The use of systematic data collection on retinopathy to assess prior glycaemic control:  The Hubble telescope equivalent of looking back in time — ASN Events
  1. Schedule
  2. ADS Clinical Young Investigator Awards
  3. The use of systematic data collection on retinopathy to assess prior glycaemic control:  The Hubble telescope equivalent of looking back in time

The use of systematic data collection on retinopathy to assess prior glycaemic control:  The Hubble telescope equivalent of looking back in time (#93)

Maria I Constantino 1 2 , Lynda Molyneaux 1 2 , Ted Wu 1 , Stephen M Twigg 1 2 , Jencia Wong 1 2 , Dennis K Yue 1 2
  1. Diabetes Centre , Royal Prince Alfred Hospital, Camperdown, NSW, Australia
  2. Discipline of Medicine, University of Sydney, Sydney, NSW, Australia

The ability of HbA1c and duration of diabetes to predict the presence of retinopathy in individual patients is only modest (14% of variance) in a cohort of 10,259 type 2 diabetic patients in our electronic database.  By contrast, this study shows that duration of diabetes is an excellent predictor of prevalence of retinopathy when individuals are grouped by duration of diabetes in 5 year bands (Figure 1).  The slope of increase in retinopathy over time is used to calculate retinopathy development/year.  We used this method to study 14 subgroups {7 ethnicities, 2 time-periods (< and >2000), 2 language cohorts (fluency in English) and 3 socio-economic strata (postcode)}. We showed that their rates of increase in prevalence of retinopathy correlated well with their respective updated HbA1c (Figure 2).

For our study, the last visit of each patient was analysed, but controlled studies showed that the use of first or last or any visit by random sampling generated rates of retinopathy development which were not different.   The relationship between retinopathy prevalence and duration up to 20 years fits best to a linear model but inclusion of patients with longer duration produced results with similar conclusion.  Controlled studies also showed that the cv of estimating retinopathy development was low at 4.6% when the total patient number was >2000, rising to 14.9% at 320 patients.

This method allows the comparison of the rates of retinopathy development of different patient cohorts.  It can be used to estimate future retinopathy burden of any population.  Perhaps more intriguing, the camera can be reversed to allow an estimate of prior glycaemic control of any population if its retinopathy prevalence at different duration of diabetes is known.  As the relationship between retinopathy prevalence and diabetes duration is self-adjusting, this method is independent of whether patients present early or late.  Only cross-sectional data is required and no long term follow up is necessary.  These features make this method feasible to be used by health care organisations with cross-sectional database on retinopathy (eg. national retinopathy screening programme) to quantify retrospectively the glycaemic control in different patient populations.

1827-Fig_1_with_%20arrow.jpg1828-Fig_2_abs.jpg