A number of studies have addressed the question whether glaucoma-impaired ganglion cell function can be (partially) restored by IOP-reducing interventions. Sehi et al. (2048) reassessed this question using the Pattern ERG (PERG) prospectively after Ventura & Porciatti's pilot study.1 The PERG is mainly generated by retinal ganglion cells and has proven a useful objective biomarker to reflect glaucoma damage.
Sehi et al. Recorded PERGs before and three months after one-sided surgery (trabeculectomy or drainage implant). The 47 eyes of 47 patients that passed their criteria had a mean MD of -11 and PSD of 8 (standard automated perimetry, SAP), average age was 70 years, mean visual acuity 0.2 logMAR or better. Their surgical intervention successfully reduced IOP reduction from 20 to 10 mmHg. The PERG amplitude increased by 24% from 0.37 ± 0.18 to 0.46 ± 0.22 µV, highly significant in a paired t-test and remaining significant after the missing correction for multiple testing. To compare their amplitudes to the normal range I assume that their setup is comparable to Porciatti & Ventura2 (2004) who report for a 70-year population 0.85 µV ± 45%. Thus Sehi et al.'s amplitudes are below the norm before and after surgery. This is no surprise as there was, interestingly, no significant change in the SAP parameters after surgery. More complicated is the phase data. Sehi and her colleagues report that the PERG phase reduces from 1.81 to 1.72 ?πrad. In their introduction they correctly state that reduction in their phase corresponds to an increase in latency. In nearly any disease the electrophysiological parameters indicate slower processing, which may explain the later mix-up in Sehi et al.'s discussion 'the reduced [..] phase after surgery suggests a reduction in latency.' For the PERG, loss of ganglion cell function reduces latency,3 probably because this un-masks a faster mechanism. Thus the reduced phase (≈0.09 πrad difference) actually means an increase in latency (≈2.8 ms), good news for ganglion cell function.
A possibly confounding factor in PERG interpretation is acuity: If acuity is reduced by optical factors (incorrect refraction, misaccommodation, media opacities) the PERG amplitude reduces.4 Here the authors carefully only included eyes with a pre-operative acuity above 20/30. Mean post-operative acuity (0.07 logMAR) was a little lower than the pre-operative one (0.06 logMAR), thus their interpretation of recovered ganglion cell function as mirrored in the PERG seems not marred by an optical artifact.