Abstract
Purpose::
To correlate the simplified International Clinical Diabetic Retinopathy Severity (ICDRS) Scale for diabetic retinopathy (DR), to the "gold standard" of stereoscopic 7-field fundus photographic gradings. The 5-stage ICDRS scale, designed to improve screening and care of DR, includes: 1) no DR, 2) mild nonproliferative DR (NPDR) 3) moderate NPDR, 4) severe NPDR, and 5) proliferative DR.
Methods::
The Family Investigation of Nephropathy and Diabetes (FIND) study is a multi-center cross-sectional genetic study of patients with diabetic nephropathy. Dilated patients examined by general ophthalmologists or retina specialists using direct and/or slit lamp ophthalmoscopy were scored on the 5-stage ICDRS scale. Standardized stereoscopic 7-field fundus photographs were also taken and graded by masked readers on the ETDRS scale and recoded to a matching 5-point scale.
Results::
1851 patients (3602 eyes) in the FIND study had a mean age of 58.0±10.8 years, and a mean duration of diabetes of 15.7 ±10.2 years. 43% were on insulin treatment, and 2/3 had some degree of diabetic retinopathy on fundus photography. Comparisons between the ICDRS and recoded ETDRS scores reveal a good degree of agreement (weighted Κ=0.79, 95% limits: 0.77-0.80), with 73.3% of eyes having complete agreement, and 91.2% agreeing within 1 stage. In cases of non-agreement, ICDRS scores tended to underestimate, rather than overestimate, disease severity (p<0.001). Using the ETDRS scoring as the "gold standard", the ICDRS scoring was able to detect the presence of any retinopathy (≥ETDRS level 20) with a sensitivity of 87.9% and a specificity of 92.4%, and the presence of proliferative disease (≥ETDRS level 60) with a sensitivity of 81.8% and a specificity of 97.3%. In 4.7% of cases, the ICDRS scoring recommended routine follow-up (stages 1 and 2) when scatter laser treatment may be appropriate (stages 4 and 5), and in 0.9% of cases, treatment would have been considered (stages 4 and 5) when only routine follow up is needed (stages 1 and 2).
Conclusions::
In this more severely affected cohort, the ICDRS scoring is effective at detecting the absence of DR but less so at detecting higher levels of DR. Additional physician education on the recognition of more advanced stages may help improve correlation to the ETDRS scale.
Clinical Trial::
www.clinicaltrials.gov NCT00342927
Keywords: diabetic retinopathy • clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology • clinical (human) or epidemiologic studies: treatment/prevention assessment/controlled clinical trials