Correlation between glycated hemoglobin A1c, urinary microalbumin, urinary creatinine, β2 microglobulin, retinol binding protein and diabetic retinopathy

Development and validation of predictive models for diabetic retinopathy using machine learning

OBJECTIVE

This study aimed to develop and compare machine learning models for predicting diabetic retinopathy (DR) using clinical and biochemical data, specifically logistic regression, random forest, XGBoost, and neural networks.

METHODS

A dataset of 3,000 diabetic patients, including 1,500 with DR, was obtained from the National Population Health Science Data Center. Significant predictors were identified, and four predictive models were developed. Model performance was assessed using accuracy, precision, recall, F1-score, and area under the curve (AUC).

RESULTS

Random forest and XGBoost demonstrated superior performance, achieving accuracies of 95.67% and 94.67%, respectively, with AUC values of 0.991 and 0.989. Logistic regression yielded an accuracy of 76.50% (AUC: 0.828), while neural networks achieved 82.67% accuracy (AUC: 0.927). Key predictors included 24-hour urinary microalbumin, HbA1c, and serum creatinine.

CONCLUSION

The study highlights random forest and XGBoost as effective tools for early DR detection, emphasizing the importance of renal and glycemic markers in risk assessment. These findings support the integration of machine learning models into clinical decision-making for improved patient outcomes in diabetes management.

QiMing granules for diabetic retinopathy: a systematic review and meta-analysis of randomized controlled trials

Objective

This study aimed to assess the efficacy and safety of QiMing granules (QM) in the treatment of patients with diabetic retinopathy (DR).

Methods

We systematically searched multiple databases, including Pubmed, Embase, Web of Science, Cochrane Library, SinoMed, Chinese National Knowledge Infrastructure (CNKI), Wanfang database, and VIP database. Randomized controlled trials (RCTs) of QM in the treatment of DR were collected, and the search time limit was from the establishment of the database to 27 March 2024. Two independent researchers were involved in literature screening, data extraction, and bias risk assessment. The risk of bias in the included studies was assessed using the Risk of Bias Assessment tool for randomized controlled trials of Cochrane Collaboration 2.0 (RoB 2.0). The main outcomes were the overall efficacy, visual acuity, retinal circulation time, macular thickness. The secondary outcomes were the levels of triglyceride (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and glycated hemoglobin (HbA1c). The adverse events was considered the safety outcome. Review Manager 5.4.1 and Stata 15.1 were used for meta-analysis. Data were pooled by random-effects or fixed-effects model to obtain the mean difference (MD), risk ratio (RR), and 95% confidence interval (CI).

Results

A total of 33 RCTs involving 3,042 patients were included in this study. Overall, we demonstrated that QM had a significant clinical effect on DR. QM alone was superior to conventional treatment (CT) in terms of overall efficacy [RR = 1.45, 95% CI: (1.34, 1.58), p < 0.00001, moderate certainty], retinal circulation time [MD = -0.56, 95% CI: (-1.01, -0.12), p = 0.01] and macular thickness [MD = -11.99, 95% CI: (-23.15, -0.83), p = 0.04]. QM plus CT was superior to CT in terms of overall efficacy [RR = 1.29, 95% CI: (1.24, 1.33), p < 0.00001], visual acuity [MD = 0.14, 95% CI: (0.11, 0.17), p < 0.00001], macular thickness [MD = -14.70, 95% CI: (-21.56, -7.83), p < 0.0001], TG [MD = -0.20, 95% CI: (-0.33, -0.08), p = 0.001, moderate certainty], TC [MD = -0.57, 95% CI: (-1.06, -0.07), p = 0.02], and LDL-C [MD = -0.36, 95% CI: (-0.70, -0.03), p = 0.03]. In terms of safety, the incidence of adverse events in the experimental group was less than that in the control group. The results of the GRADE evidence quality evaluation showed that the evidence quality of outcome indicators was mostly low.

Conclusion

QM can effectively improve overall efficacy, visual acuity, macular thickness, retinal circulation time, and reduce the levels of TG, TC, and LDL-C. However, due to the limited number of studies included, a small sample size, and a lack of high-quality literature, the possibility of publication bias cannot be excluded. Moreover, biases are present due to differences in study design, such as the absence of placebo use in the control group and a predominant use of combined intervention designs in the control group, along with deficiencies in allocation concealment and blinding methods. Therefore, more multi-center, large-sample, and rigorously designed studies are needed to substantiate this conclusion.

Systematic review registration

https://www.crd.york.ac.uk/PROSPERO/#recordDetails, identifier CRD42023465165.