although MBG and GMI were similar for GCK-MODY and T2DM, all standard glucose excursion indices were significantly lower in GCK-MODY. More importantly, by using daytime-nighttime analysis, we found that most of the excursion differences between GCK-MODY and T2DM occurred in the daytime. This suggests that postprandial regulation is a key driver of the excursion differences between GCK-MODY and A1C-matched T2DM. In addition, although glucose excursion was low in GCK-MODY, most of the excursion indices were still not as good as NGT, regardless of the daytime or nighttime. Also, it is worth noting that CV, the most discriminatory outcome between GCK-MODY and T2DM, was similar between GCK-MODY and NGT during both daytime and nighttime.

Previously, several attempts were made to use CGM data to distinguish between different glucose statuses and subtypes of diabetes (25). In this study, we found that fulfilling any four of the cutoff values in the six fluctuation parameters was the best way to distinguish GCK-MODY from T2DM, with a sensitivity of 80% and specificity of 92%. The published clinical screening criteria for GCK-MODY (fasting plasma glucose 5.4–8.3 mmol/L, A1C 5.8–7.6%, and 2-h GI <4.6 mmol/L) demonstrated a sensitivity of 71.4% and a specificity of 81.2% (26). However, more evidence is needed before comparing the predictive value of FGM parameters with FBG and A1C, because of the small sample size and correspondingly wide CIs in this study. In addition, we do not recommend screening GCK-MODY patients by FGM at present, because FGM is more expensive than the current sequencing price, and DNA sequencing is the gold standard for the diagnosis of GCK-MODY. The value of FGM lies in the GCK-MODY phenotypic confirmation. Online bioinformatic tools are commonly used to classify pathogenic mutations (10). In addition, phenotypes such as the MODY calculator are important tools (27). Phenotypic characteristics based on the FGM profile could also provide evidence in evaluating the pathogenicity of new GCK mutations.

This study has several limitations. First, the sample size was relatively small, and we could not assign the study population into discovery and validation cohorts. Moreover, because collecting more individuals with GCK-MODY within a short time is difficult, we hope that the rapid publication of our present study results will help inspire other researchers to confirm our findings, as evaluating the predictive model in a validation cohort is important. Second, in A1C-matched T2DM, two patients (8%) refused to test GCK gene sequence, but their clinical characteristics (Supplementary Material) did not support a GCK-MODY feature (26); therefore, its influence was limited. Third, GCK-MODY and T2DM were matched on A1C, but not on age, BMI, or age of diagnosis. Age and age of diagnosis were closely related in this study population. Therefore, to minimize the impact on FGM parameters, we adjusted for age and BMI in the logistic analysis. Fourth, genetic sharing was observed in the GCK-MODY group. Relatedness may reduce the effective sample size in interpreting statistical inferences. So, in this study, we added subgroup analysis in GCK-MODY probands. The results of this analysis were consistent with those of the total study population. Fifth, we did not compare FGM metrics among different subtypes of MODY, which may raise the issue of the specificity of the glucose excursion feature in GCK-MODY. In fact, another common subtype of MODY, MODY3, exhibits larger glucose fluctuations (19).

In summary, the higher TIR and lower glucose excursion shown by FGM confirmed the stability and mildness of hyperglycemia in GCK-MODY. Postprandial regulation is a key driver of the excursion difference between GCK-MODY and A1C-matched T2DM.

Acknowledgments. The authors thank the patients and healthy volunteers for agreeing to participate in this study. The authors also thank Yunyang Sun and Lin Lou (both affiliated with Shanghai iMedpower Tech. Ltd.) for their work in FGM data analysis.

Funding. X.H. has received grants from Beijing Municipal Commission of Science and Technology (grants Z201100005520012 and D131100005313008) (clinical data collection, patient recruitment), and L.J. has received grants from the National Key Research and Development Program (grant 2016YFC1304901) (genetic testing, patient recruitment).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. All authors have contributed significantly. Q.R. collected and organized the data, developed the analysis plan, conducted the statistical analyses, and prepared the initial draft of the manuscript. P.Z. and W.P. collected and organized the data. Y.M., S.G., and T.B. conducted the statistical analyses. W.L., F.Z., Xiu.Z., R.Z., X.W., Xia.Z., W.Y., F.L., and L.G. collected and organized all the data. X.H. developed the analysis plan. X.H. and L.J. conceived the idea of the study and were responsible for the design of the study. The initial draft of the manuscript was circulated amongst all authors for critical revision. All authors are in agreement with the content of the manuscript and had access to all the data in this study. L.J. 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.