Dear Editor,
To the best of our knowledge, the first study on “Prediction of future insulin deficiency in glutamic acid decarboxylase autoantibody enzyme‐linked immunosorbent assay‐positive patients with slowly progressive type 1 diabetes”, penned by Mr. Kawasaki et al., 1 was published in your honored journal. Their work provides valuable insights; however, methodological refinements could further strengthen the predictive model and clinical applicability. We highlight four key considerations:
First, the optimal cut‐off values for age, BMI, and fasting C‐peptide were derived from ROC analysis in a cohort of 60 patients (including 30 progressors). The modest sample size may affect the stability of these cut‐offs. Reporting bootstrapped confidence intervals, as recommended in diagnostic study guidelines 2 , could help assess their generalizability.
Second, defining insulin deficiency by a single F‐CPR value (<0.6 ng/mL) may not fully capture progressive β‐cell decline. Longitudinal measurements could reduce misclassification 3 .
Third, predictive values (positive predictive value [PPV] and negative predictive value [NPV]) are prevalence‐dependent. The specialized Japanese type 1 diabetes database study (TIDE‐J) registry may not reflect general slowly progressive insulin‐dependent diabetes mellitus (SPIDDM) prevalence. Reporting sensitivity, specificity, and likelihood ratios would improve interpretability across settings.
Finally, the combined predictive ability of the identified factors (age, BMI, F‐CPR, insulinoma‐associated antigen‐2 autoantibodies [IA‐2A]) is described only by an increase in positive predictive value. A formal multivariable model (such as logistic regression or a nomogram) would more rigorously quantify each variable's independent contribution and align with contemporary prognostic‐model standards 4 .
We believe that addressing these points will strengthen the conclusions of the study and provide a more reliable tool for prognostic assessment in SPIDDM.
DISCLOSURE
The author declares no conflict of interest.
Approval of the research protocol: None.
Informed consent: None.
Registry and the registration no. of the study/trial: None.
Animal studies: None.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available in PubMed at https://pubmed.ncbi.nlm.nih.gov/. These data were derived from the following resources available in the public domain: J Diabetes Investig 2024; 15: 835–842. https://doi.org/10.1111/jdi.14178.
REFERENCES
- 1. Kawasaki E, Awata T, Ikegami H, et al. Prediction of future insulin‐deficiency in glutamic acid decarboxylase autoantibody enzyme‐linked immunosorbent assay‐positive patients with slowly‐progressive type 1 diabetes. J Diabetes Investig 2024; 15: 835–842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Leeflang MM, Moons KG, Reitsma JB. Bias in sensitivity and specificity caused by data‐driven selection of optimal cutoff values: Mechanisms, magnitude, and solutions. Clin Chem 2020; 66: 529–539. [DOI] [PubMed] [Google Scholar]
- 3. Thomas NJ, Jones SE, Weedon MN. Frequency and phenotype of type 1 diabetes in the first six decades of life: A cross‐sectional, genetically stratified survival analysis from UK biobank. Lancet Diabetes Endocrinol 2021; 9: 165–174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Moons KGM, Wolff RF, Riley RD. PROBAST: A tool to assess risk of bias and applicability of prediction model studies. Ann Intern Med 2019; 170: W1–W33. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available in PubMed at https://pubmed.ncbi.nlm.nih.gov/. These data were derived from the following resources available in the public domain: J Diabetes Investig 2024; 15: 835–842. https://doi.org/10.1111/jdi.14178.