Chromatic-contrast threshold impairment in diabetes.

A prospective study was carried out to investigate acquired colour-vision deficits in diabetics using an automated, computer-controlled, cathode-ray-tube based test of chromatic contrast. Chromatic-contrast thresholds estimates were measured along both a red/ green (constant S-cone) confusion axis and a tritan (constant M/L-cone) confusion axis for 305 eyes of 305 diabetics. The diabetic data were partitioned into groups based on a clinical categorisation of retinopathy. The diabetic data were compared with both age-matched and 'lens-equated' control data obtained from a bank of 347 normal subjects. Further analysis of differences between diabetic-status groups was performed. Associations between chromatic contrast threshold estimates and age, duration of disease, and severity of both macular oedema and ischaemia were investigated. The diabetic group was found to have significantly reduced chromatic-contrast threshold estimates when compared with normal controls, even in the absence of retinopathy. This reduction in chromatic contrast was predominantly tritanopic in nature. Interestingly, no reduction in red/green chromatic-contrast threshold estimate was found in diabetics without retinopathy. The tritan deficit seen in diabetics without retinopathy was strongly correlated with duration of disease, but when adjustments were made to account for the effects of duration-dependent lens yellowing, the tritan deficit was no longer apparent. A correlation between both the severity of macular oedema and severity of ischaemia with chromatic-contrast loss was established. Acquired reductions in both red/green and tritan chromatic-contrast threshold estimates seen in diabetics are strongly correlated with the severity of retinopathy. The results provide evidence that the specific tritan deficits seen in diabetics can be explained by the effects of lens yellowing rather than by selective damage of the blue cone system as has been hypothesised by other groups. The results provide support for the potential use of automated CRT-based tests of colour vision in diabetic retinopathy screening protocols.