Data collection on retinopathy as a public health tool: The Hubble telescope equivalent of looking back in time

https://doi.org/10.1016/j.jdiacomp.2016.12.016Get rights and content

Abstract

Objective

To test whether the rate of diabetic retinopathy development in a population calculated from the prevalence of retinopathy and duration of diabetes can be used to assess their prior glycemic control.

Research Design and Methods

9281 patients with type 2 diabetes (T2DM) were grouped by duration of diabetes and plotted against the % of retinopathy in each band. The slope was used to calculate retinopathy development/year (RD/y). We correlated the RD/y with updated HbA1c within groups of different ethnicity, age of diabetes onset, year of the eye examination, socio-economic status and fluency in English.

Results

Differences in ethnicity, age of diabetes onset and year of the eye examination affect RD/y to a degree predictable from their respective updated HbA1c. No such relationship with updated HbA1c was evident when a factor has no apparent effect on RD/y.

Conclusions

This relationship between prevalence of retinopathy and duration of diabetes can be used to assess future retinopathy burden. Perhaps more intriguing, the camera can be reversed to allow an estimate of prior glycemic control of a population from its retinopathy prevalence. Health care organizations can use this method to project future needs and to assess adequacy of prior glycemic control.

Introduction

Diabetic retinopathy (DR) is an important complication of diabetes which causes considerable morbidity and also associated with significant increase in mortality of those affected. At an individual level, the presence of retinopathy is best predicted, although far from perfectly, by the duration of diabetes and the overall glycemic exposure of the individual in the previous years (Klein et al., 1994, Nathan et al., 1986, Wong et al., 2008, Nordwall et al., 2015). By contrast to the relatively poor predictability at an individual level, investigators have previously demonstrated a strong linear relationship between duration of diabetes and prevalence of retinopathy at a population level, e.g., in a Wisconsin and a Western Australian population (Harris, Klein, Welborn, & Knuiman, 1992). The slope of such relationship between prevalence of retinopathy and duration of diabetes is a measure of the rate of retinopathy development per unit time for that population. We used patient information stored in our electronic database to test the hypothesis that the rate of retinopathy development calculated in this manner can be used to estimate the prior glycemic control of the population.

Our electronic database collected information including retinopathy status, duration of diabetes and glycemic control of our patients over more than two decades (McGill, Molyneaux, Yue, & Turtle, 1993). We used these data to construct the linear relationship between duration of diabetes and prevalence of retinopathy. As duration of diabetes is already accounted for in the x-axis, the slope of the linear regression should correlate well with the prior glycemic control of the individual group being examined. The implication being that, if this assumption is proven, cross-sectional data of retinopathy prevalence at various duration of diabetes in a defined population can be used to retrospectively assess the adequacy of glycemic control of that population in the preceding years. Health Maintenance Organizations and some community screening programs could have information on the status of retinopathy and duration of diabetes of their participants but not serial measurements of HbA1c (Scanlon, Aldington, & Stratton, 2014). Their data could be analyzed by our method to serve as an indirect measurement of the average glycemic control of their patients in the preceding 1–2 decades. By examining data of patients with defined criteria (e.g., according to their ethnicity or age of diabetes onset), the prior glycemic control of specifically defined groups of diabetic patients can be compared. Our method which only requires cross-sectional and retrospective data could be a simple but useful public health tool.

Section snippets

Methods

The findings of this study were derived from a total cohort of 9281 patients with T2DM who had information collected prospectively during clinical consultations over a period of approximately two to three decades. For this study, the following data from each patient were retrieved from our electronic database: retinopathy status, HbA1c, ethnicity, date of diabetes diagnosis, date of last consultation with documented retinopathy data, socio-economic status and fluency in English. Retinopathy was

Results

Diabetes duration and updated HbA1c were relatively poor predictors for the presence of retinopathy in individuals. These two factors only accounted for 13.8% of the variance for the presence of retinopathy in individuals in the total cohort of 9281 patients. Other factors examined including age of diagnosis, blood pressure, smoking status and gender did not add to the prediction of retinopathy in individuals. By contrast, as shown in Fig. 1A and B, the relationship between duration of diabetes

Discussion

It is well accepted that glycemic control and duration of diabetes are the major determinants for the presence of DR, but their predictive power for individual patients is only modest (Nathan, 2014, Group TDCaCTR, 1995). In our cohort these two factors only account for 13.8% of the variance of retinopathy, underlying the importance of routine screening in clinical practice. A less appreciated but the fundamental basis of our study is that, in contrast to the situation for individuals, duration

References (16)

  • J.R. Brownrigg et al.

    Microvascular disease and risk of cardiovascular events among individuals with type 2 diabetes: A population-level cohort study

    The Lancet Diabetes & Endocrinology

    (2016)
  • Group TDCaCTR

    The relationship of glycemic exposure (HbA1c) to the risk of development and progression of retinopathy in the diabetes control and complications trial

    Diabetes

    (1995)
  • M.I. Harris et al.

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

    Diabetes Care

    (1992)
  • R. Klein et al.

    The Wisconsin Epidemiologic Study of Diabetic Retinopathy. XIV. Ten-year incidence and progression of diabetic retinopathy

    Archives of Ophthalmology

    (1994)
  • C.K. Kramer et al.

    Diabetic retinopathy predicts all-cause mortality and cardiovascular events in both type 1 and 2 diabetes

    Meta-Analysis of Observational Studies

    (2011)
  • J. Leal et al.

    Temporal validation of the UKPDS outcomes model using 10-year posttrial monitoring data

    Diabetes Care

    (2013)
  • M. McGill et al.

    A single visit diabetes complication assessment service: A complement to diabetes management at the primary care level

    Diabetic Medicine

    (1993)
  • D.M. Nathan

    The Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications study at 30 years: Overview

    Diabetes Care

    (2014)
There are more references available in the full text version of this article.

Cited by (2)

Conflicts of interest: The authors have no relevant conflicts of interest to declare.

View full text