Diagnosing maturity-onset diabetes of the young [MODY] and distinguishing it from type 2 diabetes in youth are critical for improving clinical care (1) but are challenging given many shared features. We tested the utility of the MODY probability calculator (2), developed and proven accurate in a cohort of mostly European adults, and commonly used clinical characteristics to differentiate MODY and type 2 diabetes in Progress in Diabetes Genetics in Youth (ProDiGY), an ancestrally diverse (82% non-White) cohort of youth with clinician-diagnosed type 2 diabetes before age 20 years from the SEARCH for Diabetes in Youth (SEARCH), Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY), and TODAY Genetics studies (3,4). Only participants negative for islet autoantibodies were included. Exome sequences were analyzed for variants in 10 genes designated by the ClinGen Monogenic Diabetes Expert Panel as having moderate or definitive evidence for causing MODY (5). Individuals harboring a pathogenic or likely pathogenic variant were classified as having MODY. Performance of the MODY calculator and clinical variables in discriminating between participants with an identified MODY variant and type 2 diabetes (i.e., those without it) was assessed at the earliest data point where participant age was greater than age at diagnosis, per MODY calculator specifications.
In our cohort, there were 100 (3%) individuals with MODY and 3,279 (97%) individuals without MODY (classified hereafter as type 2 diabetes). Pathogenic or likely pathogenic variants were identified in five genes (HNF1A [41%], GCK [27%], HNF4A [26%], PDX1 [4%], and INS [2%]). Of the participants with MODY, 38% were White, 22% Hispanic, 19% Black, and 17% multiancestry (P = 1.6 × 10−6).
There were 44 individuals with MODY and 694 with type 2 diabetes with all data available for the MODY calculator. The MODY calculator was not helpful for differentiating MODY and type 2 diabetes in pediatrics. Even with the MODY calculator variables directly fitted to the data, the MODY calculator variables (area under the curve [AUC] 0.82) performed no better than BMI alone (AUC 0.82; P = 0.607) (Fig. 1A). Of the 44 participants with MODY, 43 (98%) had a probability of MODY >25% (Fig. 1B), a threshold with good discriminating value to trigger genetic testing in European adults. However, 92% (640 of 694) of individuals with type 2 diabetes also exceeded 25% probability, leading to a positive test rate of all individuals with probability >25% of only 6.3% (43 of 683). In fact, 73% of individuals with type 2 diabetes had >75% probability of having MODY (504 of 640), but only 7% of all of those with >75% probability of MODY (40 of 544) would test positive for MODY.
A: Receiver operating characteristic (ROC) curve for differentiation of type 2 diabetes and MODY for BMI (black) and MODY calculator variables (gray) directly fit to data. There is no significant difference in performance. B: MODY calculator probability as computed from available data, for both participants with type 2 diabetes and participants with MODY. C: ROC curve for differentiation of type 2 diabetes and MODY for BMI (black) and BMI z score (gray). AUC for BMI is significantly higher than that for BMI z score. D: ROC for differentiation of type 2 diabetes and MODY for age of diagnosis and having a parent with diabetes. E: Percentage of total participants with a parent with diabetes and percentage of participants with no, one, or two parents with diabetes. F: Proportion of participants with type 2 diabetes or MODY at each BMI threshold for all participants with available BMI data. G: Proportion of participants with type 2 diabetes or MODY at each BMI threshold for White and non-White individuals with available BMI data. ns, not significant; T2D, type 2 diabetes; *P < 0.05; **P < 0.01.
A: Receiver operating characteristic (ROC) curve for differentiation of type 2 diabetes and MODY for BMI (black) and MODY calculator variables (gray) directly fit to data. There is no significant difference in performance. B: MODY calculator probability as computed from available data, for both participants with type 2 diabetes and participants with MODY. C: ROC curve for differentiation of type 2 diabetes and MODY for BMI (black) and BMI z score (gray). AUC for BMI is significantly higher than that for BMI z score. D: ROC for differentiation of type 2 diabetes and MODY for age of diagnosis and having a parent with diabetes. E: Percentage of total participants with a parent with diabetes and percentage of participants with no, one, or two parents with diabetes. F: Proportion of participants with type 2 diabetes or MODY at each BMI threshold for all participants with available BMI data. G: Proportion of participants with type 2 diabetes or MODY at each BMI threshold for White and non-White individuals with available BMI data. ns, not significant; T2D, type 2 diabetes; *P < 0.05; **P < 0.01.
BMI was the strongest predictor of MODY (AUC 0.82). Interestingly, actual BMI performed better than BMI age- and sex-adjusted z score (P = 7.4 × 10−3) (Fig. 1C). This is likely because MODY develops at an earlier median age than type 2 diabetes and BMI is influenced by age, while BMI z score abrogates the effect of age on BMI. Age of diagnosis had some discriminating capacity (AUC 0.63) (Fig. 1D), whereas hemoglobin A1c (AUC 0.51) and treatment of diabetes with any medication (AUC 0.51) did not, even though these were important components of the established MODY calculator (2). Although having a parent with diabetes significantly increased the probability of MODY in the original calculator, it had no discriminatory capacity between type 2 diabetes and MODY in this pediatric cohort with type 2 diabetes (AUC 0.5) (Fig. 1D). A parent with diabetes was present for 74% and 73% of participants with type 2 diabetes and MODY, respectively (P = 1.0) (Fig. 1E). In contrast, diabetes in both parents was more likely among participants with type 2 diabetes (19%) than those with MODY (6%; P = 8.2 × 10−3).
Overall, 50% (6 of 12) of individuals with BMI <21 kg/m2 and 20% (18 of 92) of individuals with BMI 21–25 kg/m2 had MODY (Fig. 1F). Yet, screening participants with BMI ≤25 kg/m2 would still only capture 55% of MODY cases (24 of 44). It is important to note that, for BMI ≤25 kg/m2, 62% (65 of 104) of those with MODY met pediatric criteria for being overweight or obese by BMI percentile. The rate of MODY among White individuals was even higher (Fig. 1G), with 34% (10 of 29) of individuals with BMI ≤25 kg/m2 having MODY, possibly due to lower rates of pediatric type 2 diabetes in this subgroup. Some other clinical variables were discriminatory, including fasting insulin (AUC 0.81), waist circumference (AUC 0.81), fasting C-peptide (AUC 0.80), and fibrinogen (AUC 0.77).
This study underscores the challenge in identifying individuals with MODY on the basis of clinical presentation among those diagnosed with pediatric type 2 diabetes. The MODY calculator, while helpful in differentiating MODY from type 2 diabetes in European adults, is not helpful for youth clinically diagnosed with type 2 diabetes given important clinical differences between European adults and a multiancestry cohort of children. Our data suggest that a BMI threshold may be used as a screening threshold, which may vary depending on ancestry and desired sensitivity or specificity. New biomarkers, or gene-specific biomarkers, are needed for differentiating youth-onset type 2 diabetes and MODY.
Article Information
Ackknowledgments. The TODAY and SEARCH studies are indebted to the many youth and their families, and their health care providers, whose participation made this study possible.
Funding and Duality of Interest. This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), National Institutes of Health (NIH), grant R01DK124395 (to M.J.R. and M.T.); NIDDK/NIH grant K23DK129821 (M.T.); NIDDK/NIH grant T32DK007699-41 (R.J.K.); Doris Duke Clinical Scientist Development Award 2022063 (M.S.U.); National Heart, Lung, and Blood Institute, NIH, grant K24 HL157960 (J.C.F.); Diabetes UK (21/0006328 [B.M.S.]); NIDDK/NIH grant K23DK120932 (S.S.); NIDDK/NIH grant R03DK138213 (S.S.); and Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, grant U24 HD112205 (T.I.P.). For SEARCH 3/4, the authors acknowledge the involvement of the Kaiser Permanente Southern California Marilyn Owsley Clinical Research Center (funded by Kaiser Foundation Health Plan and supported in part by the Southern California Permanente Medical Group); the South Carolina Clinical & Translational Research Institute, at the Medical University of South Carolina (National Center for Advancing Translational Sciences [NCATS], NIH, grants UL1 TR000062 and UL1 TR001450); Seattle Children’s Hospital and the University of Washington (NCATS/NIH grant UL1 TR00423); University of Colorado Pediatric Clinical and Translational Research Center (NCATS/NIH grant UL1 TR000154); the Barbara Davis Center for Diabetes at the University of Colorado Denver (Diabetes Endocrinology Research Center NIH grant P30 DK57516); the University of Cincinnati (NCATS/NIH grants UL1 TR000077 and UL1 TR001425); and the Children with Medical Handicaps program, managed by the Ohio Department of Health. SEARCH 4 (1R01DK127208-01, 1UC4DK108173) is funded by the NIDDK/NIH and supported by the Centers for Disease Control and Prevention. The Population Based Registry of Diabetes in Youth Study (1U18DP006131, U18DP006133, U18DP006134, U18DP006136, U18DP006138, and U18DP006139) is funded by the Centers for Disease Control and Prevention (DP-15-002) and supported by NIDDK/NIH. SEARCH 1, 2, and 3 are funded by the Centers for Disease Control and Prevention (PA nos. 00097, DP-05-069, and DP-10-001) and supported by NIDDK/NIH. Kaiser Permanente Southern California (U48/CCU919219, U01 DP000246, and U18DP002714), University of Colorado Denver (U48/CCU819241-3, U01 DP000247, and U18DP000247-06A1), Cincinnati Children’s Hospital Medical Center (U48/CCU519239, U01 DP000248, and 1U18DP002709), University of North Carolina at Chapel Hill (U48/CCU419249, U01 DP000254, and U18DP002708), Seattle Children’s Hospital (U58/CCU019235-4, U01 DP000244, and U18DP002710-01), and Wake Forest University School of Medicine (U48/CCU919219, U01 DP000250, and 200-2010-35171). No other potential conflicts of interest relevant to this article were reported.
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention or NIDDK. This study includes data provided by the Ohio Department of Health, which should not be considered an endorsement of this study or its conclusions.
Author Contributions. R.J.K., M.S.U., and M.J.R. were involved in the conceptualization of the study. R.J.K., B.M.S., T.I.P., M.T., J.C.F., S.S., A.T.H., M.S.U., and M.J.R. designed the study and contributed to analysis and interpretation of results. R.J.K. conducted the study analyses. T.I.P. performed variant analysis to identify those with MODY. A.S.S., A.D.L., A.B., and C.P. collected and provided data for this study. All authors edited, reviewed, and approved the final version of the manuscript. R.J.K. 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.
Prior Presentation. Parts of this study were presented in abstract form at the 84th Scientific Sessions of the American Diabetes Association, Orlando, FL, 21–24 June 2024.
Handling Editors. The journal editors responsible for overseeing the review of the manuscript were John B. Buse and Anna L. Gloyn.