Purpose
To evaluate the efficacy and safety of monthly intravitreal ranibizumab for the treatment of choroidal neovascularizations (CNV) secondary to angioid streaks (AS) in pseudoxanthoma elasticum (PXE).
Design
Twelve-month prospective, open-label, uncontrolled, nonrandomized interventional clinical trial.
Methods
In 7 patients, 1 eye with an active CNV was injected with 0.5 mg ranibizumab monthly over 1 year. Distance and reading visual acuity, reading speed, angiographic findings, and central retinal thickness (CRT) on optical coherence tomography were assessed at each visit. Central retinal light increment sensitivity (LIS) was assessed by microperimetry at baseline, at 6 months, and 3 to 4 months after the last injection.
Results
Best-corrected visual acuity increased significantly from baseline to month 12 (20/63 or 61 ETDRS letters to 20/32 or 73 ETDRS letters; P = .012). The effect was maintained 3 months later (61 ETDRS letters to 72 ETDRS letters; P = .055). Reading acuity and speed could be maintained throughout the study. Central LIS improved (6.6 dB, SD ± 5.9 at baseline to 7.4 dB, SD ± 6.2 at last follow-up; P < .001). Leakage from active CNVs subsided. Mean change in CRT from baseline to month 12 and 15 was -86 μm ( P = .074) and −65 μm ( P = .182), respectively. No serious adverse events occurred.
Conclusions
Efficacy outcomes indicate a beneficial therapeutic effect of intravitreal ranibizumab on central visual function including retinal LIS. Both the functional and morphologic response based on angiographic and OCT findings to ranibizumab treatment implicate an important pathophysiological role of vascular endothelial growth factor in CNVs secondary to AS in PXE. Intravitreal ranibizumab appears to be a safe and efficacious treatment in these patients.
Pseudoxanthoma elasticum (PXE) is a rare hereditary autosomal recessive disease with an estimated prevalence of 1:25 000 to 1:100 000 and is characterized by systemic calcification of elastic tissue. Mutations in the ABCC6 gene, which is mainly expressed in the liver and kidneys, have been found to be causative. Besides the cardiovascular system and the skin, the disease invariably manifests in the ocular fundus with alterations secondary to progressive calcification of the Bruch membrane. This eventually leads to breaks in the Bruch membrane, visible on high-resolution optical coherence tomography (OCT), which clinically may appear as angioid streaks. These breaks predispose to the development of choroidal neovascularizations (CNVs), usually as early as in the third or fourth decade of life. Pronounced visual impairment may subsequently lead to incapability to work, read, and perform daily tasks.
Various therapeutic approaches have repeatedly been shown to only delay the deterioration of visual function. Laser photocoagulation can only halt the progression of CNVs, while a high rate of recurrences, visual loss, and central scotomas occur. Similarly, photodynamic therapy (PDT) has been shown to only delay the decline of visual function, and some CNVs increased in size even after such treatments. Shortening the interval between subsequent PDTs from 3 months to 6 to 8 weeks did not preserve visual function either.
Ranibizumab—the antigen binding region fragment of a recombinant, humanized monoclonal antibody that neutralizes all active forms of vascular endothelial growth factor (VEGF)-A—was recently approved by the Food and Drug Administration for the treatment of all angiographic subtypes of neovascular age-related macular degeneration. VEGF-A has been found to be a key regulator of ocular angiogenesis and vascular permeability and is involved in the pathogenesis of several ocular diseases such as CNV secondary to AMD, diabetic macular edema, and proliferative diabetic retinopathy.
A number of retrospective case series have shown bevacizumab, a humanized monoclonal antibody against all forms of VEGF that is used off label in the treatment of ocular disease, to be effective in maintaining visual function in PXE patients with CNVs secondary to angioid streaks. Long-term outcomes up to 2 years have been reported to be favorable. Similarly, the effect of ranibizumab in CNVs secondary to AS has only been assessed retrospectively in 1 larger and several small case series, which reported good efficacy and safety outcomes.
Based on its effectiveness in conditions sharing pathophysiological similarities and preliminary data of effectiveness of bevacizumab in CNVs in patients with PXE, we designed a prospective interventional nonrandomized clinical trial to evaluate the efficacy and safety of monthly intravitreal ranibizumab in CNVs secondary to angioid streaks in PXE ( ClinicalTrials.gov number NCT00504400 ; Paul-Ehrlich-Institut/EudraCT-EMEA CRFB002ADE03).
Methods
Study Design
In a prospective, uncontrolled, nonrandomized study, 7 PXE patients with active CNVs secondary to angioid streaks were enrolled at the Department of Ophthalmology, University of Bonn, Germany, between January 1, 2006 and June 30, 2008. Initially, enrollment of 10 patients was planned for. However, within the recruitment period, only 7 patients with an active CNV who matched the inclusion and exclusion criteria could be recruited. The primary endpoint was best-corrected distance visual acuity (VA) 1 month after the last treatment (visit 13) compared to baseline (visit 1).
To be included in the study, patients had to be at least 18 years old and present with characteristic clinical findings of PXE on ophthalmoscopy and fluorescein angiography (FA), as well as general systemic characteristics of PXE. The diagnosis was confirmed by genetic analysis. Only eyes with an active macular CNV were considered for treatment within this study.
Eyes with any ocular comorbidity or previous interventions that would potentially interfere with outcome measures were not eligible for the study. Therefore, exclusion criteria encompassed other retinal vascular diseases, ocular surgery less than 6 months before enrollment, intravitreal anti-VEGF therapy less than 6 months before enrollment, uncontrolled glaucoma, ocular inflammation, or subfoveal fibrosis. Further exclusion criteria were any systemic disease such as a recent arterial thromboembolic event including stroke.
Study Treatment
Twelve intravitreal injections of 0.5 mg ranibizumab (Lucentis; Novartis Opthalmics, Nürnberg, Germany) were performed in monthly intervals (mean: 31.0 days; SD, ± 2.0 days) according to previously published guidelines. Follow-up examinations were performed in all patients 1 month after the last treatment (mean 28.0 ± 6.4 days; visit 13) and again 3 months later (mean 88.7 ± 9.5 days; follow-up).
Assessments and Endpoints
A complete ophthalmic examination was performed at each visit, including assessment of VA, stereoscopic funduscopy, fundus photography (Zeiss FF450, Zeiss, Oberkochen, Germany), FA, and spectral-domain (SD) OCT. Best-corrected distance VA in each eye was measured at 4 meters with standard Early Treatment Diabetic Retinopathy Study (ETDRS) protocols using a testing chart transilluminator (Lighthouse International, New York, New York, USA). Visual acuity was scored as the total number of letters read correctly. Best-corrected reading acuity and reading speed in words per minute (wpm) were tested using the standard Radner reading charts at a reading distance of 25 cm as previously described by Radner and associates. The critical print size (CPS) was the smallest print size that could be read with maximum reading speed. The sentences were covered with a piece of paper, and the patients were asked to uncover sentence after sentence, reading each one aloud as quickly and accurately as possible. Patients were instructed to read each sentence to the end without correcting any reading errors. Reading time was measured with a stopwatch. Reading speed in wpm was calculated based on the number of words in a sentence and the time needed to read the sentence. The maximum reading speed (MRS) was the best reading speed achieved in the test. Reading acuity was set at the smallest print size the patient was able to read completely and expressed in terms of logRAD (logarithm of the reading acuity, which is the reading equivalent of logMAR).
Fundus-controlled static threshold microperimetry with dilated pupils was performed with the MP1 (Nidek Technologies, Padova, Italy; software version 1.5.1), using Goldmann III stimuli presented for 100 msec against a white background with a luminance of 1.27 cd/m 2 . The range of target luminance was 0 to 20 dB attenuation from a maximum of 127 cd/m 2 . The fixation target was a suprathreshold 2-degree red cross. Perimetric targets were displayed using a 16-degree macular grid centered on the fovea to assess the central and paracentral visual field and foveal sensitivity. The room was dark during examination. In order to minimize potential learning effects on the outcome of microperimetry examinations, all included patients had undergone prior microperimetry testing before the baseline evaluation of the study eye was performed.
Standardized FA with a confocal scanning laser ophthalmoscope (cSLO) and SD-OCT were recorded with a combined instrument (Spectralis HRA-OCT, Heidelberg Engineering, Heidelberg, Germany). Early angiographic frames were recorded in the study eye and at 0.5, 1, 2, 5, and 10 minutes in both eyes. As described previously, the combined system allows to record averaged OCT scans in an exact anatomic correlation with the cSLO image. Single SD-OCT scans were positioned on topographic cSLO images by the examiner, and always included a horizontal scan through the foveal center. A sequence of horizontal scans recorded in the high-speed mode (768 A-scans / 30 degrees) and covering an area of 30 degrees (horizontal) × 25 degrees (vertical) with a distance of ∼120 μm between individual scans was recorded to assess changes in macular thickness. On follow-up examinations, the software allowed re-evaluation at exactly the same location. The automatic alignment of the outer and inner retinal border was reviewed and, where necessary, manually corrected. For thickness analysis, a grid with 9 ETDRS type regions was centered on the foveola. The central region encompassed an area with a radius of 0.5 mm; the inner and outer rings were segmented into 4 quadrants, with radii of 1.5 and 3 mm, respectively.
Statistical Analysis
Statistical analysis of the data was performed using SPSS (version 17.0, SPSS Inc, Chicago, Illinois, USA) to obtain frequencies and descriptive statistics. Paired t tests were used for analysis, whereby a P value of <.05 was considered statistically significant.
Results
Baseline Demographic and Ocular Characteristics
The baseline characteristics of the 7 patients enrolled into the study are presented in the Table . Mean age was 55 years (SD ± 6.8 years), BCVA at baseline 20/63 (61 ETDRS letters; SD ± 13), reading acuity 0.6 (logRAD; SD ± 0.22), and reading speed 118 words per minute (SD ± 58).
| Patient No. | Sex | Age | ABCC6 Gene Analysis a | ABCC6 Genotype b | Study Eye | Visual Acuity c | Ocular History | |||
|---|---|---|---|---|---|---|---|---|---|---|
| 1. Detected Mutation | 2. Detected Mutation | Study Eye | Fellow Eye | Study Eye | Fellow Eye | |||||
| 1 | F | 46 | CA | E16: c.1987C>T (p.G663C), ht | E25: c.3507-3C>T, ht | L | 20/50 | 20/32 | — | |
| 2 | F | 50 | CA | Large deletion of chromosomal material | nd | L | 20/25 | 20/640 | Previous treatment with bevacizumab | Disciform scar |
| 3 | M | 62 | PA | E24: c.3421C>T (p.R1141X) ht | nd | L | 20/80 | 20/800 | Previous treatment with bevacizumab, PDT, ALC | Disciform scar |
| 4 | M | 63 | PA | E24: c.3421C>T (p.R1141X), hm | nd | L | 20/50 | HM | — | Disciform scar |
| 5 | M | 49 | PA | E9: c.1132C>T (p.Q378X), ht | E24: c.3421C>T (p.R1141X) | R | 20/160 | 20/25 | Previous treatment with bevacizumab | Extramacular CNV |
| 6 | F | 59 | PA | E24: c.3421C>T (p.R1141X), ht | nd | R | 20/40 | 20/1000 | — | Disciform scar |
| 7 | F | 57 | PA | E28: c.3902C>T (p.T1301I), ht | nd | R | 20/63 | 20/32 | — | |
| Mean | 55.1 | 20/63 | 20/250 | |||||||
a ABCC6 gene analysis: CA, complete analysis of the coding regions of all exons and adjacent intron sequences, including analysis of the most frequent ABCC6 gene deletion Ex23-29del. PA, partial ABCC6 gene analysis with at least detection of the frequent mutation c.3421C>T (p.R1141X) and the deletion Ex23-29del, plus potential DHPLC analysis and direct sequencing of selected exons.
b GenBank accession no. NM_001171.2 . For cDNA numbering +1 corresponds to the A of the ATG translation initiation codon.
c Visual acuity is converted to the Snellen equivalent from ETDRS visual acuity scores. Mean visual acuity was calculated after transformation to logMAR values.
Efficacy
Visual Acuity
One month after the last treatment (12 months from baseline, visit 13), mean BCVA was significantly improved compared to baseline (visit 1, 20/63 or 61 ETDRS letters [SD ± 13] improved to 20/32 or 73 ETDRS letters [SD ± 11]; P = .012, 2-tailed paired t test), the effect being maintained 3 months later (at 15 months; 61 ETDRS letters to 72 ETDRS letters [SD ± 9]; P = .055, 2-tailed paired t test; Figure 1 , Top). After the first injection (1 month from baseline), the mean number of letters read did not change (mean change −0.71, SD ± 11.80, P = .878). Mean change after the third injection (3 months from visit 1) was +8 ETDRS letters (SD ± 9.43, P = .066). Reading acuity and speed could be maintained throughout the study (0.6 logRAD and 118 w/m at visit 1 improved to 0.55 logRAD [SD 0.42] and 129 w/m [SD 46] at follow-up; both P = .7; Figure 1 , Bottom graph depicting reading acuity).
Choroidal Neovascularization Resolution of Leakage
Based on early angiographic frames, all CNVs were classic CNVs adjacent to angioid streaks or preexisting fibrotic scars. No retinal pigment epithelial detachments were present. Leakage from active CNVs present in all study eyes subsided, intraretinal and subretinal fluid largely disappeared, and only small areas of atrophy of the retinal pigment epithelium (RPE) appeared until follow-up ( Figure 2 ). Some patients showed more marked pattern dystrophy–like changes of the posterior pole at visit 13 and follow up, which are best visible on fundus autofluorescence (FAF) ( Figure 2 ).
Central Retinal Thickness
Mean OCT thickness within the central ETDRS region decreased from 371 μm (SD ± 91) at visit 1 to 278.50 μm after the first injection (at 1 month; SD ± 21; P = .013) and was 285 μm at visit 13 (1 month after the 12th injection; SD ± 25; P = .074). Central OCT thickness increased slightly again to follow-up (at 15-16 months; 306 μm, SD ± 33; P = .182 compared to visit 1). Subretinal fluid generally disappeared after the first 1 to 3 injections, whereas intraretinal cystoid changes persisted in some patients throughout the study ( Figure 2 ). Outer retinal changes evident by deposition of subretinal and intraretinal material and disruption and loss of normal retinal OCT band configuration remained stable with only slight intra-individual changes over time. Breaks in the Bruch membrane corresponding to clinically visibly angioid streaks showed no obvious progression during the observation period.
Central Visual Field Testing (microperimetry)
All patients had preexisting absolute scotomas in the area of central atrophic or fibrotic changes and 2 patients had a reproducible generalized reduction of central retinal light increment sensitivity (LIS). Overall, retinal LIS improved over the course of the study from a mean of 6.6 dB (SD ± 5.9) of all tested points at baseline to 7.4 dB (SD ± 6.2; P < .001) until last follow-up. Improvements were located in areas corresponding to active CNVs on FA and slight reductions of LIS in areas adjacent to atrophic or fibrotic changes seen on funduscopy, color photography, and FAF (examples in Figure 3 ).
