Continuous glucose monitoring to support the diagnosis of MODY: A multicenter cross-sectional and prospective study

Abstract

Aim

Maturity onset diabetes of the young (MODY) is often misdiagnosed as type 1 diabetes (T1D). We proposed using continuous glucose monitoring (CGM) data as an additional tool to select patients in whom the fulfillment of classical MODY criteria should be reassessed, who may then be considered for genetic testing.

Method

A multicentric cross-sectional and prospective study was designed to evaluate the clinical utility of CGM in the diagnosis of MODY among adult patients initially diagnosed as T1D with active CGM data. Pre-specified CGM criteria included: glucose management indicator <7%, time in range (70–180 mg/dL [3.9–10.0 mmol/L]) >70% and coefficient of variation <36%. Those meeting these requirements were assessed for classical clinical criteria and subsequently genetic testing.

Results

A total of 503 subjects with T1D out of 5571 fulfilled the pre-established glycometric requirements. After reviewing their medical records, 42 of them met the classic clinical criteria for MODY and genetic testing was performed in 34 patients. Finally, we found 5 new cases of MODY (3 patients with HNF1A-MODY3, 1 patient with HNF1B-MODY5 and 1 patient with ABCC8-MODY12) and detected 3 patients with heterozygous mutations associated with the development of diabetes.

Conclusion

CGM added to classic clinical criteria for MODY may constitute an effective and easily implemented approach in routine clinical practice to identify patients with MODY who have been misdiagnosed as having T1D.

Resumen

Objetivo

La diabetes del adulto de inicio juvenil (MODY) suele diagnosticarse erróneamente como diabetes tipo 1 (DT1). Proponemos utilizar los datos de monitorización continua de glucosa (MCG) como herramienta adicional para seleccionar a los pacientes en quienes debería reevaluarse el cumplimiento de los criterios clásicos de MODY, y que posteriormente podrían ser considerados para estudio genético.

Método

Se diseñó un estudio multicéntrico, transversal y prospectivo para evaluar la utilidad clínica de la MCG en el diagnóstico de MODY en adultos inicialmente diagnosticados de DT1 con datos activos de MCG. Los criterios preespecificados de MCG incluyeron: indicador de gestión de glucosa (GMI) <7%, tiempo en rango (70–180 mg/dL [3,9–10,0 mmol/L]) >70% y coeficiente de variación <36%. A quienes cumplían estos requisitos se les evaluaron los criterios clínicos clásicos y, posteriormente, se les realizó estudio genético.

Resultados

Un total de 503 sujetos con DT1 de entre 5.571 cumplieron los requisitos glucométricos preestablecidos. Tras revisar sus historias clínicas, 42 de ellos cumplían los criterios clínicos clásicos de MODY, realizándose estudio genético en 34 pacientes. Finalmente, identificamos cinco nuevos casos de MODY (tres pacientes con HNF1A-MODY3, un paciente con HNF1B-MODY5 y un paciente con ABCC8-MODY12) y detectamos tres pacientes con mutaciones heterocigotas asociadas al desarrollo de diabetes.

Conclusión

La MCG, añadida a los criterios clínicos clásicos de MODY, podría constituir un enfoque eficaz y fácilmente implementable en la práctica clínica habitual para identificar pacientes con MODY previamente diagnosticados erróneamente como DT1.

Introduction

The historical term maturity-onset diabetes of the young (MODY) encompasses a set of inherited disorders of non-autoimmune diabetes mellitus (antibody-negative) present at a young age of onset. Nowadays, experts prefer simply using the affected gene name for clarity in referring to these specific monogenic diabetes forms with beta-cell dysfunction.^1–4 Although we do not know the exact prevalence of the clinical forms that form part of the monogenic diabetes group, different studies place their prevalence between 1% and 4% depending on the population studied.⁵ However, most of the patients suffering from monogenic diabetes remain undiagnosed due to being misdiagnosed as having other types of diabetes, including type 1 diabetes (T1D). In fact, it is estimated that at least 80% of all monogenic cases of diabetes remain undiagnosed.⁶

The clinical phenotype relies on genetic defects, with important implications in the optimal treatment and prognosis definition. MODY's diagnosis remains a challenge, since this group of inherited disorders comprises a large clinical spectrum and it usually overlaps with other types of diabetes, requiring a high index of suspicion even if the definitive statement demands a molecular genetic study.^7,8 Following the traditional approach, monogenic diabetes should be considered in those patients with one or more of the following features: age at diagnosis of less than 35 years, HbA1c <7.5% (58 mmol/mol) at diagnosis, one parent with diabetes, and features of specific monogenic cause (e.g., renal cysts, partial lipodystrophy, maternally inherited deafness, severe insulin resistance in the absence of obesity).⁹

On the other hand, continuous glucose monitoring (CGM) has been incorporated in recent years into routine clinical practice in different subpopulations of people with diabetes. In Spain, CGM has been publicly funded in the treatment of patients with T1D since 2019, and its use is widespread.¹⁰ CGM offers advantages over self-monitoring of capillary blood glucose as it provides information on glucose values 24 h a day on a continuous basis.^11,12 In addition, it makes available interesting data on glycaemic variables that can be assessed following the International Consensus on Time in Range. This consensus also recommends specific glycometric goals for T1D patients such as time in range of interstitial glucose between 70 and 180 mg/dL (3.9–10.0 mmol/L) >70%, glucose management indicator (GMI) <7% and coefficient of variation (CV) <36%, to name but a few.¹³

Given the limited available data on CGM patterns and outcomes in individuals with monogenic diabetes, a recent study evaluated the utility of CGM in distinguishing patients with MODY from those with other forms of diabetes. Specifically, the CV driven by postprandial glucose could separate glucokinase-MODY from well-controlled type 2 diabetes patients.¹⁴ Therefore, CGM could play a role in the diagnosis of individuals with MODY. In the present study we aim to assess the potential utility of CGM as an additional tool in an enhanced diagnosis pathway for MODY among adults initially diagnosed with T1D.

Material and method

Study design

The study was approved by the local Ethics Committee (Castilla-La Mancha Public Health Service, SESCAM, Spain) and conducted in accordance with the Declaration of Helsinki and Good Clinical Practice, and publicly registered at ClinicalTrials.gov (NCT05918484). Participants provided written informed consent before study activities commenced. This multicenter cross-sectional observational study was conducted in the Departments of Endocrinology and Nutrition of Albacete University Hospital, Ciudad Real General University Hospital, Guadalajara University Hospital, Toledo University Hospital, Virgen de la Luz University Hospital (Cuenca) and Virgen de la Victoria University Hospital (Malaga), Spain. The reporting of this study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies.

The study was designed to assess the value of integrating a combination of predefined CGM criteria as an initial step to identify patients potentially misdiagnosed with T1D and to reassess their fulfilment of classical MODY criteria, ultimately guiding mandatory genetic testing when clinically indicated. It is a three-phase, cross-sectional and prospective, study:

• (1)

  First phase: included all adult T1D patients with active CGM data (≥70% of the possible time of use) seen in the participating hospitals. Among them, those with TIR >70%, GMI <7% and CV <36% were selected for the second phase. In the absence of any CGM criteria for the degree of glycaemic control or for diagnosis in patients with MODY, these glycaemic targets for people with T1D were used.¹³

• (2)

  Second phase: clinical data of suspected MODY (diagnosis before 35 years of age, first-degree family history of diabetes, negative pancreatic autoimmunity, preserved pancreatic beta-cell function) were reviewed.¹⁵ Those patients who met all of the above criteria proceeded to the next phase.

• (3)

  Third phase: given the previous clinical suspicion, a targeted genetic study for MODY was performed. An additional specific informed consent form was signed for patients who underwent genetic study.

Participants

Initial eligible participants were all adults (≥18 years) with T1D treated with CGM and active data from the last 14-day period. Patient suffering from other types of diabetes than T1D or <70% of time of sensor use were initially excluded. Additional exclusion criteria were T1D diagnosis in the last three years and current or programmed pregnancy.

Procedures

CGM had been initiated in T1D patients according to the National Public Health Service reimbursement procedure. FreeStyle Libre (Abbott Inc, IL, USA) was the CGM device used in all patients because it was the only publicly reimbursed CGM device at the time of the study. CGM data were downloaded from the Libreview website of each participating centre.

Digital medical records were reviewed for the aforementioned classic MODY criteria. Pancreatic autoimmunity was considered to be absent when the following auto-antibodies were negative: glutamic acid decarboxylase (GAD), islet tyrosine phosphatase 2 (IA2) and zinc transporter 8 (ZNT8). Pancreatic reserve was considered to be present if C-peptide levels were >0.2 ng/mL (0.07 nmol/L) in the presence of plasma glycaemia >140 mg/dL (7.8 mmol/L).¹⁶

Finally, the detection of MODY gene mutations was achieved through the utilization of next-generation sequencing (NGS) techniques and direct sequencing for validation of specific cases, both of which exhibit a sensitivity level approaching one hundred percent.

Outcomes

The primary objective was to determine the utility of using CGM data (combination of the glycometric variables TIR >70%, GMI <7% and CV <36%) as a first-step approach to help identify individuals who may have been misdiagnosed with T1D, reassess their adherence to classical MODY criteria, and inform the need for mandatory genetic testing when clinically justified. Secondary objectives were to assess the percentage of patients with MODY with an erroneous diagnosis of T1D, and to establish the degree of glycaemic control using glycometric variables among patients finally diagnosed as MODY.

Statistical analysis

Results are presented as mean (SD) values or percentages. Comparisons between proportions were analyzed using a chi-squared test and Fisher's exact test when this correction was necessary. For comparisons of quantitative variables between groups, normality was assessed using both the Kolmogorov–Smirnov and Shapiro–Wilk tests. If the data followed a normal distribution, an independent t-test was performed; otherwise, the Mann–Whitney U test was used. Exact 95% confidence intervals (CIs) were calculated for the main proportion using the Clopper–Pearson method. A p-value <0.05 was considered statistically significant (95% CI). Statistical analyses were conducted using RStudio (R version 4.4.1).

Results

Subjects

Among 5571 adults with T1D and CGM active data initially included in the cross-sectional study, we detected 503 fulfilling the aforementioned pre-established glycometric requirements. Of these 503 pre-selected patients, 52.7% were men, diabetes duration was 21.2 (14.9) years and total insulin dose was 0.5 (0.2) IU/kg. Patients’ flowchart and baseline characteristics can be observed in Fig. 1 and Table 1, respectively.

Fig. 1.

Patients’ flowchart.

Table 1.

Baseline characteristics of patients fulfilling glycometric criteria (n = 503).

Sex (%)	Male 52.7/female 47.3
Age at diabetes diagnosis (yrs.)	27.7 (14.7)
Current age (yrs.)	49.0 (14.3)
Diabetes duration (yrs.)	21.2 (14.9)
Total daily insulin dose (UI/kg)	0.5 (0.2)
Basal insulin (%)	57.5 (17.0)
Bolus insulin (%)	42.5 (17.0)
BMI (kg/m^2)	25.2 (4.6)
Patients with chronic diabetes complications (%)	27.2

Results are presented as mean (SD) values or percentages. BMI, body mass index.

After reviewing their medical records, we identified that 42 of them met the classical clinical criteria for MODY and were offered to perform a genetic study. Eight patients refused to undergo the genetic study. Hence, genetic testing was performed during the prospective study phase in 34 patients. Finally, we found five new cases of MODY (3 patients with HNF1A-MODY3, 1 patient with HNF1B-MODY5 and 1 patient with ABCC8-MODY12) among the 34 patients who underwent genetic testing, representing a detection rate of 14.7% (95% CI: 5.0%–31.1%). We also detected three patients with heterozygous mutations associated with the development of diabetes: one patient with missense variant Chr7:80293732C>A, NM_001001547.3:c.620C>A, p.(Thr207Asn), in exon 7 of cluster of differentiation (CD) 36; one patient with a mutation of uncertain meaning c.4648G>A (p.Val1550Ile) in transcript NM_000352.3 found in exon 39 of the ABCC8 gene and another patient with missense variant Chr4(GRCh37):g.6303576G>A, NM_006005.3(WFS1):c.2054G>A, p.(Arg685His), in exon 8 of Wolfram Syndrome 1.^17,18

Glycometric variables analysis

After selecting patients according to the compound of glycometric variables, observed GMI, TIR 70–180 mg/dL (3.9–10 mmol/L) and CV were 6.77 (3.85)%, 82.17 (8.97)%, and 30.26 (4.88)%, respectively. We did not find differences in the analysed glycometric variables between patients fulfilling or not MODY classical clinical criteria. Glycometric results among patients diagnosed with MODY at the end of the study were similar to those patients fulfilling MODY classical clinical criteria or those with negative MODY genetic study. Although we did not detect statistically significant differences between patients with MODY and the other subgroups, patients with MODY had the highest levels of TIR (85.6%), and the lowest levels of the various out-of-range times or CV (26.5%). Glycometric results can be observed in Table 2.

Table 2.

Glycometric results.

	Total(n = 503)	Patients not fulfilling MODY classical clinical criteria(n = 460)	Patients fulfilling MODY classical clinical criteria(n = 42)	Negative MODY genetic study patients(n = 29)	MODY patients(n = 5)
MIG (mg/dL, mmol/L)	133 ± 147.4 ± 0.8	133 ± 147.4 ± 0.8	135 ± 147.5 ± 0.8	133 ± 127.4 ± 0.7	139 ± 137.7 ± 0.7
GMI (%)	6.8 ± 3.9	6.8 ± 4.0	6.6 ± 0.4	6.5 ± 0.3	6.7 ± 0.3
Level 2 hyperglycemia, TAR >250 mg/dL, >13.9 mmol/L (%)	2.08 ± 5.62	2.13 ± 5.91	1.85 ± 3.6	1.67 ± 2.66	0.5 ± 1.0
Level 1 hyperglycemia, TAR >180 mg/dL, >10 mmol/L (%)	13.66 ± 8.1	13.36 ± 7.39	15.45 ± 14.05	16.03 ± 16.25	12.6 ± 6.46
TIR 70–180 mg/dL, 3.9–10 mmol/L (%)	82.17 ± 8.97	82.38 ± 8.34	80.45 ± 14.89	80.52 ± 17.54	85.6 ± 6.87
Level 1 hypoglycemia, TBR <70 mg/dL, <3.9 mmol/L (%)	3.24 ± 3.48	3.3 ± 3.39	3.02 ± 4.78	2.59 ± 3.36	1.4 ± 1.34
Level 2 hypoglycemia, TBR <54 mg/dL, <3.0 mmol/L (%)	0.27 ± 0.96	0.24 ± 0.86	0.63 ± 1.75	0.17 ± 0.38	0 ± 0
CV (%)	30.3 ± 4.9	30.2 ± 4.6	31.2 ± 7.7	31.1 ± 9.4	26.5 ± 5.3

Results are presented as mean ± SD values or percentages. CV, coefficient of variation; GMI, glucose management indicator; MIG, mean interstitial glucose; MODY, maturity-onset diabetes of the young; TAR, time above range of interstitial glucose; TBR, time below range of interstitial glucose; TIR, time in range of interstitial glucose.

Discussion

This study demonstrates the usefulness of CGM data in the screening of patients with MODY among adult patients initially diagnosed as having T1D. In fact, incorporating the pre-established CGM criteria allowed us to identify 42 new patients meeting the classical criteria. Five out of 34 participants that underwent genetic testing (15%) were diagnosed with MODY in our study. When including participants with other heterozygous mutations associated with the development of diabetes, we found that nearly 1 in 4 patients (23.5%) had either MODY or, potentially, a form of diabetes different from T1D. Up to the moment, no studies have considered and analysed the usefulness of CGM in the diagnostic pathway for MODY. Our findings may have relevant implications in routine clinical practice, potentially enabling the selection of candidates for genetic testing and the identification of patients with MODY misdiagnosed with T1D through an easy and accessible approach.

Making a confirmatory diagnosis of MODY allows patients with a specific type of diabetes to be properly identified. This may have implications not only for their therapeutic management but also for the ability to provide appropriate genetic counselling. In fact, genetic testing was offered to the first-degree relatives of those who were reclassified in their diagnosis, allowing us to provide genetic counselling to the family. Genetic testing is highly specific and sensitive and represents the gold standard for diagnosing MODY. Unfortunately, the cost of genetic testing, although increasingly cost-effective, does not currently allow it to be performed in all patients who are diagnosed with diabetes.^4,19 Therefore, deciding which patients with diabetes to screen for MODY has been one of the biggest challenges in the management of this disease.⁵

Initially, Shields et al. managed to develop a clinical prediction model to calculate an individual's probability of having MODY. This system allows a more rational approach to determine who should have molecular genetic testing.²⁰ In our study, the use of CGM to identify candidates for reassessment of classical MODY criteria allowed us to correctly reclassify the diagnosis of 5 patients, with the associated personal and healthcare implications of this new diagnosis. Using this CGM-based approach, followed by the evaluation of adherence to the classical clinical criteria for MODY, we were able to identify 5 out of 34 participants (15%) with MODY, and 8 out of 34 (23.5%) participants with MODY or other heterozygous mutations associated with the development of diabetes.

Usually, patients with MODY exhibit mild, stable hyperglycaemia and a very low risk of chronic diabetic complications.^21,22 In most forms of monogenic diabetes, some degree of insulin secretion is preserved, which can be demonstrated by measuring stimulate C-peptide levels. Although this residual insulin secretion may sometimes be insufficient to eliminate the need for exogenous insulin therapy, it contributes to lower glycemic variability, and a reduced risk of hypoglycaemia compared to patient with T1D. Therefore, in the absence of information from large patient series, it could be postulated that patients with MODY may maintain better glycaemic control. CGM-measured glycaemic targets for patients with diabetes were proposed in 2019.¹³ Unfortunately, many patients do not achieve the glycaemic control targets established as optimal in international consensus to the extent that this may lead us to suspect a misdiagnosis. Although people with monogenic diabetes were not considered in these recommendations, control targets of GMI <7%, TIR <70% and CV <36%, among others, were proposed for patients with T1D. In fact, the CV was recently used to separate MODY patients from well-controlled type 2 diabetes.¹⁴ Given that MODY is frequently misdiagnosed as T1D, we used these predefined control targets to recognize those patients with T1D for those who might be more likely to potentially have MODY.⁶ The fact that the glycometric results, with the highest TIR and lowest out-of-range times or CV values, of the patients with MODY detected in this study would help to reinforce the sense of pre-selection of the glycometric criteria used for good control among people with T1D, and could also confirm the theory that patients with MODY have mild and stable hyperglycaemia. However, some cases of patients with MODY and poorer glycaemic control have been reported.²³ Hence, our proposal would not consider a genetic study targeted at these specific cases. In addition, we used clinical criteria to establish the indication for MODY genetic testing. These suspicion data, although widely used, are not standardised internationally and different authors propose differential nuances in the criteria for requesting a genetic study.^3,5,15,24 The Swedish National Cohort Study demonstrated that limiting genetic testing to the autoantibody-negative patients with HbA1c <7.5% (58 mmol/mol) at diagnosis had a 49% detection rate.⁹ Therefore, and in line with the above, we could have left out of the genetic study patients who did not meet our clinical criteria. In any case, the availability of CGM data for all patients with T1Denabled us to reassess the classical diagnostic criteria in a small subset of patients, some of whom had previously been misdiagnosed with T1D.

On the other hand, we did not find differences in the analysed glycometric variables between patients who did not meet the classical clinical criteria for suspicion and those who did among patients fulfilling glycometric criteria.

The major strength of our pathway is the integration for the first time of CGM information, practically used in all patients with T1D, with the classical clinical criteria of suspicion. This offers a simple and cheap approach that does not require the use of complex algorithms. The easy access to the analysis of CGM information through different online platforms could facilitate the use of this diagnostic pathway in any healthcare facility. We postulate that the integration of CGM information in the diagnosis pathway of patients with MODY may have clinically relevant consequences, although further research is needed.

There were a few limitations to this study. Firstly, the CGM thresholds were not established diagnostic criteria for MODY and they are based on consensus glycemic targets for T1D. Secondly, genetic testing was not performed in patients without clinical criteria for suspected MODY due to economic and ethical reasons (this study was conducted in accordance with routine clinical practice for MODY screening, incorporating a novel glycometric pathway) and this could perhaps vary the impact of the study. Hence, our highly filtered subgroup results could not be used among a broader T1D clinical population. Regarding genetic testing, our study was limited to ruling out mutations associated with MODY and neonatal diabetes; hence, patients with mitochondrial DNA mutations or mutations linked to lipodystrophies may have gone undetected. Nevertheless, important clinical implications may be derived from these diagnoses, particularly because the possibility of misclassification of long-duration T1D is rarely considered in clinical practice. Therefore, our findings may help raise awareness of this issue, especially in patients with T1D who exhibit excellent glycemic metrics, and could encourage further studies that incorporate glycometrics in the design of future MODY probability calculators. Another limitation of this study was that glycemic stability was assumed in patients with MODY for the glycometric pathway, although this requirement is not always present in monogenic diabetes. A further limitation of the present study is that detailed information regarding the specific insulin regimens used by the participants was not systematically collected, including the type of basal insulin (e.g., insulin glargine U100, glargine U300 or insulin degludec) and the type of rapid-acting insulin analogues. Therefore, potential differences in glycaemic metrics related to the insulin formulation or regimen could not be explored in our analysis. Finally, the data presented apply to adults with T1D in the Spanish population and will vary in other populations.

Conclusion

In conclusion, we show that a CGM-based approach may constitute an effective and easily implemented tool in routine clinical practice to identify patients with MODY who have been misdiagnosed as having T1D. This approach could provide useful aid for selecting patients for diagnostic testing who may benefit from a proper diagnosis. Additional studies will help to clarify the implications of our findings and establish a more definitive understanding of this diagnostic approach.

Authors’ contributions

JMF and MJP designed the study. JMF wrote the study protocol and published in a public registry. JMF, JIMM, SHA, AM, PP, MRA, JG, EM, JL and MJP gathered data. JMF, JIMM and JRMR contributed to the interpretation of the results. JMF took the lead in writing the manuscript. JRMR and JIMM reviewed statistical analysis. All authors provided critical feedback and helped shape the final manuscript.

Ethical consideration

The study protocol was reviewed and approved by the corresponding institutional ethics committee, and the research was conducted in accordance with the principles of the Declaration of Helsinki and Good Clinical Practice guidelines. All participants provided written informed consent prior to their inclusion in the study and before any study-related procedures were performed.

Funding

This study received no specific funding. JIMM was supported by a Juan Rodés grant from Instituto de Salud Carlos III, Madrid, Spain (JR24/00006).

Conflicts of interest

The authors declare that they have not received research support or compensation for this study. All authors declare that they have no conflicts of interest concerning this article.