Assessment of macular function, structure and predictive value of pattern electroretinogram parameters for postoperative visual acuity in patients with idiopathic epimacular membrane

Idiopathic epimacular membranes (EMMs) are hypocellular, largely collagen structures caused by glial proliferation through a defect in the internal limiting lamina usually in process of posterior vitreous detachment [1]. Their prevalence increases significantly with age, from <1 % in people under 50 years to 15 % or more in those over 70 years of age [2]. Treatment of EMMs includes pars plana vitrectomy (PPV) and membrane peeling (MP), with about 80 % of patients achieving visual acuity (VA) improvement of two or more lines and reduction in metamorphopsia [3].

Success of the EMMs treatment is usually assessed by visual acuity test and optic coherent tomography (OCT) imaging. Multifocal electroretinogram (mfERG) has been used to investigate the physiological changes in eyes with idiopathic EMM [4‐7]. The results of these studies suggested that this test can objectively show macular dysfunction not limited only to the fovea but might also be a valuable tool in predicting the postoperative visual function. However, mfERG originates from preganglionic elements [8], so this test does not evaluate function of ganglion cells of the macular region. The bioelectrical function of these cells, which can be also abnormal in patients with EMMs, can be assessed by pattern PERG [9]. In the available literature, only one study described macular function using PERG and focal electroretinogram (fERG) before and after epimacular membrane surgery, but with no International Society for Clinical Electrophysiology of Vision (ISCEV) standards [10]. It was shown that decrease in visual acuity is related to dysfunction of preganglionic and ganglionic macular elements and that surgical removal of EMM may induce functional improvement of outer and inner retinal layers connected with increasing of visual acuity.

The goal of present study was to evaluate macular function and structure before and after epimacular membrane surgery and to estimate the usefulness of ISCEV standard pattern ERG parameters in predicting the postoperative visual acuity.

Thirty-two eyes of 32 selected, pseudophakic patients (mean age 70.8 ± 6.7 years; 10 men and 22 women) were included in this study. The clinical diagnosis of EMM was confirmed by patient’s complaints, indirect binocular ophthalmoscopy (Volk Optical +90 D non-contact lens) and OCT (an increase in macular thickness with optically hyperreflective epiretinal echoes from membrane). Patients presenting poor cooperation and with ocular and systemic diseases with known influence on the PERG results were excluded from these study.

Electrophysiological results from EMM eyes were compared with normative data obtained from 30 eyes of 15 healthy subjects. The controls were matched so there were no significant differences between age, sex and refractive errors between them and examined patients.

Before and 12 months after surgery the following examinations were performed: distance best-corrected visual acuity (DBCVA) using logMAR—ETDRS chart, foveal thickness (Zeiss Cirrus OCT, Humphrey Instruments model 3000, Carl Zeiss Inc., Dublin, California) and transient PERG according ISCEV standards [11].

Transient PERG was recorded with the RetiPort (Roland Consult Instr.) system. Protocol of the PERG test was implemented in the original software of the system. Patient’s pupils were not dilated, monocular stimulation was used, refraction correction was applied with respect to the eye–screen distance 0.5 m, and interruptions of the test were introduced when frequent blinking or fixation loose was observed (patient was monitored with a TV camera).

Parameters of the pattern stimulation were as follows: 21″ CRT monitor with a frame rate equal to 75 fps was used; black and white reversing checkerboard (30° Field of View) was presented to the patient, with a check size equal to 1°2′; luminance for white elements: 120 cd/m², contrast: 97 %; temporal frequency was equal to 4.6 rps (2.3 Hz); and central fixation was used. Thread DTL electrode was used as active, and gold disk was placed at the ipsilateral outer canthus as reference, with ground (gold disk) electrode placed on the forehead (Fpz).

Parameters of the recording system were as follows: amplifiers sensitivity: 20 μV/div, filters: 1–100 Hz. Notch filters: off. Artifact reject threshold: 95 % (for the amplifiers range ±100 μV). Sweep time: 250 ms (time base: 25 ms/div). Average: 200 sweeps. Two consecutive waveforms were recorded, off-line averaged and then analyzed.

According to the standard, amplitude as well as time parameters of the obtained waveforms were analyzed; manual correction was applied to the automatic cursors placement. Values of all parameters were compared with the own laboratory normal values.

The epimacular membrane was successfully removed from macular region in all treated eyes. There were no serious intraoperative or postoperative complications.

The clinical characteristics of EMM eyes before and 12 months after surgery are presented in Table 1.

In eyes with EMM, the several factors (e.g., incomplete posterior vitreous detachment) may induce cell proliferation in retinal inner limiting membrane. In EMM different cell types such as astrocytes, fibrocytes, macrophages, hyalocytes were found [12, 13] and some of them have characteristics of myofibroblasts [14]. The consequences of these events are macular distortion and/or macular edema and axoplasmic stasis. The manifestations of macular edema and or mechanical traction by EMM were reduction in DBCVA, retinal functional impairment [10] and increase in foveal thickness seen in OCT what is consistent with other study results [15].

Before surgery all eyes with EMM had reduced DBCVA, increase in foveal thickness and in majority of eyes (69 %) reduced amplitudes of P50- and N95-waves in comparison with normal values from the control group. The increase in IT of P50-wave) was obtained in 28 % of eyes. Visual acuity reduction correlated with the PERG preoperative parameters.

PERG derives from retinal ganglion cells and neighboring inner retinal structures [9, 16] not specified more precisely. The data from several investigators suggest that the N95 component of the human PERG is a contrast-related component generated in relation to the retinal ganglion cells. P50 originates partially from ganglion cells and also distally to the ganglion cells, from cells that are not yet ascertained [17]. Delayed implicit time of P50-wave suggests inner retinal dysfunction distally to the ganglion cells layer.

In our study, the detection of PERG abnormalities (reduced amplitudes of P50- and N95-waves) in EMM eyes indicates that dysfunction is present in ganglion cells layer and also in the inner retinal layers distally to the ganglion cells layer (delayed IT of P50-wave). The results are consistent with of Parisi et al. [10] study. The occurrence of functional impairment of outer retinal layer in patients with EMM was also documented by multifocal ERG or focal ERG in the other studies [4‐7, 10, 18]. However, in some patients selective dysfunction of innermost macular layers was detected [10]. The cause of the PERG abnormalities may be macular edema as well as traction of ganglion cells layer by EMM, leading clinically to reduction in visual acuity.

Twelve months after surgery, significant increase in DBCVA, reduction in foveal thickness, increase in amplitudes of P50- and N95-waves and decrease in P50 implicit time were observed (Table 2). Seventy-two percent of eyes achieved visual acuity improvement of two and more Snellen lines. About 65 % of eyes achieved final DBCVA of 0.32 or less logMAR (0.5 or better Snellen). The obtained results are consistent with results from other studies that reported success rate of 70–80 % after surgery estimated by VA improvement [18‐20].

The postoperative mean foveal thickness was reduced but was still higher than upper limit of normal value what is with agreement with other study results [21].

In PERG test, in the majority of operated eyes (78 %) the most prominent change was significant [22] increase in P50- and N95-waves amplitudes exceeding 20 % of initial values. It is worth noting that most often postoperative improvement of N95 amplitude was achieved (45.5 % of eyes). In patients with initially prolonged IT of P50-wave, shortening of IT was observed in 78 % of eyes, postoperatively. Postoperative DBCVA significantly correlated with the PERG preoperative parameters (Fig. 1).

The results of our study suggest that significant improvement of DBCVA, reduction in foveal thickness and improvement of PERG parameters were associated with postoperative partial restoration of macular structure and function (reduction in edema and mechanical traction of ganglion cells and their fibers by EMM). It seems reasonable to conclude that innermost retinal layers may have significant influence on macular function recovery after pars plana vitrectomy with EMM and ILM peeling. Our results are consistent with the results Parisi et al. [10].

ROC curves for amplitudes of P50- and N95-waves were used to establish cutoff values for good postoperative DBCVA of 0.3 or less (0.5 or better Snellen), (Figs. 2, 3).

The areas under curve (AUC) (commonly used predictor for postoperative acuity) were statistically significant when amplitudes of P50- and N95-waves were evaluated and indicate that these parameters may be useful to identify patients with favorable visual prognosis in clinical practice. Although, the AUC for preoperative DBCVA (Fig. 4) was higher than for PERG amplitudes, but the difference was not statistically significant and it may be affected by limited group size. To eliminate the bias, further studies with a larger number of eyes should be conducted.

Many prognostic factors for postoperative visual acuity after EMM surgery were analyzed in previous studies [21], but only preoperative VA was the variable consistently associated with postoperative VA. In these studies, IS/OS integrity in OCT was probably associated with postoperative VA. It was possible that severity of metamorphopsia, cone outer segment tips integrity and fundus autofluorescence also influenced postoperative VA. Central foveal thickness was a variable not consistently associated with postoperative VA. The studies with adequate methodological quality are necessary to confirm these findings. We think that PERG parameters should also be included in the overall prediction model as it has been shown that multiple regression models including both presurgical acuity and PERG amplitudes P50 (R ² = 0.65) or N95 (R ² = 0.67) explain better postsurgery acuity (in 65 and 67 %, respectively) than simple regression with: presurgery acuity R ² = 0.55; or presurgical PERG amplitudes P50 R ² = 0.46 or N95 R ² = 0.53. Although presurgical visual acuity remains slightly but statistically insignificant), stronger predictor of the postsurgical acuity compared to PERG amplitudes, PERG amplitudes have also significant predictive value and application of both parameters presents better predictive value as shown in example presented in Fig. 5a, b.

In conclusion, removal of idiopathic epimacular membranes with internal limiting membrane peeling not only provided visual acuity increase and reduction in foveal thickness but also was a cause of innermost retinal layer function improvement in macular region. Pattern ERG test might be a valuable tool in predicting the postoperative visual acuity.

Further study with larger group of patients should be performed to confirm our presented results.