Subretinal Tissue Plasminogen Activator Injection Using a Nanovitreoretinal Device for Massive Submacular Hemorrhage: A Retrospective Case Series

Article information

Abstract

Materials and Methods

This retrospective observational interventional study was conducted at a single tertiary referral center and included patients who presented with massive SMH and received subretinal tPA injections between 2020 and 2024. This study evaluated 18 eyes of 15 patients. This study protocol was approved by the Institutional Review Board of Gangnam Severance Hospital (No. 2025-0873-001), and the requirement for informed consent was waived due to the use of deidentified data and the retrospective nature of the study. The study adhered to the tenets of the Declaration of Helsinki.

Patients were eligible for inclusion if they had massive SMH involving the foveal center confirmed by optical coherence tomography (OCT), symptom duration of less than 4 weeks, and underwent subretinal tPA administration using the NVR subretinal gateway device with or without pars plana vitrectomy. Exclusion criteria included eyes with a history of prior vitreoretinal surgery in the study eye, concomitant retinal pathology that could affect visual outcomes (such as diabetic retinopathy, retinal vein occlusion, or severe geographic atrophy), and inadequate imaging or incomplete clinical follow-up of less than 3 months. In this study, massive SMH was defined as a fovea-involving hemorrhage measuring ≥1 disc area in size or >500 μm in subfoveal thickness on OCT, based on previous reports [2,11].

All patients underwent comprehensive ophthalmological examinations before and after the procedure, including best-corrected visual acuity (BCVA) measurement, intraocular pressure assessment, slit-lamp examination, ultrawide- field fundus photography (Optos plc, Dunfermline), and Spectralis-OCT imaging (Heidelberg Engineering GmbH). Vitrectomy was performed using the CONSTELLATION Vision System (Alcon Laboratories Inc.). The procedure was performed under local anesthesia, using a combination of topical and sub-Tenon anesthesia. After conjunctival disinfection using betadine solution, 25-gauge trocars were inserted to facilitate submacular tPA injections. Recombinant tPA (Actilyse, Boehringer Ingelheim) was used and prepared by diluting tPA to a concentration of 12.5 μg/0.1 m L. A n injection volume of 0.1 to 0.2 m L was administered into the subretinal space at the site of the hemorrhage using a NVR subretinal gateway device (NanoSubRet, Vortex Surgical Inc.). The NVR system consists of a 28-gauge outer cannula for trans-pars plana insertion and a retractable 41-gauge inner cannula for subretinal access. After 28-gauge trocar placement 3.5 to 4.0 mm posterior to the limbus, the NVR device was inserted through the trocar. Under direct visualization using BIOM (OCU-LUS Surgical Inc.), the outer cannula was advanced to the retinal surface at the hemorrhage margin. The inner 41-gauge cannula was then extended to penetrate the neurosensory retina creating a self-sealing retinotomy, and enter the subretinal space. tPA solution was injected slowly (over 30–60 seconds) using foot-pedal-controlled infusion connected to the vitrectomy machine’s viscous fluid control unit. After tPA delivery, the inner cannula was retracted and the device was withdrawn. Following subretinal tPA injection, 0.1 mL of sterile filtered air was injected subretinally via the previously created retinotomy site to facilitate pneumatic displacement of liquefied hemorrhage. At the end of the procedure, intravitreal bevacizumab (1.25 mg/0.05 mL) was administered in all eyes. Laser treatment was not performed at the retinotomy site. Consequently, patients were not instructed to maintain face-down positioning, but were allowed to remain in a sitting position approximately two hours after the procedure was completed.

For etiologic classification of massive SMH, each eye was assigned to one mutually exclusive clinical diagnostic category based on multimodal imaging findings, including fundus photography, OCT, fluorescein angiography, and indocyanine green angiography when available. The categories included typical neovascular AMD, polypoidal choroidal vasculopathy, retinal angiomatous proliferation, and peripheral exudative hemorrhagic chorioretinopathy. Anatomical neovascularization subtypes were not used as the primary etiologic classification in order to avoid overlap with clinical entities.

The decision to perform concomitant pars plana vitrectomy was based on predefined clinical criteria. Pars plana vitrectomy was performed in cases with (1) established poor fundus visualization due to pre-existing breakthrough vitreous hemorrhage; (2) a high risk of intraoperative or postoperative breakthrough vitreous hemorrhage that could compromise visualization, including massive subretinal hemorrhage with a central macular thickness (CMT) ≥700 μm and long-term systemic anticoagulant use; (3) coexisting vitreoretinal conditions requiring vitrectomy, such as intraocular lens dislocation; and (4) poor visualization unrelated to vitreous hemorrhage, including dense cataract.

The patients were monitored postoperatively at regular follow-up visits. During these visits, repeat assessments of BCVA and intraocular pressure, slit-lamp examination, ultrawide- field fundus photography, and OCT scans were performed to evaluate anatomical and functional outcomes. Anatomical outcomes were assessed based on changes in CMT and SMH thickness on OCT. CMT improvement was defined as a ≥20% reduction in CMT from baseline, based on previously reported OCT-based criteria for clinically meaningful anatomical response. Because previous SMH studies have generally defined favorable anatomical outcomes by complete foveal displacement of hemorrhage rather than by a standardized CMT reduction threshold, the 20% cutoff was adopted from previously reported OCT-based criteria for clinically meaningful anatomical response. CMT was evaluated at baseline, 1 month, and 6 months after treatment when available. The primary anatomical analysis was based on the change in CMT from baseline to 6 months. For the favorable anatomical outcome analysis, the ≥20% reduction in CMT was assessed at 6 months when available; if 6-month OCT data were unavailable, the 1-month OCT measurement was used. BCVA was assessed at baseline, 1 month, 6 months, and the final follow-up visit when available. Final follow-up was defined as the last available postoperative visit after subretinal tPA injection and was used for the assessment of final BCVA, recurrence, and safety outcomes. Missing values were not imputed and were excluded from the corresponding analyses. Primary outcome measures included changes in CMT and improvements in BCVA. Secondary outcomes included changes in SMH thickness and exploratory subgroup analyses stratified according to whether patients underwent combined air injection or concomitant pars plana vitrectomy in addition to subretinal tPA injection. Anti-VEGF therapy was generally continued at approximately monthly intervals during the early postoperative period, with subsequent injections individualized based on disease activity.

Results

The study included 18 eyes from 15 patients, with a mean age of 67.2 ± 11.4 years (range, 43.0–88.0 years) (Table 1). Seven eyes (38.9%) were from male patients and 11 eyes from female patients (61.1%). Neovascular AMD was the leading etiology of massive SMH, accounting for 17 of 18 eyes (94.4%): typical neovascular AMD was identified in nine eyes (50.0%), polypoidal choroidal vasculopathy in seven eyes (38.9%), retinal angiomatous proliferation in one eye (5.6%), and peripheral exudative hemorrhagic chorioretinopathy in the remaining eye (5.6%). Representative preoperative multimodal imaging findings are shown in Fig. 1A–1C. Mean baseline BCVA was 1.79 ± 0.77 logMAR (range, 0.22–2.60 logMAR), reflecting severe visual impairment at presentation. Baseline CMT was 673.8 ± 346.7 μm (range, 166.0–1,427.0 μm; CMT was unavailable at baseline in one eye with breakthrough vitreous hemorrhage that precluded OCT measurement). Mean SMH thickness was 536.9 ± 345.0 μm. The mean interval from symptom onset to tPA injection was 9.7 ± 8.4 days (range, 0.0–29.0 days). Baseline demographic and clinical characteristics are summarized in Table 1. Because of the small and imbalanced subgroup sizes, baseline characteristics were interpreted descriptively rather than as evidence of true comparability between groups.

Table 1

Baseline characteristics (n = 18)

Fig. 1

Retinal images before the tissue plasminogen activator injection without vitrectomy in a 58-year-old woman. (A) Optical coherence tomography image of the macula. (B) Fluorescein angiography (left) and indocyanine green angiography (right) images. (C) Ultra- widefield fundus photograph (Optos plc, Silverstone). Massive subretinal hemorrhage is observed within the vascular arcade involving the macula.

Statistically significant improvements were observed in both anatomical and functional outcomes following tPA injections. CMT showed a trend toward reduction at 1 month (381.1 ± 318.2 μm; range, 80.0–1,083.0 μm; p = 0.214 vs. baseline) and decreased significantly at 6 months (266.0 ± 327.1 μm; range, 47.0–844.0 μm; p = 0.019 vs. baseline) (Table 2, Figs. 2A–2C and 3A–3D). SMH thickness showed a substantial reduction from 536.9 ± 345.0 to 221.4 ± 443.4 μm (p = 0.020).

Table 2

Baseline and postoperative outcomes after subretinal tPA injection

Fig. 2

Serial ultra-widefield fundus and optical coherence tomography images after the tissue plasminogen activator (tPA) injection with vitrectomy, obtained from a 78-year-old man. (A) Before the tPA injection, massive submacular hemorrhage is observed. (B) Six months after the tPA injection, minimally remaining submacular hemorrhage with signs of intraretinal fluid and subretinal fluid was noted. (C) Three years after the tPA injection, no evidence of recurrence is noted.

Fig. 3

Serial ultra-widefield fundus and optical coherence tomography (OCT) images after the tissue plasminogen activator (tPA) injection without vitrectomy, obtained from a 58-year-old woman. (A) Three days after the tPA injection, submacular hemorrhage is seen as shifted slightly downward after the treatment. (B) Ten days after the tPA injection, vitreous hemorrhage is noted with some remaining signs of submacular hemorrhage. (C) Two weeks after the tPA injection, minimal subretinal fluid volume is observed on the OCT images. (D) Three months after the tPA injection, no evidence of recurrence is noted.

Statistically significant improvement in visual acuity was also observed following subretinal tPA injection. Mean BCVA improved from 1.79 ± 0.77 logMAR at baseline to 1.67 ± 0.79 logMAR at 1 month (p = 0.515), 1.28 ± 0.83 logMAR at 6 months (p = 0.010), and 1.18 ± 0.85 logMAR at final follow-up (p = 0.005) (Table 2). Of the 18 eyes, 10 eyes (55.6%) achieved a ≥3-line improvement in BCVA, one eye (5.6%) achieved ≥2- to <3-line improvement, and the remaining seven eyes (38.9%) remained stable or experienced visual deterioration. In total, 11 eyes (61.1%) demonstrated a ≥2-line improvement in BCVA.

With respect to the role of combined subretinal air injection, among eyes with evaluable CMT data, eyes that received air tamponade demonstrated a favorable anatomical outcome (CMT ≥20% reduction) compared with those that did not (90.9% [10/11] vs. 60.0% [3/5]; odds ratio [OR], 6.67; 95% confidence interval [CI], 0.44–101.70; p = 0.214), though this difference did not reach statistical significance. Similarly, in terms of visual outcomes, four of six eyes (66.7%) that did not receive subretinal air injection showed <2-line improvement in BCVA, compared with four of 12 eyes (33.3%) in the air injection group.

Comparison of outcomes between the vitrectomy and nonvitrectomy groups revealed no statistically significant differences in functional or anatomical parameters (Table 3). Final BCVA was 1.24 ± 0.77 logMAR in the vitrectomy group and 1.37 ± 1.01 logMAR in the nonvitrectomy group (p = 0.850), with mean BCVA improvement of 0.66 ± 0.76 and 0.20 ± 0.58 logMAR, respectively (p = 0.241). The proportion of eyes achieving ≥3-line BCVA improvement was comparable between groups (58.3% [7/12] vitrectomy vs. 50.0% [3/6] nonvitrectomy, p > 0.999). Anatomically, CMT at 6 months was 210.2 ± 253.2 μm in the vitrectomy group and 333.0 ± 420.8 μm in the nonvitrectomy group (p = 0.329), with favorable CMT outcomes in 81.8% (9/11) and 80.0% (4/5) of eyes, respectively (p > 0.999).

Table 3

Comparison of outcomes between vitrectomy and nonvitrectomy subretinal tPA injection groups

Univariate logistic regression analysis identified no statistically significant predictors of favorable anatomical outcome in this cohort (Table 4). Among the factors evaluated, subretinal air injection demonstrated a numerically higher OR (OR, 6.67; 95% CI, 0.44–101.70; p = 0.214), and shorter time to surgery showed a trend toward favorable outcome (OR, 1.21/day; p = 0.217). Multivariable analysis was not performed given the limited number of unfavorable events (n = 3), which precludes reliable multivariable estimation.

Table 4

Prognostic factors associated with favorable post-sub-retinal tPA injection outcomes

The mean follow-up duration was 23.9 ± 10.9 months (range, 1.5–37.0 months). During follow-up, two eyes (11.1%) developed breakthrough vitreous hemorrhage at 23 and 24 days postprocedure, respectively; both were successfully managed with pars plana vitrectomy. Three eyes (16.7%) experienced recurrent SMH—at 3 months, 3 years, and 4 years following the initial procedure—and all underwent repeat subretinal tPA injection. No serious procedure- related complications were observed, including retinal detachment, tPA-related retinal toxicity, endophthalmitis, RPE tear, or epiretinal membrane formation.

Discussion

In this study, we evaluated the outcomes of subretinal tPA injection in patients with massive SMH using a 41-gauge NVR device. The results of our study demonstrate that subretinal tPA injection with or without pars plana vitrectomy was associated with improvements in anatomical and functional outcomes in patients with massive SMH. CMT and SMH thickness were reduced, and BCVA improved after treatment. No procedure-related complications were observed across the entire follow-up period, supporting the safety profile of the procedure.

The etiology of SMH primarily involves neovascular AMD, which was the leading cause of SMH in our study, affecting 94.4% of the cases. SMH severely impairs visual outcomes due to its pathophysiology. An SMH occurs when fragile choroidal neovascular membranes rupture, leading to extensive bleeding beneath the macula. This hemorrhage exerts mechanical pressure on the photoreceptors and RPE, causing immediate and severe vision loss. Additionally, the SMH impairs nutrient delivery and waste removal between the retina and RPE [12]. Traction from fibrin membranes formed within clots may also damage the photoreceptors. Finally, the accumulated blood can lead to iron toxicity and further damage to the retinal structures if not promptly addressed [11,16]. Previous studies have highlighted the poor visual prognosis associated with SMH regardless of the treatment modality used [13]. This highlights the need for effective treatment strategies in managing these challenging conditions.

The mechanism of action of tPA in the treatment of SMH involves its fibrinolytic properties [12–14,17]. tPA catalyzes the conversion of plasminogen to plasmin, an enzyme that dissolves fibrin clots. By breaking down the fibrin network within the hemorrhage, tPA facilitates the displacement and resorption of blood, thereby relieving the toxic effects of the hemorrhage on the macula and allowing for potential visual recovery (Fig. 3). This enzymatic action is particularly beneficial for the management of massive hemorrhages that are otherwise resistant to spontaneous resolution. When combined with a pneumatic agent, either subretinal or intravitreal air, this liquefied material is further propelled inferiorly by buoyancy forces, away from the foveal center, relieving photoreceptor compression and restoring the metabolic exchange between the retina and RPE. The rapid clinical response observed in the present study, with SMH thickness declining by approximately 59% and BCVA improving significantly within the early postoperative period, is consistent with this mechanism and underscores the importance of prompt intervention before irreversible photoreceptor damage occurs [7,11].

A notable finding of the present study is the temporal dissociation between early hemorrhage resolution and delayed total macular thickness recovery. SMH thickness and BCVA both improved significantly in the early postoperative period (p=0.020 and p=0.010, respectively), reflecting prompt fibrinolysis and displacement of the hemorrhagic mass. In contrast, total CMT did not reach statistical significance at 1 month postinjection (381.1 ± 318.2 μm, p = 0.214 vs. baseline) yet improved significantly by 6 months (266.0 ± 327.1 μm, p = 0.019). This pattern is mechanistically coherent: although tPA rapidly lyses the fibrin clot and air displaces the liquefied hemorrhage, residual subretinal fluid, reactive intraretinal edema, and RPE perturbation are expected to persist in the early postoperative period, elevating total CMT independent of the primary hemorrhage. These secondary structural changes resolve progressively over subsequent months as the RPE recovers its pump function and the foveal architecture normalizes. Importantly, the nonsignificant 1-month CMT result should therefore not be interpreted as an absence of early treatment effect—the significant reductions in SMH thickness and BCVA at this timepoint confirm that tPA provides immediate and clinically meaningful benefit. Rather, these findings suggest that CMT alone is an insufficient metric for early outcome assessment after subretinal tPA injection, and that a minimum follow-up of 6 months is advisable for comprehensive anatomical evaluation.

Despite its efficacy, subretinal tPA administration carries potential risks, although clinically significant complications were not observed in the present study. For example, retinal toxicity is a concern because tPA can induce apoptosis in retinal cells if administered inappropriately [18]. Irvine et al. [19] and Johnson et al. [20] demonstrated the occurrence of retinal toxicity owing to tPA in a rabbit model. Yamamoto et al. [21] reported a case of retinal toxicity after an intravitreal tPA injection. However, the concentration and volume used in our study (12.5 μg/0.1 mL) were within the safe range, as evidenced by the absence of tPA-related complications over a mean follow-up of 23.9 ± 10.9 months.

Several features of the NVR system provide advantages over the traditional vitrectomy-assisted approach. The beveled 41-gauge cannula creates a self-sealing retinotomy, thereby minimizing complications such as retinal detachment or hemorrhage and reducing the need for intraocular tamponade. Coupling the injection to the viscous fluid control unit of a standard vitrectomy machine allows gentle and foot-pedal-controlled delivery of tPA, eliminating the need for an assistant and reducing the risk of tremor-related complications. Because the injection can be performed through one or two small ports under local anesthesia, the procedure is minimally invasive, efficient, and can often be completed within 5 to 10 minutes, thereby lowering surgical time, cost, and overall burden for both patients and physicians [15,22]. In our study, postoperative BCVA and SMH thickness did not differ significantly between the nonvitrectomy and vitrectomy groups. Considering these practical and clinical advantages, it is worth considering performing subretinal tPA injection using the NVR device without vitrectomy for patients with massive SMH in a carefully selected group of patients.

One of the central contributions of the present study is its systematic comparison of outcomes between vitrectomy and nonvitrectomy subretinal tPA injection, despite the small number of eyes in the study. No statistically significant differences were identified between the two groups across any outcome measure, including final BCVA (1.24 ± 0.77 logMAR vs. 1.37 ± 1.01 logMAR, p = 0.850), mean BCVA improvement from baseline (0.66 ± 0.76 logMAR vs. 0.20 ± 0.58 logMAR, p = 0.241), ≥3-line improvement rate (58.3% vs. 50.0%, p > 0.999), 6-month CMT (210.2 ± 253.2 μm vs. 333.0 ± 420.8 μm, p = 0.329), and CMT ≥20% reduction rate (81.8% vs. 80.0%, p > 0.999) (Table 3). These findings may be clinically relevant, as vitrectomy carries well-recognized disadvantages in this context: increased surgical complexity and anesthetic risk, accelerated intraocular pharmacokinetic clearance of subsequently administered anti-VEGF agents, potential induction of nuclear sclerotic cataract in phakic eyes, and greater operative time and associated costs. The NVR device allows subretinal access through a single minimally invasive trans-pars plana port under local anesthesia, typically within 5 to 10 minutes, substantially reducing procedural burden for both patient and surgeon [15,22]. Given these practical advantages and the favorable outcomes demonstrated here, nonvitrectomy subretinal tPA injection via the NVR device may represent a feasible option in eyes where the vitreous is clear and surgical access is not otherwise indicated.

A secondary finding of interest is the observed numerical difference in anatomical outcomes in eyes that received combined subretinal air injection. Among the prognostic factors evaluated in univariate analysis, subretinal air injection showed a higher OR, with 90.9% of eyes receiving combined air tamponade achieving ≥20% CMT reduction compared with 60.0% of those treated with tPA alone (OR, 6.67; 95% CI, 0.44–101.70; p = 0.214) (Table 4), although the CI was wide. While this difference did not reach statistical significance, attributable to the limited sample size, the magnitude of the observed effect is clinically relevant and mechanistically plausible. Subretinal air exerts a direct pneumatic displacement force on the tPA-liquefied hemorrhagic material, actively propelling it inferiorly away from the fovea, and may additionally serve as a physical scaffold that maintains displacement during the early postoperative period until the liquefied blood is reabsorbed. This potential adjunctive role of subretinal air warrants prospective evaluation in larger cohorts, and the combined tPA and air injection may be considered where feasible.

Throughout the follow-up period, two cases of vitreous hemorrhage occurred in 18 eyes. However, these vitreous hemorrhages might not be injection-related complications but rather breakthrough vitreous hemorrhages, which are part of the natural course of massive SMH. This is particularly noteworthy, given the traditionally poor prognosis associated with SMH and the potential risks associated with invasive procedures.

This study has several limitations. First, this was a single- center, retrospective interventional case series without a control group, which may limit the generalizability of the findings. Second, the sample size was relatively small, with only 18 eyes included in the analysis. In particular, subgroup comparisons between the vitrectomy and nonvitrectomy groups were based on small and imbalanced sample sizes (6 eyes vs. 12 eyes), which may have resulted in insufficient statistical power to detect between-group differences. Therefore, the absence of statistically significant differences between the two groups should be interpreted with caution and does not necessarily indicate equivalence between the treatment approaches. In addition, because vitrectomy was preferentially performed in eyes with poor visualization, greater hemorrhagic burden, or coexisting vitreoretinal conditions, selection bias may have influenced the comparison between the vitrectomy and nonvitrectomy groups. This baseline imbalance, including differences in the interval from presentation to surgery and baseline CMT, should be considered when interpreting the comparative outcomes. Third, multivariable regression analysis was not performed due to the limited sample size, as inclusion of multiple covariates such as baseline visual acuity, hemorrhage size, SMH duration, etiology, and age would have carried a substantial risk of overfitting and unreliable estimates. Accordingly, the statistical analyses were limited to univariable comparisons to describe overall anatomical and functional outcomes. Finally, although long-term follow-up was available for most patients, the retrospective design and heterogeneity of underlying etiologies may have influenced visual and anatomical outcomes. Larger prospective studies are warranted to further validate these findings.

In summary, our study demonstrates that subretinal tPA injection using a 41-gauge NVR device, with or without vitrectomy, is associated with anatomical and functional improvements in patients with massive SMH. These findings suggest that subretinal tPA injection via the NVR device may represent a minimally invasive treatment option for the management of massive SMH. Further research and larger studies are warranted to confirm these findings and refine treatment protocols for massive SMH.

Article information Continued