Intertest variability of mfVEP amplitude: Reducing its effect on the interpretation of sequential tests

Purpose: The multi-focal visual evoked potential (mfVEP) has been recently introduced as an alternative to subjective perimetry in detecting visual field defects. This study examines the source of variability in the mfVEP amplitude, and determines the relationship of this variability to the strength of the signal itself across the visual field. It also investigates possible means to reduce the effects of this variability on between-test interpretation to allow for easier detection of progression. Methods: 85 normal subjects participated in the study. The mfVEP was recorded using Accumap™ (ObjectiVision Pty Ltd, Sydney, Australia). Each subject was tested twice with an interval between visits of 3-4 weeks. Comparison between tests was performed using coefficient of variability (CV). Variability was also analysed using scaling and clustering procedures. Results: In the majority of the retinal areas CV fell within 15-20%. Variability increased with eccentricity, but there was no age dependency. There was a significant reduction of variability (by 15.8 ± 6%, Student's t-test p < 0.0001) when a scaling procedure was applied and this was consistent at all eccentricities. A clustering procedure reduced variability on average by a further 18.5 ± 4.5% (Student's t -test p < 0.0001). This result was also consistent at all eccentricities. Conclusion: Between test comparisons of raw mfVEP traces is limited by a variability of at least 15%. While this variability required the amplitude of the individual VEP signal to change by 30-40% in order to detect progression, scaling and clustering procedures were able to reduce the required change to 20-25%, thus making an interpretation of consecutive test results more clinically viable.