PURPOSE: To analyse and compare the prevalence of epiretinal membranes (ERMs) obtained using either standard retinal colour images or spectral-domain optical coherence tomography (SD-OCT) in a population-based setting of French elderly subjects. METHODS: Six hundred twenty-four subjects of the Alienor cohort aged 75 years or older underwent both colour fundus imaging and SD-OCT examinations. The ERMs were graded from retinal images and SD-OCT macular scans in a masked fashion. On SD-OCT images, the early ERMs, mature contractile ERMs without foveal modifications and mature contractile ERMs with foveal alterations were distinguished. RESULTS: 610 (97.8%) subjects had gradable SD-OCT examinations, and 511 (81.9%) had gradable fundus images in at least one eye. According to colour photographs, 11.6% of participants had definite ERMs. From SD-OCT images, 52.8% of the subjects had early ERMs, 7.4% had mature ERMs without foveal involvement, and 9.7% had mature ERMs with foveal alterations. Regardless of the imaging method used, the ERMs were more often observed in pseudophakic eyes than in phakic eyes. Comparison of ERM assessment using fundus photographs versus SD-OCT images demonstrated that the specificity of retinal colour images was good (>89.3%), whereas the sensitivity remained low even though it increased with ERM severity on SD-OCT images. CONCLUSIONS: Spectral-domain optical coherence tomography (SD-OCT) examinations have high feasibility in this elderly population and are much more sensitive than standard colour images for ERM assessments, especially in the early stages of the disease. Our results further highlight the need to use SD-OCT instead of colour retinal photographs for the classification of ERMs in epidemiological studies.
PURPOSE: To analyse and compare the prevalence of epiretinal membranes (ERMs) obtained using either standard retinal colour images or spectral-domain optical coherence tomography (SD-OCT) in a population-based setting of French elderly subjects. METHODS: Six hundred twenty-four subjects of the Alienor cohort aged 75 years or older underwent both colour fundus imaging and SD-OCT examinations. The ERMs were graded from retinal images and SD-OCT macular scans in a masked fashion. On SD-OCT images, the early ERMs, mature contractile ERMs without foveal modifications and mature contractile ERMs with foveal alterations were distinguished. RESULTS: 610 (97.8%) subjects had gradable SD-OCT examinations, and 511 (81.9%) had gradable fundus images in at least one eye. According to colour photographs, 11.6% of participants had definite ERMs. From SD-OCT images, 52.8% of the subjects had early ERMs, 7.4% had mature ERMs without foveal involvement, and 9.7% had mature ERMs with foveal alterations. Regardless of the imaging method used, the ERMs were more often observed in pseudophakic eyes than in phakic eyes. Comparison of ERM assessment using fundus photographs versus SD-OCT images demonstrated that the specificity of retinal colour images was good (>89.3%), whereas the sensitivity remained low even though it increased with ERM severity on SD-OCT images. CONCLUSIONS: Spectral-domain optical coherence tomography (SD-OCT) examinations have high feasibility in this elderly population and are much more sensitive than standard colour images for ERM assessments, especially in the early stages of the disease. Our results further highlight the need to use SD-OCT instead of colour retinal photographs for the classification of ERMs in epidemiological studies.