Transcorneal Electrical Stimulation for Patients With Retinitis Pigmentosa: A Prospective, Randomized, Sham-Controlled Follow-up Study Over 1 Year

Abstract

Purpose: After promising results of an exploratory study, this study was designed to assess the safety and efficacy of transcorneal electrical stimulation (TES) over 1 year in patients with retinitis pigmentosa (RP).

Methods: We included 52 RP patients in this prospective, randomized, partially-masked study which was done in accordance with rules of good clinical practice. Transcorneal electrical stimulation by Okustim was applied monocularly for 30 minutes per week for 52 consecutive weeks. Patients were assigned randomly to the groups: sham, 150%, or 200% of individual electrical phosphene threshold (EPT). Visual acuity (retroilluminated Snellen charts), visual field (VF; Octopus 900), electroretinography (ERG), rod and cone full-field stimulus threshold, EPTs and IOP were assessed. Primary outcome measures were VF area; secondary outcome measures were development of ERG parameters.

Results: The application of TES was tolerated well. Dry eye symptoms (31 of 52 patients) were registered as the main adverse event. Throughout the study period, the VF area showed only a trend for prevention of VF loss in the 200% group (P = 0.24). A significant improvement of light-adapted single flash b-wave was noted for the 200% (P < 0.0001) and 150% (P = 0.006) groups compared to the sham group. Tendencies of improved function were observed for scotopic b-wave amplitude for the 200% group (P = 0.097). Other examination methods did not reach statistical significance.

Conclusions: The safety and acceptable tolerability of weekly TES self-administered by patients at home was confirmed over the course of 1 year. Objectively measured improvements in retinal function with ERG provide support for the potential benefits of TES for RP patients.

Transcorneal electrical stimulation (TES) has demonstrated beneficial effects in patients with retinitis pigmentosa (RP).^1,2 Transcorneal electrical stimulation exerts its effects most probably through the release of various neurotrophic factors supporting remaining retinal cells,^2–8 including so-called dormant cells.⁹ A detailed review of the current understanding has been given in the introduction section on results of our first exploratory study.¹⁰ Since then more studies have found support of beneficial effects in transgenic 347L rabbits,¹¹ in a rodent model of nonarteritic ischemic optic neuropathy,¹² and in patients with Best's macular dystrophy.¹³ Effects also have been demonstrated in proteomic studies through upregulation of proteins associated with neuronal transmission, metabolic proteins, immunological factors, and structural proteins¹⁴ and by “suppressing the proinflammatory effect of microglia and enhancing the neurotrophic potential of Müller cells.”¹⁵ 

In our initial pilot study, we have shown that TES was safe in 16 patients whose eyes were stimulated for 30 minutes per week over a course of 6 weeks. There was a tendency in RP patients for almost all functional parameters to improve or to remain constant in the subgroup stimulated with the highest current (150% of their individual electrical phosphene threshold [EPT]). Visual field (VF) area and scotopic b-wave amplitude reached statistical significance.¹ There was no obvious trend in the group treated with 66% of their individual EPT. Based on these results it seemed justified to pursue the potential of TES in RP patients in more depth, that is in a larger cohort and over a longer period of time. In this follow-up study the 66% stimulation group was abandoned in favor of a 200% group. While in the pilot study DTL electrodes were used¹⁶ for stimulation and all stimulations took place under the supervision of an investigator and an assisting nurse, stimulation at home done by patients and/or relatives themselves seemed more appropriate. Okuvision GmbH (Reutlingen, Germany) developed a stimulator and electrodes that allowed safe and repeatable stimulation at home with parameters based on the results of the pilot study.¹ 

Materials and Methods

Patient Selection

A total of 52 eyes of 52 patients were included; all patients suffered from advanced RP (rod-cone-dystrophy). Visual acuity (VA; as primary inclusion parameter) and VF (as secondary inclusion parameter) results at beginning of the study were used to choose the worse eye for inclusion. In patients having equal VA values in both eyes, the eye with the smaller VF was selected. An exception was made with patients who participated in the previous study as well: in those patients the untreated eye in the first study was selected for treatment in the present study, except for one patient. Since the fellow eye received an indeterminable amount of stimulation current, this eye was not included for analysis. Diagnosis was established by detailed history, funduscopy, electroretinography (ERG), and VF examination. Inclusion criteria were age 18 to 80 years, VA 0.02 to 0.9 (decimal notation), recordable ERG (>5 μV in amplitude), and reliable VF results (>150°² in area). For home stimulation, the patient or a relative had to be able to attach the electrodes and start the stimulation reliably. Stimulation parameters, duration, exact timing, and frequency of all stimulation sessions, including impedance values, were recorded by the device in an individualized USB stick for off-line evaluation by the study team. Exclusion criteria were other ocular diseases than RP (e.g., diabetic retinopathy, retinal or choroidal neovascularization, exudative age-related macular degeneration, glaucoma, history of retinal detachment, macular edema), silicon oil endotamponade, or severe systemic disease; patients with history of dry eye were not excluded. 

Patients were recruited from the hospital's hereditary retinal degeneration clinic (Centre for Ophthalmology, University of Tübingen, Tübingen, Germany; 12 of them had been included in the pilot study; all patients, except one, were included with the other eye than the one in the pilot study) and were randomized to one of the three treatment arms: sham-stimulation, or stimulation with 150% or 200% of their individual EPT at 20 Hz. Stimulation was performed monocularly. 

Study Design

Patients and technicians who performed ERG, VF, and rod and cone thresholds were masked for the treatment group during the entire study period. The investigator who performed VA, EPT, IOP, slit-lamp examination, and fundus photography was not masked, as he/she was responsible for adjusting the stimulation parameters. The study was approved by the local ethics committee and local agencies; and was registered at clinicaltrials.gov (NCT01837901). All patients gave written informed consent and all procedures adhered to the Declaration of Helsinki. The study was performed according to the standards of good clinical practice, and the EU directive for medical devices and German Medical Product Law (MPG). The study was supervised by the STZ eyetrial, the clinical trial centre at the Centre for Ophthalmology, University of Tübingen, which is a certified member of the European Vision Institute, Clinical Research Network (www.EVICR.net; available in the public domain). The STZ eyetrial provided study documents, performed regular monitor visits, and controlled all study documentation including adverse events (AE). All documents, especially informed consents and study logs, were reviewed; all electronic case report form (eCRF) data were monitored and source data verification was conducted according to the International Conference on Harmonisation (ICH) - Good Clinical Practice (GCP) guidelines (ICH, Geneva, Switzerland; www.ich.org; available in the public domain). 

Results were recorded in an eCRF system (Open Clinica version 3.2; OpenClinica, LLC, Waltham, MA, USA). 

Patients were seen at 13 times: for screening (visit 1 for rod and cone thresholds and ERG), for the baseline visit (visit 1 for VA, IOP, VF, and EPT), and thereafter at 3 weekly visits (visits 2–4) to monitor the patients' ability to properly handle the stimulation device, followed by 8 consecutive visits every 6 weeks (visits 5–12). At all visits VA, IOP, and EPT examinations were performed; at visits 1, 6, 8, 10, and 12 ERG and VF examinations were performed (VF examinations were performed at screening for training); at visits 1, 6, 8, and 12 rod and cone threshold examinations were performed. All examination methods were performed once before stimulation, except of VA, VF, IOP, and EPT (twice before stimulation). After visit 1 TES was performed for 30 minutes per week for 52 consecutive weeks. The stimulation strength was adjusted at each visit according to the individual EPT and the treatment arm. 

System for Electrical Stimulation

A commercially available stimulation system was used consisting of OkuStim, OkuSpex, and OkuEl (CE approved; Okuvision GmbH, Reutlingen, Germany). 

OkuStim is the stimulation unit that delivers pulses of 20 Hz with current-balanced 5 ms positive immediately followed by 5 ms negative deflections. Only the study team had a software to determine EPTs and upload stimulation parameters for home use onto a patient's individual USB stick; the patient used the USB stick to start stimulation by plugging it into the OkuStim. The USB stick recorded time, date, electrical parameters, and duration of stimulation until the next visit to the study center. During stimulation the device checked the impedance of the attached electrodes and alerted when impedance was too high. 

OkuSpex is the special frame to be adjusted to the patient's face and to accept the electrodes. 

OkuEl are the electrodes based on the DTL-type described originally by Dawson et al.¹⁶ The electrodes have been constructed to be positioned on the lower eyelid on the ocular surface when the patient is wearing the lens frame. A red dot electrode from 3M (3M Europe, Diegem, Belgium) was attached to the ipsilateral temple as counter electrode. 

Determination of Electrical Phosphene Thresholds

To assess EPTs an alternative forced choice method was used as described previously^1,17 at each study visit for adaptation of the stimulation strength according to the treatment arm. 

Examination Techniques

Electroretinography.

Ganzfeld ERGs were recorded and analyzed according to the ISCEV standard¹⁸ using a ColorDome controlled by an Espion E² (Diagnosys LLC, Cambridge, UK). After 30 minutes of dark adaptation and application of two drops of tropicamide 0.5% (Mydriatikum Stulln, Stulln, Germany) self-constructed DTL-electrodes^16,17 were applied and gold-cup electrodes (VIASYS Healthcare, Warwick, UK) were positioned on the forehead and the temples as ground and reference electrode respectively. The ERG protocol consisted of 4 steps with 4 ms duration. A single flash (3 cd.s/m²; white 6500 K) response and a 9 Hz flicker (0.16 cd.s/m²; blue LED 470 nm as described previously¹⁹) were used as rod protocol. A single-flash cone response and a 30 Hz flicker (both with 3 phot cd.s/m² stimulation intensity in presence of a background illumination of 30 phot cd/m²) were chosen as cone protocol. Impedance level was <10 kΩ. A band-pass filter was applied from 0.3 to 300 Hz using the machine's built-in software algorithm. ERG potentials <5 μV for single-flash responses were excluded from the analysis. Flicker ERG potentials were analyzed using a Fourier analysis and previously described markers for significance.²⁰ 

Visual Field.

An Octopus 900 perimeter (Haag-Streit, Inc., Koeniz, Switzerland) was used with background luminance was 10 cd/m². For semiautomatic kinetic perimetry up to 90° eccentricity white stimuli (Goldmann III4e and V4e with a constant angular velocity of 3°/s) were used. Stimuli were presented at meridians separated by 15°. Isopter and scotoma areas (in deg²) were quantified using the built-in software algorithm. As quality criterion for the kinetic perimetry the blind spot was assessed with at least 5 stimuli Goldmann size I4e at 2°/s. Two experienced VF technicians performed the data acquisition throughout the study after teaching on study protocol. 

Ophthalmologic Examination.

Best-corrected VA was assessed using retroilluminated Snellen charts (Snellen, Visual Acuity Tester; Steinbeis-Transferzentrum, Tübingen, Germany) at 4 m distance. Intraocular pressure was tested using Goldmann applanation tonometry (AT 900; Haag Streit, Koeniz, Switzerland). 

Measurement of Rod and Cone Threshold.

The full-field stimulus threshold test (FST; Diagnosys LLC, Cambridge, UK) was performed using a ColorDome controlled by an Espion E² to determine final thresholds for cones (red test light, 625 nm) and rods (blue test light, 470 nm) after 30 minutes of dark-adaptation; pupils were maximally dilated. The 0 dB illumination was fixed at 0.01 cd.s/m². The range of decibels available for the test varies from −75 (dim) to 15 dB (bright), with 10 dB equaling 1 log unit. 

Statistical Analysis

Ten percent of all data extracted from the eCRF software were randomly chosen to be double-checked manually before statistical analysis with JMP (version 11.1.1; SAS Institute, Inc., Cary, NC, USA). Descriptive statistics were used to summarize data with illustrations of means and 95% confidence intervals (CI). To analyze the influence of treatment intraindividual ratios were calculated for each subject between baseline and follow-up visits. To estimate the development of parameters under treatment during visits for each group and to achieve unbiased estimates of variance parameters the method of restricted maximum likelihood (REML)²¹ was used. To compare groups, the Tukey-Kramer post hoc test analysis was applied with a global level of significance set at P < 0.05. Baseline values were compared using ANOVA and the Tukey-Kramer post hoc test analysis (P < 0.05). 

Results

A total of 52 patients completed the study per protocol. Number and mean age in each group were: sham n = 20, age 48 ± 15 years (± SD); 150% n = 15, age 42 ± 13 years; and 200% n = 17, 49 ± 16 years; there was no statistically significant difference between groups' age (P = 0.37; ANOVA). 

The treatment protocol was tolerated well. Temporary dry eye symptoms related to the treatment sessions were the most frequent adverse events (31 of 52 patients) resolving within less than 1 day with artificial tear application. No patient had a superficial punctate keratitis. No serious adverse events (SAEs) or study dropouts related to the treatment were observed. All patients conducted the stimulation at home after four training sessions in the hospital. Patient compliance was very well. The patients missed in total 65 sessions during the study period, which corresponded to 2.4% of all stimulation sessions, and 46.2% of all patients missed no single stimulation session. The highest proportion of missed sessions was 17.3% in one patient. Due to a good compliance in the current study, a dose-response relationship related to compliance and cumulative exposure is not assessable. 

Results at Baseline

No significant differences between groups existed at baseline, except for the photopic standard flash (SF) b-wave amplitude (ANOVA P = 0.025), which showed significantly lower amplitudes in the 200% group in comparison to the sham group (data are in Table 1); there was no significant difference between the sham and 150% groups, nor between the 150% and 200% groups. 

Table 1

 

Means and 95% CI (Lower and Upper 95% CI) of Tested Parameters in the Three Stimulation Groups (Sham, 150%, and 200% of Individual Phosphene Threshold Current at 20 Hz) for Baseline, Visit 6, Visit 8, and the Last Visit (Visit 12)

 

Results During Study Period

The scotopic ISCEV standard-flash a-wave amplitude changed by +3.8 μV (+11%) in the sham group (recordable n = 11–16), by +20.1 μV (+55%) in the 150% group (recordable n = 9–11), and by +10.92 μV (+80%) in the 200% group (recordable n = 14–16; REML P = 0.58; Table 2, Fig. 1). The scotopic standard-flash b-wave amplitude changed by +0.5 μV (+1%) in the sham group, by −7.3 μV (−%) in the 150% group, and by +4.1 μV (+11%) in the 200% group (REML P = 0.097; Table 2, Fig. 1). The scotopic 9 Hz flicker amplitude changed by +0.59 μV (+26%) in the sham group, by +0.24 μV (+14%) in the 150% group, and by +0.09 μV (+6%) in the 200% group (REML P = 0.78; Table 2, Fig. 2). No significant differences were observed for implicit time and phase (REML P = 0.66, P = 0.19, and P = 0.14 for a-wave, b-wave and 9 Hz flicker, respectively; Table 2, Figs. 1, 2). 

Table 2

 

Values of the REML Estimation Are Given as Mean (Grey Background) and Upper and Lower Borders of 95% CI for All Tested Parameters in the Three Stimulation Groups (Sham, 150%, and 200% of Individual Phosphene Threshold at 20 Hz)

 

Figure 1

 

Scotopic a- and b-wave amplitudes and implicit times in the different treatment arms (Ganzfeld ERG; standard flash, 3 cd.s/m²). Left: data from the different treatment arms of the study at all visits that included ERG examinations (visits 1, 6, 8, 10, and 12) in relation to baseline (visit: baseline; sham, 150%, and 200% stimulation strength of individual phosphene threshold at 20 Hz). Dots: mean values of all patients in one group. Right: the estimated mean change in each group over all visits, as calculated by a REML model. Scale bars: 95% CI). The P values calculated by this method describe the probability that the strength of TES influenced development of the respective parameter.

Figure 2

 

Scotopic 9 Hz and photopic 31 Hz flicker ERG amplitudes and phases in the different treatment arms (Ganzfeld ERG using 9 Hz with 0.01 cd.s/m² and 31 Hz with 3 cd.s/m²). The format is analogous to that of Figure 1.

Figure 3

 

(A) Photopic b-wave amplitudes and implicit times in the different treatment arms (Ganzfeld electroretinography; standard flash, 3 cd.s/m²). Significant changes are marked by stars. The format is analogous to that of Figure 1. (B) Examples of ERG examinations revealed from one stimulated subject with corresponding ERG waves at visit 1 and at the end of the study (visit 12).

The 31 Hz flicker amplitude changed by −0.3 μV (−6%) in the sham group, by − μV (−7%) in the 150%, and by +1.8 μV (+42%) in the 200% group (REML P = 0.30; Table 2, Fig. 2). No significant differences were observed for 31 Hz flicker phases (REML P = 0.80; Table 2, Fig. 2). The photopic ISCEV standard-flash b-wave amplitude was highly significantly increased by +6.9 μV (+37%) in the 200% group (recordable n = 13–16) compared to the 150% group (recordable n = 8–11; changed by +3.5 μV; +17%; REML P = 0.027) and the sham group (recordable n = 10–13; changed by −6.5 μV; –19%; REML P < 0.0001; Table 2, Fig. 3). The b-wave amplitude was significantly increased in the 150% group compared to the sham group (P = 0.009). The photopic b-wave implicit time was significantly shorter in the 150% group (−1.9 ms; −5%) compared to the sham group (+1.1 ms; +3%) and the 200% group (+0.7 ms; +2%; REML P = 0.006; Table 2, Fig. 3A). 

Examples of ERG waves as derived from ERG examinations in the current study are illustrated in Figure 3B. 

Evaluation of EPT at 20 Hz revealed a mean change by −0.04 mA (−9%) in the sham group, by +0.01 mA (+2%) in the 150% group, and by −0.004 mA (−1%) in the 200% group (REML P = 0.35; Table 2, Fig. 4). 

Figure 4

 

Electrical phosphene thresholds in the different treatment arms (thresholds were assessed using DTL electrodes at a stimulation frequency of 20 Hz). The format is analogous to that of Figure 1.

The final rod light sensitivity using the FST test changed by −1.3 dB (−2%) in the sham group, by +0.58 dB (+1%) in the 150% group, and remained unchanged (0%) in the 200% group (REML P = 0.51; Table 2, Fig. 5). The final cone light sensitivity changed by +2.7 dB (+8%) in the sham group, by +2.6 dB (+8%) in the 150% group, and by +3.6 dB (+11%) in the 200% group (REML P = 0.62; Table 2, Fig. 5). 

Figure 5

 

Cone and rod sensitivity in the different treatment arms using Ganzfeld stimulations (blue test light for rods and red test light for cones). The format is analogous to that of Figure 1.

Visual field area for the III4e mark changed by −387 deg² (−8%) in the sham group, by −449 deg² (-9%) in the 150% group, and by −125 deg² (−2%) in the 200% group (REML P = 0.24; Table 2, Fig. 6). For the V4e testing mark the area changed by −357 deg² (−5%) in the sham group, by −380 deg² (−5%) in the 150% group, and by −249 deg² (−3%) in the 200% group (REML P = 0.69; Table 2, Fig. 6). 

Figure 6

 

Visual field area (deg²) was tested by semiautomatic kinetic perimetry up to 90° eccentricity (Goldmann III4e and V4e; background luminance of 10 cd/m²) in the different treatment arms. The format is analogous to that of Figure 1.

Visual acuity (decibel notation) examined with Snellen charts changed by +0.04 (+6%) in the sham group, by −0.01 (−2%) in the 150% group, and by +0.04 (+6%) in the 200% group (REML P = 0.42; Table 2, Fig. 7). 

Figure 7

 

Visual acuity using retroilluminated Snellen testing chart at 4 m distance; IOP using applanation tonometry in the different treatment arms. The format is analogous to that of Figure 1.

The IOP remained unchanged in all groups (REML P = 0.79; Tables 1, 2; Fig. 7). 

All data (mean and 95% CI) are shown in Table 1, all values of the REML analyses in Table 2. 

Discussion

While the first prospective, randomized, sham-controlled study found significant enlargement of VF area and an improved scotopic b-wave amplitude after 6 weeks of treatment¹ the current study did not reproduce these exact findings after 1 year of stimulation in a larger cohort. A reduction of VF area was observed in all groups (Fig. 6), as is expectable in the natural course of the disease. The reduction was lowest in the 200% group but not statistically significant (P = 0.24; Table 2). 

We detected a significant improvement of cone function objectively in the presented study rather than rod function as in the previous study. The significant improvement was not detected for subjective findings. The improved cone function was seen in significantly increased photopic standard-flash b-wave amplitudes and shorter implicit times in the stimulated groups (Fig. 2). The lowest mean amplitude at baseline was found in the 200% group compared to the sham and 150% groups (Table 1; ANOVA P = 0.025). At the end of the study the mean values had increased in the 200% group by +30% compared to a decrease in the sham group by −29% and in the 150% group by −22% (Table 1). This finding may indicate that patients with more advanced cone function deterioration are more likely to benefit from TES, but this hypothesis should be proven by further studies. Table 2 illustrates the changes in light adapted b-wave amplitudes of all groups (200% group +37%, 150% group +17%, and sham group −19%), which led to a significant finding in the current study. In a previous study Fishman et al.²² described a significance border of 38% to 61% for the light adapted single-flash b-wave amplitude due to the test–retest variability in patients with RP. It should be noted that the data in this study were derived from 15 patients with RP. Our finding of 37% above baseline examination for the 200% group was at the lower border of significance. However, our study design comprised a comparison of intraindividual ratios between all groups in 4 visits. Due to multiple testing and the group comparison a single test–retest border is not applicable in our study. However, the difference between the 200% and sham groups is 56% in the current study, which is in concordance with the previously mentioned study.²² However, a significant improved cone function was not found by other examination methods. Especially the FST test, as a very reliable test for patients with RP,^23,24 reveals the highest improvement in the 200% group, but without reaching significance. As the ERG test was highly correlated with the FST test,²⁵ some differences between the two tests may account for the finding in our study. The ERG is the only objective test measuring retinal function in the current study. In contrast to that, the FST provides subjective information about retinal sensitivity to light. On the other hand, and probably more likely, the FST test provides information about the most sensitive areas of the retina,^23,26 while the ERG demonstrates functional information of the whole retina. It seems possible that retinal areas improved after TES, which were not the most sensitive areas at baseline. In this case the ERG would show increased retinal function without any change in the FST test, as in the current study. Another reason with probably less importance is the difference of spectral properties of both tests. The ERG was performed using a mixed white light (6500 K), which was a mixture of all LEDs. The FST test was performed using a blue light LED with 470 nm and a red light LED with 635 nm. The white stimulation light has a broader spectral efficiency compared to a particular LED. Therefore, the mixed white light stimulation is able to stimulate more retinal cells as the blue LED, which aims on the rod system, and the red LED, which aims on the red wavelength cones. These differences may account for the inability of the FST results to support findings of the ERG in the current study. 

Scotopic retinal function improved not statistically significant in the new study as seen in increased scotopic b-wave amplitudes and shortening of implicit times in the stimulated groups (Fig. 1; P = 0.097 and P = 0.19, respectively). Morimoto et al.²⁷ showed in 2007 a preservation of retinal function in RCS rats induced by TES with larger scotopic b-wave amplitudes in young rats and larger scotopic threshold responses in older rats compared to a sham stimulated group. However, photopic responses were not recorded. A study from our group in rats exposed to bright light as a model for mild retinal degeneration showed significant higher scotopic responses for TES-treated animals compared to the sham group but without any significant differences in photopic ERG responses.¹⁰ Ni et al.⁷ reported about a stronger light damage protocol and also showed a significant higher a- and b-wave amplitude in TES-treated rats in comparison with the control group, but photopic ERG responses were not recorded. All these findings were not able to support our results with predominantly effects on photopic ERG. A possible explanation could be that rats and mice are nocturnal with higher rods-to-cones ratio than in human retinas (99%–97% rods and 1%–3% cones).^28–30 

Using rhodopsin P347L transgenic rabbits as model for RP, Morimoto et al.¹¹ showed higher dark-adapted b-wave amplitudes in TES-treated animals compared to the sham group. In the same study significantly higher a-wave and b-wave amplitudes were shown under photopic conditions for TES-treated animals compared to sham-stimulated animals. However, in this study ERG results showed that TES preserved the cone components better than rod components of the treated rabbits. These findings, therefore, support our results in the current study of effects on the cone pathway in patients with RP. The highest density of rods and cones in rabbits is in the visual streak,³¹ where cones represent approximately 5% of all photoreceptors.³² The rods-to-cones ratio in rabbits is more similar to the human eye compared to rats and possibly reflect effects of TES better compared to rats. 

The incongruous finding between the two studies in regard to VF and ERG must probably be explained in a multifaceted way. As the previous study¹ was performed over a period of 17 weeks with a stimulation period of 6 weeks in a smaller group of patients results might represent short-term effects of TES. An enlargement of VF was not evident in our data, but it could have been missed by our evaluation strategies. On the other hand, the current study provides more detailed results from a longer study period with a larger sample size. Therefore, the other, less favorable, explanation certainly is that the effect was much overestimated in the first study, although it has shown such strong statistical significance (P < 0.001) with a presumably clinically somewhat relevant area of enlargement of 10% (from 4681.1–5176.7 deg²). However, effects due to test–retest variability were not considered in the pilot and the current study and, therefore, could not be ruled out. Although test–retest variability was not critical in the current study compared to many other studies due to the study design, it is a limitation of both studies that should be considered and investigated in future studies. 

It was not possible to detect a strong dose–response relationship in the current study. However, a possible dose relationship might be assumed between the 150% and 200% groups for the scotopic a-wave amplitude (Fig. 1), the photpic b-wave amplitude (Fig. 2), and for the VF V4e mark (Fig. 6). Other examination results showed no dose–response relationship and were not able to confirm this assumption. Future studies might be able to define a dose–response relationship and possibly a critical dose needed to treat for patients with degenerative retinal diseases. 

More than 200 genetic backgrounds (OMIM database) have been defined in RP, and more and more are being identified. Different genetic backgrounds often represent different disease mechanism only resulting in a seemingly similar disease. It can be hypothesized that different genetic pathomechanisms will respond differently to a treatment, such as TES, as the disease is located in different ophthalmic structures, like RPE cells, retinal cells, and so forth. Animal studies using electrical stimulation in different gene models for RP support this assumption, as shown by photoreceptor preservation for Royal College of Surgeons (RCS) rats using electrical stimulation,³³ but no preservation of photoreceptors or rod outer segment length in the P23H-1 rat.³⁴ While genetic analyses are under way for our study population, we have not been able to describe our population other than by the various parameters described in the protocol. 

In the progression of RP and accordingly at different ages of patients, stages in the clinical course of RP are characterized by certain functional and morphologic changes with each stage presumably responding differently to treatments, such as TES. The study population in the present was significant younger than in the first study (54 ± 12 years, mean ± SD, in the first and 46 ± 15 years in the second study; ANOVA P = 0.023). Admittedly, it seems difficult to imagine why a younger population with (at least in average) less progressed disease would benefit less in terms of VF and more in terms of cone function rather than rod function. Quite contrary, it seems difficult to imagine a beneficial effect on rods in a very late stage of the disease when most peripheral rods already have been afunctional and are certainly not in a “dormant” state anymore. An effect on cone function, which is preserved over a longer period in the disease course of RP, would respond better and was detected in our patients. Since mean VA of all included patients was 0.63 (decimal notation) at baseline, which reflects early to moderate disease stages, further studies should address or examine effects of TES in patients with more advanced stages of RP. 

In an attempt to minimize variations through group heterogeneity we have, as in the previous study, chosen intraindividual ratios to compare the behavior of parameters during the study between all groups that allows recognizing of group differences independently from absolute values. 

As in the pilot trial we have again used a stimulation protocol, which employed adjustment of the stimulation strength according to individual EPT at each study visit. So, in the patients for whom TES improves retinal function, EPT decreases (Fig. 8). Our primary intention using individualized stimulation strengths was that we hypothesized that more degenerated retinae would need more current and healthier retinae less current. Although the perception of phosphenes helps to ensure that the current has a defined effect on retinal cells, the mere perception of phosphenes might have nothing to do with the release of neurotrophic factors. In contrast, a constant dosing of release of endogenous factors by constant current may be preferable and better achieved by fixed stimulation currents. Reduction of stimulation current due to improved retinal function, as previously shown,¹⁷ could result in stimulation strengths below the “necessary” current for release of neurotrophic factors and/or improving function. This fact may account for the significant improvement only in a few parameters of all performed examinations. 

Figure 8

 

Scatterplot of improved ERG amplitudes (ratios to baseline) in relation to levels of EPT (ratios to baseline) shows decreasing EPT thresholds with increasing photopic b-wave amplitudes.

The equipment used for TES in the present study was designed to allow application of TES by the patients themselves at home. The question may be taken that stimulation has been applied incorrectly, not sufficiently, or not for the necessary time. To avoid such risk factors, the OkuStim devices have been designed to signal ineffective electrode positions by continuous impedance measurement, that is voltage control and to record the stimulation duration, the stimulation time, and date for evaluation by the study team. No conspicuities were found in these files pointing to mal- or dysfunction. 

Analyzing individual data indicated that some patients were responding better to TES than others (Naycheva L, et al. IOVS 2015;56:ARVO E-Abstract 3804). However, no significant correlation of high or low responders was found with any of our outcome parameters. 

In conclusion, this study has again proven safety of TES in our cohort of patients using a new stimulator device. We were not able, however, to reproduce the highly significant enlargement of VF area and, thus, the primary endpoint of our study was not met as the slower progression of VF deterioration in the 200% group was not statistically significant. The secondary endpoint was met in form of a significant improvement of the photopic ERG in the 200% group demonstrating an effect of TES on cone function. There was a discrepancy to the first study where only the scotopic ERG was significantly improved in the treated group. We have no conclusive explanations to offer which could account for the differences between the two studies. The concordance of (not significantly) VF area preservation and (significant) improved photopic ERG results supported the assumption of beneficial effects on cones, with both, visual field size and cone ERG are well known to be highly significantly linked in RP.³⁵ There are several bystander effects rods exert on cones, indispensable for the latter's survival³⁶ such as rod-derived cone viability factors that stimulate aerobic glycolysis,³⁷ shown to protect photoreceptors from degeneration.³⁸ Therefore, we see a potential for TES in helping patients with RP, but more studies, which already are under way, will have to clarify optimal stimulation themes and prove the definitive role of TES in treating patients with retinal degenerations. 

Acknowledgments

Supported by Okuvision GmbH, Reutlingen, Germany. 

Disclosure: A. Schatz, None; J. Pach, None; M. Gosheva, None; L. Naycheva, None; G. Willmann, None; B. Wilhelm, None; T. Peters, None; K.U. Bartz-Schmidt, None; E. Zrenner, None; A. Messias, None; F. Gekeler, None