Artificial silicon retina (ASR).The model used here is 2 mm in diameter and 25 µm thick and containsapproximately 5000 negative intrinsic layer–positive microphotodiodepixels electrically isolated from each other and separated by 5 µm.Each pixel is 20 × 20 µm square and is fabricated with a 9 ×9-µm iridium oxide electrode deposited and electrically bonded to eachpixel. Pixel current was 8 to 12 nA with approximately 800 foot-candles ofillumination. The ASR microchip was placed within a fabricated Teflon sleeveand secured intraoperatively to a saline-filled syringe injector; it was thendeposited within the retina by fluid flow. A, The ASR's size relative to apenny. B, The ASR microchip (original magnification ×36). C, The ASRpixels (original magnification ×1400). D, Subretinal location of theimplanted ASR microchip.
Fundus photographs and fluoresceinangiograms of an implanted artificial silicon retina microchip in the superiortemporal retina. Photograph number indicates the patient number; the fluoresceinangiogram (right) is from patient 3. Top to bottom: early, middle, and latephases.
A, Electroretinograms from patients1, 2, and 3 at 1 year after implantation, showing persistent electrical activityof the artificial silicon retina (ASR) microchip. B, Early Treatment DiabeticRetinopathy Study (ETDRS) chart visual acuity measurements at 0.5 m in 3 patients.Patients 3, 5, and 6 demonstrated improvement in their ETDRS visual acuityin the ASR-implanted right eye. Patient 3 read no letters preoperatively butat 12 to 18 months was able to read several letters.
Results of Humphrey central visualfield tests with the V white static spot size for patient 5, demonstratingconsistently improved central and paracentral visual fields in the right eyepostoperatively compared with the preoperative measurements. Whereas almostall of the visual field outside the 15° radius in both eyes was preoperativelyless than a 0-dB threshold (unrecordable with threshold sensitivity >10 000apostilbs) (A, top, and B, top), large portions of the visual field in theright eye were recordable postoperatively at 0 dB or better (A, middle andbottom). The Humphrey visual field test results of the unoperated on lefteye were substantially unchanged (B, middle).
Results of 9-sector testing. A,The right eye of patient 1 showed improvement in light thresholds of 1000%to 1500% in all sectors. No persistent changes were noted in the control (left)eye. B, The right eye of patient 3 demonstrated an improvement in light thresholdsof 5000% to 10 000% in the right-middle, right-lower, and middle-lowersectors. No persistent changes were noted in the control (left) eye. The thresholdimprovements in the indicated sectors of the right eye in patient 3 were consistentwith this patient's subjective impressions.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Chow AY, Chow VY, Packo KH, Pollack JS, Peyman GA, Schuchard R. The Artificial Silicon Retina Microchip for the Treatment of VisionLoss From Retinitis Pigmentosa. Arch Ophthalmol. 2004;122(4):460–469. doi:10.1001/archopht.122.4.460
To determine the safety and efficacy of the artificial silicon retina(ASR) microchip implanted in the subretinal space to treat vision loss fromretinitis pigmentosa.
The ASR microchip is a 2-mm-diameter silicon-based device that containsapproximately 5000 microelectrode-tipped microphotodiodes and is powered byincident light. The right eyes of 6 patients with retinitis pigmentosa wereimplanted with the ASR microchip while the left eyes served as controls. Safetyand visual function information was collected.
During follow-up that ranged from 6 to 18 months, all ASRs functionedelectrically. No patient showed signs of implant rejection, infection, inflammation,erosion, neovascularization, retinal detachment, or migration. Visual functionimprovements occurred in all patients and included unexpected improvementsin retinal areas distant from the implant.
Main Outcome Measures
Subjective improvements included improved perception of brightness,contrast, color, movement, shape, resolution, and visual field size.
No significant safety-related adverse effects were observed. The observationof retinal visual improvement in areas far from the implant site suggestsa possible generalized neurotrophic-type rescue effect on the damaged retinacaused by the presence of the ASR. A larger clinical trial is indicated tofurther evaluate the safety and efficacy of a subretinally implanted ASR.
RETINITIS PIGMENTOSA (RP) IS a prevalent and devastating cause of bothcentral and peripheral vision loss.1-3 Thisgenetically diverse disease commonly affects both eyes and is progressive.No treatment is effective at restoring vision once it is lost. Although otherpatterns are observed, vision loss typically develops first in the midperipheryand then progresses to involve the peripheral and finally the central visualfields.
Retinitis pigmentosa results from outer retinal degeneration. This causesthe outer portion of the inner anatomical retina (outer retina), composedprimarily of photoreceptor outer and inner segments and their cell bodies,to become damaged, the inner portion (inner retina), comprising the remainingbipolar, horizontal, amacrine, and ganglion cells and nerve fiber layer, canbe substantially spared.4,5 Thepresence of these relatively intact remaining retinal layers has led investigatorsto study the effect of electrical stimulation on these structures to improvevision.
Electrical stimulation applied to external structures of the eye hasproduced visual sensations called phosphenes in healthy subjects6-8 andblind patients with RP.9 An electrophysiologicalcorrelate of this finding has been demonstrated in blind RCS [Royal Collegeof Surgeons] rats, a model of photoreceptor degeneration in which extraocularelectrical stimulation has produced visual evoked potentials.10,11 Intraocularelectrical stimulation of the retinal nerve fiber layer also produced phosphenesin patients with RP 12 and visual evoked potentialsin animals.13 In vitro and in vivo animal studieshave shown that retinal and cortical electrical activity can be induced byelectrical stimulation of the outer retinal area.14,15
As a result of these observations, we investigated in a pilot safetyand feasibility study whether a subretinal prosthesis could produce electricalstimulation and phosphenes.16-23 A2-mm-diameter semiconductor microphotodiode array chip, 25 µm in thickness(artificial silicon retina [ASR] microchip), was designed for implantationinto the subretinal space. This chip is composed of approximately 5000 independentlyfunctioning electrode-tipped microphotodiodes and is powered solely by incidentlight. The electrical charge produced by these microphotodiodes is designedto alter the membrane potentials of contacting retinal neurons and to simulatehow light would normally activate these cells to form retinotopic visual images.Because the implant would stimulate the outer retina at an early functionalstage, subsequent visual signal processing by the remaining neuroretinal networkswould theoretically be possible.
In the cat, pig, and rat models, placement of the solid ASR disc intothe subretinal space produced a model of outer retinal degeneration that histologicallyresembled that of RP.17,18,20,24,25 Theimmunohistochemistry of the overlying retina also showed an appearance similarto that seen in patients with hereditary retinal degeneration.20 Additionally,ASR microchips functioned within the subretinal space17-19 anddemonstrated continued electrical activity for more than 3 years after implantation.26 Functionally, ASR microchips induced retinal andpossible cortical responses in the animal models.
Because of these findings in animal models and the substantial degenerationof the outer retina in patients with late-stage RP, we believed that the placementof a small ASR microchip in the subretinal space of a patient with late-stageRP would not cause further substantial injury to the retina. Furthermore,if the ASR microchip was placed in a midperipheral retinal location, the safetyand efficacy of the device could be evaluated with a minimal risk of damagingthe macular area.
To determine the safety and efficacy of the ASR microchip for possiblehuman application, we conducted a pilot clinical trial, implanting the ASRinto the right eyes of patients with RP and using the left eyes as controls.
Between June 2000 and July 2001, Food and Drug Administration and institutionalreview board approval were obtained to enroll 6 patients into an ASR safetyand feasibility clinical trial. Informed consent was obtained from all patientsprior to entry. Eligible patients were aged 40 years and older, had RP, andwere free of other significant eye or medical diseases such as uveitis, diabetes,glaucoma, or cardiac conditions. They had to have a Snellen visual acuitymeasurement of 20/800 OU or worse and/or 15° or less of the remainingcentral visual field as determined by Humphrey automatic visual field testing27 (loss >10 dB, size III white static, and 31.5 apostilbsof background illumination). Finally, the patients had to be able to perceiveelectrically induced phosphenes produced by contact lens electrical stimulation.In this test, current and voltage were provided by 1 to 6 serially connectedphotodiodes, each illuminated by a 940-nm infrared light-emitting diode affixedabove the photodiode and powered by 50-mA current. The voltage and currentproduced per photodiode were approximately 0.40 V and 200 µA with approximately5 kΩ of measured impedance between the corneal contact lens electrodeand the ipsilateral temple return electrode. Stimulation pulses consistedof 50% duty-cycle 5-Hz pulses with a polarity change every second and a totalduration not exceeding 15 seconds. Thresholds for phosphene recognition inthe 6 patients varied from 2 to 5 photodiodes electrically connected in series.The initial current generated varied depending on the impedance from approximately200 µA for 1 photodiode to 600 µA for 5 photodiodes. Exclusioncriteria included unrealistic expectations of the study, unstable personality,or other significant psychiatric conditions.
After enrollment, each patient provided a complete medical and ophthalmichistory, and a complete medical examination was performed. As part of theophthalmic history, a quality-of-life questionnaire was administered. To ruleout cystoid macular edema, fundus photography and fluorescein angiographywere performed unless contraindicated by allergy to fluorescein dye.
Patients answered questions regarding their visual function outsideof the physician's office. They were asked to describe their visual perceptionsfor 7 aspects of visual function and to give a comparison rating of one eyerelative to the other. These perceptions included brightness, contrast, color,shape, resolution, movement, and visual field size. Because the ASR was implantedin the right eye, patients were instructed to use their left eye as the basisfor comparison, to assign a fixed rating value of 10 to the left eye, andthen to compare the right eye with the left. For example, if the brightnessof the 2 eyes was equal, both would receive a value of 10. If the brightnessof the right eye was subjectively twice that of the left, the right eye wouldbe rated a 20; if it was half that of the left eye, the right eye would berated a 5. If the patient had no perception, a value of 0 was assigned. Inthe latter case, if the left eye had perception but the right eye did not,the left would be assigned a value of 10 and the right would receive a 0.Perceptions of the 2 eyes were again compared postoperatively and were alsocompared with their preoperative values when possible. For example, if theright eye developed subjective perception even though it previously had none,it would be compared with the left because a ratio comparison with a preoperativevalue of 0 in the right eye would not be possible.
Preoperative visual acuity testing was performed at least twice usingstandard back-illuminated charts from the Early Treatment Diabetic RetinopathyStudy28 (ETDRS) at 0.5 m, with the patientundergoing cycloplegia (1% cyclopentolate hydrochloride, 1% tropicamide, and2.5% phenylephrine hydrochloride) and best-corrected visual acuity testingwith a retinoscopic refraction at 0.5 m. Total ETDRS letters correctly identifiedwere counted until 1 line (5 letters) was completely missed. If neither ofthe top 2 lines of ETDRS letters could be identified at 0.5 m, we recordeda visual acuity of counting fingers or hand motions (HM) at the associateddistance as well as light perception (LP) in 9 visual field sectors.
Testing with the Humphrey Visual Field Analyzer II (Zeiss Humphrey Systems,Dublin, Calif) was conducted using the III and V white static spot sizes inthe 30-2 (30° radius) and 60-4 (30°-60° radius) protocols, aswell as a custom protocol with a 30° radius and a 4° spot separation,in both preoperative and postoperative test sessions.
Because Humphrey visual field testing was limited by the brightnessof the instrument test target (10 000 apostilbs), additional visual fieldlight-threshold testing was conducted in 9 visual field sectors in a 3 ×3 grid with less than 0.1 foot-candle (ft-c) of background room illumination.This was accomplished by using a 0.5-in-diameter optical fiber halogen lightsource placed 10 cm from the patient's eye at the following 9 locations fromthe patient's perspective: right-upper, right-middle, right-lower, middle-upper,middle-middle, middle-lower, left-upper, left-middle, and left-lower. Allpositions except middle-middle were located approximately 45° from theoptical axis (middle-middle position). Using stacked neutral-density filtersin slide holders, illuminations from 300 ft-c down to 1 e-4 ft-c in 5-dB stepswere used for threshold testing. The threshold was established in each sectorby crossing it at least 3 times in an ascending and descending staircase paradigm.The testing was continued until all 9 sectors were completed. Both the implantedand control eyes were tested during the sessions. In patients 1, 2, and 3,this test was implemented by 4 to 6 months postoperatively and in patients4, 5, and 6, by 2 months postoperatively. The test is referred to as the nine-sectortest.
Electroretinograms and visual evoked potentials were performed preoperativelyand postoperatively using an LKC (LKC Technologies, Gaithersburg, Md) or DiagnosysEspion (Diagnosys LLC, Littleton, Mass) computer signal averaging system.White or infrared light (940 nm supplied by light-emitting diodes) was appliedvia handheld Ganzfeld stimulators (Optobionics Corporation, Naperville, Ill).The infrared handheld Ganzfeld stimulator allowed the determination of isolatedimplant electrical responses and patient perceptions to infrared light inthe area of the implant.
The ASR (Figure 1) was implantedin the superior to superior temporal subretinal space (approximately 20°off axis from the macula) in the right eyes of all patients, who were givengeneral anesthesia. A standard 3-port vitrectomy (irrigation cannula, lightpipe, and aspiration vitreous cutter) was performed with pars plana lensectomy.A retinal bleb was created using a cannula and hydrostatic dissection. Theretinotomy was extended to 2.5 mm using vitreoretinal scissors. The ASR wasinserted through the retinotomy into the subretinal space, and air-fluid exchangewas performed to flatten the retina. Laser or thermal cautery was not requiredin most patients. The scleral incisions were closed with absorbable sutures,and antibiotic steroid medication was applied. Postoperative follow-up examinationswere conducted according to the study protocol. Patients visits were scheduledas follows: postoperative days 1, 2, and 4; weeks 1, 2, 4, 6, and 8; and months3, 4, 6, 9, 12, 15, 18, 21, and 24. Fluorescein angiograms were performedat 6 months, and electroretinograms were done at multiple visits, including1 year postoperatively.
Fifteen patients with RP were screened for our investigation. Thirteenpatients were able to perceive phosphenes, and 6 were selected for ASR implantation.Patient 1 had isolated RP without a significant family history. Patient 2had an extensive vertical autosomal dominant family history with multipleaffected family members. Patient 3 had autosomal dominant RP with an affectedbrother and daughter. Patient 4 had type 2 Usher syndrome with no family historyof this condition. Patients 5 and 6 were brothers who had autosomal dominantRP and a vertical family history.
In the immediate postoperative period, the most common adverse effectrequiring intervention was elevation of the intraocular pressure (IOP) tohigher than 25 mm Hg. This occurred in patients 1, 5, and 6. The IOP elevationgenerally occurred toward the end of the first week. This elevation was believedto be related to the steroid contained in the postoperative antibiotic steroiddrops (dexamethasone with either tobramycin, neomycin sulfate, or polymyxinB sulfate) because the IOP decreased rapidly when treatment with the dropswas stopped but increased when their administration was restarted. ElevatedIOP was treated with IOP-lowering medication and steroid tapering. After approximately3 weeks, when the steroid antibiotic drops regimen was stopped, the IOP returnedto preoperative values. Scratchiness in the eye that was operated on was notedby several patients and resolved after approximately 6 weeks when the externalabsorbable sutures dissolved. Patient 5 noted aniseikonia between his aphakicASR-implanted eye and his unoperated on eye when using glasses. A subsequentanterior chamber intraocular lens relieved those symptoms. Another patientnoted syneresis of images seen from the implanted eye, which was believedto be related to syneresis of a previously implanted posterior chamber intraocularlens. These symptoms substantially improved after replacement of the synereticposterior chamber intraocular lens with a stable anterior chamber intraocularlens.
No patient experienced infection, prolonged inflammation or discomfort,undesirable visual symptoms, intraocular or retinal hemorrhage, neovascularization,implant rejection, migration, or erosion through the retina.
Patients 1, 3, and 6 were pseudophakic before ASR implantation. Preoperatively,patient 2, who had bare to no LP, had a 3+ posterior subcapsular cataract(<20/200 view in the affected eye). Patient 4, who had a visual acuityof HM at 1 ft, had a 1+ anterior subcapsular cataract, 1+ nuclear sclerosis,and 1+ posterior subcapsular cataract (20/30 view in the affected eye). Patient5, who had a visual acuity of counting fingers at 1 to 2 ft, had a 1 to 2+anterior subcapsular cataract, 1+ posterior cortical cataract, and 0 to 1+nuclear sclerosis cataract (20/30 view in the affected eye). To facilitateviewing of the implant during the procedure, the cataracts were removed frompatients 2, 4, and 5 during the ASR operation. Patients 2 and 4 were leftaphakic, and patient 5 underwent secondary anterior chamber intraocular lensimplantation approximately 1 month after ASR implantation.
Table 1 summarizes the clinicalcharacteristics and results. At the last follow-up visit, there were no ASR-relatedcomplications. The retina overlying the implant remained clear with patentvessels (Figure 2). Fluoresceinangiograms showed no signs of neovascularization, vascular dropout, disruption,or leakage. In all patients, the anterior and posterior segments of the eyeappeared quiet. All devices were functioning electrically, as demonstratedby electroretinographic recordings of ASR electrical spikes to infrared stimuli(Figure 3A).
Before implantation, only 2 (patients 5 and 6) of 6 patients were ableto read ETDRS letters in either eye at 0.5 m. Preoperatively, patient 5 read16 to 25 letters OD and 24 to 28 letters OS, and patient 6 read 0 lettersOD and 0 to 3 letters OS. These 2 patients demonstrated postoperative improvementsin the total number of ETDRS letters read (Figure 3B) that were consistent with their subjective impressionof improved central perception of contrast, shape, and resolution. Six monthsafter implantation surgery, patient 5 read 35 to 41 letters OD and 21 to 28letters OS, and patient 6 read 25 to 29 letters OD and 0 letters OS. The smallestletters read in the right eye improved from a Snellen equivalent of approximately20/800 to 20/200 OD for patient 5 and from worse than 20/1600 (no lettersread) to approximately 20/400 OD for patient 6. Patient 3 was unable to readany of the ETDRS letters preoperatively (<20/1600) in either eye but postoperativelywas able to see some of the largest letters with the right eye only (approximately20/1280-20/1600 OD) at 12 to 18 months (Figure3B). On multiple tests, positive responses from preoperative centralHumphrey visual field testing with the size V white static target could beobtained consistently only for patients 5 and 6. Postoperatively, only patient5 demonstrated improved central and paracentral visual fields (30-2) in theright eye on multiple tests (Figure 4).
Compared with the unoperated on eye, 2 eyes (patients 1 and 3) withthe ASR showed improvement on the 9-sector test at 6 months to 1 year aftersurgery. In patient 1, threshold sensitivity improved by approximately 1000%to 1500% in all sectors and was consistent with the patient's impression thathis entire visual field was brighter in the eye with the implant comparedwith the same eye before surgery as well as the unoperated on eye (Figure 5). In patient 3, threshold sensitivitiesin the right-middle, right-lower, and middle-lower sectors of the 9-sectortest improved at 18 months by approximately 5000% to 10 000% (Figure 5). These visual field areas of improvementon the 9-sector test were consistent with the patient's subjective impressionthat his best vision for objects directly in front of him was achieved whenhe elevated his chin and used his inferior visual fields to look straightahead. Patient 2 showed consistent LP in multiple sectors of the operatedeye on the nine-sector test compared with her subjective bare to no LP inthose same sectors preoperatively. These perceptions were in keeping withthe patient's postoperative impression that she developed consistent LP inthe right eye and noticed shadows of people given the proper lighting conditions.This patient's 9-sector thresholds did not improve further beyond 1 year aftersurgery.
No patient was able to perceive or discriminate color on preoperativepseudoisochromatic plate color testing. Postoperatively, patient 5 reportedsubstantial improved color perception of his environment such as seeing thegreen and white of highway signs, red and white of stop signs, red and whitechecks on a tablecloth, green grass, and multiple colors in his environment.These perceptions were consistent with his ability to correctly identify theblue and orange dots of the control isochromatic plate and the red and greendots of the test plate using the operated on eye. The unoperated on controleye was never able to perceive colors in the pseudoisochromatic plates.
In the first group of 3 patients, at 18 months after surgery, theirimpressions were that visual function improvements had stabilized. In thesecond group of 3 patients, at 6 months after surgery, the impressions of2 patients (patients 5 and 6) were that their visual function changes hadgenerally stabilized, but patient 4 reported continuing improvement.
Patient 1 had LP in both eyes before surgery. The preoperative right-leftself-reported comparison ratio for brightness was 5:10, and for visual fieldsit was 2:10. Postoperatively, the ratios stabilized at 7:10 and 15:10, respectively,at 18 months. The visual field size in the right eye was subjectively about750% larger compared with the same visual field before surgery. Functionally,the patient reports not having to turn his head to see light coming from theright side.
Patient 2 had bare to no LP in the right eye with LP in the left eyebefore surgery. Preoperatively, only the left eye had subjective perceptionsof brightness, contrast, shape, and visual field size. Postoperatively, sheis still unable to read any letters on the ETDRS chart. However, she subjectivelyreports substantial visual function improvement in the right eye, particularlyin the inferior nasal visual field, that has persisted at 18 months. The self-reportedpostoperative right-left ratios were as follows: brightness, 8:10; contrast,10:10; shape, 10:10; and visual field size, 8:10. Functionally, this patientreports being able to see shadows of people with her right eye
Patient 3 had a visual acuity of HM to LP OU before surgery. At 18 monthsafter surgery, the patient noted that preoperatively the right-left ratioshad been 7:10 for brightness and 10:10 for shape, resolution, movement, andvisual field size. He indicated that postoperatively these ratios were 30:10,35:10, 50:10, 50:10, 50:10, and 50:10, respectively. Functionally, the patientreports regaining the ability to use nightlights for navigation at night andcan now see movement on television.
Patient 4 had a visual acuity of HM OU before surgery. Preoperatively,the self-reported right-left ratios were 10:10 for brightness, contrast, shape,and visual field size. Postoperatively, the ratios were variable but improvedin the right eye compared with the left: 15:10, 17:15, 17:10, and 13:10, respectively,for brightness, contrast, shape, and visual field size. Postoperative perceptionof movement was noted to be 2:10 relative to what the patient remembered fromhis youth. Subjectively, this patient indicates that when both eyes are used,his overall visual function is substantially improved from a rating of 10preoperatively to approximately 25 after surgery. Functionally, the patientreports now being able to navigate his yard without a cane and that he canreadily tell which lights are on at night in his house.
Patient 5 had a visual acuity before surgery of approximately countingfingers at 1 to 2 ft OU, with the smallest ETDRS letters recognized translatingto a Snellen equivalent of approximately 20/800 OU. He noted equal visualfunction in both eyes in all perceptions (10:10) preoperatively. Postoperatively,the right-left ratios were as follows: brightness, 17:10; contrast, 30:12;color, 17:10; shape, 15:10; resolution, 35:10; movement, 13:10; and visualfield size, 11:10. Functionally, the patient reports that he can more easilydiscern denominations of paper money, sees well enough to use eating utensils,and recognizes faces again, something he has not been able to do for approximately10 years.
Patient 6 had a preoperative visual acuity of HM OU and noted equalvisual function in both eyes in all perceptions (10:10) before surgery. Preoperatively,he recognized no ETDRS letters with the right eye (<20/1600 OD) and a maximumof 3 letters with the left (20/1600 OS). Postoperatively, the right-left ratioswere variable between days but appeared to maximize as follows: brightness,20:10; contrast, 25:10; color, 20:10; shape, 20:10; resolution, 20:10; movement,20:10; and visual field size, 18:10. Functionally, the patient reports thathe can sometimes recognize denominations of paper money. At times, he is ableto differentiate the color of traffic lights. He also sees well enough tolocate cars in the street and to find his coffee cup at meals.
This pilot clinical trial supports the hypothesis that ASR retinal prostheticchips can be safely and consistently implanted into the subretinal spacesof patients with RP. The microchips were well tolerated without discomfort,and patients showed no signs of rejection, infection, inflammation, neovascularization,vessel disruption, retinal detachment, migration, or erosion of the implantthrough the retina. These results are consistent with previously reportedfindings from animal studies showing similar biocompatibility of the implantmaterials (silicon, silicon oxide, titanium, and iridium oxide).17-20 Thecontinued electrical activity of the ASR microchip is also consistent withsimilar observations from animal studies.26
Regarding subjective responses, 4 of 6 patients (patients 2, 3, 4, and5) indicated perception of light sensation to infrared light in the projectedvisual field of the implant during testing. Typically, the first test of asession resulted in perception of light but not subsequent tests. This responsemay be associated with an electrical capacitive block in the retina that resultsfrom the initial monophasic electrical stimulus, which prevents repeated acuteresponses (the repetitive light flashes observed by all patients preoperativelyas a result of external contact lens electrical stimulation were caused bybiphasic stimulation, which would prevent a capacitive block).
Substantial and persistent visual function improvements were noted inall patients who underwent implantation with the ASR. These improvements spannedsubjective impressions, lifestyle and quality-of-life changes, task performance,ETDRS letter recognition, color recognition, Humphrey visual field testing,27 and the custom 9-sector test of visual fields. Theretinal areas and levels of improvement, however, were greater than thoseexpected from a small ASR chip implanted in the superior to superior temporalretina and stimulating a small portion of the retina. Although phospheneswere perceived in the visual fields corresponding to the ASR in 4 of 6 patients,improvements in visual function also occurred in retinal visual fields distantfrom the implant, including the macular region. These improvements were firstnoted about 1 week to 2 months after surgery and continued until approximately6 to 12 months postoperatively.
The mechanism of visual function improvement in the retinal areas distantfrom the implant is unlikely to be caused by direct ASR electrical stimulationfrom the pixels to the retinal cells. The improved perceptions of contrast,color, resolution, movement, and visual field size are too great and too complexto be explained by a direct electrical effect of the implant. A possible explanationof this improvement may be that it is due to an indirect, generalized neurotrophiceffect on the retina from ASR electrical stimulation.
Consistent with this theory is the observation that visual functionimprovements did not appear immediately. Improvements began from 1 week to2 months after ASR implantation and continued for approximately 1 year. Patients3 and 5 complained of worsened vision during the first month after surgerybefore improvement was noted. Patient 2, who had no subjective LP before surgery,noted inconsistent LP during the first week after surgery and then a "quarter-size"light at several feet in the projected visual field of the implant. In thesucceeding weeks, the spot of light increased to a vertical oval that coveredthe left and middle visual fields.
Data from other studies have suggested growth and neurotrophic effectsfrom electrical stimulation. The application of electrical currents to a varietyof organ systems may promote and maintain certain cellular functions. Thesefunctions include bone growth,29,30 spinalcord growth,31 and cochlear spiral ganglioncell preservation.32,33 Recently,deep brain electrical stimulation of the subthalamic nucleus and globus pallidusinterna in patients with Parkinson disease significantly relieved tremorsand spasticity in these patients.34 The mechanismof improvement has been hypothesized to involve improved neurotransmitterbalance and the up-regulation of a variety of growth and neurotrophic factors.35,36
Neurotrophic factors have been widely reported to promote and maintainretinal cellular functions. Brain-derived neurotrophic factor, neurotrophin4, neurotrophin 5, fibroblastic growth factor, and glial cell line–derivedneurotrophic factor have been shown to enhance neurite outgrowth of retinalganglion cells and to increase their survival in cell cultures.37 Glialcell line–derived neurotrophic factor has been shown to preserve rodphotoreceptors in an animal model of retinal degeneration,38 andciliary neurotrophic factor has slowed photoreceptor degeneration in micewith retinal degeneration, and the Q344ter rhodopsin mutation with photoreceptordegeneration.39 Nerve growth factor injectedinto the intraocular space of the C3H mouse with retinal degeneration resultsin a temporary rescue of photoreceptor cells compared with controls.40,41
Mechanical injury stimuli, such as a penetrating wound of the scleraand retina, up-regulate messenger RNA expression of basic fibroblast growthfactor and ciliary neurotrophic factor and are accompanied by a transientincrease in fibroblast growth factor receptors. These factors are hypothesizedto exert photoreceptor-protective and rescue effects after injury.42 We hypothesize that chronic low-level electricalstimulation to a partially degenerated retina with RP induces a similar up-regulationof protective neurotrophic survival factors that improve the function of remainingbut inadequately functioning photoreceptors.
Some limitations of this pilot study should be addressed. Our studyinvolved only limited controls and validation of the newly developed 9-sectortest. This consisted primarily of using 1 main examiner (A.Y.C.) to performalmost all of the 9-sector examinations, with assistants recording the results.A few evaluations were performed by other examiners but generally with thesupervision and guidance of the main examiner. All examinations of the eyewith the implant were accompanied by evaluations of the unoperated on controleye to help reveal potential intersession variability and placebo effects.Although multiple preoperative repetition of the 9-sector test was performedon some of the later-enrolled patients to establish a preoperative baseline,repetition was not universally performed with the earlier patients.
Caution is appropriate in interpreting patients' subjective comparisonsof visual function between their two eyes preoperatively and postoperatively;these perceptions could be affected by their belief in whether a surgicalintervention (ie, ASR implantation) would help them. Nevertheless, we feltthat the comparison of this type of information with data obtained from othervisual function tests would be useful.
Finally, 3 of the 6 patients who underwent implantation had cataractsof varying severities, which were subsequently removed during ASR surgery.Although removal of mild cataracts may improve visual acuity at higher spatialfrequencies in normally sighted individuals, it is generally acknowledgedthat the removal of mild cataracts (20/30 view in the affected eye) wouldunlikely affect visual acuity in the range of patient 5 (20/200 to 20/800)or patient 4 (HM). In addition, removal of a 3+ posterior subcapsular cataractprobably would not improve vision from subjective no-LP to LP with form recognitionin a patient with retinal injury.
Questions for future research include the following: Can similar safetyresults and efficacy responses be obtained from a larger group of rigorouslytested preoperative and postoperative patients? If so, can more optimal ASRstimulation parameters be used (eg, voltage, current, duration, charge, phase,and chronicity of stimulation)? Would the implantation of multiple devicesbe more effective than a single device? If ASR implantation exerts a neurotrophiceffect, would earlier implantation in specific types of retinal degenerativedisease be more effective? Finally, would patients with other forms of outerretinal degeneration, such as age-related macular degeneration, also benefit?
In summary, ASR microchips containing approximately 5000 microelectrode-tippedmicrophotodiodes were implanted into 6 eyes of 6 patients in a pilot safetyand feasibility study. After 6 to 18 months of follow-up, all ASRs functionedelectrically, and no patient showed signs of implant rejection, infection,inflammation, erosion, neovascularization, retinal detachment, or migration.Visual function improvements occurred in all patients and included unexpectedvision improvements in retinal areas distant from the implant. Further studyis required to verify these findings, to assess the optimal settings for ASRstimulation, and to determine the groups of patients most likely to benefitfrom ASR implantation.
Corresponding author and reprints: Alan Y. Chow, MD,191 Palamino Pl, Wheaton, IL 60187 (e-mail: email@example.com).
Submitted for publication July 3, 2002; final revision received February11, 2003; accepted March 20, 2003.
This study was supported in part by Optobionics Corporation, Naperville,Ill.
We thank Neal Peachey, PhD, and Sherry Ball, PhD, of the Cleveland VeteransAdministration Rehabilitation Research and Development Service and the ClevelandClinic, Cleveland, Ohio, for scientific and basic science research support;and Machelle Pardue, PhD, of the Atlanta Veterans Administration RehabilitationResearch and Development Service and Emory University School of Medicine,Atlanta, Ga, without whom this study would not have been possible. We alsothank Donald Fong, MD, for reviewing the manuscript.
Dr A. Chow has full access to all of the data in the study and takesresponsibility for the integrity of the data and the accuracy of the dataanalysis.