The Sensitivity of the Photopic Negative Response of the Flash ERG for the Detection of Glaucomatous Optic Neuropathy is Comparable to Global Visual Field Indices

Abstract

Purpose: : Although standard automated perimetry (SAP) remains the gold standard functional test for the identification of glaucomatous optic neuropathy (GON), the photopic negative response (PhNR) of the flash ERG has shown promise for this purpose. PhNR amplitudes and linear measures of SAP both correlate linearly with structural measures of GON. Comparisons of the diagnostic accuracy of the PhNR relative to perimetry are often confounded by the use of study inclusion criteria based on SAP. The aim of this study was to compare the PhNR with global indices from SAP for the detection of glaucoma for a population of healthy and glaucomatous eyes defined on the basis of optic nerve head structure and intra-ocular pressure (IOP).

Methods: : One eye each of 32 patients with GON and 23 healthy volunteers was tested. GON was defined on the result of the Moorfields Regression Analysis (MRA) of the HRTII as ‘outside normal limits’ in at least one of the 6 pre-defined optic disc sectors and IOP greater than 21 mmHg, irrespective of the results of SAP. For the ERG, a red flash was presented on a blue background following pupillary dilatation and light adaptation. The PhNR was measured from the baseline to the trough of the negative response following the b-wave. SAP mean sensitivity (MS) was obtained by averaging sensitivities for the whole visual field in linear units. Receiver Operator Curves (ROC) for the detection of GON were plotted for the PhNR, Mean Deviation, MD (dB) and MS (L^-1).

Results: : The Mean (± SD) absolute PhNR amplitude, MS and MD were reduced (t >3.0, p<0.001) in glaucomatous eyes (PhNR = -17.0±12.5µV; MD=-5.9±6.5 dB; MS=650±300 L^-1) compared to healthy eyes (PhNR = -28.4± 10.3µV; MD=-0.12 ± 1.2 dB; MS=1000±300 L^-1). The total ROC area was similar (Z>0.7, p>0.2) for the PhNR (0.78) and both visual field measures (0.84 for MD and 0.89 for MS). Likewise, the sensitivity at a fixed specificity of 90% was comparable (Z>1.10, p>0.10) for the PhNR (≈56%) and SAP global indices (MD≈69%; MS≈72%).

Conclusions: : The diagnostic accuracy of the PhNR was comparable to global SAP measures, although the power to detect differences was low in this small study. The results suggest that evaluation of factors which may further improve the diagnostic accuracy of the PhNR for clinical testing of patients with GON should be pursued.

Keywords: electrophysiology: clinical • visual fields