Latency of the mfVEP to diagnose glaucoma?

Short abstract

The multifocal approach is a truly innovative technique

Correct identification of early dysfunction, before the onset of irreversible glaucomatous damage, is a holy grail in glaucoma research. Rodarte et al, in this issue of BJO (p 1132), present a predominantly negative report in that respect.¹ While disappointing, negative results are vitally important and often fall prey to publication bias.²,³ Of the many approaches including morphological imaging and functional testing, these authors used an advanced electrophysiological technique, specifically the multifocal visual evoked potential (mfVEP).⁴ Visual electrophysiology assesses, step by step, the function of the visual processing chain. It has progressed from its classic arsenal of “three letter examinations” (ERG, electroretinogram⁵; EOG, electro‐oculogram⁶; VEP, visual evoked potential⁷) to four (PERG, pattern ERG⁸) and five letter methods (mfERG⁹/mfVEP, mf representing “multifocal”).

Although each of these techniques test a different and specific stage of the visual processing chain (the EOG—for example, the function of the retinal pigment epithelium), all of them have been applied in glaucoma, reportedly successful at that, even the EOG (seen on an ARVO poster, but a Diamox response has been demonstrated¹⁰). Significant findings may, of course, represent spurious chance or tiny effects, raised to arbitrary significance values either by choosing advanced disease stages or by including a very high number of participants in group comparisons.

The multifocal approach, mostly applied to the ERG,¹¹ but also to pupillary responses,¹² is a truly innovative technique¹³,¹⁴ to probe the central plus or minus 30° of the visual field quasi‐simultaneously, resulting in a functional map of responses either locally on the retina (mfERG) or of the entire visual pathway (mfVEP). To aid the non‐specialist readers, who by now will have had their measure of acronyms: when summing over all local responses, the mfVEP becomes rather similar to the normal, or “classic” VEP.

The multifocal VEP has been repeatedly demonstrated to be of some use in glaucoma as an objective form of perimetry.¹⁵ Given that the amplitudes or an appropriate signal to noise measure is the main variable, what motivated Hood's group¹ to look at latencies of the mfVEP responses in glaucoma patients? Parisi et al¹⁶ recently reported 100% sensitivity and 100% specificity in diagnosing glaucoma patients based on (classic) VEPs. Strong stuff. Early local damage might even be more apparent at appropriate locations in the mfVEP, so Rodarte and colleagues evaluated the mfVEP in 47 normal controls, 25 high tension glaucoma, and 25 normotensive glaucoma patients. Their main outcome variable was latency, looking at differences between groups and interocular differences.

Correct identification of early dysfunction, before the onset of irreversible glaucomatous damage, is a holy grail in glaucoma research

The results show rather little increase of latency in glaucoma, certainly not in the range that would markedly benefit in individual assessment. Of the receiver operating space they report a point with 30% sensitivity and 87% specificity. (I will not discuss the interocular comparison, which performs slightly better, but depends on chance differences between the eyes.)

What an exciting situation: a major discrepancy in the reports from two well known groups! Glaucoma stage does not seem to have differed too much between these two studies. Of course, there are major methodological differences: in visual field extent of the stimuli, in their spatial frequency content in the temporal parameters, etc. None of these can consistently explain the differing results to me: the higher temporal frequency used in the multifocal approach should make the method more sensitive to glaucomatous changes.¹⁷,¹⁸ In addition to the VEP, Parisi et al¹⁶ also recorded the PERG in their patients. The PERG, as a direct correlate of ganglion cell function, would be expected to be a good surrogate marker for glaucoma. Indeed, they also report 100% sensitivity and 100% specificity for the P50‐N95 amplitude of the PERG. In my laboratory, the PERG indeed is a strong predictor of progression in glaucoma but never with perfect accuracy.¹⁸,¹⁹,²⁰ While there may be shortcomings in our methods, this suggests to me that Parisi et al had an auspicious patient group. We eagerly await further developments in this field, especially to discover whether studies combining PERG, VEP, and mfVEP, in well defined patient groups, may bring us nearer to said holy grail.