Pambianco · Diabetes 2006 · Prospective observational cohort study · n=906

The 30-year natural history of type 1 diabetes complications: the Pittsburgh Epidemiology of Diabetes Complications Study experience.

Cited 493 times in the scientific literature.

Level 3 - non-randomized controlled study

Prospective non-randomized cohort study evaluating longitudinal outcomes across historical diagnosis cohorts.

PubMed 16644706 · doi:10.2337/db05-1423 · record verified 2026-08-29

What was done

Participants with childhood-onset type 1 diabetes from the prospective Pittsburgh Epidemiology of Diabetes Complications Study (n = 906) were stratified into five cohorts by year of diagnosis: 1950-1959, 1960-1964, 1965-1969, 1970-1974, and 1975-1980. Mortality, renal failure, and coronary artery disease (CAD) were assessed in the complete cohort at 20, 25, and 30 years of duration. Overt nephropathy, proliferative retinopathy, and neuropathy were evaluated at 20 and 25 years in the subset of participants who underwent clinical examination.

What was found

Mortality, renal failure, and neuropathy showed significant downward trends across successive diagnosis cohorts (P < 0.05). The 1950-1959 cohort had a fivefold higher mortality rate at 25 years of diabetes duration compared to cohorts diagnosed in the 1970s. In contrast, proliferative retinopathy (P < 0.16) and overt nephropathy (P < 0.13) showed non-significant declines at 20 years and no change at 25 years. CAD event rates were lower overall and showed no significant secular trend across cohorts.

Why it matters

This study provides US-based longitudinal data demonstrating substantial reductions in mortality, end-stage renal disease, and neuropathy across diagnosis eras, while highlighting that CAD and microvascular complications remain persistent long-term challenges.

Limits

The study is restricted to a single geographic cohort with childhood-onset diabetes. The sample size of the subset receiving clinical examinations is not reported in the abstract, and numerical incidence rates or confidence intervals for specific endpoints are largely omitted.

Cited by