Harris · Diabetes care 1992 · cross-sectional observational study with linear extrapolation modeling · n=?

Onset of NIDDM occurs at least 4-7 yr before clinical diagnosis.

Cited 1314 times in the scientific literature.

Level 4 - case-series / case-control

Cross-sectional observational study using regression-based extrapolation

PubMed 1516497 · doi:10.2337/diacare.15.7.815 · record verified 2026-08-27

What was done

Researchers evaluated two population-based cohorts of white individuals with non-insulin-dependent diabetes mellitus (NIDDM) in the United States and Australia. They measured the prevalence of diabetic retinopathy and the duration of diabetes following clinical diagnosis. Weighted linear regression was used to model the relationship between diabetes duration and retinopathy prevalence and to extrapolate back to zero prevalence.

What was found

Prevalence of retinopathy at clinical diagnosis was estimated at 20.8% in the United States and 9.9% in Australia, increasing linearly with duration of diabetes. Back-extrapolation indicated that detectable retinopathy began approximately 4 to 7 years before clinical diagnosis. Accounting for external data suggesting diabetes is present 5 years before retinopathy becomes detectable, the onset of NIDDM was estimated to occur 9 to 12 years prior to clinical diagnosis.

Why it matters

This paper demonstrates that type 2 diabetes often remains clinically unrecognized for up to a decade, during which microvascular complications actively develop without intervention.

Limits

The abstract does not report the total sample size (n). The study population was restricted to white patients in two countries. Findings rely on linear back-extrapolation and external assumptions regarding preclinical disease timelines rather than direct prospective longitudinal screening.

Cited by