Test-retest variability for standard automated perimetry and short-wavelength automated perimetry in diabetic patients

Purpose: To assess limits for significant improvement or deterioration of visual fields in diabetic patients based on short-term test-retest variability in subjects with different degrees of retinopathy. Methods: Fifty patients with diabetic retinopathy ranging from level 10 to 75 [according to the Early Treatment Diabetic Retinopathy Study (ETDRS) severity scale] were tested repeatedly with both standard automated perimetry (SAP) and short-wavelength automated perimetry (SWAP) with short intervals. The association between visual field loss and degree of retinopathy outside fovea was analysed. Test-retest variability of global and local visual field indices and prediction limits for significant change were calculated. Results: The amount of visual field loss was significantly associated to the degree of retinopathy, with a correlation coefficient of -0.51 for SAP (P = 0.0003) and -0.45 for SWAP (P = 0.002). Global test-retest variability was smaller with SAP than with SWAP (P < 0.0001). For both SAP and SWAP, local test-retest variability was considerably smaller at test points with normal sensitivity than at test points with reduced sensitivity (P < 0.0001). Paracentral test points within 10 degrees of eccentricity had less variability than peripheral points (P < 0.0001), implying that smaller change is required to reach statistically significant improvement or deterioration at initially normal and paracentral points than at depressed points and peripherally located test points. Conclusion: Our results propose that SAP, as well as SWAP, can be useful for monitoring visual function outside fovea in diabetic patients with various degrees of retinopathy. We report a preference for SAP because of less variability generally. Limits for significant improvement or deterioration have been assessed but need future validation in a longitudinal study.