Meta-Analysis of Cognitive Impairment Prevalence and Risk Factors in Diabetes Patients

41 studies involving 18,768 diabetes patients were analyzed to estimate cognitive impairment prevalence and risk factors. The meta-analysis found a 40.80% prevalence of cognitive impairment among diabetes patients and identified key risk factors including age, HbA1c, and diabetic complications. Eight studies contributed to the assessment of predictive models, which reported a c-index of 0.83 in training sets.

Use Cases

• Analyze the association between demographic factors like age, low income, and unmarried status and the odds ratio for cognitive impairment in diabetes patients.
• Investigate the impact of disease-related factors such as Hemoglobin A1c levels and hypoglycemia on cognitive impairment risk using reported odds ratios.
• Evaluate the performance of predictive models for cognitive impairment based on the reported c-index of 0.83 from training and validation sets.
• Examine the relationship between diabetic complications, including nephropathy and retinopathy, and their respective odds ratios for cognitive impairment.

Strengths

• Analysis is based on 41 observational studies encompassing data from 18,768 patients with diabetes mellitus.
• Identifies 10 specific risk factors for cognitive impairment, each with a quantified odds ratio and 95% confidence interval.
• Assesses predictive model performance from 8 studies, reporting a pooled c-index of 0.83 in training sets.

Limitations

• The underlying data is aggregated from published studies; the raw patient-level data is not available for novel analysis.
• The predictive model assessment is based on only 8 studies, which the authors note limits clinical applicability.
• The dataset is a 12.9 MB PDF document, requiring text extraction for computational analysis.

Provenance

Source

Systematic review and meta-analysis of studies from Embase, Web of Science, Cochrane Library, and PubMed.

Collection Method

Observational studies were identified, screened, and their data was synthesized via meta-analysis using R 4.5.0.

Time Range

Literature search covered studies published up to May 2025.

Freshness

Literature was searched up to May 2025; the record was last updated in March 2026.

Geography

Global scope, as the review aimed to estimate global prevalence.