This narrative review synthesizes 50 years of observational cohort research on type 2 diabetes (T2D), drawing on data from general population-based, disease-based, intervention-based, and registry-based cohorts across every continent, organized around four themes: diagnosis/screening, complications, risk factors, and pathophysiology.
Key findings include: ≥50% of adults with T2D were previously undiagnosed in multiple cohorts; HbA1c ≥48 mmol/mol (6.5%) is validated as a diagnostic threshold via retinopathy association; screen-detected T2D carries a ~twofold lower all-cause mortality than clinically diagnosed T2D; T2D confers ~double the CVD risk in men and ~triple in women vs. those without T2D; 91% (95% CI 83–95%) of T2D cases were attributed to modifiable lifestyle factors in the Nurses' Health Study; and five reproducible T2D subtypes (ANDIS/German Diabetes Study) show distinct complication profiles.
- Narrative (not systematic) review; cohort selection was a convenience sample, so many cohorts and findings may be omitted. - Historical cohorts systematically under-represent young adults, workers, low-SES groups, and non-European ethnicities, limiting generalizability and risking lead-time bias in screening studies. - Observational designs are inherently prone to unmeasured confounding; causal inference requires complementary RCT evidence.
Cohort evidence supports targeted T2D screening using validated risk scores (e.g., Indian Diabetes Risk Score), with HbA1c ≥48 mmol/mol as a reliable diagnostic cut-off. Clinicians should address the full spectrum of T2D comorbidities—including cognitive decline, cancer, mental health, and musculoskeletal disease—not just cardiovascular and microvascular outcomes.
Explore related topics
