Abstract
Purpose
Effectiveness of ocriplasmin for vitreomacular traction (VMT) varies depending on the presence of common ocular conditions and patient selection criteria. We carried out a systematic literature review and meta‐analysis of ocriplasmin studies conducted in real‐world settings (RWS) and compared outcomes with those from randomized controlled trials (RCTs).
Methods
We included prospective and retrospective studies from RWS documenting effectiveness of ocriplasmin in patients with VMT with or without MH, and RCTs of ocriplasmin versus control. Key end‐points were vitreomacular adhesion resolution (VMAR), nonsurgical MH closure, need for vitrectomy and safety. We conducted meta‐regression on pooled results to evaluate effects of baseline covariates and study design on outcomes.
Results
Thirty RWS (2402 patients) and 5 RCTs (737 patients) were included epiretinal membrane (ERM) and broad VMA were more prevalent in RCTs. Primary VMAR, vitrectomy and MH closure rates were comparable between RWS and RCTs. Rates of nsVMAR were significantly higher in RWS than RCTs (odds ratio 1.66; 95% confidence interval [CI]: 1.18–2.34). nsVMAR rates were inversely associated with ERM prevalence (odds ratio 0.20; 95% CI: 0.08–0.51). Compared with the recent OASIS trial, RWS reported a higher incidence of new/worsening subretinal fluid cases and less photophobia, photopsia, vitreous floaters, electroretinogram abnormalities and MH progression.
Conclusions
Ocriplasmin was significantly more effective in achieving nsVMAR in RWS than in RCTs. Lower ERM prevalence in RWS was the single significant explanatory variable for this difference. Conclusions on ocriplasmin safety in RWS are limited due to inconsistent reporting.
Keywords: Ocriplasmin, vitreomacular traction, effectiveness, real world
Introduction
The human vitreous is a clear, gel‐like structure that consists of primarily of water (98%), collagen and hyaluronan. Over decades during midlife, the vitreous liquefies while it progressively separates from the posterior vitreous cortex, eventually leading to full posterior vitreoretinal detachment (PVD) (Sebag 2009; Duker et al. 2013). In rare cases, this process does not complete, and the vitreous remains attached, mostly in areas where the vitreoretinal interface is strongest, including the vitreous base, the large retinal vessels, the optic disc margin and the 500‐µm diameter foveola (Johnson 2010). The tractional effects of incomplete PVD can result in deformities at the level of the fovea, giving rise to metamorphopsia, blurring and decreased visual acuity (Ezra 2001; Johnson 2010). Continued traction may lead to the development of epiretinal membrane (ERM) (Joshi et al. 2013) or full‐thickness macular hole (MH) (Ezra 2001).
Ocriplasmin (JETREA, Oxurion, Leuven, Belgium) is a truncated, recombinant form of human plasmin that has proteolytic activity against protein components of the vitreous body and the vitreoretinal interface such as laminin, fibronectin and collagen (European Medicines Agency 2013). Ocriplasmin is approved for the treatment of symptomatic vitreomacular adhesion (VMA)/vitreomacular traction (collectively: ‘VMT’) with or without macular hole (MH) ≤400 µm. Some authors have characterized its efficacy as modest (European Medicines Agency 2013; Jackson et al. 2013; Kim et al. 2013; Lim et al. 2017). However, post hoc analyses of these trials have shown that several ocular and patient characteristics, including presence of ERM, broad VMA adhesion size (>1500 µm), older age\, and non‐phakic lens status, were associated with lower rates of VMA resolution (VMAR), while larger MHs (>250 µm) were associated with lower closure rates (Haller et al. 2015; Jackson et al. 2016). We hypothesized that careful patient selection based on these risk factors may increase the effectiveness of ocriplasmin in real‐world settings (Khanani et al. 2019). We conducted a systematic literature review and meta‐analysis of ocriplasmin use in clinical practice to evaluate the effectiveness of a single 125‐µg dose in patients with symptomatic VMA or VMT with or without MH in real‐world settings (RWS). We compared our findings with results from randomized controlled trials (RCTs), which we identified through a separate systematic literature review.
Materials and methods
Eligibility criteria for considering studies for this review
We performed two separate systematic literature searches. First, we included prospective or retrospective studies documenting the effectiveness of single doses of ocriplasmin (125 µg) in the treatment of adult patients with VMT with or without MH who had received treatment in RWS. The second literature search identified double‐blind RCTs using single‐dose (125 µg) ocriplasmin versus control in the treatment of adults with VMT/symptomatic VMA. The main end‐point was the incidence of VMAR, which we reported as primary VMAR (pVMAR, defined as VMAR occurring after ocriplasmin injection, regardless of subsequent events), and nonsurgical VMAR (nsVMAR, defined as VMAR without the need for subsequent vitrectomy). Secondary end‐points were nonsurgical MH closure, the need for vitrectomy, and safety. To qualify for inclusion, studies had to report either pVMAR or nsVMAR outcomes and have a mean follow‐up of at least 4 weeks after injection.
Search methods for identifying studies
To identify publications of ocriplasmin use in RWS, we searched PubMed (National Institutes of Health) and EMBASE (Elsevier) using general terms describing VMT, combined with the term ‘ocriplasmin’ or ‘Jetrea’ without any time period restriction (see Table S1 for EMBASE search terms). We excluded RCTs, safety‐only reports, conference abstracts and series smaller than 12 patients. We did not contact authors for further information.
To identify RCTs, we searched EMBASE, PubMed and the Cochrane CENTRAL database for articles with ocriplasmin as a major topic and applied search filters without any time period restriction (Robinson & Dickersin 2002; Wong et al. 2006). Table S2 lists the EMBASE search terms for this search. We excluded conference abstracts and series smaller than 12 patients. Oxurion provided the clinical study databases for the identified trials.
The RWS and RCT searches were first executed on 28 October 2018 and 04 January 2019, respectively, and were updated on 5 June 2019; no new studies were found. We used EndNote X8 to collate the records from the respective literature searches and remove duplicates. We used the Covidence web app (Covidence, Melbourne, Australia) to review records and assign eligibility status.
Study selection
Two investigators (authors KHB, BL) independently reviewed title, abstract and full text. Discrepancies were resolved by discussion.
Data collection and risk of bias assessment
Data from included RWS were extracted by KHB into a Microsoft Excel spreadsheet and independently verified by BL. KHB checked all suggested corrections against the full text. Discordant readouts were resolved by discussion. For RWS, we extracted average study follow‐up and clinically relevant baseline characteristics including the number of patients/treated eyes, age, sex, MH status, ERM prevalence, focal VMA, lens status and best‐corrected visual acuity (BCVA). We extracted data on nonsurgical and primary VMAR, overall vitrectomy, vitrectomy after successful pVMAR, MH closure, and ≥10 Early Treatment Diabetic Retinopathy Study (ETDRS) letter gain in BCVA for the overall population and by MH status, as applicable.
Safety end‐points included the incidence of retinal breaks/detachment, dyschromatopsia, new or worsening electroretinogram (ERG), subretinal fluid (SRF) and ellipsoid zone (EZ) abnormalities, MH progression, BCVA loss of ≥10 ETDRS letters and lens subluxation. If publications contained discrepancies in results presented in text or in tables, results from tables were upheld. KHB and BL independently assessed studies using the Quality Assessment Tool for Case Series Studies (National Heart, Lung and Blood Institute, Bethesda, MD) (National Heart Lung & Blood Institute 2019). Discrepant evaluations were resolved by discussion. Details on the study selection process and quality assessment for the RWS, and the full Risk of Bias spreadsheet for the RCTs are provided in Supplemental files 1, 2 and 3, respectively.
For the RCTs, we converted the original study databases into Stata data sets. KHB evaluated risk of bias using the Cochrane Risk of Bias Tool 2 (Higgins et al. 2016), using clinical study protocols, clinical study reports and statistical analysis plans. Dr Timothy Jackson independently reviewed the resulting output, and Dr Jackson’s amendments were integrated into the final output.
Data synthesis and analysis
Our analysis used all treated eyes as the denominator. We evaluated effectiveness overall and separately for eyes with MH (MH group) and eyes with VMT without MH (VMT group). If outcomes were reported for more than one time‐point, those from the last observation time‐point were deemed most relevant and were included in the analysis. For the RCTs, we chose Month 6 as the relevant evaluation time‐point for comparison with the RWS.
We analysed all data in Stata 16.0. Missing means and standard deviations (SD) of baseline characteristics were imputed using the method by Hozo et al (Hozo et al. 2005). The dichotomous treatment effectiveness estimates were calculated using a mixed effects logistic regression model with study as random effect and reported as proportions with their 95% confidence intervals. Forest plots show exact confidence intervals for individual study estimates and Wald confidence intervals for pooled estimates. Between‐study heterogeneity was assessed by the χ2 statistic of the likelihood ratio (LR) test comparing the random and fixed effects model (Nyaga et al. 2014).
We conducted a mixed effects meta‐regression to assess the prognostic value of study‐level covariates on pVMAR, nsVMAR and vitrectomy. Study was included as random effect, while study design (RCT = 0, RWS = 1), age (years), length of follow‐up (months) and within‐study proportions of ERM, MH, pseudophakia and women were included as fixed effects. Covariates were initially tested one by one, and those that were significant at the 0.05 level in the univariate analysis were included in a multivariate model. Funnel plots were used to evaluate the risk of publication bias among the studies. All statistical tests were two‐sided.
The RCT systematic review protocol is registered with PROSPERO (CRD42019121138). It was not possible to register the RWS review because data extraction had already started when registration was attempted (a new PROSPERO rule that went into effect in January 2019).
Results
Study population
For the RWS (Fig. 1, Panel A), we identified 392 records, 117 of which were duplicates. We screened the titles and abstracts of 275 publications and assessed the full text for 47 publications. At this stage of our analysis, we excluded 17 studies; 6 of these exclusions were due to part or all of the data being included in a later, collaborative paper (Paul et al. 2018). We retained 30 publications (2402 patients, 2416 eyes) for analysis.
Fig. 1.
PRISMA flow chart for (A) real‐world studies and (B) randomized controlled trials. PRISMA = Preferred Reporting Items for Systematic Reviews and Meta‐Analyses.
For the RCTs (Fig. 1, Panel B), 335 potentially eligible records were identified. After removing duplicates, 215 were screened at the title and abstract level. Full text was assessed for 31 records. In total, 15 publications were included, which described 5 unique RCTs, all sponsored by Oxurion, the manufacturer of ocriplasmin. These included the dose‐ranging Microplasmin for Intravenous Injection (MIVI) IIT trial (NCT00435539), the pivotal Microplasmin for Intravitreous Injection‐Traction Release without Surgical Treatment (MIVI‐TRUST) trials MIVI 006 (NCT00781859) and MIVI 007 (NCT00798317), and the long‐term Ocriplasmin for Treatment for Symptomatic Vitreomacular Adhesion Including Macular Hole (OASIS) trial (NCT01429441). A fifth trial (NCT01889251) codenamed J‐12‐075 was conducted in Japan and had not been published as of March 2019. Oxurion made individual participant data available for all 5 trials.
Table 1 summarizes the design and patient baseline characteristics for each study. Mean age was 70.8 (standard deviation [SD] 3.59) years, 1439 participants (68.1%) were women, and 600 (25.0%) presented with MH at baseline. Most RWS were single‐centre retrospective series with 21 (70%) having fewer than 50 patients. Two studies only included patients with MH while 3 studies included only patients with VMT. Five studies were prospective in nature, including two Oxurion‐sponsored observational studies (Tadayoni et al. 2018; Khanani et al. 2019).
Table 1.
Patient baseline characteristics of included real‐world studies
| Author (year) | Design | Patients | Age, mean (SD) | Women, N (%) | Eyes | MH, N (%) | ERM, N (%) | Focal VMA, N (%) | Phakic, N (%) | Baseline logMAR, mean (SD) |
Follow‐up (weeks), mean (SD) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| (Barca et al. 2018) | R | 74 | 71.0 | 46 (62.2) | 74 | 20 (27.0) | 12 (16.2) | 71 (95.9) | 61 (82.4) | 0.48 (0.31) | 16 (9) |
| (Cacciamani et al. 2017) | P | 15 | 55.1 (4.6) | 9 (60.0) | 16 | 0 (0.0) | 0 (0.0) | 16 (100.0) | 12 (75.0) | 0.12 | 26 (0) |
| (Cereda et al. 2017) | R | 15 | 70.0 | 10 (66.7) | 15 | 5 (33.3) | 0 (0.0) | 15 (100.0) | 15 (100.0) | 0.40 | 26 (0) |
| (Chatziralli et al. 2016c) | P | 24 | 71.4 (8.7) | 16 (66.7) | 24 | 7 (29.2) | 3 (12.5) | 24 (100.0) | 15 (62.5) | 0.28 (0.24) | 8 (3) |
| (Feng et al. 2018) | R | 47 | 71.0 (10.0) | 33 (70.2) | 49 | 16 (32.7) | 12 (24.5) | 43 (87.8) | 32 (65.3) | 0.64 (0.28) | 45 (9) |
| (Figueira et al. 2016) | R | 78 | 70.9 (8.7) | 54 (69.2) | 83 | 12 (14.5) | 9 (10.8) | 83 (100.0) | 60 (72.3) | 0.40 | 20 (17) |
| (Haynes et al. 2017) | R | 241 | . | 241 | 112 (46.5) | . | . | ||||
| (Heider et al. 2016) | R | 96 | 70.8 (9.2) | 55 (57.3) | 96 | 5 (5.2) | 16 (16.7) | 78 (81.2) | 62 (64.6) | 0.37 (0.21) | 31 (28) |
| (Itoh et al. 2014) | R | 19 | 69.6 | 14 (73.7) | 19 | 12 (63.2) | 0.30 | 12 (0) | |||
| (Khanani et al. 2019) | P | 539 | 71.1 (8.3) | 345 (64.0) | 539 | 137 (25.4) | 129 (23.9) | 449 (83.3) | 359 (66.6) | . | 32 (16) |
| (Kim et al. 2013) | R | 19 | 71.2 (6.4) | 11 (57.9) | 19 | 6 (31.6) | 8 (42.1) | 16 (84.2) | 15 (78.9) | 0.44 (0.21) | 8 (5) |
| (Lenk et al. 2018) | R | 21 | 71.1 (10.7) | 14 (66.7) | 23 | 4 (17.4) | 1 (4.3) | 23 (100.0) | 16 (69.6) | 0.39 (0.25) | . |
| (Lim et al. 2017) | R | 208 | . | 143 (68.8) | 208 | 75 (36.1) | 179 (86.1) | 0.52 (0.34) | 29 | ||
| (Manousaridis et al., 2017) | R | 20 | 68.5 | 20 | 6 (30.0) | 1 (5.0) | 20 (100.0) | 0.51 | 4 | ||
| (Mastropasqua et al. 2016) | R | 14 | . | 11 (78.6) | 14 | 0 (0.0) | 4 (28.6) | 13 (92.9) | 0.33 (0.24) | 4 (0) | |
| (Meyer et al. 2015) | R | 22 | 72.4 (10.2) | 14 (63.6) | 22 | 3 (13.6) | 4 (18.2) | 13 (59.1) | 0.51 | 11 (6) | |
| (Michalska‐Malecka et al. 2016) | R | 16 | 75.6 (6.0) | 16 | 4 (25.0) | 10 (62.5) | 0.39 (0.15) | 4 (0) | |||
| (Muqit et al. 2018) | R | 25 | 71.0 | 17 (68.0) | 25 | 6 (24.0) | 0 (0.0) | 25 (100.0) | 21 (84.0) | 0.52 | 28 (6) |
| (Nudleman et al. 2016) | R | 35 | 74.1 | 25 (71.4) | 36 | 9 (25.0) | 12 (33.3) | 17 (47.2) | 0.49 | 51 (6) | |
| (Paul et al. 2018) | R | 167 | 72.7 (8.9) | 118 (70.7) | 167 | 0 (0.0) | 40 (24.0) | 111 (66.5) | . | . | |
| (Quezada‐Ruiz et al. 2015) | R | 25 | 74.0 | 17 (68.0) | 25 | 8 (32.0) | 2 (8.0) | 24 (96.0) | 14 (56.0) | 0.50 | 25 (10) |
| (Scholz et al. 2017) | R | 14 | 73.0 (10.0) | 12 (85.7) | 14 | 1 (7.1) | 14 (100.0) | 11 (78.6) | 0.36 (0.22) | 13 | |
| (Sharma et al. 2015) | R | 56 | 72.0 | 38 (67.9) | 58 | 15 (25.9) | 23 (39.7) | 58 (100.0) | 37 (63.8) | 0.51 (0.37) | 37 (21) |
| (Singh et al. 2014) | R | 17 | 68.8 (9.0) | 11 (64.7) | 17 | 3 (17.6) | 3 (17.6) | 13 (76.5) | 12 (70.6) | 0.36 | . |
| (Steel et al. 2015) | P | 12 | 72.6 (4.0) | 9 (75.0) | 12 | 12 (100.0) | 2 (16.7) | 12 (100.0) | 10 (83.3) | . | 12 (0) |
| (Steel et al. 2016) | R | 33 | 71.0 (6.0) | 22 (66.7) | 33 | 31 (93.9) | 0 (0.0) | 31 (93.9) | 26 (78.8) | 0.72 (0.22) | . |
| (Tadayoni et al. 2018) | P | 466 | 71.7 (8.3) | 344 (73.8) | 466 | 86 (18.5) | 0 (0.0) | 466 (100.0) | 0.29 | 26 | |
| (Tschuppert & Gerding 2016) | R | 12 | . | 12 | 1 (8.3) | 8 (66.7) | . | . | |||
| (Warrow et al. 2015) | R | 35 | 69.4 (9.2) | 23 (65.7) | 35 | 6 (17.1) | 6 (17.1) | 33 (94.3) | 25 (71.4) | 0.46 (0.35) | 14 (7) |
| (Willekens et al. 2015) | R | 37 | 71.2 (9.5) | 28 (75.7) | 38 | 10 (26.3) | 2 (5.3) | 37 (97.4) | 32 (84.2) | . | . |
| Summary a | 2402 | 70.8 (3.59) | 1439 (68.1) | 2416 | 600 (25.0) | 289 (15.2) | 1730 (91.6) | 1019 (68.8) | 0.43 (0.13) | 20.9 (12.99) |
ERM = epiretinal membrane; logMAR = logarithm of the minimal angle of resolution; MH = macular hole; P = prospective; R = retrospective; SD = standard deviation; VMA = vitreomacular adhesion.
Percentages in this row refer to the proportion of patients with a given characteristic (e.g. MH) among the studies that reported this characteristic.
Epiretinal membranes, reported in 23 studies, were present in 289 patients (15.2%). Mean weighted BCVA (24 studies) was 0.43 logMAR (63.5 ETDRS letters). Average follow‐up was reported for 23 studies and ranged from 4 to 52 weeks (unweighted average: 20.9 weeks). Minimum follow‐up (reported in 23 studies) ranged from 1 to 52 weeks.
Reporting quality of treatment outcomes in RWS was variable and often incomplete. The quality assessment grid (Fig. S1) indicates that lack of information on statistical methodology and incompleteness of reporting (for the purpose of our analysis) were the most common deficiencies.
Patient baseline characteristics of the ocriplasmin RCTs are described in Table 2. Mean age and BCVA were comparable to those of the RWS, as were the proportions of women, phakic eyes and eyes with MH. On the other hand, the proportion of ERM was twofold greater than in the RWS, while focal VMA was less frequent than in the RWS.
Table 2.
Baseline characteristics of ocriplasmin‐treated patients in randomized controlled trials.
| Trial | Patients | Age (years), mean (SD) | Women, N (%) |
MH, N (%) |
ERM, N (%) |
Focal VMA, N (%) |
Phakic, N (%) |
Baseline logMAR, mean (SD) | Follow‐up (weeks) |
|---|---|---|---|---|---|---|---|---|---|
| MIVI IIT (Stalmans et al. 2010) | 13 | 74.2 (5.7) | 8 (61.5) | 1 (7.7) | N/A | N/A | 7 (53.8) | 0.51 (0.20) | 26 |
| MIVI 006 (Stalmans et al. 2012) | 219 | 71.5 (10.2) | 148 (67.6) | 57 (26.0) | 86 (39.3) | 145 (66.2) | 128 (58.4) | 0.41 (0.22) | 26 |
| MIVI 007 (Stalmans et al. 2012) | 245 | 72.6 (7.6) | 166 (67.8) | 49 (20.0) | 98 (40.0) | 169 (69.0) | 164 (66.9) | 0.43 (0.27) | 26 |
| OASIS (Dugel et al. 2016) | 145 | 69.4 (10.0) | 102 (70.3) | 50 (34.5) | 33 (22.8) | 128 (88.3) | 106 (73.1) | 0.43 (0.18) | 104 |
| J‐12‐075 a | 115 | 68.1 (7.3) | 59 (51.3) | 43 (37.4) | 35 (30.4) | 104 (90.4) | 104 (90.4) | 0.39 (0.18) | 26 |
| Summary | 737 | 70.6 (9.44) | 483 (65.5) | 200 (27.1) | 252 (34.2) | 546 (74.1) | 509 (69.1) | 0.42 (0.23) |
For all percentages shown in the table, the denominator is number of patients; this is the same as number of eyes for all RCTs.
ERM = epiretinal membrane; logMAR = logarithm of the minimal angle of resolution; MH = macular hole; MIVI = Microplasmin for Intravenous Injection; MIVI‐TRUST = Microplasmin for Intravitreous Injection‐Traction Release without Surgical Treatment; N/A = not assessed; OASIS = Ocriplasmin for Treatment for Symptomatic Vitreomacular Adhesion Including Macular Hole; SD = standard deviation; VMA = vitreomacular adhesion.
Study J‐12‐075 (NCT01889251) has not been published to date.
Overall risk of bias in the RCTs was low (Fig. S2). The MIVI 006 and MIVI 007 trials were placebo‐controlled, while the other trials used sham injection as control, administered by a trial site member who did not participate in the evaluation of patients. All outcomes were prespecified, intention‐to‐treat analysis was used throughout, and missing values were imputed using the ‘last observation carried forward’ method.
Effectiveness in RWS
Table 3 shows the VMAR, vitrectomy and MH closure outcomes by study. While only 2 publications explicitly reported nsVMAR outcomes (Lim et al. 2017; Tadayoni et al. 2018), 13 more studies contained sufficient information to derive nsVMAR outcomes. Nonsurgical VMAR rates were therefore available for 15 studies, 14 of which detailed outcomes in VMT patients and 14 in MH patients. Six studies did not report separate VMAR incidence rates for the VMT and MH subgroups, and 8 studies did not report the incidence of vitrectomy.
Table 3.
Main efficacy outcomes of included real‐world studies.
| Author (year) | Nonsurgical VMAR | Primary VMAR | Vitrectomy | MH closure | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| All | VMT | MH | All | VMT | MH | All | VMT | MH | MH | |
| (Barca et al. 2018) | 44/74 (59.5) | 12/74 (16.2) | 8/20 (40.0) | |||||||
| (Cacciamani et al. 2017) | 11/16 (68.8) | 11/16 (68.8) | 11/16 (68.8) | 11/16 (68.8) | 0/16 (0.0) | 0/16 (0.0) | ||||
| (Cereda et al. 2017) | 10/15 (66.7) | 7/10 (70.0) | 3/5 (60.0) | 12/15 (80.0) | 7/10 (70.0) | 5/5 (100.0) | 2/15 (13.3) | 0/10 (0.0) | 2/5 (40.0) | 3/5 (60.0) |
| (Chatziralli et al. 2016c) | 16/24 (66.7) | 2/7 (28.6) | ||||||||
| (Feng et al. 2018) | 14/49 (28.6) | 10/33 (30.3) | 4/16 (25.0) | 20/49 (40.8) | 11/33 (33.3) | 9/16 (56.2) | 25/49 (51.0) | 14/33 (42.4) | 11/16 (68.8) | 4/16 (25.0) |
| (Figueira et al. 2016) | 47/83 (56.6) | 39/71 (54.9) | 8/12 (66.7) | 4/83 (4.8) | 0/71 (0.0) | 4/12 (33.3) | 6/12 (50.0) | |||
| (Haynes et al. 2017) | 44/129 (34.1) | 31/112 (27.7) | ||||||||
| (Heider et al. 2016) | 40/96 (41.7) | 38/96 (39.6) | 0/5 (0.0) | |||||||
| (Itoh et al. 2014) | 9/19 (47.4) | |||||||||
| (Khanani et al. 2019) | 283/539 (52.5) | 198/383 (51.7) | 85/137 (62.0) | 137/539 (25.4) | 67/383 (17.5) | 70/137 (51.1) | 38/137 (27.7) | |||
| (Kim et al. 2013) | 7/19 (36.8) | 4/13 (30.8) | 3/6 (50.0) | 8/19 (42.1) | 4/13 (30.8) | 4/6 (66.7) | 12/19 (63.2) | 9/13 (69.2) | 3/6 (50.0) | 3/6 (50.0) |
| (Lenk et al. 2018) | 6/23 (26.1) | 4/19 (21.1) | 2/4 (50.0) | 8/23 (34.8) | 4/19 (21.1) | 4/4 (100.0) | 8/23 (34.8) | 6/19 (31.6) | 2/4 (50.0) | 2/4 (50.0) |
| (Lim et al. 2017) | 90/208 (43.3) | 60/133 (45.1) | 30/75 (40.0) | 65/208 (31.2) | 25/133 (18.8) | 33/75 (44.0) | 30/75 (40.0) | |||
| (Manousaridis et al., 2017) | 12/20 (60.0) | 5/20 (25.0) | 2/14 (14.3) | 3/6 (50.0) | 2/6 (33.3) | |||||
| (Mastropasqua et al. 2016) | 7/14 (50.0) | 7/14 (50.0) | ||||||||
| (Meyer et al. 2015) | 12/22 (54.5) | 9/19 (47.4) | 3/3 (100.0) | 14/22 (63.6) | 11/19 (57.9) | 3/3 (100.0) | 2/22 (9.1) | 2/19 (10.5) | 0/3 (0.0) | 1/3 (33.3) |
| (Michalska‐Malecka et al. 2016) | 12/16 (75.0) | 10/12 (83.3) | 2/4 (50.0) | |||||||
| (Muqit et al. 2018) | 14/25 (56.0) | 10/19 (52.6) | 4/6 (66.7) | 14/25 (56.0) | 10/19 (52.6) | 4/6 (66.7) | 2/25 (8.0) | 0/19 (0.0) | 2/6 (33.3) | 4/6 (66.7) |
| (Nudleman et al. 2016) | 15/36 (41.7) | 7/27 (25.9) | 8/9 (88.9) | 15/36 (41.7) | 7/27 (25.9) | 8/9 (88.9) | 3/36 (8.3) | 2/27 (7.4) | 1/9 (11.1) | 7/9 (77.8) |
| (Paul et al. 2018) | 79/167 (47.3) | 79/167 (47.3) | ||||||||
| (Quezada‐Ruiz et al. 2015) | 9/25 (36.0) | 7/15 (46.7) | 2/8 (25.0) | 11/25 (44.0) | 7/15 (46.7) | 4/8 (50.0) | 6/25 (24.0) | 0/15 (0.0) | 6/8 (75.0) | 2/8 (25.0) |
| (Scholz et al. 2017) | 7/14 (50.0) | 6/13 (46.2) | 1/1 (100.0) | 1/1 (100.0) | ||||||
| (Sharma et al. 2015) | 27/58 (46.6) | 23/43 (53.5) | 4/15 (26.7) | 29/58 (50.0) | 23/43 (53.5) | 6/15 (40.0) | 10/58 (17.2) | 0/43 (0.0) | 10/15 (66.7) | 4/15 (26.7) |
| (Singh et al. 2014) | 8/17 (47.1) | 6/14 (42.9) | 2/3 (66.7) | 8/17 (47.1) | 6/14 (42.9) | 2/3 (66.7) | 1/17 (5.9) | 0/14 (0.0) | 1/3 (33.3) | |
| (Steel et al. 2015) | 3/12 (25.0) | 3/12 (25.0) | 7/12 (58.3) | 7/12 (58.3) | 9/12 (75.0) | 9/12 (75.0) | 3/12 (25.0) | |||
| (Steel et al. 2016) | 19/33 (57.6) | 19/31 (61.3) | 11/31 (35.5) | |||||||
| (Tadayoni et al. 2018) | 230/466 (49.4) | 184/380 (48.4) | 46/86 (53.5) | 240/466 (51.5) | 184/380 (48.4) | 56/86 (65.1) | 56/466 (12.0) | 25/380 (6.6) | 31/86 (36.0) | 36/86 (41.9) |
| (Tschuppert & Gerding 2016) | 5/12 (41.7) | 4/11 (36.4) | 1/1 (100.0) | 5/12 (41.7) | 4/11 (36.4) | 1/1 (100.0) | 3/12 (25.0) | 3/11 (27.3) | 0/1 (0.0) | 1/1 (100.0) |
| (Warrow et al. 2015) | 15/35 (42.9) | 12/29 (41.4) | 3/6 (50.0) | 8/35 (22.9) | 3/29 (10.3) | 4/6 (66.7) | 1/6 (16.7) | |||
| (Willekens et al. 2015) | 27/38 (71.1) | 18/28 (64.3) | 9/10 (90.0) | 1/38 (2.6) | 4/10 (40.0) | |||||
| Random effects pooled estimate, % (95% CI) | 44.5 (39.7, 49.4) | 44.5 (38.3, 50.7) | 47.5 (35.6, 59.6) | 51.8 (49.6, 54.0) | 47.3 (42.4, 52.3) | 64.0 (59.0, 68.7) | 19.1 (12.7, 27.6) | 6.9 (2.7, 16.6) | 47.6 (39.8, 55.4) | 35.4 (30.4, 40.7) |
CI = confidence interval; MH = macular hole; SD = standard deviation; VMAR = vitreomacular adhesion resolution; VMT = vitreomacular traction.
Figures 3 and 4 show the random effect forest plots of VMAR outcomes in VMT and MH patients, respectively. Among VMAR outcomes, pVMAR was most commonly reported. For the overall population, its pooled estimate of 51.8% (95% CI: 49.6% to 54.0%; 28 studies, 1967 eyes) was approximately 7% higher than the nsVMAR estimate (44.5%; 95% CI: 39.7% to 49.4%; 15 studies, 1003 eyes). This difference was mostly due to the MH group, where initial (primary) VMAR success rates were reduced by 16.5% due to post‐VMAR vitrectomies (Table 3, Fig. 2). Nonsurgical MH closure occurred in 35.4% of MH eyes (95% CI: 30.4% to 40.7%) – Fig. 4.
Fig. 3.
Incidence of primary (A) and nonsurgical (B) VMA resolution in patients with MH, in real‐world settings. CI = confidence interval; MH = macular hole; VMA = vitreomacular adhesion; VMAR = vitreomacular adhesion resolution; VMT = vitreomacular traction.
Fig. 4.
Macular hole closure in real‐world settings. CI = confidence interval; MH = macular hole; RWS = real‐world settings.
Fig. 2.
Incidence of primary (A) and nonsurgical VMA (B) resolution in patients with VMT, in real‐world settings. CI = confidence interval; VMA = vitreomacular adhesion; VMAR = vitreomacular adhesion resolution.
Effectiveness comparison between RWS and RCTs
We subsequently performed an analysis comparing outcomes between RWS and the 5 Oxurion‐sponsored RCTs. Given the average length of follow‐up observed in the RWS (20.9 weeks), we chose the 26‐week evaluation time‐point of the RCTs to compare their results with the RWS.
Nonsurgical VMAR outcomes were significantly better in RWS than in RCTs (OR 1.66; 95% CI: 1.18–2.34; p = 0.004) while other outcomes tended to be similar (Table 4). We used meta‐regression to evaluate the prognostic value of baseline covariates on the effectiveness outcomes in the pooled RWS and RCT populations. We initially limited the multivariate meta‐regression to those studies that had information on all covariates that were significant in the univariate regression (results not shown). Since relatively few RWS reported all covariates of interest, this approach led to the exclusion of the majority of RWS and a preponderance of RCTs in the analyses. We therefore reran the univariate and multivariate meta‐regression without this limitation. The final results (shown in Table 5) indicated that in this broader analysis, the same covariates remained significant in the multivariate meta‐regression, with similar ORs. For nsVMAR, study design and proportion of ERM in the study were significant in univariate analysis, but only ERM remained significant when the two covariates were combined in a multivariate model. Higher proportions of MH were associated with higher vitrectomy rates. Average length of follow‐up was not significantly associated with improved VMAR outcomes. Study design (RWS, RCT) was not significantly associated with pVMAR, nsVMAR or vitrectomy outcomes when other covariates were taken into account.
Table 4.
Effectiveness outcomes in real‐world studies versus randomized trials
| Outcome | RWS | RCT | OR (95% CI) | p |
|---|---|---|---|---|
| Nonsurgical VMAR | 44.5 (39.7, 49.4) | 32.5 (26.6, 39.1) | 1.66 (1.18, 2.34) | 0.004 |
| Vitrectomy | 19.1 (12.7, 27.6) | 23.3 (15.2, 33.9) | 0.85 (0.32, 2.23) | 0.74 |
| MH closure | 35.4 (30.4, 40.7) | 35.5 (29.0, 42.5) | 1.00 (0.67, 1.50) | 0.99 |
CI = confidence interval; MH = macular hole; OR = odds ratio; RCT = randomized controlled trial; RWS = real‐world settings; VMAR = vitreomacular adhesion resolution.
Table 5.
Meta‐regression of prognostic covariates on effectiveness outcomes in RWS and RCTs.
| Outcome | Covariate | OR (95% CI) | p | Studies | Eyes |
|---|---|---|---|---|---|
| Nonsurgical VMAR | Design | 1.29 (0.93, 1.78) | 0.13 | 17 | 1507 |
| ERM | 0.20 (0.08, 0.51) | 0.001 | |||
| Vitrectomy | MH | 42.13 (5.02, 353.24) | 0.001 | 27 | 2625 |
Meta‐regression was carried out using a mixed effects logistic regression model with study as random effect. Study design (RCT = 0, RWS = 1), proportion of ERMs, MHs, pseudophakic status and women in the study, and age and length of follow‐up were included as fixed effects in a univariate model. Covariates that were significant at the .05 level in the univariate analysis were included in a multivariate model (not shown).
CI = confidence interval; ERM = epiretinal membrane; MH = macular hole; OR = odds ratio; RCT = randomized controlled trial; RWS = real‐world settings; VMAR = vitreomacular adhesion resolution.
Risk of publication bias
The funnel plots constructed for nsVMAR outcomes for VMT (Fig. 5) and MH patients (Fig. 6) show that only RCTs fell well below the 95% CI of the expected effect. Among the larger studies, one Oxurion‐sponsored observational study that excluded eyes with ERM (Tadayoni et al. 2018) exceeded the expected 95% CI of the effect size for VMT patients. For the MH subgroups, the industry‐sponsored RCTs remained at the low end of the efficacy estimates, while 3 RWS (Meyer et al. 2015; Nudleman et al. 2016; Tschuppert & Gerding 2016) with low numbers of MH patients reported higher than expected efficacy.
Fig. 5.
Funnel plot of nsVMAR in VMT patients (RWS studies and RCTs). Squares denote RCTs, while circles denote RWS. Blue markers indicate industry sponsorship. MIVI = Microplasmin for Intravitreous Injection; MIVI‐TRUST = Microplasmin for Intravitreous Injection‐Traction Release without Surgical Treatment; nsVMAR = nonsurgical vitreomacular adhesion resolution; OASIS = Ocriplasmin for Treatment for Symptomatic Vitreomacular Adhesion Including Macular Hole; RCT = randomized controlled trial; RWS = real‐world studies; VMT = vitreomacular traction (without macular hole).
Fig. 6.
Funnel plot of nsVMAR in MH patients (RWS studies and RCTs). Squares denote RCTs, while circles denote RWS. Blue markers indicate industry sponsorship. MH = macular hole; MIVI = Microplasmin for Intravitreous Injection; MIVI‐TRUST = Microplasmin for Intravitreous Injection‐Traction Release without Surgical Treatment; OASIS = Ocriplasmin for Treatment for Symptomatic Vitreomacular Adhesion Including Macular Hole; RCT = randomized controlled trial; RWS = real‐world settings; nsVMAR = nonsurgical vitreomacular adhesion resolution.
Safety outcomes in RWS studies
In general, adverse drug reactions (ADR) were inconsistently reported. To put the reported ADRs into perspective, we compared them with those from ocriplasmin‐treated patients in the OASIS trial (Dugel et al. 2016). We considered the OASIS trial to be the most appropriate comparator study to the RWS because the prevalence of ERM and focal VMA was closer to that observed in the RWS, and because it was the only published RCT that used SD‐OCT, and prospectively looked at inner/outer segment abnormalities, changes in photoreceptor area and thickness, (Dugel et al. 2016), full‐field ERG changes (Birch et al. 2018) and microperimetry changes (Sadda et al. 2017). Table 6 shows the incidence (95% CI) of selected ADRs in the RWS studies and the corresponding incidence in the OASIS trial. The presence of photopsia and vitreous floaters, the most common adverse reactions after ocriplasmin injection (Stalmans et al. 2010; Stalmans et al. 2012; Dugel et al. 2016), was reported in only 16 and 10 studies, respectively, while the presence (or absence) of retinal detachment was reported in only 17 publications. New or worsening SRF and EZ abnormalities were reported in 22.8% (95% CI: 13.7%–35.6%) and 26.6% (95% CI: 18.0%–37.5%) of patients, respectively. In publications that reported on resolution rates of new/worsening SRF, 100 of 104 cases (96.2%) resolved. Similarly, out of 112 new/worsening EZ abnormalities, 106 (94.6%) resolved. A BCVA loss of ≥10 ETDRS letters at any time during follow‐up was reported in 11 studies for an average rate of 16.1% (95% CI: 9.5%–25.9%). Lens instability was reported in 6 studies, 5 of whom reported no cases in a total of 678 eyes (Meyer et al. 2015; Quezada‐Ruiz et al. 2015; Manousaridis et al., 2017; Barca et al. 2018; Khanani et al. 2019), while one study of 241 eyes (Haynes et al. 2017) reported 4 cases.
Table 6.
Safety outcomes as reported in real‐world settings, compared with the OASIS trial
| Event | Studies reporting event | Events reported | Patient total |
RWS studies % (95% CI) |
OASIS % (95% CI) |
|---|---|---|---|---|---|
| Photophobia | 7 | 54 | 1368 | 3.9 (3.0, 5.1) | 13.0 (8.0, 19.6) |
| Photopsia | 16 | 362 | 1901 | 22.3 (16.3, 29.6) | 29.5 (22.2, 37.6) |
| Floaters | 10 | 120 | 827 | 22.4 (13.6, 34.6) | 37.7 (29.8, 46.1) |
| Dyschromatopsia | 12 | 102 | 1705 | 9.5 (4.4, 19.4) | 11.0 (6.4, 17.2) |
| Retinal detachment | 17 | 42 | 1812 | 2.2 (1.4, 3.3) | 1.3 (0.2, 4.9) |
| Retinal tear | 9 | 21 | 1097 | 1.2 (0.3, 5.4) | 1.3 (0.2, 4.9) |
| Subretinal fluid a | 16 | 152 | 1060 | 22.8 (13.7, 35.6) | 11.6 (6.9, 18.0) |
| EZ abnormal a | 15 | 156 | 732 | 26.6 (18.0, 37.5) | 21.4 (14.9, 29.2) |
| ERG abnormal a | 3 | 9 | 250 | 3.6 (1.9, 6.8) | 40.0 (24.9, 56.7) |
| MH progression a | 5 | 71 | 795 | 10.3 (4.3, 22.4) | 15.8 (10.3, 22.7) |
| BCVA ≥ 2‐line loss b | 12 | 246 | 1277 | 16.1 (9.5, 25.9) | 12.5 (7.6, 19.0) |
BCVA = best‐corrected visual acuity (Early Treatment Diabetic Retinopathy Study letters); CI = confidence interval; ERG = electroretinogram; EZ = ellipsoid zone; MH = macular hole; OASIS = Ocriplasmin for Treatment for Symptomatic Vitreomacular Adhesion Including Macular Hole; RWS = real‐world settings.
New or worsening.
Numbers for OASIS represent patients who had ≥10‐letter BCVA loss at any point prior to vitrectomy, regardless of subsequent improvement.
Discussion
To our knowledge, this meta‐analysis of 30 real‐world studies comprising 2402 patients and 2416 eyes is the largest published systematic literature review on ocriplasmin use in real‐world settings to date. A similar literature review (Chatziralli et al., ,2016a, 2016b, 2016c) with a cut‐off date of 30 June 2015 included 19 studies. However, that study also included eyes treated in RCTs (Stalmans et al. 2012), with only 414 of 878 eyes (47.2%) being treated in real‐world settings. In addition, that meta‐analysis did not take into account that the included RCTs (Stalmans et al. 2012) published nsVMAR end‐points, while the RWS studies did not.
We have shown in our analysis that the nsVMAR rate was significantly higher in RWS than in RCTs and that this difference was most likely due to the proportion of ERM patients in the studies. The crude prevalence of ERM in RCTs was 263 out of 737 eyes (35.7%) while the RWS studies included 261 patients with ERM out of 1771 (14.7%).
The funnel plot comparisons of nsVMAR across RWS studies and RCTs were not suggestive of publication or industry bias. In patients with VMT, the MIVI 006 and MIVI 007 trials were well below the pooled effectiveness estimate, while the OVIID‐1 trial (Tadayoni et al. 2018) had a higher than expected effectiveness. However, given the prevalence of baseline ERM in these trials (39.7% in MIVI 006/007 and 0% in OVIID‐1), these results are not unexpected. The estimates in the MH group showed similar trends, with outliers typically generated by studies with very small sample sizes.
Our ADR estimates from RWS are based on a limited number of studies and eyes. Common ‘nuisance’ ADRs such as photophobia and vitreous floaters were infrequently reported and were reported at a much lower frequency than what was observed in the OASIS trial. In contrast, new or worsening SRF was reported to be twice as frequent in RWS than the OASIS study, which seems to suggest a data collection/reporting bias towards ADRs that are deemed to be more clinically relevant, which can be especially important in retrospective chart reviews. In a study that prospectively monitored ADRs (including OCT findings) among 24 ocriplasmin‐treated patients (Chatziralli et al., ,2016a, 2016b, 2016c), the incidence of EZ disruption (17%), SRF (17%), dyschromatopsia (4%) and MH progression (4%) was not significantly different from those observed in the OASIS trial. In that study, 6 patients (25%) developed acute vision loss, although all but 2 (8%) had recovered at the last follow‐up examination.
This meta‐analysis has several limitations. The vast majority of RWS were retrospective in nature and therefore had no preplanned data collection or follow‐up schedule. The reporting of VMAR outcomes, as well as that of established prognostic risk factors such as age, sex and ERM/MH/phakic status in RWS, was often incomplete. These factors limited our meta‐regression to the overall population since risk factor prevalence was rarely available for MH/VMT subgroups. Also, adding multiple covariates into the model quickly resulted in decreasing numbers of studies and eyes in the analysis.
A challenge in any meta‐analysis of retrospectively collected data is the variability of the time‐point at which outcomes are measured. We tried to mitigate this problem by excluding studies with a mean follow‐up duration of less than 4 weeks, since it has been shown that that in a mixed population of VMT and MH patients, nsVMAR quickly reaches a plateau at 28 days that remains relatively stable thereafter (Stalmans et al. 2012; Dugel et al. 2016). It was therefore not entirely unexpected that no association was observed between length of follow‐up and nsVMAR incidence. Given the average length of follow‐up in the RWS (20.9 weeks), we chose the 6‐month time‐point for the analysis of the RCTs. Had we chosen the 3‐month time‐point instead, more than 40% of vitrectomies would not have been taken into account, which would have biased the results towards a higher nsVMAR incidence, as well as lower vitrectomy numbers.
Similarly, reporting of VMAR outcomes was inconsistent, with most authors not mentioning the incidence of vitrectomy after successful pVMAR. Eight out of 30 RWS (27%) did not mention vitrectomy at all. The lack of attention to this common outcome is concerning and should be addressed in future studies. Lastly, we were unable to evaluate the effectiveness of ocriplasmin in improving BCVA in RWS studies, since the reported data did not allow for such an analysis.
The main strength of this meta‐analysis resides in the size of the dataset and the diverse origin of patients. Despite the fact that the two largest RWS (comprising 1005 of the total of 2416 eyes – 41.6%) (Tadayoni et al. 2018; Khanani et al. 2019) were sponsored by the manufacturer of ocriplasmin, the pooled overall nsVMAR rate for the independent RWS studies was not different from that of the industry‐sponsored RWS studies in a mixed effects model with study as random effect and baseline ERM as fixed effect: OR 1.19 (95% CI: 0.82–1.73; p = 0.35). It is therefore unlikely that industry sponsorship resulted in a favourable effectiveness bias for ocriplasmin.
Our study supports the hypothesis that improved patient selection (i.e. avoiding treatment of eyes with ERM) is associated with higher rates of nsVMAR in RWS compared with RCTs. The incidence of vitrectomy and MH closure in RWS is consistent with that seen in RCTs. This review shows also that there is a need for improved reporting standards for VMT disease and relevant treatment outcomes.
Supporting information
Fig. S1 RWS quality assessment grida.
Fig. S2 Risk of bias in RCTs.
Table S1 EMBASE search terms for real‐world ocriplasmin studies (June 5, 2019).
Table S2 EMBASE search te 1 rms for ocriplasmin randomized controlled trials (June 5, 2019).
We thank Dr Timothy Jackson for independently reviewing the risk of bias assessment outputs for the randomized trials. Editorial assistance was provided by Fishawack Communications, Inc., and was funded by Oxurion, NV (formerly Thrombogenics). Oxurion had no role in the design or conduct of the study, other than to review the manuscript. These data have been previously presented, in part, at the 2019 EURETINA meeting, September 5–8, Paris, France and at the 2019 Association for Research in Vision & Ophthalmology (ARVO) annual meeting in Vancouver, Canada. AMK reports personal fees from Allergan, Genentech, Kodiak, Notal, Novartis, Opthea, Orbit Biomedical, Oxurion, Polyphotonix, Recens Medical, Regenxbio, Santen; grant funding from Adverum, Allergan, Gemini Therapeutics, Genentech, Graybug, Gyroscope, Kodiak, Novartis, Opthea, Oxurion, Regenxbio, and Recens Medical. RNC reports personal fees from Oxurion, outside the submitted work. KHB reports personal fees from Oxurion during and outside the conduct of the submitted work. BL reports personal fees from Oxurion during and outside of the conduct of the submitted work. PS reports no conflict of interest.
Contributor Information
Arshad M. Khanani, Email: arshad.khanani@gmail.com.
Koenraad H. Blot, Email: koenraad.blot@xintera.com.
References
- Barca F, Mucciolo DP, Caporossi T, Virgili G, Tartaro R & Rizzo S (2018): Italian real‐life experience on the use of ocriplasmin. BMJ Open Ophthalmol 3: e000110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Birch DG, Benz MS, Miller DM, Antoszyk AN, Markoff J, Kozma P, Meunier E & Sergott RC (2018): Evaluation of full‐field electroretinogram reductions after ocriplasmin treatment: results of the OASIS trial ERG substudy. Retina 38: 364–378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cacciamani A, Gelso A, Simonett JM, Ripandelli G, Pileri M, Stirpe M & Scarinci F (2017): Longitudinal microperimetry evaluation after intravitreal ocriplasmin injection for vitreomacular traction. Retina 37: 1832–1838. [DOI] [PubMed] [Google Scholar]
- Cereda MG, Preziosa C, D'Agostino I, Cozzi M, Bottoni F, Pellegrini M, Giani A & Staurenghi G (2017): Ocriplasmin for vitreomacular traction: Looking outside the macula: A wide‐field optical coherence tomography study. Retina 38: 1541–1548. [DOI] [PubMed] [Google Scholar]
- Chatziralli IP, Theodossiadis GP, Parikakis E, Datseris I & Theodossiadis P (2016c): Complications of intravitreal ocriplasmin for vitreomacular traction and macular hole: a prospective spectral‐domain optical coherence tomography study. Cutan Ocul Toxicol 35: 263–269. [DOI] [PubMed] [Google Scholar]
- Chatziralli I, Theodossiadis G, Parikakis E, Datseris I & Theodossiadis P. (2016a): Real‐life experience after intravitreal ocriplasmin for vitreomacular traction and macular hole: a spectral‐domain optical coherence tomography prospective study. Graefes Arch Clin Exp Ophthalmol = Albrecht von Graefes Archiv fur klinische und experimentelle Ophthalmologie 254: 223–233. [DOI] [PubMed] [Google Scholar]
- Chatziralli I, Theodossiadis G, Xanthopoulou P, Miligkos M, Sivaprasad S & Theodossiadis P. (2016b): Ocriplasmin use for vitreomacular traction and macular hole: A meta‐analysis and comprehensive review on predictive factors for vitreous release and potential complications. Graefes Arch Clin Exp Ophthalmol = Albrecht von Graefes Archiv fur klinische und experimentelle Ophthalmologie 254: 1247–1256. [DOI] [PubMed] [Google Scholar]
- Dugel PU, Tolentino M, Feiner L, Kozma P & Leroy A (2016): Results of the 2‐year ocriplasmin for treatment for symptomatic vitreomacular adhesion including macular hole (OASIS) randomized trial. Ophthalmology 123: 2232–2247. [DOI] [PubMed] [Google Scholar]
- Duker JS, Kaiser PK, Binder S et al. (2013): The International Vitreomacular Traction Study Group classification of vitreomacular adhesion, traction, and macular hole. Ophthalmology 120: 2611–2619. [DOI] [PubMed] [Google Scholar]
- European Medicines Agency (2013): EMA/CHMP/74766/2013: JETREA European public assessment report. 13 Jan 2013.
- Ezra E (2001): Idiopathic full thickness macular hole: natural history and pathogenesis. Br J Ophthalmol 85: 102–108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Feng HL, Roth DB, Hasan A et al. (2018): Intravitreal ocriplasmin in clinical practice: Predictors of success, visual outcomes, and complications. Retina 38: 128–136. [DOI] [PubMed] [Google Scholar]
- Figueira J, Martins D, Pessoa B et al. (2016): The Portuguese experience with ocriplasmin in clinical practice. Ophthal Res 56: 186–192. [DOI] [PubMed] [Google Scholar]
- Haller JA, Stalmans P, Benz MS et al. (2015): Efficacy of intravitreal ocriplasmin for treatment of vitreomacular adhesion: subgroup analyses from two randomized trials. Ophthalmology 122: 117–122. [DOI] [PubMed] [Google Scholar]
- Haynes RJ, Yorston D, Laidlaw DAH, Keller J & Steel DHW (2017): Real world outcomes of ocriplasmin use by members of the British and Eire Association of Vitreoretinal Surgeons. Eye 31: 107–112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Heider A, Dimopoulos S, Szurman P & Januschowski K (2016): Real‐life experience with intravitreal ocriplasmin on vitreomacular traction and full‐thickness macular holes. Acta Ophthalmol 96: e890–e891. [DOI] [PubMed] [Google Scholar]
- Higgins JP, Sterne J, Savović J, Page MJ, Hróbjartsson A, Boutron I, Reeves B & Eldridge S. (2016): A revised tool for assessing risk of bias in randomized trials. In: Chandler JMJ & Welch V (eds.) Cochrane Methods. John Wiley and Sons; 29–31. 10.1002/14651858.CD201601 [DOI] [Google Scholar]
- Hozo SP, Djulbegovic B & Hozo I (2005): Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol 5: 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Itoh Y, Kaiser PK, Singh RP, Srivastava SK & Ehlers JP (2014): Assessment of retinal alterations after intravitreal ocriplasmin with spectral‐domain optical coherence tomography. Ophthalmology 121: 2506–2507.e2502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jackson TL, Nicod E, Angelis A, Grimaccia F, Prevost AT, Simpson AR & Kanavos P (2013): Pars plana vitrectomy for vitreomacular traction syndrome: a systematic review and metaanalysis of safety and efficacy. Retina 33: 2012–2017. [DOI] [PubMed] [Google Scholar]
- Jackson TL, Regillo CD, Girach A & Dugel PU (2016): Baseline predictors of vitreomacular adhesion/traction resolution following an intravitreal injection of ocriplasmin. Ophthal Surg Lasers Imag Retina 47: 716–723. [DOI] [PubMed] [Google Scholar]
- Johnson MW (2010): Posterior vitreous detachment: evolution and complications of its early stages. Am J Ophthalmol 149: 371–382 e371. [DOI] [PubMed] [Google Scholar]
- Joshi M, Agrawal S & Christoforidis JB (2013): Inflammatory mechanisms of idiopathic epiretinal membrane formation. Mediat Inflamm 2013: 192582. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Khanani AM, Duker JS, Heier JS et al. (2019): Ocriplasmin treatment leads to symptomatic vitreomacular adhesion/vitreomacular traction resolution in the real‐world setting: The phase IV ORBIT study. Ophthalmol Retina 3: 32–41. [DOI] [PubMed] [Google Scholar]
- Kim BT, Schwartz SG, Smiddy WE, Doshi RR, Kovach JL, Berrocal AM, Moshfeghi AA & Fortun JA (2013): Initial outcomes following intravitreal ocriplasmin for treatment of symptomatic vitreomacular adhesion. Ophthal Surg Lasers Imag Retina 44: 334–343. [DOI] [PubMed] [Google Scholar]
- Lenk J, Matthé E, Ventzke S, Pillunat LE & Sandner D (2018): Initial clinical experiences using ocriplasmin for the treatment of vitreomacular traction with or without a macular hole. Klin Monatsbl Augenheilkd 235: 73–80. [DOI] [PubMed] [Google Scholar]
- Lim JI, Glassman AR, Aiello LP, Chakravarthy U, Flaxel CJ, Singerman LJ, Spaide RF & Roth DB, & MSOfVTS Group (2017): Macula Society collaborative retrospective study of ocriplasmin for symptomatic vitreomacular adhesion. Ophthalmol Retina 1: 413–420. [DOI] [PubMed] [Google Scholar]
- Manousaridis K, Peter‐Reichart S & Mennel S. (2017): Ocriplasmin treatment for vitreomacular traction in real life: can the indication spectrum be expanded? Graefes Arch Clin Exp Ophthalmol = Albrecht von Graefes Archiv fur klinische und experimentelle Ophthalmologie 255: 1907–1916. [DOI] [PubMed] [Google Scholar]
- Mastropasqua R, Di Antonio L, Ciciarelli V, Aharrh‐Gnama A, Rispoli M & Carpineto P (2016): Comparison of guided and unguided ocriplasmin injection for the ttreatment of vitreomacular traction: A preliminary study. J Ophthalmol 2016: 6521304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Meyer JC, Shah GK, Blinder KJ, Waheed NK, Reichel E, Stalmans P, Singer M & Tewari A (2015): Early evolution of the vitreomacular interface and clinical efficacy after ocriplasmin injection for symptomatic vitreomacular adhesion. Ophthal Surg Lasers Imag Retina 46: 209–216. [DOI] [PubMed] [Google Scholar]
- Michalska‐Malecka K, Witek K, Sierocka‐Stepien J, Wygledowska‐Promienska D & Nowak M (2016): Novel possibility of vitreomacular traction treatment. Acta Ophthalmol 94: e818–e819. [DOI] [PubMed] [Google Scholar]
- Muqit MMK, Hamilton R, Ho J, Tucker S & Buck H (2018): Intravitreal ocriplasmin for the treatment of vitreomacular traction and macular hole‐ A study of efficacy and safety based on NICE guidance. PLoS One 13: e0197072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- National Heart Lung and Blood Institute (2019): Study quality assessment tools: Quality assessment tool for case series studies. National Heart Lung en Blood Institute.
- Nudleman E, Franklin MS, Wolfe JD, Williams GA & Ruby AJ (2016): Resolution of subretinal fluid and outer retinal changes in patients treated with ocriplasmin. Retina 36: 738–743. [DOI] [PubMed] [Google Scholar]
- Nyaga VN, Arbyn M & Aerts M (2014): Metaprop: a Stata command to perform meta‐analysis of binomial data. Arch Public Health = Archives belges de sante publique 72: 39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paul C, Heun C, Muller HH et al. (2018): Calculating the individual probability of successful ocriplasmin treatment in eyes with VMT syndrome: a multivariable prediction model from the EXPORT study. Br J Ophthalmol 102: 1092–1097. [DOI] [PubMed] [Google Scholar]
- Quezada‐Ruiz C, Pieramici DJ, Nasir M et al. (2015): Outer retina reflectivity changes on sd‐oct after intravitreal ocriplasmin for vitreomacular traction and macular hole. Retina 35: 1144–1150. [DOI] [PubMed] [Google Scholar]
- Robinson KA & Dickersin K (2002): Development of a highly sensitive search strategy for the retrieval of reports of controlled trials using PubMed. Int J Epidemiol 31: 150–153. [DOI] [PubMed] [Google Scholar]
- Sadda SR, Dugel PU, Gonzalez VH, Meunier E & Kozma P (2017): THE OASIS MP‐1 SUBSTUDY: characterization of the effect of ocriplasmin on microperimetry parameters. Retina Publish Ahead of Print: 1. [DOI] [PMC free article] [PubMed]
- Scholz P, Sitnilska V, Hess J, Becker M, Michels S & Fauser S (2017): Comparison of resolution of vitreomacular traction after ocriplasmin treatment or vitrectomy. Retina 39: 180–185. [DOI] [PubMed] [Google Scholar]
- Sebag J (2009): Vitreous: the resplendent enigma. Br J Ophthalmol 93: 989–991. [DOI] [PubMed] [Google Scholar]
- Sharma P, Juhn A, Houston SK, Fineman M, Chiang A, Ho A & Regillo C (2015): Efficacy of intravitreal ocriplasmin on vitreomacular traction and full‐thickness macular holes. Am J Ophthalmol 159: 861–867 e862. [DOI] [PubMed] [Google Scholar]
- Singh RP, Li A, Bedi R, Srivastava S, Sears JE, Ehlers JP, Schachat AP & Kaiser PK (2014): Anatomical and visual outcomes following ocriplasmin treatment for symptomatic vitreomacular traction syndrome. Br J Ophthalmol 98: 356–360. [DOI] [PubMed] [Google Scholar]
- Stalmans P, Benz MS, Gandorfer A, Kampik A, Girach A, Pakola S & Haller JA (2012): Enzymatic vitreolysis with ocriplasmin for vitreomacular traction and macular holes. N Engl J Med 367: 606–615. [DOI] [PubMed] [Google Scholar]
- Stalmans P, Delaey C, de Smet MD, van Dijkman E & Pakola S (2010): Intravitreal injection of microplasmin for treatment of vitreomacular adhesion: results of a prospective, randomized, sham‐controlled phase II trial (the MIVI‐IIT trial). Retina 30: 1122–1127. [DOI] [PubMed] [Google Scholar]
- Steel DH, Parkes C, Papastavrou VT et al. (2016): Predicting macular hole closure with ocriplasmin based on spectral domain optical coherence tomography. Eye (Lond) 30: 740–745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Steel DH, Sandinha MT & White K (2015): The plane of vitreoretinal separation and results of vitrectomy surgery in patients given ocriplasmin for idiopathic macular hole. Invest Ophthalmol Vis Sci 56: 4038–4044. [DOI] [PubMed] [Google Scholar]
- Tadayoni R, Holz FG, Zech C, Liu X, Spera C & Stalmans P (2018): Assessment of anatomical and functional outcomes with ocriplasmin treatment in patients with vitreomacular traction with or without macular holes: Results of OVIID‐1 trial. Retina 39: 2341–2352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tschuppert S & Gerding H (2016): Transient subfoveal fluid and visual loss after ocriplasmin. Klin Monbl Augenheilkd 233: 453–455. [DOI] [PubMed] [Google Scholar]
- Warrow DJ, Lai MM, Patel A, Raevis J & Berinstein DM (2015): Treatment outcomes and spectral‐domain optical coherence tomography findings of eyes with symptomatic vitreomacular adhesion treated with intravitreal ocriplasmin. Am J Ophthalmol 159: 20–30. e21. [DOI] [PubMed] [Google Scholar]
- Willekens K, Abegao Pinto L, Vandewalle E, Stalmans I & Stalmans P (2015): Improved efficacy of ocriplasmin for vitreomacular traction release and transient changes in optic disk morphology. Retina 35: 1135–1143. [DOI] [PubMed] [Google Scholar]
- Wong SS, Wilczynski NL & Haynes RB (2006): Developing optimal search strategies for detecting clinically sound treatment studies in EMBASE. J Med Library Assoc 94: 41–47. [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Fig. S1 RWS quality assessment grida.
Fig. S2 Risk of bias in RCTs.
Table S1 EMBASE search terms for real‐world ocriplasmin studies (June 5, 2019).
Table S2 EMBASE search te 1 rms for ocriplasmin randomized controlled trials (June 5, 2019).