Normal Central Retinal Function and Structure Preserved in Retinitis Pigmentosa

Abstract

Purpose.: To determine whether normal function and structure, as recently found in forms of Usher syndrome, also occur in a population of patients with nonsyndromic retinitis pigmentosa (RP).

Methods.: Patients with simplex, multiplex, or autosomal recessive RP (n = 238; ages 9–82 years) were studied with static chromatic perimetry. A subset was evaluated with optical coherence tomography (OCT). Co-localized visual sensitivity and photoreceptor nuclear layer thickness were measured across the central retina to establish the relationship of function and structure. Comparisons were made to patients with Usher syndrome (n = 83, ages 10–69 years).

Results.: Cross-sectional psychophysical data identified patients with RP who had normal rod- and cone-mediated function in the central retina. There were two other patterns with greater dysfunction, and longitudinal data confirmed that progression can occur from normal rod and cone function to cone-only central islands. The retinal extent of normal laminar architecture by OCT corresponded to the extent of normal visual function in patients with RP. Central retinal preservation of normal function and structure did not show a relationship with age or retained peripheral function. Usher syndrome results were like those in nonsyndromic RP.

Conclusions.: Regional disease variation is a well-known finding in RP. Unexpected was the observation that patients with presumed recessive RP can have regions with functionally and structurally normal retina. Such patients will require special consideration in future clinical trials of either focal or systemic treatment. Whether there is a common molecular mechanism shared by forms of RP with normal regions of retina warrants further study.

Retinitis pigmentosa (RP) is a genetically heterogeneous group of retinal degenerations that share the features of progressive loss of rod and cone photoreceptors and an associated loss of vision. ^1,2 Patients with RP are said typically to develop night blindness in adolescence, lose peripheral visual field in young adulthood, and eventually have only a residual central island of vision. ³ Abnormal electroretinograms (ERGs) are reported to be present in the first decade of life. ¹ The early detection by ERG taken together with the many histopathologic abnormalities found in postmortem human RP retinas ^4,5 suggests that the RP retina is abnormal from early life and only gets more severely affected. Bolstering this assumption are the critical roles in photoreceptor function and structure played by many of the genes now known to cause RP (RetNet; http://www.sph.uth.tmc.edu/ retnet/, provided in the public domain by the University of Texas Houston Health Science Center, Houston, TX). 

Recently, we defined in detail the retinal disease in Usher syndrome 1B (USH1B), an autosomal recessive syndromic form of RP, for the purpose of planning a clinical trial. An unexpected finding in some of the patients with USH1B was retained normal photoreceptor layer structure by optical coherence tomography (OCT) and normal rod- and cone-mediated visual function over a wide expanse of central retina. At the boundary of normal retina, there was a transition zone into abnormal retina with the more expected results of reduced vision and a thinned photoreceptor layer. ⁶ Our previous studies of USH2A and a more recent study of USH1C also showed that some of these patients could retain a region of normal structure and function. ^7–9 The presence of normal function and structure in these syndromic forms of RP prompted the question of whether such patterns of regional normality were also present in patients with nonsyndromic simplex, multiplex, or recessive RP. Previously published psychophysical data in autosomal dominant (ad)RP have indicated that some patients can have normal rods and cones. ¹⁰  

We surveyed our population of patients with RP and asked whether preserved normal retinal structure by OCT and normal function by rod- and cone-mediated sensitivities were unique to Usher syndrome or could occur in nonsyndromic autosomal recessive, simplex, or multiplex RP. The results suggest that wide regions of central retina can indeed be normal by these measurements of photoreceptor-mediated function and microstructure in a percentage of patients with RP. Retained normal retina should be factored into the design of clinical studies of natural history and therapeutic trials. The lack of a retinal degeneration phenotype in some animal models of RP may also have to be understood better, rather than dismissed as probably not relevant to the study of the human disease or ascribed to species variation. 

Methods

Subjects

A population of patients with inherited retinal degenerations (n = 321; ages 9–82 years) in the disease category of RP (n = 238) or USH (n = 83) were included in the study. In the category of RP were patients with most or all of the following features: retinal pigment epithelium (RPE) disturbances causing either hypo- or hyperpigmentation, granularity, or bone spicule-like changes; narrowed retinal vessels; an appearance of waxy pallor of the optic nerve head; reduced to nondetectable standard electroretinograms; and symptoms of nightblindness and visual field loss but fewer complaints of visual acuity loss. ¹⁰ Those with USH not only had these retinal and visual abnormalities but also bilateral auditory disturbances to different degrees. All patients had been evaluated by clinical ophthalmic examinations (by one clinician, SGJ) and ancillary testing was performed in one clinical research setting—a Center for Hereditary Retinal Degenerations. The reported results are retrospective analyses. Excluded were the following patient diagnostic groups: adRP and X-linked RP, cone or cone-rod dystrophy, ¹¹ late-onset retinal degeneration, ^12,13 acquired retinopathies (e.g., Ref. 14), the heterozygous state of X-linked RP, ¹⁵ inherited macular degenerations, choroideremia, and the Bardet-Biedl syndromes. A subgroup of the patients (n = 145) had follow-up visits. This retrospective study had no fixed interval for follow-up and the timing of the visits mainly depended on a patient's request to return for evaluation. A group of normal subjects (n = 28, ages 5–58 years) were also studied. Informed consent was obtained for all subjects; procedures adhered to the Declaration of Helsinki and were approved by the institutional review board. 

Psychophysics

Patients underwent kinetic visual fields and dark- and light-adapted chromatic static threshold perimetry (200-ms duration, 650- and 500-nm stimuli in dark and 600 nm in light, 1.7° diameter target). Central sensitivity was measured at 2° intervals along the horizontal meridian. Peripheral sensitivity was measured with a full-field test of 72 loci on a 12° grid. ¹⁶ We used the difference between dark-adapted sensitivities to 500 and 650 nm to determine whether rods or cones or both mediated vision. ^16,17 To quantify the extent of zones with normal rod vision, we measured the width of clusters of contiguous rod-mediated locations having sensitivities to 500 nm within two standard deviations of the normal mean. Similarly, we measured the width of long/middle wavelength (L/M) cone function using the 600-nm target on a white background. Other details of the techniques, data analyses, and normal results have been described. ^16,17  

Optical Coherence Tomography (OCT)

Retinal cross-sections along the horizontal meridian crossing the fovea were obtained with OCT in 28 of the patients (RP, n =14; USH, n =14). Fifteen patients were examined with Fourier-domain (FD) OCT imaging (RTVue-100; Optovue Inc., Fremont, CA). Other patients had imaging with either OCT3 (n = 8) or OCT1 (n = 5) (Carl Zeiss Meditec, Inc., Dublin, CA) imaging systems. The principles of the method and our recording and analysis techniques have been published. ^18–21 Briefly, overlapping OCT scans that were 4.5 mm in length were used to cover the horizontal meridian up to 9 mm eccentricity from the fovea. At least three OCTs were obtained at each retinal location. Post-acquisition processing of OCT data were performed with custom programs (MATLAB 6.5; MathWorks, Natick, MA). Longitudinal reflectivity profiles (LRPs) making up the OCT scans were aligned by straightening the major RPE reflection. ^18,19 Nuclear layers were defined as previously published. ^18–21 Specifically, the outer photoreceptor nuclear layer (ONL) was defined as the major intraretinal signal trough delimited by the signal slope maxima of the LRPs. ¹⁸ ONL thickness in each patient was quantified along the horizontal meridian from 3 mm nasal to 9 mm temporal retina, plotted as a function of eccentricity, and compared with the normal range (mean ± 2 SD; n = 28; ages 5–58 years). The extent (in mm) of retina along this region where ONL thickness remained within normal limits was defined and used to relate to the extent of colocalized measures of normal rod- and cone-mediated vision. 

Results

Normal Rod and Cone Function in the Central Retina of Patients with RP

We identified 57 patients with nonsyndromic RP who showed contiguous central loci having normal rod- and cone-mediated sensitivities. This subset of patients with RP constitutes 24% (57/238) of the entire group of patients with RP included in the study (Table 1). Twenty-four patients with USH also had contiguous loci with normal rod and cone sensitivities, thereby confirming and extending our previous reports of normal structure and function in different USH genotypes. ^6,7,9,22 Representative dark- and light-adapted profiles of sensitivity across the horizontal meridian are shown in four patients with nonsyndromic RP (Fig. 1A, described from left to right). A 50-year-old patient with simplex RP had best-corrected visual acuity of 20/25 and superotemporal field loss by kinetic perimetry; rod and cone ERGs were abnormally reduced in amplitude. ^15,23 Dark- and light-adapted sensitivity profiles across the central 18 mm of field were within normal limits, but peripheral rod and cone sensitivity (data not shown) was not detectable in the superotemporal field and there was ∼1–2 log units of sensitivity loss at all other peripheral loci. A patient with simplex RP, age 23, also had normal central rod and cone function but to a more limited extent. Visual acuity was 20/20 and kinetic perimetry showed a central island separated from a peripheral island by an annular midperipheral absolute scotoma; rod ERGs were not detectable and cone ERGs were markedly reduced. With dark-adapted perimetry, there was only detectable peripheral rod- and cone-mediated function in nasal and temporal peripheral islands and these were at least 1 to 2 log units reduced in sensitivity. A very similar central extent of normal rod- and cone-mediated function was present in a 12-year-old patient with simplex RP. Visual acuities were 20/20; the kinetic field was reduced to a central island only; and ERGs were not detectable. Peripheral rod sensitivity by perimetry was reduced on average by 2.5 log units and cone sensitivity was reduced ∼1 log unit. The 40-year old patient with simplex RP retained a small central island of normal rod and cone function and 20/20 visual acuity but no measurable peripheral function by kinetic or static perimetry; ERGs were not detectable. 

Table 1.

 

Characteristics of the Patients

RP and Usher Syndrome, Cross-sectional Data	Sex* F/M	Age Range, y† (mean)	Visual Acuity† (range)
Pattern 1‡
RP (n = 57)	33/24	10–69 (37)	20/20–20/100§
Usher syndrome (n = 24)‖	14/10	10–45 (22)	20/20–20/40§
Pattern 2‡
RP (n = 79)	41/38	9–82 (42)	20/20–20/100§
Usher syndrome (n = 32)‖	21/11	10–59 (30)	20/20–20/80§
Pattern 3‡
RP (n = 102)	52/50	11–81 (45)	20/20–20/500
Usher syndrome (n = 27)‖	15/12	21–69 (40)	20/20–20/400

RP, Longitudinal Data	Age Range, y (mean)	Follow-up Interval¶ Range (mean)
Pattern 1 at first visit (n = 33)
Remained in P1 (n = 23)	10–56 (33)	1–16 (6)
Progressed to P2 (n = 7)	14–57 (37)	4–20 (10)
Progressed to P2, then P3 (n = 3)	19–26 (24)	16–23 (20)
Pattern 2 at first visit (n = 33)
Remained in P2 (n = 25)	13–68 (42)	1–22 (5)
Progressed to P3 (n = 8)	12–75 (28)	3–23 (12)
Pattern 3 at all visits (n = 32)	11–81 (44)	1–22 (7)

*  Number of patients, from a total of 321.

†  Age and best-corrected visual acuity in eyes used for psychophysical testing at first visit.

‡  Pattern of central retinal function determined with psychophysical sensitivity profiles. Pattern 1 (P1) is defined as contiguous central regions of normal rod and cone function; Pattern 2 (P2) is abnormal central rod function; Pattern 3 (P3) is cone function only.

§  Visual acuity 20/32 or better in more than 50% of the patients.

‖  Of the 83 patients with Usher Syndrome, 15 with pattern 1, 16 with pattern 2, and 16 with pattern 3 had known genotypes (Schwartz et al., ²² Herrera et al., ⁷ Jacobson et al., ⁸ Jacobson et al. ⁶ ).

¶  Time between first and last visit, in years.

Figure 1.

 

Rod- and cone-mediated central function in patients with RP. (A) Dark-adapted (top, rod function) and light-adapted (bottom, cone function) sensitivity profiles horizontally across the central 18 mm of retina in four patients with RP with different extents of normal function. Shaded area: normal limits (mean ± 2 SD) for the dark-adapted 500-nm stimulus condition (top) and light-adapted 600-nm stimulus condition (bottom). Insets: kinetic fields (with V-4e, I-4e test targets) from the same eye and visit in each patient. Magnified view (box, top right, A) of central island isopters from kinetic perimetry in a 40-year-old patient with RP. Hatched bar: physiological blind spot. (B) Extent of contiguous normal rod-mediated sensitivity from horizontal profiles in a group of patients with RP and a group with USH, ranked from left to right by the extent of visual field with normal results. (C) Relationship between the extent of normal rod-mediated function and that of normal cone-mediated function in the patients with RP or USH. Dashed line: linear regression. N, nasal; T, temporal visual field.

Extents of the central zones with normal rod sensitivities in the RP group ranged from 12 mm (full width of the studied zone: 9 mm temporal to 3 mm nasal) to 1.2 mm. Patients with USH showed a similar range of extents of normal function (Fig. 1B). There was a correlation between the extent of contiguous zones with normal rod sensitivities and the extent of zones with normal cone sensitivities as measured by light-adapted perimetry (r² = 0.90, Fig. 1C). We asked whether the patients with RP with greater extents of normal function were younger than those with reduced extents. There was no correlation between age (at time of visit) and extent of retained normal central rod- or cone-mediated function (r² = 0.0048 and 0.00005, respectively). Another question to ask is whether the extent of retained central normal function in RP reflects better general retinal health in the tested eye. Kinetic perimetry extent using the V-4e target ¹⁵ was used as a surrogate measure of general retinal health and did not show a relation with the extent of normal central rod or cone function (r ² = 0.023 and 0.039, respectively). Peripheral retinal function is therefore not an easy predictor of central retinal health. The genotypes of 53 of the 57 patients with normal rod and cone sensitivities in the central retina are currently unknown. Two patients had two disease-causing variants in the USH2A gene (F12, P1; F16, P1 in Ref. 7) and two have only one USH2A allele identified to date (F13, P1; F17, P1 in Ref. 7). 

Other Patterns of Central Rod and Cone Function in RP: Cross-sectional and Longitudinal Data

Sensitivity profiles in all patients with RP identified as having contiguous central loci with normal rod and cone function are shown overlaid in a single panel (Fig. 2A, left panel, blue profiles); this group of 57 patients (ages 10–69) has been labeled pattern 1 (Table 1). The remaining 181 patients with RP did not have contiguous normal sensitivities in the central 18 mm of retina and were sorted into one of two other patterns of dysfunction. The subcategorization scheme was chosen not only to simplify the data set but also to be of value in future clinical trials with rod- or cone-targeted therapies. Patients with reduced but detectable rod function within the sampled central region (n = 79, ages 9–82) were designated as having pattern 2 (Fig. 2A, middle panel, green profiles). Patients with no detectable rod function and only cone function remaining (n = 102, ages 11–81) were considered to have pattern 3 (Fig. 2A, right panel, red profiles). It was also possible to subdivide the 83 patients with USH by using the same scheme, for comparison with the RP group (Table 1; data not displayed). 

Figure 2.

 

Patterns of rod- and cone-mediated central function in RP. (A) Cross-sectional dark-adapted sensitivity profiles in patients with RP with contiguous loci having normal central rod function (pattern 1, left, 500 nm), detectable but abnormal rod function (pattern 2, middle, 500 nm) and cone-mediated function only (pattern 3, right, 650 nm). Shaded area: normal limits (mean ± 2SD) for the dark-adapted 500-nm stimulus; pink: dark-adapted normal data at cone plateau with a 650-nm stimulus. The number of patients is provided in Table 1. (B) Longitudinal sensitivity profile data in two representative patients with RP with patterns 1, 2, and 3 (P1, P2, and P3) at different visits, spanning nearly two decades in each patient. Progression from P1 to P3 in each patient is shown from left to right. Mediation (R, rod; M, mixed rod and cone; C, cone) at each locus is near the top of the sensitivity profiles. (C) Change in pattern or lack thereof in those patients with RP with longitudinal data (each patient depicted as a line). P1→P1, normal rod function on first visit and remained within P1; P1→P2, progressed from P1 to P2; P1→P3, sequential progression from P1 to P2 and then P3. Also shown are number of patients that on first visit were categorized as P2 and on subsequent visits remained as P2 (P2→P2) or progressed to P3 (P2→P3). Patients with longitudinal data that were cone-mediated central islands on the first visit (P3) and remained so are also depicted (P3→P3). (D) Longitudinal sensitivity profiles spanning 16 years in a patient with autosomal recessive RP caused by PDE6B mutations, exemplifying change within category P3. N, nasal; T, temporal.

The simplest hypothesis for interpreting the differences in pattern of these cross-sectional psychophysical data is that the patterns represent stages sampled along a continuum of disease progression. ^24,25 Almost by definition, RP would be expected to proceed from pattern 1 (with normal rods) to pattern 2 (with abnormal rods) and then to pattern 3 (with no rods and only residual cones). ⁴ If the first visit by the patient occurred at a stage when there were already only abnormal rods in the central retina, as in pattern 2, this could be expected to progress to pattern 3. Once within pattern 3, further cone function loss would be anticipated. The different rates of change within and between patterns cannot be determined from this study. Another hypothesis to explain differences in pattern is that they represent different phenotypes from early in life. For example, patients with pattern 3 could begin life without detectable rods and there would be change only within the pattern (e.g., PDE6B form of arRP). ²⁶  

Longitudinal data permit testing the hypothesis that patterns represent stages of progression (Table 1). Of the 57 patients with RP with pattern 1, 33 (58%) had longitudinal data. Serial dark-adapted horizontal profiles from two patients with RP illustrate changes in visual function over a period of years to decades. There was transition from pattern 1 to pattern 2 and then to pattern 3 (left to right, Fig. 2B). At 34 years of age, a patient with RP simplex showed a central island with some normal rod function (Fig. 2B, top row of three panels connected by lines and arrows). On a visit 7 years later, there was reduced but measurable rod function; and 17 years from the first visit, central function was only cone-mediated. A 26-year-old patient with RP simplex showed a similar progression over two decades from a few contiguous normal rod loci centrally, to abnormal rods, to a cone-only central island (Fig. 2B, bottom row of three panels). 

Of the 33 patients with RP with longitudinal data and pattern 1 at first visit, 23 (70%) continued to show this same pattern on return visits (Fig. 2C; Table 1). Seven patients (21%) progressed from pattern 1 to pattern 2, and 3 patients (9%) progressed from pattern 1 through pattern 2 to pattern 3. Of the 33 patients with RP who had longitudinal follow-up and pattern 2 at first visit, 25 (76%) continued to show pattern 2 on return visits, whereas 8 (24%) progressed to pattern 3. Thirty-two of the 102 patients with RP with pattern 3 had longitudinal follow-up. All of these patients remained within this pattern on repeat visits (Fig. 2C); and most patients showed reduction of the extent and sensitivity of their remaining islands of central cone function on follow-up visits. Longitudinal data in patients with USH showed behavior similar to that reflected in the data in patients with RP. Most of the patients with USH with pattern 1 and multiple visits (15/17, 88%) remained in this subgroup on subsequent visits; one progressed to pattern 2, and one progressed from pattern 1 through pattern 2 to pattern 3. Of the 18 patients with USH with pattern 2 and follow-up visits, 11 (61%) maintained that classification, with the remainder progressing to pattern 3 in later visits. 

The alternative hypothesis of a single phenotype pattern, such as an early-onset rod dysfunction and only cone-mediated vision at first visit explaining pattern 3, is exemplified by a patient with RP with known PDE6B mutations. ²⁶ There was mainly cone-mediated vision at age 18 years (Fig. 2D) and subsequent visits over 16 years showed further reduction of cone function. This sequence may be expected from the literature on large and small animals with PDE6B mutations that exhibit early-onset severe rod disease and residual cones that slowly degenerate. ^27,28  

Photoreceptor Layer Thickness Underlying the Normal Rod and Cone Function

In a subset of patients with RP with normal rod and cone function over an extent of central retina, we studied the underlying retinal structure with OCT. Representative scans across the horizontal meridian from temporal retina through the fovea to the edge of the optic nerve are shown for a subject with normal vision and for three patients with RP (Figs. 3A–D). Above each scan is the extent of psychophysically measured function that is normal for both rod- and cone-mediated sensitivity (gray bars). Highlighted in the scans is the extent of photoreceptor nuclear layer (ONL) thickness that is within normal limits. Function and structure in the normal subject extended across the entire length of the scan (Fig. 3A). Each of the patients with RP had a limited extent by comparison, from a maximum of ∼6 mm shown in a 58-year-old patient (Fig. 3B) to a minimum of ∼3 to 4 mm in a 25-year-old patient (Fig. 3D). There appeared to be correspondence between the extent of normal function and normal laminar structure. Eccentric to the normal extent of scan in each patient is a transition zone to abnormal and disorganized retina. ⁶  

Figure 3.

 

Retinal structure of patients with RP with normal central rod and cone function and relationship of structure and function. (A–D) Representative scans across the horizontal meridian from temporal retina through the fovea to the edge of the optic nerve in three patients (B–D) compared with a scan of a normal subject (A). Gray bars above each scan show the extent of psychophysically measured function that is normal for both rod- and cone-mediated sensitivity. Photoreceptor (ONL) layer, which is within normal limits is highlighted in blue. (E) Relationship between photoreceptor layer structure and photoreceptor-mediated function. Extent of ONL thickness that is within normal limits plotted against the extent of psychophysically measured, co-localized rod function that is also within normal limits for a group of 14 patients with RP and 14 with USH. Solid line: linear regression.

The relationship between the extent of normal ONL and that of normal photoreceptor-mediated function is plotted (Fig. 3E) for 14 patients with RP. Also plotted for comparison are data from 14 patients with USH. ^6–8 There is a linear relationship between the extents of normal structure and function (slope, 1.05; r² = 0.92). 

Discussion

The present study documented with psychophysical testing that almost 25% of patients from a larger group of those with presumed recessive forms of RP have central regions of normal rod and cone function. A normal photoreceptor layer thickness, measured with OCT, co-localized with the normal function. The concept of regional variation in retinal degeneration has long been known and discussed. ²⁹ Areas of normal retina, defined by measurements of function and structure, however, have not been reported in groups of simplex, multiplex, or autosomal recessive RP. Histopathologic studies have not reported any RP donor retinas with normal rods and cones (reviewed in Refs. 4, 5), including young ungenotyped XLRP or patients with adRP ^30,31 and class B (regional, type 2) adRP due to rhodopsin gene mutations. ^30–33  

For half a century, the surrogate for retinal histopathology in RP has been rod- and cone-mediated visual profiles, which have been considered to represent “the functional counterpart of a histologic cross-section.” ³⁴ In certain patients with adRP, normal rod- and cone-mediated visual thresholds have been reported: for example, “rod and cone function in type 2 patients is first lost in the mid-periphery at a time when the rest of the retina is relatively normal”. ¹⁰ The central 20° in some patients with adRP have normal retinal sensitivity (Fig. 21 in Ref. 10), or “at least one zone with near normal rod function” has been described. ³⁵ AdRP of known genotype has more recently been shown to have areas of normally functioning rod and cone systems. ^25,36 Use of psychophysical and multifocal ERGs in different genetic types of RP have also suggested that there can be normal rod and cone function. ³⁷  

Until recent advances in cross-sectional imaging OCT technology, the unanswered question was whether normal psychophysically measured rod and cone function was indeed the counterpart of a histologic section in RP. ³⁴ AdRP caused by rhodopsin mutations was shown to abide by the early prediction; normal retina and normal photoreceptor-mediated function colocalized in class B. ³⁸ OCT studies in RP (without co-localized function) have shown examples of normal central structure (for example, Refs. 39–42). One recent study compared results of rod- and cone-mediated profiles to OCT-determined outer retinal thickness and found different patterns among six patients, including normal function and structure in a patient with adRP. ⁴³ Patients with dominant and presumed recessive inheritance with an unusual phenotype termed concentric RP had normal OCT retinal thickness and normal rod- and cone-mediated function in the more central retina and abrupt loss of structure and function peripherally. ⁴⁴ Our recent observations of large expanses of normal central rod- and cone-mediated vision co-localizing with normal ONL in certain stages of USH2A, USH1B, and USH1C ^6–9 led to the inquiry of whether this feature is unique to these syndromic forms of RP or whether other patients with nonsyndromic recessive RP might also present with normal retina. 

Our survey of dark- and light-adapted horizontal perimetric profiles across the central retina in a population of patients with the clinical diagnosis of simplex, multiplex, or recessive RP indicated that nearly 25% had contiguous regions of normal rod- and cone-mediated function. The remaining patients could be subdivided into two further groups by their different patterns of central rod and cone dysfunction. The three patterns defined in these cross-sectional studies may be explained by simple hypotheses: (1) Progression hypothesis: Normal central retinal rod and cone function progresses to reduced rod function and eventually to cone-only function. We were able to test this hypothesis in a few patients for whom we had longitudinal follow-up, and the data supported the notion that there can be progression from contiguous normal rod function to cone-only vision. (2) Distinct phenotype hypothesis: Molecular subgroups may represent diseases with early-onset severe rod loss or diseases of abnormal rod development with only preserved cone function that persists to some degree throughout life. This hypothesis, as demonstrated in a patients with autosomal recessive RP who had the PDE6B mutation and cone-only vision, is consistent with data from relevant large and small animal models. ^27,28  

There are clinical implications of finding normal central function and structure in some patients with RP. The observation adds to the understanding of the diversity of functional and structural abnormalities in RP. The finding also may be worthy of the attention of those planning treatment trials for different forms of RP. For example, focal treatment in forms of RP, such as gene therapy, will seek focal outcomes after a therapy is delivered (at least in early trial phases) to relatively small regions of retina. It may be ill-advised in early-phase safety trials to treat relatively normal regions of retina (despite a projected future loss of normality of uncertain natural history); the outcome would be either no effect within the time of the trial or loss of vision and structure from the study agent or from subretinal surgical trauma or an accelerated natural history. ⁴⁵ Safety studies in the central retina of normal nonhuman primates could provide relevant although not identical information. ⁴⁵ Another strategy to determine safety and efficacy in upcoming focal trials would seem to be needed in patients with expanses of normal retina. Recently, we defined a transition zone between the central region of useful vision and mid-peripheral retina with little or no function and remodeled structure. For USH1B, we suggested that certain subzones of the transition area may be locations for introducing genetic material. ⁶  

The information we provided on different patterns of central dysfunction in RP may be useful for ongoing or planned therapies that are not gene specific. Therapies intended to affect rods should be developed with the awareness of whether there are detectable rods in patients with only central visual islands remaining. Patients with pattern 3 would not be readily identifiable on clinical examination, fundus photos, visual acuity measurements, or certain visual field tests and would not be appropriate candidates for rod-specific treatments, assuming that there is evidence that the remainder of the retina is also devoid of rod function. Patients with patterns 1 or 2, on the other hand, would be more appropriate for inclusion in a rod treatment trial. In a recent example, there was no visual benefit reported in patients with RP after intravitreal implantation of an encapsulated cell delivery system for ciliary neurotrophic factor (CNTF). ⁴⁶ A biological effect involving thickening of the macular retina by OCT measurement, however, was announced (http://www.neurotechusa.com/news_events/pr_2009-05-28.asp). It would be highly instructive if the pattern of dysfunction in the central retina of such patients with RP was defined before and after therapy so it could be determined whether any patients with pattern 2 were involved and showed improved rod function, or whether those with pattern 1 were involved and showed the well-known dysfunction associated with certain doses of CNTF. ^47–49  

Footnotes

 Supported by grants from the National Neurovision Research Institute, Hope for Vision, Foundation Fighting Blindness, Macula Vision Research Foundation, and the Chatlos Foundation.

Footnotes

 Disclosure: S.G. Jacobson, None; A.J. Roman, None; T.S. Aleman, None; A. Sumaroka, None; W. Herrera, None; E.A.M. Windsor, None; L.A. Atkinson, None; S.B. Schwartz, None; J.D. Steinberg, None; A.V. Cideciyan, None