Visual Acuity Variability: Comparing Discrepancies between Snellen and ETDRS Measurements among Subjects Entering Prospective Trials

Abstract

Purpose: To compare the variability between Snellen visual acuity (VA) and Early Treatment Diabetic Retinopathy (ETDRS) best-corrected VA (BCVA) measurements.

Design: Retrospective chart review.

Participants: Eyes from subjects entering 12 prospective retinal trials in a large, urban retina practice.

Methods: Eyes were included if a Snellen VA measurement was performed at the visit preceding trial screening and VA was better than counting fingers. Snellen VA and ETDRS BCVA were then converted to logarithm of the minimum angle of resolution (logMAR) units, and the variability between measurements was calculated.

Main outcome measures: Outcome measures include VA variability among disease states, absolute VA, and central subfield thickness (CST).

Results: A total of 773 eyes of 413 subjects were identified with a mean of 27.2 days (median, 19; 95% confidence interval [CI], 25.1-29.3) between measurements. Mean Snellen and ETDRS measurements were 0.40 (20/50) and 0.27 logMAR (20/40), respectively. Overall, mean ETDRS BCVA was 6.1 letters better than Snellen VA (median, 5.8; 95% CI, 5.3-7.0; P < 0.05); 76.6% of eyes had improved ETDRS. Restricting the number of days between measurements did not result in any meaningful differences from this directionality. Among eyes with worse VA, variation was more pronounced than among eyes with better VA; eyes 20/25 or better were a mean +1.9 letters better on ETDRS testing (P < 0.05) and eyes 20/160 or worse were a mean +12.6 letters better on ETDRS testing (P < 0.05). Subgroup analyses by disease state found statistically significantly better vision measurements with the ETDRS protocol compared with Snellen in 4 of the 5 disease states studied. Although lens status did not affect the extent of discrepancy between ETDRS and Snellen measurements, amount of retinal edema (CST) did: increased CST correlated with increased variability.

Conclusions: The ETDRS protocol BCVA measurement resulted in significantly better scores when compared with Snellen VA measurements. This difference was more pronounced among eyes with worse VA. Additionally, specific retinal disease states and anatomic variables such as extent of retinal edema (CST) may have a meaningful impact on the anticipated variability between ETDRS and Snellen VA measurement.