Artificial intelligence (AI) is increasingly shaping the landscape of diabetes mellitus (DM) care by enabling data-driven approaches to prediction, diagnosis, classification, and management. The growing availability of continuous biometric data from wearable technologies, combined with advances in machine learning, has created new opportunities to overcome limitations of traditional diabetes metrics and population-based care models. This review summarises current clinical applications of AI in DM, with a focus on translational relevance and, critically, on the level of validation and clinical readiness of each application. AI-based models have demonstrated improved performance in predicting DM onset and diabetes-related complications by integrating electronic health records, imaging data, and longitudinal clinical variables. In diagnosis and screening, AI offers non-invasive and scalable tools, including automated retinal screening systems and emerging digital biomarkers. Data-driven clustering approaches have further refined diabetes classification. In diabetes management, AI supports personalised glucose monitoring, insulin titration, pharmacological decision-making, and lifestyle interventions. Crucially, these applications differ markedly in maturity: autonomous retinal screening and closed-loop insulin delivery are approaching or have reached routine clinical use, whereas voice-based biomarkers and simulation-evaluated lifestyle systems remain proof-of-concept. Addressing challenges of generalisability, data quality, interpretability, and ethical governance through robust validation and explainable AI will be essential to realise the full potential of AI in precision, patient-centred diabetes care.
Aims We compared associations of CGM-derived time in range (TIR) and glucose management indicator (GMI) vs. HbA1c with CVD and eGFR decline in type 2 diabetes. Methods We studied 1363 participants with ≥10-day blinded CGM in the GRADE-CGM Substudy. Exposures were TIR and GMI, with HbA1c as comparator. Cox models estimated hazard ratios (HR) for the composite of any CVD or all-cause death. Linear mixed-effects models estimated eGFR slope. Models were adjusted for age, sex, race/ethnicity, treatment group, diabetes duration, and baseline eGFR, then mutually adjusted each CGM metric with HbA1c. Results Over median 2.3 years, 74 CVD or death occurred, eGFR slope was −2.37 mL/min/1.73m2/year. GMI, HbA1c, and TIR were each significantly associated with CVD/death (adjusted HR 1.44 [1.21–1.71] per 1% GMI; 1.37 [1.16–1.61] per 1% HbA1c; 0.73 [0.61–0.87] per 20% TIR). Higher GMI and HbA1c were associated with faster eGFR decline (−0.35 [−0.63, −0.07] and − 0.38 [−0.65, −0.11] mL/min/1.73m2/year); TIR was not significantly associated with eGFR slope. After mutual adjustment, most associations of CGM metrics and HbA1c with outcomes were attenuated. Conclusions In type 2 diabetes, GMI and HbA1c were comparably associated with cardiovascular and kidney outcomes, whereas TIR was associated with cardiovascular outcomes only.
Aims This study investigated the impact of hyperglycemia on systemic and skeletal muscle metabolite profiles. Methods Serum and gastrocnemius metabolites of fructose-fed, streptozotocin-injected rats (T2D) were analyzed by 1H NMR and related to markers of energy metabolism and mitochondrial remodeling. Results T2D rats exhibited hyperglycemia and impaired insulin secretion without obesity or dyslipidemia. Hyperglycemia induced a generalized catabolic state, reflected by reduced body weight and relative gastrocnemius mass. T2D rats presented increased phosphorylated AMP-activated protein kinase (pAMPK) and peroxisome proliferator-activated receptor alpha (PPARα) muscle levels and pAMPK/AMPK ratio, which correlated with muscle mass loss, indicating AMPK signaling and fatty acid oxidation activation. 1H NMR results suggested muscle proteolysis, with elevated serum levels of isoleucine, alanine, glycine, and threonine in T2D rats. Reduced muscle levels of all three branched-chain amino acids (BCAAs) were also detected in T2D rats, suggesting enhanced BCAA oxidation/impaired uptake. T2D rats also exhibited elevated 3-hydroxybutyrate and acetate serum levels. No changes were observed in glucose, glucose transporter 4 (GLUT4), and 6-phosphofructokinase muscle type (PFKM) muscle levels, or ATP synthase activity. Conclusions These findings demonstrate an association between hyperglycemia and muscle loss, with muscle loss being associated with metabolic reprogramming toward the preferential utilization of non-glucose substrates, possibly to preserve energy homeostasis at the expense of protein maintenance.
Aims To assess the cumulative risk of diabetic foot complications in a real-world type 1 diabetes (T1DM) population to examine whether observed risk patterns support risk-adapted screening intervals. Methods This observational cohort study included T1DM patients in the Central Denmark Region (2020–2025). Patients were stratified into low, high, or very high risk using modified IWGDF criteria. Cumulative risks of a composite endpoint (ulcers, amputations, or Charcot foot) were estimated, accounting for death as a competing event and stratifying by albuminuria and advanced retinopathy. Results Among 7329 individuals with T1DM, 2074 (28.3%) were low-risk, 3562 (48.6%) high-risk, 437 (6.0%) very high-risk, and 1256 (17.1%) had unknown status. Over a mean follow-up of 3.26 years, the three-year cumulative risk was 0.55% (95% CI: 0.35–0.83) for low-risk, 3.45% (2.84–4.15) for high-risk, and 20.86% (16.67–25.38) for very high-risk patients. Among low-risk patients without albuminuria or advanced retinopathy, the risk was 0.37% (0.15–0.77). Conclusions Low-risk T1DM patients exhibit a very low incidence of foot complications. These findings support consideration of extended, risk-adapted screening intervals, such as every three years, in this subgroup, with the potential to optimize healthcare resource utilization. While these data support the feasibility of such an approach, further studies are warranted to validate the safety of extended screening intervals in clinical practice.
