The diagnostic accuracy of photopic negative responses evoked by broadband and chromatic stimuli in a clinically heterogeneous population

Abstract

Abstract Purpose To compare the diagnostic accuracy of the photopic negative response (PhNR) elicited by red-blue (RB) and white-white (WW) stimuli, for detection of retinal ganglion cell (RGC) dysfunction in a heterogeneous clinical cohort. Methods Adults referred for electrophysiological investigations were recruited consecutively for this single-centre, prospective, paired diagnostic accuracy study. PhNRs were recorded to red flashes (1.5 cd·s·m−2) on a blue background (10 cd·m−2) and to white flashes on a white background (the latter being the ISCEV standard LA 3 stimulus). PhNR results were compared with a reference test battery assessing RGC/optic nerve structure and function including optical coherence tomography (OCT) retinal nerve fibre layer thickness and mean RGC volume measurements, fundus photography, pattern electroretinography and visual evoked potentials. Primary outcome measures were differences in sensitivity and specificity of the two PhNR methods. Results Two hundred and forty-three participants were initially enrolled, with 200 (median age 54; range 18–95; female 65%) meeting inclusion criteria. Sensitivity was 53% (95% confidence intervals [CI] 39% to 68%) and 62% (95% CI 48% to 76%), for WW and RB PhNRs, respectively. Specificity was 80% (95% CI 74% to 86%) and 78% (95% CI 72% to 85%), respectively. There was a statistically significant difference between sensitivities (p = 0.046) but not specificities (p = 0.08) of the two methods. Receiver operator characteristic (ROC) area under the curve (AUC) values were 0.73 for WW and 0.74 for RB PhNRs. Conclusion PhNRs to red flashes on a blue background may be more sensitive than white-on-white stimuli, but there is no significant difference between specificities. This study highlights the value and potential convenience of using white-on-white stimuli, already used widely for routine ERG assessment.