RESUMEN
BACKGROUND: The reproducibility of an individual's full-field ERG between centres has not previously been investigated. METHODS: ERGs were recorded using both silver thread and skin electrodes from the same two normal adult subjects at 15 UK centres using routine, local protocols and a highly standardised, 'ISCEV-specified' protocol matching the values specified in the ISCEV standard; where the ISCEV standard allows options, a single value was chosen. RESULTS: Inter-ocular differences were small, and amplitudes were smaller for skin than silver thread electrodes. No centre produced outlying data points, and ERGs across all 15 centres were remarkably similar. Amplitude variability was less for local protocols (using LED flashes) than for the ISCEV-specified protocol using xenon flashes (22 vs. 24 %, p = 0.01), but peak time variability was less for the ISCEV-specified protocol (6.1 vs. 7.4 %, p = 0.001). Only the DA 0.01 ERG correlated with photometric variability. The bifidity of the DA 3 a-wave doubled its peak time variability compared with the DA 10 a-wave. CONCLUSIONS: Inter-centre amplitude variability was typically within clinically significant thresholds, suggesting that inter-centre variability with suitable standardisation may not add more to total variability than inter-subject variability. Variability improvements gained by the tighter specifications of the ISCEV-specified protocol were possibly more than lost due to imprecisions of xenon flashtubes. Peak time variability was far lower than amplitude variability, corresponding with acceptable variability of biochemical assays. These results represent a vindication of the existence of an ERG standard and suggest that further standardisation would lend itself to greater reproducibility of ERGs worldwide.
