Structural and Functional Recovery Following Limited Iatrogenic Macular Detachment for Retinal Gene Therapy

Key Points

Question  Does limited iatrogenic detachment of the macula have a detrimental short-term effect on retinal structure and function?

Findings  In this prospective interocularly controlled study of 5 patients who underwent gene therapy for choroideremia, structural recovery occurred within 1 week of iatrogenic retinal detachment, while functional recovery occurred within 1 month. However, subtle functional changes to color matching—consistent with reduced cone photopigment optical density—persisted at 1 month in 1 patient.

Meaning  Retinal structure and function appears to recover rapidly following iatrogenic macular detachment.

Importance  The early decline and recovery of retinal structure and function following iatrogenic macular detachment for retinal gene therapy is not well characterized in those with relatively preserved central visual function. Here, the recovery of retinal structure and function over the first month following iatrogenic retinal detachment for the delivery of adeno-associated viral vector encoding Rab Escort Protein 1 is described as a part of gene therapy for choroideremia.

Objective  To study changes in both retinal structure and function during the first month following iatrogenic macular detachment surgery.

Design, Setting, and Participants  This prospective interocularly controlled study was conducted between February 1 and December 31, 2015. Treatment consisted of a subretinal injection of 0.1 mL of a gene therapy solution containing 1 × 10¹¹ viral particles performed unilaterally. The participants were 5 males, aged 23 to 71 years, with a clinical and genetic diagnosis of choroideremia.

Main Outcomes and Measures  Retinal structure and function were assessed at baseline, 1 week, and 1 month using optical coherence tomography, logMAR visual acuity, microperimetry, the Farnsworth-Munsell (FM) 100-hue test, and the Rayleigh match.

Results  Five white male patients aged 23 to 71 years underwent unilateral subretinal gene therapy for genetically confirmed choroidermeia. Optical coherence tomographic images demonstrated a complete resolution of the resulting iatrogenic retinal detachment by 1 week in all 5 patients. At 1 month, the mean (SE) change in central foveal thickness was +9.6 (7.2) μm in treated eyes and +8.8 (12.6) μm in control eyes. The mean (SE) change in visual acuity was +5.4 (3.3) letters in treated eyes and +0.8 (3.1) letters in control eyes. At 1 month, the mean (SE) threshold sensitivity changes were −1.2 (2.1) dB in treated eyes and −1.0 (1.2) dB in control eyes. Color discrimination at the FM 100-hue changed little at 1 month (mean [SE] change in C-index, −0.2 [0.4] in treated eyes and 0.1 [0.2] in control eyes). Rayleigh matches in 1 patient were consistent with a diagnosis of pseudoprotanomaly, suggesting decreased effective optical density of the cone photopigments.

Conclusions and Relevance  Retinal structural recovery—as assessed by optical coherence tomography—occurs soon after iatrogenic detachment. Similarly, visual acuity recovers or improves within 1 month of the procedure and may not be accompanied by improvements in threshold sensitivity or color discrimination. Changes in color matching in 1 patient suggest decreased optical density of the cone photopigments in the early postoperative period.

Retinal detachment occurs as the result of pathological processes that are broadly classified by their etiology into exudative, tractional, or rhegmatogenous forms. Deliberate iatrogenic detachment represents a unique etiological mechanism; its use reached an initial zenith more than a decade ago in the surgical management of choroidal neovascular membranes.¹ More recently, limited iatrogenic detachment has been used in retinal gene therapy.^2-4 This maneuver appears to only rarely compromise retinal function,⁵ and no cases of proliferative vitreoretinopathy (PVR)—a known complication of more invasive procedures, such as retinal translocation and retinal pigment epithelium (RPE) transplantation¹—have been reported. However, such observations come with caveats. First, any visual decline produced by iatrogenic detachment may be defrayed by improvements resulting from successful retinal transduction. Second, an asymptomatic retinal hole has been reported in 1 patient with Leber congenital amaurosis (LCA) following iatrogenic detachment.³ Third, central retinal atrophy is noted in approximately 50% of patients with LCA following iatrogenic detachment as part of gene therapy.^6,7 Other forms of macula involving retinal detachment are, as a rule, associated with declines in spatial,⁸ temporal,⁹ and spectral resolution.⁸ The effects on visual function of the most common cause of limited detachment of the macula—central serous retinopathy—have been studied in detail.^10,11 Although best-corrected visual acuity (BCVA) may be unaltered in central serous retinopathy, patients commonly exhibit acquired tritan color-vision deficiency combined with a red-shifted and widened Rayleigh match known as pseudoprotanomaly.^10,11 The latter is the result of reduced cone photopigment optical density, possibly through altered outer-segment alignment.¹¹

There is limited histological evidence of the effects of iatrogenic macular detachment on photoreceptor structure. Simian models demonstrate cone outer-segment shortening after extensive iatrogenic detachment that resolves after 5 months.¹² However, experiments involving limited retinal detachment as part of RPE transplantation procedures in humans suggest that M- and L-cone outer segments are not discernably affected.¹ Recent analysis of spectral domain optical coherence tomographic (OCT) scans has demonstrated that disruption and thinning of the retina follow macula-off rhegmatogenous retinal detachment. This largely resolves, with a small but statistically significant residual thinning of the RPE to ellipsoid zone distance observed at 12 months.¹³ This lengthy period of retinal remodeling mirrors the natural history of functional recovery.⁸ It should be noted that the height of macula-off detachment has been demonstrated to correlate with final visual acuity following reattachment.^14,15 A further issue is the duration of detachment, which has been demonstrated in natural history studies to correlate with loss of function.^15,16 These combined observations may explain the differences in findings between simian and human histopathological studies.

Retinal gene therapy for choroideremia necessitates iatrogenic retinal detachment in patients with well-preserved central vision. Given the evidence for the deleterious effects of retinal detachment, we aimed to study changes in both retinal structure and function during the first month following surgery.

The male participants included in this study were the final 5 consecutively treated patients in a phase 1/2 clinical trial of gene therapy for choroideremia (clinicaltrials.gov identifier NCT01461213).⁴ This study was approved by the National Research Ethics Committee (Nottingham Research Ethics Committee Centre, Nottingham, United Kingdom). Clinical assessment and surgery were performed at Oxford Eye Hospital, and written informed consent was obtained for all participants. Each patient underwent a full ophthalmic assessment before enrollment and had a confirmed causative CHM (OMIM 303100) gene (coding for Rab Escort Protein 1 [REP1]) mutation (eTable in the Supplement).

Full details of the protocols used in the phase 1/2 clinical trial have been published.⁴ Clinical examination consisted of ophthalmic and medical history, anterior segment and posterior segment examination, and measurement of intraocular pressure by applanation tonometry. Assessments were undertaken before surgery and then at 1 week and 1 month following the surgical procedure. Baseline investigations included BCVA using a Bailey-Lovie style letter chart.¹⁷ Threshold sensitivity under mesopic conditions (1.3 candela/m²) was assessed by MAIA (Macular Integrity Assessment) microperimetry (CentreVue). This instrument combines a scanning laser ophthalmoscope with perimetry; Goldmann size III (0.7° diameter) targets are projected onto the fundus via a 2-mm entrance pupil. Sensitivities were determined for the central 10° (10-2 test) of the visual field using an interleaved rapid 4-dB to 2-dB step bracketing procedure.⁴ Results for the 4 most sensitive points closest to the point of fixation were averaged to eliminate variability, caused by small shifts in fixation and sparse sampling elements, inherent in including all points. Color-vision assessment consisted of the Farnsworth-Munsell (FM) 100-hue test performed under standardized lighting conditions (D₆₅ illuminant) and the Rayleigh equation¹⁸ using a Neitz OT-2 anomaloscope. The FM 100-hue test¹⁹ consists of a total of 85 caps divided into 4 boxes; each cap subtends 1.5° at 50 cm. The patient’s task is to arrange the caps in a gradual progression in perceived color. The results were scored using a program written by one of us (M.P.S.), which implements the “moment of inertia analysis” proposed by Vingrys and King-Smith.²⁰ The Rayleigh equation is a metameric color match in which the patient must match a mixture of green light (545 nm) and red light (679 nm) to monochromatic yellow light (589 nm).¹⁸ The Neitz anomaloscope presents the Rayleigh match as a circular bipartite 2° field viewed through a Keplerian eyepiece: The top half of the circular field is the green-red mixture and the bottom half is monochromatic yellow. As described by Smith and colleagues,¹¹ results were expressed as the log₁₀ of the anomalquotient (AQ):

AQ = (G_p/R_p) × (R_n/G_n)

where G_p is the green setting at the midmatching point of the patient, R_p is the red setting, and G_n and R_n are the green and red settings at the midmatching points of healthy patients (normal log₁₀AQ = 0 ± 0.2). All key numerical variables were averaged across patients and expressed as means with standard errors (SEs).

Retinal structure was assessed using Heidelberg SPECTRALIS spectral domain OCT (Heidelberg Engineering GmbH) in accordance with the manufacturer’s protocols. A 5-line horizontal raster and “dense volume” fovea-centered OCT images were assessed for the presence of subretinal fluid at 1 day, 1 week, and 1 month following surgery. In addition, we measured changes in retinal thickness. Baseline outer-retinal degeneration in choroideremia precluded accurate assessments of RPE–ellipsoid zone thickness. Instead, we measured total central foveal thickness and average retinal thickness in the central 1 mm². Central foveal thickness estimates were obtained after locating the foveal center of the baseline scans. This point was marked using the Heidelberg Eye Explorer software (Heidelberg Engineering) such that the same position could be assessed in subsequent scans. Measurements were made using ×400 zoom and the built-in caliper function. In addition, we measured average retinal thickness within the central 1 mm² after first checking and, where appropriate, modifying manually the segmentation of acquired images. In each instance, the retinal thickness was measured from the internal aspect of the internal limiting membrane to the outer border of the RPE-Bruch’s membrane complex.

Each patient underwent unilateral gene therapy treatment with the fellow eye acting as an internal control. The gene therapy procedure was undertaken by a single surgeon (R.E.M.) following a standard 3-port 23-gauge trans pars plana vitrectomy (Alcon Accurus System). Subretinal gene therapy occurred as a 2-step procedure.⁴ In the first step, a subretinal bleb of approximately 6 disc areas and which included the fovea was raised by injecting balanced salt solution through a 41 G Teflon, blunt-tipped cannula (DORC) at a maximal pressure of 12 psi from the viscous fluid controller port of the vitrectomy system. The retinotomy site was always superior to the macula and at the edge of an island of well-preserved retina and RPE to aid in the process of finding and defining the correct surgical plane.⁴ In the second step, 0.1 mL of an adeno-associated viral vector encoding REP1 (AAV.REP1) containing 1 × 10¹¹ viral particles was injected into the bleb through the same retinotomy.⁴ (Note that the treatment dose in this extension study is 1 log-unit higher than previously reported.) The eye was left with balanced salt solution as a vitreous substitute at the completion of intraocular surgery, and the sclerostomies were closed with 8-0 polyglactin 810 sutures. In 1 patient (patient 10), the surgical procedure was modified because of the presence of a thin fovea. In this patient, the subretinal bleb was raised around the fovea while a small bubble of n-perfluoro-octane protected the thinned fovea from stretch, maneuvers similar to those described previously.²¹ After n-perfluoro-octane aspiration, the subretinal fluid containing the vector suspension was observed to have extended passively subfoveally. This patient also underwent partial fluid-air exchange at case completion to protect the fovea from opening into a macular hole.

Five white male patients aged 23 to 71 years underwent unilateral subretinal gene therapy for genetically confirmed choroidermeia (eTable in the Supplement). All 5 patients (patients 10-14) underwent successful surgery, with the macula being detached by subretinal balanced salt solution and AAV.REP1 vector at the completion of surgery.

Optical coherence tomographic images could be obtained for only 4 patients at day 1 (as patient 10 had air as the initial vitreous substitute following surgery). These images demonstrated resolution of the subretinal fluid; furthermore, no evidence of fluid reaccumulation was noted in any of the patients at 1 week or at 1 month (Figure 1). In contrast to the OCT changes observed in patients with rhegmatogenous macula-off retinal detachment—where significant loss of central foveal thickness is noted in the early postoperative period¹³—such changes were not observed in our patients at any point (mean [SE] change in central foveal thickness, +9.6 [7.2] μm in AAV.REP1-treated eyes vs +8.8 [12.6] μm in control eyes at 1 month; mean [SE] change in mean central 1 mm² thickness, +16.2 [12.9] μm in AAV.REP1-treated eyes vs −5.2 [4.4] μm in control eyes at 1 month), although it will be noted that patient 10 demonstrated a decline in central foveal thickness at 1 month (Figure 2 and eFigure 1 in the Supplement) equivalent to approximately 12.5% of patient 10’s baseline caliper and 7% average central 1-mm thicknesses.

At 1 week, visual acuity was at, or better than, baseline in the treated eyes of 3 patients, but declines were noted in the remaining patients (Figure 3; mean [SE] change, −0.8 [5.3] letters in treated eyes vs +0.4 [3.8] letters in control eyes). At 1 month, BCVA was at baseline in the treated eyes of 2 patients and had improved in the remaining patients (mean [SE] change, +5.4 [3.3] letters in treated eyes vs +0.8 [3.1] letters in control eyes). Although patient 13 appeared to have improved visual acuity in the control eye following treatment, his visual acuities at baseline with this eye were variable because of his variable fixation strategy given his poor visual acuity. Because this was his worse eye, it had originally been scheduled to undergo gene therapy; however, at the baseline visit, he was unable to read the required chart line to meet the entry criterion (20/200) with his right eye despite having read the same line several times previously. After lengthy discussion and counseling, he agreed to proceed with treatment to his better (left) eye, which had stable BCVA and which met the criterion for treatment. The apparent improvement in visual acuity in the control eye, therefore, reflects the variability in his fixation strategy in this eye and his 1-week and 1-month acuities are similar to the estimates leading up to the trial.

Microperimetry did not show any postoperative shift in preferred locus of retinal fixation in any patient. Declines in paracentral mesopic threshold sensitivity were noted in all 5 patients (Figure 4) at 1 week (mean [SE] change, −3.0 [2.0] dB in treated eyes vs −1.2 [0.8] dB in control eyes); by 1 month, threshold sensitivity was greater than baseline in 3 patients, with declines (SEs) of −0.25 (3.6) dB and −7.5 (4.2) dB in the remaining 2 patients (mean [SE] change at 1 month, −1.2 [2.1] dB in treated eyes vs −1.0 [1.2] dB in control eyes). Color discrimination at the FM 100-hue was unaltered at 1 month in 2 of 4 patients; in patient 10, an exacerbation of a preexisting Verriest type III acquired tritan color-vision deficiency was noted (eFigure 2 in the Supplement).²² Color discrimination improved in patient 13. The mean (SE) change in the Vingrys and King-Smith²⁰ confusion index (negative values indicate improved color discrimination) at 1 month for the FM 100-hue was −0.2 (0.4) in treated eyes and +0.1 (0.2) in control eyes. Two of 5 patients were able to perform the Rayleigh matches; the remaining patients were unable to perform matches because of difficulties in maintaining alignment with the Keplerian eyepiece due to their limited visual field. In patient 12, no shift in Rayleigh match was noted at either 1 week or 1 month, although his baseline matches were consistent with a diagnosis of congenital protanomaly (a congenital color-vision deficiency affecting about 1% of the male European population).²³ In patient 10, pseudoprotanomaly was noted at 1 week but showed signs of partial resolution at 1 month (Figure 5).

Previous descriptions of visual recovery following iatrogenic retinal detachment for the delivery of subretinal gene therapy have chiefly concentrated on medium- to long-term visual function.^2-4 Furthermore, the majority of such reports are of patients with profound changes in central retinal structure and loss of central visual function.^2,3 Our study, by contrast, describes the short-term recovery of retinal structure and function in patients with comparatively well-preserved central retinal architecture and central visual function.

As anticipated, we observed a total resolution of subretinal fluid in all patients by 1 week. Although we cannot definitively exclude some regurgitation of subretinal fluid via the 41 G retinotomy site, the supposition is that resolution occurs actively through other routes, particularly because small subretinal air bubbles are likely to prevent egress of subretinal fluid through superiorly placed retinotomies. In the healthy eye, extensive bullous retinal detachments commonly resolve within several hours via RPE-facilitated choroidal drainage in patients treated with pneumatic retinopexy. Although patients with choroideremia have loss of peripheral RPE, this mechanism in the central residual chorioretinal island may be sufficient to reabsorb the small volumes of subretinal fluid delivered in the process of gene therapy. Alternative routes in our patients could also include reabsorption through the retina or the scleral vasculature.

In contrast to macula-off rhegmatogenous retinal detachment, where significant thinning is observed,¹³ the effects of the iatrogenic detachment procedure on OCT appearance are minimal in the early postoperative period. Several reasons may account for this observation. First, the patients with choroideremia had baseline alterations in central retinal structure that may have masked alterations due to iatrogenic macular detachment. Second, the limited iatrogenic macular detachments resulted in minimal retinal elevation: Natural history studies suggest that the negative effect of macular detachment correlates with detachment height.¹⁴ Third, the iatrogenic macular detachment resolved overnight: Natural history studies suggest that the negative effects of macula-off detachments correlate with their duration.²⁴ The fact that other patients in the low-dose subgroup of the phase 1/2 clinical trial⁴ also demonstrated recovery of visual function following treatment further supports the use of the subretinal approach in patients with milder or earlier disease. (Some of the study patients in this early subgroup had BCVAs of ≥20/20.)

In keeping with our previous observations, visual acuity returned to, or improved over, baseline levels within 1 month of treatment. Color discrimination at the FM 100-hue similarly recovered to baseline by 1 month in 2 of the 4 patients tested. In 1 patient (patient 13), an overall improvement in error score was noted at 1 month in the treated eye; this improvement was accompanied by a subjective increase in saturation of colors. However, in the remaining patient (patient 10), an exacerbation of a type III, acquired tritan color-vision deficiency was noted at the FM 100-hue test undertaken at 1 month; this was accompanied by pseudoprotanomaly at 1 week and 1 month. The shift in Rayleigh match at 1 week is equivalent to a loss of 50% to 60% in cone photopigment optical density,²⁵ and there are several possible explanations for this observation. One possibility is photoreceptor misalignment, which has been proposed to account for the pseudoprotanomaly observed in central serous retinopathy.¹¹ However, given that directional photoentrainment may occur over 5 days²⁶ to 3 weeks,²⁷ photoreceptor misalignment may not solely account for the effect observed in patient 10, unless we suppose that the detachment procedure or the disease itself has had a negative effect on this process. It is also possible that this patient’s pseudoprotanomaly resulted from a decrease in photopigment concentration, decreased outer-segment length, or both. Given previous histological observations of retinal detachment in simian models,¹² combined with the OCT observations of natural history studies of patients with macula-off rhegmatogenous detachment,¹³ it is likely that alterations in cone outer-segment length may play a role. Baseline retinal changes made it impossible for us to accurately determine if the RPE–ellipsoid zone distance has decreased following surgery, although patient 10 was the only patient in whom central foveal thickness decreased at 1 month. Regardless of the mechanism of pseudoprotanomaly in patient 10, our findings suggest that color discrimination and color matching may provide a suitable means of tracking subtle changes in retinal function caused by subretinal injections in patients with well-preserved central vision.

Our study had limitations. Although our patient group demonstrated rapid structural and functional recovery following iatrogenic macular detachment, it is unknown whether similar recovery would occur in patients with other inherited retinal diseases. However, given that we anticipate mildly dysfunctional retinal homeostasis in our patients, our data would support the use of this approach in other retinal dystrophies in which central retinal structure and function are well preserved.

We observed discordance between BCVA results and threshold sensitivity and color vision in our patients, with a general trend of changes in BCVA being superior to changes in threshold sensitivity and color vision at 1 week and 1 month. A possible explanation for this phenomenon is that gene therapy has restored function to severely dysfunctional photoreceptors that had previously not contributed to spatial discrimination at baseline, as some expression of REP1 is anticipated by 1 month. This gain of spatial resolution is expected to be offset by some loss of function of healthier cells as the result of iatrogenic macular detachment; therefore, test results of contrast detection (ie, microperimetry) and color discrimination (FM 100-hue) do not improve concomitantly. Future studies using advanced imaging techniques, such as adaptive optics scanning laser ophthalmoscopy and OCT, may successfully test this hypothesis.

Corresponding Author: Matthew P. Simunovic, MB BChir, PhD, FRANZCO, Nuffield Laboratory of Ophthalmology, Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford OX3 9DU, United Kingdom (enquiries@eye.ox.ac.uk).

Accepted for Publication: December 5, 2016.

Published Online: February 2, 2017. doi:10.1001/jamaophthalmol.2016.5630

Author Contributions: Drs Simunovic and Xue had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Simunovic, Xue, MacLaren.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Simunovic, MacLaren.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Simunovic, Xue, MacLaren.

Obtained funding: Simunovic, MacLaren.

Administrative, technical, or material support: All authors.

Study supervision: MacLaren.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr MacLaren is coinventor on a UK patent application filed on February 22, 2011, and owned by the University of Oxford. No other disclosures were reported.

Funding/Support: This study was supported by the Foundation Fighting Blindness (Dr Simunovic), RANZCO Eye Foundation (Dr Simunovic), the Bayer Global Ophthalmology Awards Program (Dr Simunovic), the Wellcome Trust (Dr MacLaren), the UK Department of Health (Dr MacLaren), the Wellcome Trust (Dr MacLaren), and the NIHR Biomedical Research Centres of Oxford University Hospitals (Dr MacLaren), and the Moorfields Eye Hospital NHS Foundation Trust (Dr MacLaren).

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.