Repeatability of a novel chromatic perimeter in a cohort of people with and without glucose dysfunction.

SIGNIFICANCE

Sensitive diagnostic techniques are crucial for the early detection of diabetes. Here, we tested a novel color perimeter and found it repeatable in control subjects. It also has the potential to detect functional changes in diabetic eye disease.

PURPOSE

Type 2 diabetes (T2DM) causes retinal dysfunction before the appearance of retinopathy. Our group developed a color perimeter that measures chromatic sensitivity at discrete retinal locations. This study evaluates the repeatability of this perimeter in detecting chromatic sensitivity changes in normal and glucose-dysfunctional states.

METHODS

Eighteen subjects, 49.4 (7.8) years, nine controls (HbA1c ≤5.6%), nine with glucose dysfunction (HbA1c ≥5.7%, no/mild retinopathy), participated. Congenital color defects, diseases affecting color vision, vision >20/32, and those with tinted glasses were excluded. Visual acuity, L'Anthony D-15, and color perimetry were performed at two visits within 9 months. Lens and fundus photographs were taken. Chromatic thresholds were measured at four diagonal locations 3° from fixation in the right eye using 0.5° diameter red, green, blue, and yellow stimuli through a randomly interleaved staircase method. Thresholds were converted to log cone contrast and averaged. The test-retest repeatabilities were analyzed using Bland-Altman plots. Trends were analyzed using linear regression, and Wilcoxon matched-pairs signed rank test and t tests for comparisons.

RESULTS

Clinical tests did not differ between visits (all p ≥ 0.05). Bland-Altman control group plots showed bias lines and measurement differences near zero for all colors (all p > 0.53). Trend lines were horizontal and near zero for all colors (all p > 0.22). In the glucose dysfunction group, the mean of measurement differences was not significantly different from zero for all colors (all p > 0.13). However, green, blue, and yellow stimuli had significant or trending negative-sloped trend lines (p = 0.03, 0.04, and 0.06, respectively). Detailed review of individual clinical data revealed that subjects with relatively longer duration of T2DM, higher HbA1c, and/or mild retinopathy often performed worse on their second visit.

CONCLUSIONS

The novel color perimeter demonstrated both good repeatability and potential for detecting subtle visual changes due to glucose dysfunction. Further testing is necessary to compare performance with other structural and functional tests.