Abstract
Background:
Submacular hemorrhage (SMH) is an uncommon complication of neovascular age-related macular degeneration (AMD).
Objectives:
The study aim was to evaluate the efficacy of combination therapy vitrectomy with subretinal recombinant tissue plasminogen activator (rtPA) given under the retina and anti-VEGF given into the vitreous cavity in patients with SMH based on duration of symptoms and to identify biomarkers to help predict potential visual acuity gain after the procedure.
Design:
Retrospective case series.
Methods:
Patients with SMH secondary to AMD were treated with quadruple therapy: (1) vitrectomy; (2) rtPA administration; (3) bevacizumab injection into the vitreous cavity; followed by (4) monthly intravitreal double-dose bevacizumab 2.5 mg.
Results:
Forty-eight patients were treated (68.8% women; mean age, 80 years). A median of 14 days elapsed between SMH onset and surgical intervention (range, 2–120 days), with vitrectomy performed after ⩾30 days in 39.6% of patients. A significant improvement from baseline in mean best-corrected visual acuity (BCVA) was observed 6 months after SMH treatment (1.6 (baseline) vs 0.9 (Month 6) logMAR; p < 0.001). Overall, 46 (95.8%) patients had improved BCVA and two (4.2%) had stable (unchanged) BCVA in the treated eye. Improvement in BCVA at Month 6 was similar between the cohort with only SMH and patients with coexisting subretinal pigment epithelium hemorrhage (p = 0.11). The main biomarkers predicting surgical outcome were the state of the retina in the foveal center point and hemorrhage thickness. Treatment ⩾30 days after SMH onset resulted in poorer BCVA at Month 6 compared with treatment occurring <30 days after onset (p < 0.05).
Conclusion:
A combination of vitrectomy, subretinal rtPA, and bevacizumab, performed in patients with submacular hemorrhage due to neovascular AMD, followed by monthly intravitreal double-dose bevacizumab for SMH, was an effective intervention to achieve visual acuity improvement.
Introduction
Submacular hemorrhage (SMH) is an uncommon condition secondary to neovascular (wet) age-related macular degeneration (AMD).1–3 Presence of the SMH causes progressive anatomical and functional damage to photoreceptors through various mechanisms, including iron/hemosiderin-related toxicity, interference with nutrient/waste removal, and clot contraction–related forces.1,2 Given the progressive course of SMH-related damage, 1 it is believed that early intervention is important, although guidelines related to specific treatment and timing have not been established, and large, randomized trials are lacking.
The goal of SMH treatment is the removal of the clot to retain vision in the affected eye, and therapeutic strategies have evolved over time.1,2 These approaches have included single and multiple mechanical and enzymatic modalities, including surgery, recombinant tissue plasminogen activator (rtPA) injection, and anti-vascular endothelial growth factor (anti-VEGF) drug administration. 1 There have been reports on the use of a combination of pars plana vitrectomy, rtPA and anti-VEGF administration and partial fluid-air exchange for the treatment of SMH secondary to neovascular AMD.4–9 However, several of these studies restricted patient inclusion to those with a history of symptoms of ⩽2 weeks,4,5,9 and not all patients received anti-VEGF treatment during and/or routinely after the procedural intervention (e.g., limited to those with posttreatment neovascularization).4,5,7,8 Data on outcomes for patients with SMH treatment are lacking, and further data are needed on the efficacy and safety of a combined treatment strategy. The aim of this large case series was to assess the benefit of standardized surgical intervention performed by the same surgeon using a quadruple therapy strategy (vitrectomy, subretinal rtPA, intravitreal anti-VEGF (bevacizumab) injection, followed by a high monthly dosage of bevacizumab) in all patients. In addition, analyses were conducted to identify potential relationships between timing of the surgery, select SMH characteristics like size, thickness, location and state of the retina in the macular region on the final visual acuity outcome.
Material and methods
Medical records were retrospectively reviewed for patients with SMH secondary to neovascular AMD who were treated at the Warsaw Eye Hospital (Warsaw, Poland) by a single ophthalmic surgeon between January 2018 and March 2023.
Patients with SMH were consecutively included in the analysis if they had a dense hemorrhage >1 disk diameter obstructing underlying anatomical structures and had experienced a sudden and severe decrease in visual acuity due to hemorrhage. The SMH comprised the foveolar region.
Exclusion criteria encompassed any clinically significant opacification of media, cataract or vitreous hemorrhage, which obstructed the possibilities of examining the fundus and/or required extension of the surgery beyond vitrectomy.
Patients with other concomitant ophthalmic conditions potentially influencing visual acuity were not included into the study.
Patients and the public were not involved in the design, conduct, reporting, or dissemination plans for this research.
The treatment protocol was standardized across patients. Under retrobulbar anesthesia, all patients received a pars plana 23-gauge vitrectomy (without suture unless the surgeon believed there was a risk of severe postsurgical hypotony).
Recombinant tissue plasminogen activator (Actilyse, Boehringer Ingelheim International GmbH, Ingelheim am Rhein, Germany) was mixed in a balanced salt solution to a concentration of 12.5 µg/0.1 mL and up to 0.15 mL was administered into the subretinal space via a 41-gauge subretinal cannula (DORC BV, Zuidland, the Netherlands). The rtPA volume delivered was left for self-absorption, and there was no aspiration of any subretinal fluid during surgery. Subsequently, bevacizumab 2.5 mg/0.1 mL (Avastin, Roche Pharm AG, Grenzach-Whylen, Germany) was injected intravitreally, followed by a partial fluid-air exchange to reduce risk of postsurgical hypotony. Patients remained in a supine position for 30–40 min after surgery completion, with no additional postsurgical posturing required. In the treated eye, patients received an intravitreal injection of a fixed-dose protocol of bevacizumab 2.5 mg administered 4 weeks after surgery and continued once monthly, regardless of the state of the retina.
Patients were examined at baseline (presentation) and monthly for 6 months after treatment. Ophthalmic assessments included best-corrected visual acuity (BCVA) measured on the ETDRS chart (Precision Vision, USA), intraocular pressure assessed with Goldmann applanation tonometry, and optical coherence tomography (OCT) scanning (DRI OCT-1 Triton, Topcon Europe Medical BV, Capelle a/d IJssel, the Netherlands). The diagnosis of AMD was made based on the comparative analysis of fundoscopic and OCT findings in both eyes.
Patients were divided into subgroups based on state of the retina due to wet-AMD: eyes with minor no alteration in retinal layers integrity seen on OCT, eyes with some degree of intraretinal edema and eyes with scar tissue formation in the macular region. Data was fully retrieved and no patient was lost to follow-up.
Anonymized data were analyzed using the IBM SPSS (Statistical Package for the Social Sciences) version 25 (IBM, Armonk, NY, USA). Changes from baseline were compared using the Wilcoxon test with a significance level of 0.05. Spearman correlation analyses (coefficient (r)) were conducted to explore relationships between select baseline characteristics and BCVA.
Results
A total of 48 patients (48 eyes) were included in the study; the majority (68.8%) of whom were women; the overall mean age was 80 ± 7 years (Table 1). Prior to the onset of SMH, 32 (66.7%) patients had not taken anticoagulant therapy compared with 16 (33.3%) who had (p = 0.02). Of the 16 patients who had exposure to anticoagulant therapy, 11 (68.8%) were taking aspirin 150 mg/d (Table 1). Therefore, no statistically significant correlation between the onset of SMH and anticoagulation therapy was seen.
Patient demographic and baseline characteristics.
Hemorrhage areas of 10 mm × 10 mm or more (n = 7) were removed from calculation.
From expression to intervention.
AMD, age-related macular degeneration; BCVA, best-corrected visual acuity; logMAR, logarithm of the minimum angle of resolution; SMH, submacular hemorrhage.
Only 7 (14.6%) out of 48 patients were phakic, but the lens did not require removal. Less than half (39.6%) of patients had previously received treatment in the eye currently affected (mean duration between last injection and onset of SMH, 481 days; Table 1), and for the majority (60.4%) of patients, SMH was the first sign of neovascular AMD.
Based on OCT assessment, 34 (70.8%) patients had both a subretinal and subretinal pigment epithelium (RPE) hemorrhage (χ2 = 6.15; p = 0.01), and 14 (29.1%) patients had a subretinal hemorrhage only (χ2 = 2.57; p = 0.11). Seven (14.6%) of the 48 patients had an SMH of 10 mm × 10 mm or more, clinically extending over temporal arcades (Table 1). Approximately one-third of patients had some degree of scarring in the macula as a result of advanced-stage wet AMD; additionally, one-third had some changes due to wet AMD, such as intraretinal edema (Table 1). A median of 14 days elapsed between SMH onset and treatment (range, 2–120 days).
For the study purposes, the follow-up period of 6 months was used as the indicator of long-term treatment result.
A significant improvement from baseline in BCVA (mean, 1.6 logMAR) was observed at Month 6 (mean, 0.9 logMAR; p < 0.001; Figure 1), with 46 (95.8%) of the 48 patients having improved BCVA and two (4.2%) having stable (unchanged) BCVA in the treated eye. Fundus photographs and OCT scans of a representative case before and after quadruple therapy are shown in Figure 2(a) and (b). Improvement in BCVA at Month 6 was similar for SMH patients and those with coexisting sub-RPE hemorrhage (p = 0.11).
Visual acuity at baseline and Month 6.
Patient presenting with BCVA 0.8 logMAR with SMH and retinal detachment shown on fundus photograph and OCT scan (a) with improvement in BCVA 0.4 logMAR and improvement in SMH and retinal architecture at 6 months after treatment (b).
A comparison of baseline parameters in patients grouped by the quality of retina in the foveal center point showed no significant differences between groups (Supplemental Table).
BCVA outcomes at Month 6 were better for eyes with conserved retinal layer integrity at the time of SMH onset compared to those with scarring (p = 0.01) or intraretinal edema in the macula (p < 0.001; Figure 3(a)). In addition, a significant difference in BCVA at Month 6 was observed between patients with a time interval between SMH expression and treatment of <30 days compared with those with a gap ⩾30 days (p < 0.05; Figure 3(b)), Mean BCVA at Month 6 was similar between patients with previous treatment with anti-VEGF therapy (0.7 (0.4) logMAR) and those who had not previously received anti-VEGF therapy (0.6 (0.5) logMAR; p = 0.24) in the affected eye.
Visual acuity at Month 6 by baseline fovea characteristics (a) and by time interval between SMH expression and treatment (b).
Correlation analyses were conducted to identify potential relationships between select baseline characteristics and visual acuity. There was a weak, non-significant correlation between time interval (SMH expression to intervention) and BCVA in the affected eye (r = −0.25; p = 0.09; Figure 4(a)) and the fellow eye (r = 0.10; p = 0.50) at baseline. However, there was a positive correlation between baseline SMH area and baseline BCVA (logMAR) in the affected eye (r = 0.32; p = 0.03) and between the baseline SMH area and baseline BCVA in the fellow eye (r = 0.33; p = 0.03). Hemorrhage spreading beyond the vascular arcades is usually considered to be extensive and is associated with a poor prognosis. 1 Therefore, correlation analyses were conducted, excluding seven patients with a minimum hemorrhage area of 10 mm × 10 mm. With this exclusion, there was no longer a correlation between SMH area and BCVA at baseline in the affected eye (r = 0.03; p = 0.88); however, a correlation between SMH area and baseline BCVA in the fellow eye was maintained (r = 0.42; p = 0.008; Figure 4(b)). Retinal thickness in the fovea moderately correlated with BCVA (logMAR) both at baseline (r = 0.43; p = 0.003) and after intervention (r = 0.36; p = 0.01). Moreover, a correlation was observed between baseline BCVA and retinal thickness in the highest (uppermost) point (r = 0.3; p = 0.04) and the foveal center point thickness (r = 0.41; p = 0.004).
Correlation between time interval (SMH expression to treatment) and BCVA in affected eye prior to treatment (a) and correlation between SMH area and visual acuity in the fellow eye prior to affected eye treatment (b).
Fourteen (29.2%) of 48 patients experienced postoperative complications, including foreign body sensation and subconjunctival hemorrhage, which were considered by the investigator to be minor, with no need for treatment. Three patients developed macular holes, all of which were successfully closed after an additional vitrectomy with internal limiting membrane peeling and gas tamponade.
Discussion
Submacular hemorrhage leads to sudden vision loss and treatment is critical to minimize further loss of visual acuity.1,2 Treatment of SMH should be initiated as soon as possible; however, data have been lacking on the impact of timing between SMH onset and treatment on visual acuity outcomes. Some surgeons are reluctant to treat the condition if the patient presents with an SMH duration of onset >7 days.10–13 In the current study, significant improvements in BCVA compared to baseline were observed in patients who had a broad range of time between SMH expression and treatment. Of note, BCVA at 6 months was statistically significantly lower in patients in whom treatment was performed ⩾30 days after SMH onset, compared with <30 days after SMH onset. Another study reported a negative correlation between baseline hemorrhage duration and postoperative visual acuity at 6 months post-treatment (r = 0.53; p = 0.01). 8 However, that study was limited to 30 patients, and the mean duration between SMH and treatment was shorter, at 13.7 (±8.0) days, with a maximum interval between onset and treatment of 30 days. 8 In the current study, the mean timeframe was longer—at 22 days, with a maximum interval between onset and treatment of 120 days. In addition, improvement in visual acuity at Month 6 was not impacted by prior anti-VEGF therapy, as there was no statistically significant difference between those with and without a history of its use in the affected eye. Correlations were observed between visual acuity and retinal thickness in the fovea, at baseline and after treatment, and thickness in the foveal center point, suggesting that the thickness of the hemorrhage may be responsible for the degree of visual acuity loss rather than the area of the SMH. Surprisingly, there was a correlation between SMH area and baseline BCVA in the fellow eye, with a larger area resulting in poorer BCVA (r = 0.42; p = 0.008).
Global consensus guidelines on the treatment of SMH secondary to neovascular AMD are lacking. Various modalities in multiple combinations have been considered, with the goal of removing blood from the submacular region and, subsequently, treating macular neovascularization with anti-VEGF therapy. 1 A 2016 literature review that assessed data from multiple trials suggested that combination intravitreal rtPA and pneumatic displacement, followed by anti-VEGF therapy, may be considered an appropriate treatment strategy. 1 A 2024 randomized, controlled trial reported that ranibizumab followed by adjunct use of perfluoropropane (C3F8) did not provide added benefit for SMH, as there was no statistically significant difference between patients who did versus did not receive perfluoropropane. 14 However, intravitreal rtPA adjunctive therapy after ranibizumab did significantly improve visual acuity (p = 0.03). 14 A study of vitrectomy, subretinal rtPA, and 3 monthly ranibizumab injections for SMH reported visual acuity improvement in 73.3% of 45 patients (mean follow-up, 12.9 months), 15 which was lower than that observed herein (95.8% of 48 patients). In the current study, SMH was treated using a combination of vitrectomy, subretinal rtPA, and bevacizumab injections, followed by a monthly double-dose of bevacizumab. This combination is the treatment of choice for SMH at the Warsaw Eye Hospital. Administration of rtPA during vitrectomy allows direct delivery into the subretinal space at a concentration lower than that required for intravitreal injection (i.e., nontoxic dose (e.g., ⩽50 μg) 6 ). The double-dose (2.5 mg) of bevacizumab was administered during surgery to reduce subretinal neovascularization, given that the pharmacokinetics (e.g., half-life) of intravitreal bevacizumab in this treatment scenario (i.e., postvitrectomy) were unknown.6,16 Additionally, data were lacking on the appropriate concentration of anti-VEGF medication in vitrectomized eyes and, therefore, the higher dose of bevacizumab (2.5 mg) was continued monthly to minimize the risk of hemorrhage recurrence, stabilize the visual acuity regained post-surgery in these patients, and minimize a change in anti-VEGF medication that might have confounded interpretation of the results. It has been proposed that anti-VEGF therapy should be administered only when symptoms of clinical activity of subretinal proliferation are present in patients previously treated for SMH. 17 At the Warsaw Eye Hospital, this approach was abandoned, with a preference for a fixed-dose prophylactic protocol to minimize the risk of an additional SMH due to waiting for clinical symptoms of wet AMD before anti-VEGF therapy is administered. In addition, inclusion of vitrectomy allows the surgeon an opportunity to immediately address treatment-related adverse events that may arise, or other comorbid conditions within the eye (e.g., epiretinal membrane or vitreous hemorrhage).
This treatment strategy has been considered previously, and the results herein are consistent with those observations.4–9 However, the current procedure included several modifications compared with the general technique employed in other case series.4,17 The modifications in the current study included a low dose of rtPA injected into the subretinal space to further minimize any toxicity risk, a higher dosage of one specific anti-VEGF therapy (i.e., bevacizumab 2.5 mg) with ongoing monthly treatment in all patients, and a shortened supine positioning requirement after the procedure (i.e., 30–40 min), without further postsurgical posturing required. In addition to treatment differences, the population (e.g., SMH characteristics) evaluated in the current study differed from that in other published studies.4,5,9,17 For example, size was not a limiting factor in the current study, which included SMH that had spread over the vascular arcades, with seven patients having a spread of ⩾ 10 mm × 10 mm. As well, there was no limit on the duration of time between SMH expression and treatment, while other studies have restricted this interval (e.g., ⩽2 weeks).4,5,9,17 The current study is unique in this respect, as data are lacking on outcomes in patients with SMH with a delay in treatment. Ideally, treatment should be administered as early as possible. It is encouraging, however, that visual acuity can still be improved in patients with SMH who experience a delay in treatment with use of the protocol described herein.
This study also conducted analyses to help identify baseline factors that might impact treatment success, such as baseline visual acuity, SMH size, retinal thickness and retinal damage. Unsurprisingly, several parameters were found to impact visual acuity prior to intervention (i.e., at baseline), including thickness of the retina, together with underlying hemorrhage, measured on OCT. However, there was no correlation between size or location and visual acuity outcomes in the affected eye after treatment, supporting that neither of these factors should hinder implementation of the interventional protocol established in this study. However, another study of vitrectomy (mean follow-up, 12.9 ± 10.8 months), subretinal rtPA, and monthly anti-VEGF prophylaxis identified a relationship between subretinal hemorrhage area and visual acuity outcomes (p = 0.001). 15
Furthermore, in the current study, there was a positive correlation between baseline SMH area in the affected eye and baseline visual acuity in the fellow eye, indicating that patients with a larger SMH area will have poorer visual acuity in the fellow eye. Thus, given that patients with AMD and SMH often have a less favorable prognosis, with poorer visual acuity in the fellow eye, SMH treatment is crucial, particularly when patients cannot rely on the visual acuity in the unaffected eye. Results of the current study also highlight the importance of monitoring. Given that more than one-third (39.6%) of patients had previously been treated with anti-VEGF therapy in the affected eye, patients with a history of macular neovascularization should be closely monitored for recurrence, including those with apparent quiescent disease. In addition, patients with end-stage neovascular AMD should be monitored because SMH can worsen visual acuity.
Limitations of the study include the retrospective nature of the analysis, the relatively small sample size, the wide variation in time to treatment, the limited duration of follow-up and the absence of another control group with anti-VEGF treatment or pneumatic displacement. Strengths of the study were the diverse patient population, including those with SMH in eyes previously treated with anti-VEGF therapy, and the standardized treatment protocol implemented by a single expert surgeon. The standardized treatment protocol included post-surgery anti-VEGF therapy as an integrated part of the treatment. Given that no anti-VEGF therapies are currently approved for administration during surgery, bevacizumab was selected based on cost-effectiveness, availability, and investigator experience, and was the only anti-VEGF therapy administered in the current study. An additional strength was that the study analyzed both biomarkers seen on OCT examination (e.g., hemorrhage thickness and location) and the state of the retina on SMH presentation to assess their impact on BCVA after 6 months.
Submacular hemorrhage due to wet-AMD is a very rare disorder. It might seem that a sample size of 48 patients might be the limitation of this study due to the relatively small sample size; however, in a real-world scenario, this group of 48 patients treated with the same protocol is rather unique in the literature.
These study results provide further support and guidance on a treatment methodology for SMH, and the outcomes reinforce that subretinal injection, rather than intravitreal injection, of rtPA should be considered as part of a first-line strategy for SMH. In conclusion, vitrectomy, subretinal rtPA and bevacizumab injections are an effective combination for improving visual acuity, including for patients who may experience a delay between SMH expression and treatment.
This case report series was designed following the STROBE guidelines. 18
Supplemental Material
sj-docx-1-oed-10.1177_25158414261447184 – Supplemental material for Impact of treatment gap on visual acuity in patients with submacular hemorrhage secondary to neovascular age-related macular degeneration treated with quadruple intervention strategy
Supplemental material, sj-docx-1-oed-10.1177_25158414261447184 for Impact of treatment gap on visual acuity in patients with submacular hemorrhage secondary to neovascular age-related macular degeneration treated with quadruple intervention strategy by Agnieszka Joanna Nowosielska and Grzegorz Rotuski in Therapeutic Advances in Ophthalmology
Supplemental Material
sj-pdf-2-oed-10.1177_25158414261447184 – Supplemental material for Impact of treatment gap on visual acuity in patients with submacular hemorrhage secondary to neovascular age-related macular degeneration treated with quadruple intervention strategy
Supplemental material, sj-pdf-2-oed-10.1177_25158414261447184 for Impact of treatment gap on visual acuity in patients with submacular hemorrhage secondary to neovascular age-related macular degeneration treated with quadruple intervention strategy by Agnieszka Joanna Nowosielska and Grzegorz Rotuski in Therapeutic Advances in Ophthalmology
Footnotes
Acknowledgements
Technical editorial and medical writing assistance were provided under the direction of the author by Mary Beth Moncrief, PhD, Synchrony Medical Communications, LLC, West Chester, PA, USA.
Declarations
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
