Screening for sight threatening diabetic retinopathy using non-mydriatic retinal camera in a primary care setting: to dilate or not to dilate?

Objectives: To evaluate three different strategies (without or with mydriasis) for community based screening for sight threatening diabetic retinopathy (STDR).
Design: A masked cross-sectional comparative study.
Setting: Primary care centre.
Participants: 216 type-1 and type-2 diabetic patients (432 eyes).
Interventions: Screening for STDR was performed with the Topcon CRW6S non-mydriatic camera (NMC), compared with the grading of the seven-field standard stereoscopic photographs, 30 degrees field photographs with mydriasis as the gold standard method. Three fields were studied: Field 1 centred on the macula; field 2 centred on the disc and the macula and field 3 centred on the disc, on the macula and temporal upper to the macula all without and with mydriasis.
Main Outcome Measures: Sensitivity and specificity of screening method and prevalence of STDR. Strategies: field 1 (macular), addition of field 2 (nasal), addition of field 3 (superior temporal) all without and with mydriasis.
Results: The prevalence of STDR was 14.3%. When screening without mydriasis is performed, the percentage of referred patients increases from 14% (STDR) to almost 33% because of ungradable photographs. Sensitivity of detection of STDR by NMC using single field with mydriasis was 82% (95% confidence interval 72-92%) and without mydriasis 67% (54-80%). Specificity was 99% (97-100%) for single field with mydriasis and 99% (98-100%) without mydriasis. Sensitivity of detection using two fields with mydriasis was 95% (89-100%); without mydriasis (54-80%), slightly better than single field with mydriasis (72-92%). Specificity was similar.
Conclusions: Screening for STDR with NMC can be performed effectively with one non-stereoscopic retinal photograph per eye with mydriasis using tropicamide 0.5%. Screening without mydriasis increases the number of patients to be referred.