Journal of Diabetes最新文献

β-cell dedifferentiation in humans is considered as a mechanism underlying β-cell failure, which is characterized by the loss of their differentiated identity and reversion to a fetal-like state. In humans, significant increases in the hormones-negative/Chromogranin A-positive endocrine cells were detected in islets from T2D individuals, which were accompanied by decreased expressions of the β-cell identity transcription factors (NKX6.1, MAFA, etc.) and reactivation of the progenitor marker, NEUROGENIN 3, and dramatic inductions of the dedifferentiation marker, ALDH1A3 and BACH2. The observation of increased dysfunctional dedifferentiated β-cells in non-diabetic patients with chronic pancreatitis or pancreatic ductal adenocarcinoma as well as in non-diabetic aged subjects proposed that β-cell dedifferentiation is not only a major mechanism of pancreatogenic diabetes and age-related diabetes but also an initial trigger for human β-cell failure before hyperglycemia occurs. Indeed, the finding that the ratio of dedifferentiated β-cells is inversely related to HOMA-β and 2h-C Peptide in subjects without a diabetes history further supports the contribution of β-cell dedifferentiation to β-cell failure in humans. Considerable efforts have been made and are still required to identify targets and pathways (in)dependent of hyperglycemia that trigger dedifferentiation in human β-cells. These findings hold promise for the development of effective therapeutic strategies aimed at preventing or reversing this process, thereby maintaining or restoring functional β-cell mass.

The figure shows the market share of prescriptions filled from 2008 to 2023. Over this period, the GLP-1 prescription market shifted from predominantly Byetta to a broader distribution across Victoza, Trulicity, and Ozempic, with smaller shares of Rybelsus and Mounjaro by 2023.

Aim: This study aimed to determine the effectiveness of deep learning models for detecting and predicting diabetes related kidney disease (DRKD) using two-dimensional retinal images.

Methods: A systematic review was conducted across eight databases from inception to July 14, 2026. Eligible cross-sectional and longitudinal studies assessed deep learning models for DRKD detection or prediction using retinal images alone, clinical data alone, or hybrid models. Data on study design, model development, validation, and performance metrics were extracted, with study quality appraised using QUADAS-2 and post hoc, PROBAST+AI. Results were synthesized narratively, with random effects meta analysis for hybrid models.

Results: Of 367 studies, six met inclusion criteria. Three cross-sectional studies of hybrid models contributed to the meta-analysis for DRKD detection, yielding pooled AUC 0.88, sensitivity 0.80, and specificity 0.81, though heterogeneity was high (I2 93.7%-99.8%). Three further studies were synthesized narratively rather than pooled: two cross-sectional studies used more severe or narrower CKD case definitions (AUC 0.86 and 0.96 respectively), and only one longitudinal study predicted the 5-year risk of DRKD (AUC 0.81, sensitivity 0.70, and specificity 0.77). External validation showed hybrid models consistently outperformed retinal-only (ΔAUC 0.03-0.11), but their advantage over clinical-only models was inconsistent.

Conclusion: Deep learning models may be useful for detecting and predicting DRKD from retinal images, with or without clinical data. However, the small number of heterogenous studies restricts generalizability. Future work should prioritize robust model development, and validation using large, diverse populations.

We thank the authors for their thoughtful comments on our randomized controlled trial of glycated albumin (GA)-guided therapy in newly diagnosed type 2 diabetes mellitus (T2DM) [1]. We appreciate the opportunity to clarify several key points.

First, we fully agree that the primary endpoint—achievement of HbA1c<7% at 12 weeks—was neutral, and we explicitly reported and discussed this negative result. We had acknowledged that secondary analyses were exploratory and hypothesis-generating, and we have interpreted them with appropriate caution, as stated in the manuscript.

Second, regarding the HbA1c<6.5% result, we agree that intention-to-treat inference is primary. The difference was marginal in the full analysis set (35.7% vs. 23.8%, p = 0.065) and nominally significant in the per-protocol set (40.5% vs. 25.3%, p = 0.034). We reported both transparently and did not base our main conclusion on the per-protocol finding; rather, we view the consistent direction of effect as a signal warranting further investigation.

Third, we agree that the greater improvements in HbA1c reduction, HOMA-β, HOMA-IR, and anthropometric parameters in the GA-guided group may partly reflect earlier and more intensive combination therapy. However, this is precisely the intended clinical utility of GA: to provide a timely, actionable trigger for treatment intensification when HbA1c remains insensitive to short-term changes. In addition, according to the protocol, in the standard therapy group, apart from withholding the GA measurement data, we did not impose any additional restrictions on blood glucose monitoring; capillary blood glucose testing was permitted as per routine clinical practice. So, the relevant question is not whether GA has intrinsic therapeutic effects, but whether GA measurement facilitates earlier and more effective treatment decisions compared with routine clinical practice. Our trial provides preliminary evidence.

Finally, we agree that short-term changes in HOMA-β and HOMA-IR should not be equated with durable β-cell recovery or sustained improvement in insulin sensitivity. We interpreted these as short-term metabolic improvements rather than evidence of long-term disease modification.

We thank the authors for their insightful comments.

Q.R. wrote the manuscript. L.J. reviewed and edited the manuscript. Q.R. and L.J. contributed significantly and in keeping with the latest guidelines of the International Committee of Medical Journal Editors.

This work was supported by Asahi Kasei Pharma Corporation and National Clinical Key Specialty Construction Program of China (Department of Endocrinology, Peking University People's Hospital) with support from the central government budget.

The authors declare no conflicts of interest.

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

Aims: To evaluate the longitudinal prognostic value of small- and large-fiber peripheral nerve assessments for accelerated renal decline in patients with Type 2 diabetes (T2D).

Methods: An initial cross-sectional analysis evaluated concurrent neuro-renal associations, while a subsequent longitudinal cohort assessed accelerated renal progression (eGFR loss > 3 mL/min/1.73 m²/year) using paired baseline and follow-up data. Prognostic predictors were selected via LASSO regression. Model utility was assessed using area under the curve (AUC), net reclassification improvement (NRI), and decision curve analysis (DCA).

Results: Cross-sectional sudomotor small-fiber dysfunction accurately mirrored concurrent renal impairment (β = 0.18, 95% CI: 0.10-0.26, p < 0.001). Longitudinally, ESCF demonstrated no independent prognostic value for accelerated renal decline, and its apparent longitudinal association with future eGFR was completely attenuated after adjusting for baseline renal function. LASSO regression retained right peroneal motor amplitude, serum albumin, platelet count, and SGLT2i use as robust prognostic factors. Notably, subgroup analyses revealed this large-nerve predictive capacity was significant primarily in patients with established DKD (p for interaction = 0.006). Incorporating this neurofunctional phenotype into standard clinical models yielded a modest AUC improvement but achieved a clinically relevant continuous NRI of 30.35%, successfully recalibrating risk profiles. DCA confirmed a superior net clinical benefit.

Conclusion: A distinct prognostic dissociation exists within the diabetic neuro-renal axis. Sudomotor small-fiber dysfunction primarily reflects established microvascular impairment, whereas large motor nerve degeneration demonstrates independent prognostic value for future accelerated renal decline. Incorporating this neurofunctional phenotype into traditional clinical models yields substantial individualized risk reclassification, thereby potentially aiding in identifying high-risk individuals.

Background: Cardiovascular-kidney-metabolic (CKM) risk stratification may identify patients with type 2 diabetes mellitus (T2DM) who require more intensive primary prevention, but real-world data from Vietnam are limited. We aimed to describe advanced CKM risk and risk-factor control in Vietnamese patients with T2DM without established cardiovascular disease (CVD).

Methods: This multicenter cross-sectional study enrolled 1436 adults with T2DM without CVD in Vietnam. Advanced CKM risk was defined as 10-year total CVD risk ≥ 20% or eGFR < 60 mL/min/1.73 m². Guideline-based control of HbA1c, blood pressure, LDL-C, triglycerides, and non-HDL-C was evaluated. Logistic regression assessed associations between advanced CKM risk and target achievement.

Results: Advanced CKM risk was identified in 551 patients (38.4%). Despite a relatively low prevalence of BMI ≥ 25 kg/m² (26.9%), hypertriglyceridemia was the most common metabolic abnormality, affecting 88.0% of participants. Risk-factor control was generally low, with 33.3% achieving glycemic control, 33.5% blood pressure control, 18.8% LDL-C control, 12.0% triglyceride control, and 21.1% non-HDL-C control. After adjustment, advanced CKM risk was associated with poorer blood pressure control (adjusted OR = 0.466, 95% CI = 0.359-0.603, p < 0.001) and triglyceride control (adjusted OR = 0.677, 95% CI = 0.473-0.969, p = 0.033), but not with achievement of HbA1c, LDL-C, non-HDL-C, or combined lipid control.

Conclusions: Advanced CKM risk was common among Vietnamese patients with T2DM without established CVD and was consistently associated with poorer blood pressure control. CKM-based risk assessment may identify high-risk patients earlier and support guideline-directed risk-factor management. Frequent hypertriglyceridemia with relatively low obesity supports a CKM-based strategy tailored to this metabolic risk pattern.