Evaluation of a screening algorithm for detection of diabetic retinopathy based on quantitation of retinal pathological elements


Purpose : Diabetic retinopathy (DR) screening is generally recommended once yearly. However, among diabetes patients with glycemic dysregulation, it is recommended to reduce the interval length. At Steno Diabetes Center Copenhagen (SDCC), a diabetes clinic in Denmark, screening intervals are based on quantitation of pathological findings in both the macula and peripheral retina, resulting in individualized screening intervals, not necessarily following the recommended guidelines. We followed 6000 patients between 1/1-2003 – 1/5-2017 at SDCC and assessed whether the individualized screening interval lengths were adequate for optimal patient outcomes. Methods : All screening intervals with planned interval lengths ≥12 months were assessed for occurrence of adverse events, defined as; 1)DR progression requiring treatment, at the following screening date 2)DR-related hospital contact or photocoagulation treatment within the planned interval According to guidelines, patients were considered dysregulated if they had HbA1c levels ≥53 mmol/mol and/or hypertension (>130/80 mmHg) before the beginning of an interval. Results : The average interval length was longer than recommended; 19.9 (SD 7.6) months among well-regulated patients and 19.4 (SD 7.1) months among dysregulated patients. Out of 20,251 screening intervals, 17,534 (87%) followed the planned interval. Among screenings following the planned interval, only 0.5% (n=88) resulted in an adverse event. Dysregulated patient intervals (82% of all intervals) experienced a significantly higher proportion of adverse events compared with intervals of well-regulated patients (0.55% vs. 0.28%, p<0.05). Adverse events were more frequent among type 2 diabetes compared with type 1 diabetes (0.71% vs. 0.40%, p<0.05), if screened within the planned interval. 1.62% experienced an adverse event if screened beyond the planned interval. Conclusions : The present study supports the validity of the SDCC algorithm, introducing individualized screening intervals for patients with no to moderate DR, even though this resulted in longer than recommended screening intervals for both well-regulated and dysregulated patients. Although intervals among dysregulated patients more frequently resulted in an adverse event, the proportion of adverse events was still less than 1%. Further analyses will assess the optimal screening interval for the individual patient
Original languageEnglish
JournalInvestigative Ophthalmology & Visual Science
Issue number9
Pages (from-to)4798
Number of pages1
Publication statusPublished - Jul 2019


Dive into the research topics of 'Evaluation of a screening algorithm for detection of diabetic retinopathy based on quantitation of retinal pathological elements'. Together they form a unique fingerprint.

Cite this