The purpose of this review is to synthesize literature investigating the relationship between type 2 diabetes (T2D) and obstructive airway diseases and to identify implications for clinical care.

Type 2 diabetes is a common and challenging comorbidity in patients with asthma and chronic obstructive pulmonary disease (COPD). Basic, translational and clinical studies support a bidirectional association between T2D and the lung. In animal models and human studies, insulin resistance and hyperglycemia are associated with pulmonary inflammation, respiratory exacerbation risk and disease severity. Corticosteroids are a mainstay for respiratory disease control and exacerbation treatment but promote ongoing metabolic dysregulation. Randomized, placebo-controlled trials of glucose-lowering medications for asthma are actively ongoing. Additional studies addressing clinical pathways to co-manage respiratory and metabolic risk are needed. Patients with comorbid T2D and asthma or COPD are at risk for worse outcomes. There are opportunities to improve cross-disciplinary care, potentially reducing risk and multimorbidity associated with both conditions.

To evaluate the impact of a structured tele-education program on hospital hyperglycemia and diabetes, focusing on residents' medical knowledge and inpatient care.

This open-label, multicenter, randomized clinical trial enrolled internal medicine residents from four university hospitals in southern Brazil. Teams were block-randomized to an intervention group that received an online lecture plus 30 days of tele-education via WhatsApp, or to a control group with no intervention. The primary outcome was medical knowledge, assessed with a validated 10-item questionnaire. Secondary outcomes included quality of insulin prescriptions, hypoglycemia and hyperglycemia rates, and hospital length of stay (LOS). Analyses were performed using SPSS v29 (5% significance).

Fifty residents completed the study. The intervention group achieved higher post-intervention knowledge scores than the control group (median 8 vs. 6 correct answers; p = 0.005) and showed significant improvement from preto post-test (6 to 8; p < 0.001), with consistent gains across centers. Clinical data from 149 hospitalized patients were analyzed (mean age 67.8 years; 55% female); 56% had diabetes, and 44% had hospital-related hyperglycemia. There was a nonsignificant trend toward more appropriate NPH (p = 0.107) and regular insulin (p = 0.203) prescriptions in the intervention group. Median LOS was longer in the intervention group (19 vs. 13 days; p = 0.009).

The tele-education program improved residents' knowledge of inpatient hyperglycemia. Larger studies are needed to confirm clinical effects and long-term outcomes of tele-education in hospital glycemic management.

Research on the association between TyG-GGT index and diabetes mellitus (DM) risk remains scarce. This study aimed to investigate the relationship between TyG-GGT and DM incidence.

This retrospective cohort investigation enrolled 8,678 participants who underwent comprehensive health screenings at Kuichong People's Hospital in Shenzhen from 2018 through 2023. Cox proportional hazards regression models were employed to assess the association between TyG-GGT and DM risk, and Cox proportional hazards regression model with restricted cubic spline functions was used to evaluate non-linear relationships. Subgroup analyses and sensitivity analyses further verified the stability of these findings. Finally, receiver operating characteristic (ROC) curve methodology and time-dependent ROC analysis were performed to determine the predictive capacity of TyG-GGT for incident DM within a 5-year period.

Following multivariable adjustments, higher TyG-GGT levels were found to be associated with elevated DM risk, demonstrating an HR of 1.116 (95% CI: 1.041-1.196) per 50-unit increase in TyG-GGT. Additionally, a non-linear association between them was observed, exhibiting a threshold value at 380. When below this inflection point, the HR per 50-unit increase in TyG-GGT was 1.723 (95% CI: 1.500-1.979), while above this value the association was not statistically significant. Additionally, in predicting DM risk, TyG-GGT had the highest AUC value (0.732), while the AUC values of TG (0.635), GGT (0.649), FPG (0.660), and TyG (0.675) were all lower than this value. Time-dependent ROC analysis revealed that the AUC values of TyG-GGT remained stable between 0.7292-0.7338 over a prediction horizon of 1.0 to 5.0 years. The stability of these results was further corroborated via sensitivity analysis.

This study found that TyG-GGT demonstrated an independent positive association and non-linear relationship with DM risk, with an inflection point at 380. TyG-GGT below 380 was associated with higher observed DM risk. Additionally, TyG-GGT exhibits discriminatory performance for DM risk assessment and may serve as a clinically useful predictor, thereby aiding clinicians in early identification of high-risk individuals and providing a novel perspective for optimizing clinical prevention and management of DM.

Chronic kidney disease (CKD) affects 8.2% of China's population and is a major global health concern. While insulin resistance (IR) is linked to CKD, the relationship between insulin resistance (HOMA-IR) and CKD risk remains unclear, especially in diabetic and non-diabetic populations.

This cross-sectional study analyzed data from 32,055 Chinese adults in the REACTION study. Logistic regression and generalized additive models assessed the association between HOMA-IR and CKD risk in diabetic (DM) and non-diabetic (Non-DM) populations, with nonlinear relationships explored using two-piecewise logistic regression.

The overall CKD prevalence was 16.09%(95% CI: 15.68%-16.49%). In the Non-DM group, HOMA-IR was positively associated with CKD risk (OR = 1.037, 95% CI: 1.010-1.066, P = 0.008), while no significant association was found in the DM group (OR = 0.991, 95% CI: 0.952-1.032, P = 0.667). Both groups showed an n-shaped relationship, with inflection points at HOMA-IR values of 2.581 (Non-DM) and 2.587 (DM). Below these thresholds, CKD risk increased with HOMA-IR; above them, risk decreased.

Elevated HOMA-IR is independently associated with an increased risk of CKD in non-diabetic individuals, whereas this association is not significant in diabetic patients. These findings strongly highlight the clinical value of HOMA-IR as an early predictor of CKD risk, particularly in non-diabetic populations, emphasizing the importance of monitoring insulin resistance for early risk stratification and tailored management.

In-hospital hypoglycemia remains a serious and potentially life-threatening complication among adults with type 1 diabetes mellitus (T1DM), yet reliable and interpretable prediction tools for Chinese inpatients are lacking. We aimed to develop and validate an interpretable machine learning model using multicenter inpatient data to predict the risk of in-hospital hypoglycemia in adults with T1DM, and to enhance clinical understanding of key predictors.

This multicenter retrospective cohort study enrolled adult inpatients with T1DM from five tertiary Grade A hospitals in China between January 1, 2019 and September 30, 2025. From the same multicenter cohort, the total dataset was randomly split 7:3 into a development set (n = 1,048) and an independent external validation set (n = 450). Within the development set, we performed 5-fold stratified cross-validation for hyperparameter tuning, and both internal cross-validation and external validation remained fully independent throughout model development. Machine learning models were trained to predict in-hospital hypoglycemia and evaluated for discrimination, calibration, clinical utility, and interpretability.

The study enrolled 1,498 patients, of whom 580 (38.7%) experienced in-hospital hypoglycemia. The random forest model demonstrated superior predictive performance in the external validation cohort, achieving an AUC of 0.831 (95% CI: 0.798-0.873), sensitivity of 0.793, specificity of 0.748, and a Brier score of 0.149. Hemoglobin, potassium, sodium, low-density lipoprotein cholesterol, and age at onset were identified as the top predictors. Hemoglobin, potassium, sodium, and BMI exhibited U-shaped associations with hypoglycemia risk, where both low and high values increased risk. Exploratory analysis of joint biomarker status showed that patients with abnormalities in two or more of these core predictors had a non-significant trend toward higher event rates, while the complexity of their combined effects was better captured by the non-linear model. The model enabled effective risk stratification into four quartiles, and decision curve analysis confirmed its consistent net clinical benefit across relevant probability thresholds.

The interpretable random forest model using routine inpatient data showed strong discrimination, good calibration and useful risk stratification for in-hospital hypoglycemia in Chinese adults with T1DM, which may help identify high-risk patients early and guide targeted preventive interventions in clinical practice.

To develop a machine learning-based classification model to aid in the early diagnosis of diabetic microvascular complications.

This study analyzed clinical and laboratory data from 1,498 patients, categorized into two groups: diabetes alone and diabetes with microvascular complications. Independent risk factors for complications were identified through intergroup comparison, collinearity analysis, and logistic regression. Nine machine learning models were subsequently developed and compared. A comprehensive evaluation of the binary classification performance of the Gradient Boosting Decision Tree (GBDT) model was performed.

Urea, fibrinogen (FIB), prothrombin time (PT), D-dimer (DD), creatine kinase MB isoenzyme (CKMB), lipoprotein(a) (Lpa), activated partial thromboplastin time (APTT), triglycerides (TG), and cholinesterase (CHE) were identified as independent risk factors for diabetic microvascular complications. Among the nine predictive models constructed, the GBDT model demonstrated superior performance across multiple metrics, including the area under the receiver operating characteristic curve (AUC) and sensitivity, indicating strong generalization ability on the validation set. Further evaluation confirmed its consistent and robust predictive performance across training, validation, and test datasets. Calibration curve analysis showed good agreement between predicted probabilities and actual outcomes. Decision curve analysis demonstrated the model's clinical utility, and the Kolmogorov-Smirnov (KS) curve indicated excellent discriminatory power.

The GBDT model, constructed based on the identified risk factors, exhibits outstanding predictive performance and promising application potential. It provides important theoretical support and a practical tool for the early identification and targeted intervention of diabetic microvascular complications.

Diabetic foot ulcers (DFUs) represent a severe and prevalent complication of diabetes, contributing to substantial disability and elevated mortality. This study aimed to develop and validate a diagnostic nomogram for Wagner Grade ≥2 DFUs in hospitalized patients with type 2 diabetes (T2DM).

This retrospective cohort study included 510 hospitalized patients with T2DM treated at the Second Affiliated Hospital of Fujian Medical University between January 2023 and December 2025, of whom 248 had Wagner Grade ≥2 DFUs. Patients were randomly divided into a training set (n=357) and an internal validation set (n=153) in a 7:3 ratio. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors and construct a diagnostic nomogram. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration curves, decision curve analysis (DCA), and clinical impact curves. External validation was performed in an independent cohort of 154 patients from Quanzhou Southeast Hospital, including 86 with Wagner Grade ≥2 DFUs.

The nomogram incorporated four independent predictors: Angle α, K time, platelet count (PLT), and lymphocyte count. The model exhibited excellent discrimination in the training set (area under the curve [AUC] = 0.940, 95% confidence interval [CI]: 0.916-0.965) and internal validation set (AUC = 0.914, 95% CI: 0.870-0.956), with modest discrimination in the external validation set (AUC = 0.690, 95% CI: 0.604-0.775). Calibration curves demonstrated strong concordance between predicted and observed probabilities. DCA and clinical impact curves confirmed substantial clinical utility across all cohorts.

This nomogram, integrating thromboelastography parameters and hematological indicators, provides a practical tool for identifying the presence of Wagner Grade ≥2 DFUs in hospitalized patients with T2DM, supporting early risk stratification and informed clinical decision-making.

Cardiometabolic multimorbidity (CMM) has become an increasingly serious public health problem. Patients with type 2 diabetes mellitus (T2DM) often present with multiple cardiometabolic disorders and carry a significantly higher risk of CMM. Insulin resistance (IR) is the core mechanism of T2DM and atherosclerotic cardiovascular disease. The triglyceride-glucose index (TyG index) can serve as a reliable alternative for evaluating IR. The pericoronary fat attenuation index (FAI) is a non-invasive biomarker of coronary inflammation based on coronary CT angiography. However, the combined associations of the TyG index and FAI with CMM among patients with T2DM remain unknown. Therefore, this study aims to evaluate the TyG index, RCA-FAI, LAD-FAI, and LCX-FAI in relation to CMM among middle-aged and elderly patients with T2DM in China.

We conducted a cross-sectional study and enrolled 497 middle-aged and elderly patients (aged ≥45 years) with T2DM who underwent coronary CT angiography for clinical indications. We defined CMM as the concurrent presence of T2DM together with coronary heart disease or stroke. We used a multivariate logistic regression model to analyze the association between the TyG index and the FAI in each coronary segment (including RCA-FAI, LAD-FAI, and LCX-FAI) with CMM. We presented the study results as odds ratios (ORs) with their corresponding 95% confidence intervals (CIs). We employed restricted cubic splines to analyze the nonlinear relationship and used receiver operating characteristic (ROC) curves to assess the discriminatory capacity of each index in identifying CMM.

After fully adjusting for confounding factors, the TyG index (OR = 2.07, 95% CI: 1.44-2.99), RCA-FAI (for each increase of 1 unit: OR = 1.19, 95% CI: 1.14-1.23), LAD-FAI (OR = 1.16, 95% CI: 1.12-1.21), and LCX-FAI (OR = 1.11, 95% CI: 1.07-1.15) were all significantly and positively associated with CMM (all P < 0.001).Dosage-response analysis revealed nonlinear associations of the TyG index and LAD-FAI with CMM (P for nonlinearity < 0.05), whereas RCA-FAI and LCX-FAI showed linear relationships. Receiver operating characteristic (ROC) curve analysis was further performed to evaluate the discriminatory performance of each indicator for CMM. Among these indices, adding the RCA-FAI showed the most pronounced improvement, with a C-statistic of 0.900 (95% CI: 0.873-0.926, P < 0.001), a net reclassification improvement (NRI) of 0.749 (95% CI: 0.585-0.913, P < 0.001), and an integrated discrimination improvement (IDI) of 0.141 (95% CI: 0.110-0.171, P < 0.001). In contrast, adding the TyG index did not meaningfully improve the predictive value of the baseline clinical model.

This study confirms that among middle-aged and elderly Chinese patients with T2DM, both the TyG index and FAI, including RCA-FAI, LAD-FAI, and LCX-FAI, are independently and positively associated with CMM. However, only coronary FAI indices significantly improve the discriminatory capacity for CMM, with RCA-FAI showing the strongest association and incremental predictive value. These findings suggest that FAI could serve as a useful imaging biomarker for identifying CMM status in patients with T2DM.

Patients with hypertension commonly exhibit a persistent chronic inflammatory state. Accumulating evidence suggests that inflammation can induce insulin resistance, which is a key pathological mechanism in the development of diabetes. However, whether composite inflammatory indices are independently associated with the risk of incident diabetes in hypertensive patients remains insufficiently supported by current evidence.

Multivariable-adjusted Cox proportional hazards regression models were used to assess the associations between composite inflammatory indices and the risk of incident diabetes in patients with hypertension. Kaplan-Meier (KM) curves were generated to visually depict the cumulative incidence of diabetes across different levels of these inflammatory indices. Furthermore, comparative analyses were conducted to identify the most predictive composite inflammatory marker.

Cox regression analysis revealed that elevated composite inflammatory indices were significantly associated with an increased risk of incident diabetes in patients with hypertension. KM curves further demonstrated that individuals with higher inflammatory levels exhibited a significantly higher cumulative incidence of diabetes during follow-up compared to those with lower levels. Moreover, comparative analyses among the inflammatory markers identified the inflammatory burden index (IBI) as the most effective predictor of diabetes risk.

This study demonstrates that elevated composite inflammatory indices are closely associated with an increased risk of future diabetes in patients with hypertension, with a notable threshold effect. These findings suggest that actively controlling inflammatory levels may help reduce the risk of diabetes in this population. Furthermore, the IBI holds promise as a simple and accessible risk assessment tool for screening and early identification of individuals at high risk for diabetes.

This study aimed to evaluate the independent discriminative performance of elevated Lipoprotein(a) [Lp(a)] in identifying prevalent cardiovascular disease (CVD), specifically in patients with Diabetes Mellitus (DM), and to determine the optimal cutoff values for identifying CVD in the DM population.

A stratified analysis was conducted across general, DM, and non-DM patient groups. Correlation analysis was employed to assess the association between elevated Lp(a) levels and metabolic factors in different age groups. Furthermore, binary regression was used to calculate combined risk scores. Receiver operating characteristic (ROC) curve analysis was utilized to determine the area under the curve (AUC) for the discriminative performance of Lp(a) alone, traditional parameters (excluding Lp(a)), and the combined model. This analysis identified the optimal cutoff values for each group.

Comparisons of Lp(a) variations showed that Lp(a)>300mg/L was associated with an increased prevalence of CVDs in general and non-diabetic patients, while it was insignificant in patients with DM, unless the cutoff was set as low as 70mg/L; Correlation analyses showed that, regardless of minor nuances between the general and DM groups, Lp(a) was significantly related to Low-density lipoprotein (LDL) and Apolipoprotein B (ApoB), but negatively related to Glycated Hemoglobin A1c (HbA1c), Triglycerides (TG), and free triiodothyronine (FT3); when stratified by age, no correlation was associated with Lp(a), but an association was found with CVD in the 65-75 age group, while Non-alcoholic fatty liver disease (NAFLD) prevalence was higher in the <65 age group across groups except for Lp(a)>70mg/L in the DM group; correlational analyses revealed that Lp(a) was positively related with CVD in the <65 age group compared to the general group. Regression analyses revealed that HbA1c and age significantly contributed to increased CVD in DM and that Lp(a) was determined by FT3 and albumin (Alb) in DM; ROC curves demonstrated that the combination of Lp(a) with traditional parameters significantly enhanced the AUC for CVD in DM.

Elevated Lp(a) levels are significantly associated with CVD and demonstrate strong discriminative utility, particularly in patients with DM. These findings suggest that more stringent Lp(a) thresholds may be warranted in the clinical management of diabetic patients to better identify individuals at high risk for cardiovascular outcomes.