Management of submacular hemorrhage with intravitreous tissue plasminogen activator injection and pneumatic displacement

Management of Submacular Hemorrhage with Intravitreous Tissue Plasminogen Activator Injection and Pneumatic Displacement Adam S. Hassan, MD,1 Mark W. Johnson, MD,1 Todd E. Schneiderman, MD,2 Carl D. Regillo, MD,3 Paul E. Tornambe, MD,4 Lon S. Poliner, MD,4 Barbara A. Blodi, MD,5 Susan G. Elner, MD1 Objective: To investigate the efficacy and safety of treating thick submacular hemorrhages with intravitreous tissue plasminogen activator (tPA) and pneumatic displacement. Design: Retrospective, noncomparative case series. Participants: From 5 participating centers, 15 eligible patients had acute (⬍3 weeks) thick subretinal hemorrhage involving the center of the macula in eyes with pre-existing good visual acuity. Hemorrhages were secondary to age-related macular degeneration in 13 eyes and macroaneurysm and trauma in 1 eye each. Methods: The authors reviewed the medical records of 15 consecutive patients who received intravitreous injection of commercial tPA solution (25–100 ␮g in 0.1– 0.2 ml) and expansile gas (0.3– 0.4 ml of perfluoropropane or sulfur hexafluoride) for thrombolysis and displacement of submacular hemorrhage. After surgery, patients maintained prone positioning for 1 to 5 days (typically, 24 hours). Main Outcome Measures: Degree of blood displacement from under the fovea, best postoperative visual acuity, final postoperative visual acuity, and surgical complications. Results: In 15 (100%) of 15 eyes, the procedure resulted in complete displacement of thick submacular hemorrhage out of the foveal area. Best postprocedure visual acuity improved by 2 lines or greater in 14 (93%) of 15 eyes. After a mean follow-up of 10.5 months (range, 4 –19 months), final visual acuity improved by 2 lines or greater in 10 (67%) of 15 eyes and measured 20/80 or better in 6 (40%) of 15 eyes. Complications included breakthrough vitreous hemorrhage in three eyes and endophthalmitis in one eye. Four eyes developed recurrent hemorrhage 1 to 3 months after treatment, three of which were retreated with the same procedure. Conclusions: Intravitreous injection of tPA and gas followed by brief prone positioning is effective in displacing thick submacular blood and facilitating visual improvement in most patients. The rate of serious complications appears low. Final visual outcomes are limited by progression of the underlying macular disease in many patients. Ophthalmology 1999;106:1900 –1907 Although there is significant variability in the clinical course of submacular hemorrhage, natural history studies suggest that factors predictive of poor visual outcomes include thick blood under the fovea and the presence of Originally received: November 10, 1998. Revision accepted: June 16, 1999. Manuscript no. 98556. 1 W. K. Kellogg Eye Center, Department of Ophthalmology, University of Michigan School of Medicine, Ann Arbor, Michigan. 2 Pacific Eye Care, Poulsbo, Washington. 3 Retina Service, Wills Eye Hospital, Philadelphia, Pennsylvania. 4 Retina Consultants, San Diego, California. 5 Department of Ophthalmology, University of Wisconsin, Madison, Wisconsin. Presented in part at the Association for Research in Vision and Ophthalmology annual meeting, Ft. Lauderdale, Florida, May 1998, and at the American Academy of Ophthalmology annual meeting, New Orleans, Louisiana, November 1998. Address correspondence to Mark W. Johnson, MD, W. K. Kellogg Eye Center, 1000 Wall Street, Ann Arbor, MI 48105.

1900

age-related macular degeneration (ARMD).1– 4 These retrospective reports show that the visual outcome in patients with ARMD with thick blood under the fovea is rarely better than 20/200. Such clinical observations, in combination with experimental evidence for retinal damage by subretinal blood,5–7 stimulated interest in experimental surgical approaches to remove thick blood from under the fovea.7–12 Pilot studies in human eyes have shown that surgical removal of the hematoma, with or without tissue plasminogen activator (tPA) thrombolysis, may improve the visual outcome in selected patients.13–19 Unfortunately, the overall surgical results have been disappointing in patients with ARMD, with only 16% of reported patients achieving a visual outcome of 20/80 or better. Heriot recently described the management of submacular hemorrhage with intravitreous tPA injection and pneumatic displacement of blood from under the fovea (Heriot WJ, presented at the AAO Annual Vitreoretinal Update, San Francisco, October 1997). His initial experience suggested a high anatomic success rate (blood displacement in 19 of 20

Hassan et al 䡠 Intravitreous tPA/Pneumatic Displacement of Submacular Hemorrhage eyes) with few complications. The purpose of this study was to further investigate the efficacy and safety of treating thick submacular hemorrhage with this minimally invasive, office-based procedure.

Patients and Methods We reviewed the medical records of 15 consecutive patients who had undergone intravitreous injection of commercial tPA solution and expansile gas for thrombolysis and displacement of submacular hemorrhage. Patients presenting at several centers to one of the authors (M.W.J., T.E.S., C.D.R., P.E.T., L.S.P., B.A.B., or S.G.E.) were accumulated retrospectively and accrued between January 1, 1997, and October 15, 1997. All patients had acute hemorrhages (⬍3 weeks old), thick blood under the fovea, and reading vision in the involved eye immediately before the hemorrhage. To be classified as thick, the blood under the fovea had to be sufficient to cause an obvious elevation of the retina from the retinal pigment epithelium on biomicroscopic examination. Patients in whom thick blood was apparent beneath the retinal pigment epithelium in the foveal area were not considered candidates for the procedure. No patient who underwent the procedure at one of our centers during the indicated period was excluded from this retrospective analysis. Each patient underwent complete ophthalmologic examination. Visual acuity was obtained using Snellen charts with manifest refraction or the patient’s spectacle correction and pinhole. Visual acuities less than 20/400 on the Snellen chart were categorized as 5/200, 2/200, or hand motions. Standardized refraction and visual acuity testing protocols were not used. Informed consent was obtained before performing the procedure. The procedure was performed in an outpatient setting with the patient under topical anesthesia. After the bulbar conjunctiva was prepared with 5% Betadine solution (Purdue Frederick, Norwalk, CT), 25 to 100 ␮g of commercial tPA solution (Activase; Genentech, Inc, San Francisco, CA) in a volume of 0.1 to 0.2 ml was injected into the midvitreous cavity through a 30-gauge needle introduced 3 to 4 mm posterior to the limbus superotemporally. Sterile balanced salt solution (BSS; Alcon Laboratories, Ft. Worth, TX) was used to dilute the tPA solution where applicable. After an aqueous tap to reduce intraocular pressure, 0.3 to 0.4 ml of perfluoropropane or sulfur hexafluoride gas was injected into the vitreous cavity in a similar fashion. The patient was instructed to begin prone positioning within 3 hours of the procedure and to maintain positioning for 24 hours or as directed by the treating ophthalmologist. The follow-up schedule was nonstandardized. Patients typically were examined 1 day and 1 week after surgery, as well as at the conclusion of prone positioning if this was maintained for longer than 24 hours. Where choroidal neovascularization was suspected, most patients underwent fluorescein or indocyanine green angiography or both within the first 2 weeks after surgery. Thereafter, follow-up was determined by the treating ophthalmologist depending on the clinical course. Inclusion in the study was not dependent on a minimum follow-up period. For several patients, final follow-up data were obtained through telephone calls to local ophthalmologists. Primary outcome measures were (1) degree of blood displacement from under the fovea, (2) best postoperative visual acuity, and (3) final postoperative visual acuity. The degree of blood displacement was measured ophthalmoscopically by the treating physician and was graded as complete, partial, or no displacement. Complete displacement was defined as no blood or only a thin layer of blood within 1 disc diameter of the foveal center after

completion of prone positioning. Secondary outcome measures included operative complications and anatomic status of the macula at final follow-up. Two-sample Student’s t tests were used to compare patients with postoperative Snellen visual acuities (both best and final) of 20/80 or better to those with acuities less than 20/80. Variables examined included duration of submacular hematoma before treatment, diameter of hematoma, and dose of tPA administered.

Results Fifteen eyes of 15 patients (12 women) were included in the study. Patient age ranged from 13 to 91 years (median, 78 years). The mean duration of submacular hemorrhage was 5.7 days (range, 1–21 days), and the maximum diameter of the hematoma ranged from 2 to 13 disc diameters. The cause of the hemorrhage was choroidal neovascularization complicating ARMD in 13 eyes and macroaneurysm and trauma in 1 eye each (Table 1). The procedure resulted in complete displacement of thick blood from under the fovea in 15 of 15 eyes. In 14 of 15 eyes, submacular blood displacement to extramacular locations was identified by biomicroscopy or indirect ophthalmoscopy or both at the conclusion of prone positioning. Typically, the blood had shifted to subretinal areas temporal or inferotemporal to the macula. In one case (patient 11), the determination of complete displacement was made echographically, because vitreous hemorrhage precluded a view of the macular region. The duration of postoperative prone positioning typically was 24 hours (Table 1). In five patients, blood displacement out of the fovea was partial after 24 hours and complete after several additional days of prone positioning. Compared with initial visual acuity, the best postoperative visual acuity improved by 2 Snellen lines or greater in 93% (14 of 15) of eyes. The average time to best postoperative acuity was 4.6 months. Over extended follow-up, the acuity in five eyes declined by two Snellen lines or greater from the best postoperative level because of neovascular or atrophic progression of ARMD. Compared to initial acuity, the final postoperative visual acuity improved by 2 Snellen lines or greater in 67% (10 of 15) and worsened by 2 lines or more in 7% (1 of 15) of eyes (Fig 1). The final visual acuity was 20/80 or better in 40% (6 of 15) of eyes, with a mean follow-up of 10.5 months (range, 4 –19 months). A comparison of patients with a best postoperative visual acuity of 20/80 or better to those with visual acuity less than 20/80 revealed no statistically significant differences in hematoma duration, hematoma diameter, and dose of tPA injected. Similarly, no statistically significant associations were found between these variables and a final postoperative visual acuity of 20/80 or better. After blood displacement, fluorescein or indocyanine green angiography revealed a subfoveal choroidal neovascular membrane (CNVM), typically a fibrovascular pigment epithelial detachment, in 8 of 15 eyes. The remaining studies revealed juxtafoveal CNVM in one eye, no detectable CNVM in three eyes, and a macroaneurysm in one eye. In the remaining two eyes (including the eye with traumatic submacular hemorrhage), no angiography was performed. Laser treatment for choroidal neovascularization was performed in two eyes. Four eyes developed recurrent submacular hemorrhage 1 to 3 months after the procedure. Three of these four eyes received repeat treatment with intravitreous tPA injection and pneumatic displacement (Fig 2). On final follow-up, ophthalmoscopy revealed atrophic macular changes in five eyes, subfoveal fibrovascular pigment epithelial detachment in four eyes, fibrous disciform scarring in four eyes, and an extrafoveal choroidal rupture in one eye (Fig 3). In one eye, the macular status was unknown because of vitreous hemorrhage.

1901

Ophthalmology Volume 106, Number 10, October 1999 Table 1. Patient Data Patient No./ Sex/Age (yrs)/Eye 1/F/83/OD 2/M/13/OS 3/F/84/OS 4/F/77/OS 5/M/74/OS 6/F/70/OD 7/F/72/OD 8/F/81/OS 9/F/73/OD 10/F/70/OS 11/F/78/OD 12/F/88/OD 13/F/83/OD 14/M/84/OD 15/F/91/OS

Hemorrhage Diagnosis

Duration (days)

Diameter (DD)

Dose tPA ␮g/ml

ARMD Trauma ARMD ARMD ARMD ARMD ARMD MA/ARMD ARMD ARMD ARMD ARMD ARMD ARMD ARMD

4 5 2 21 7 2 1 4 6 2 4 6 14 1 5

3 4 2 4 3 3 3 5 4 3 13 2 2 4 4

75/.15 75/.15 75/.15 75/.15 50/.10 75/.15 40/.20 25/.10 100/.10 100/.10 75/.10 100/.10 100/.10 100/.10 100/.10

Visual Acuity

Gas

Duration of Positioning (days)

Initial

Best

Final

Follow-up (mos)

C3F8 SF6 SF6 SF6 SF6 SF6 SF6 C3F8 C3F8 SF6 SF6 C3F8 C3F8 C3F8 C3F8

1 3 1 1 1 3 1 3 5 1 3 1 1 1 1

20/300 20/200 20/400 2/200 20/200 2/200 5/200 2/200 2/200 5/200 HM 2/200 2/200 20/200 20/200

20/30 20/25 20/40 20/80 20/25 20/300 20/200 20/200 20/60 20/200 HM 20/70 20/200 20/60 20/30

20/30 20/30 20/40 20/400 20/25 20/300 20/300 20/200 20/80 20/200 HM 2/200 5/200 20/400 20/30

19 9 17 5 18 11 10 8 12 7 12 6 8 4 11

Complications

VH VH VH Endophthalmitis

Final Macular Status FVPED Choroidal rupture FVPED Disciform scar FVPED Disciform scar Atrophic changes Atrophic changes Atrophic changes Disciform scar Unknown (VH) Disciform scar Atrophic changes FVPED Atrophic changes

OD ⫽ right eye; OS ⫽ left eye; ARMD ⫽ age-related macular degeneration; MA ⫽ macroaneurysm; tPA ⫽ tissue plasminogen activator; HM ⫽ hand motion; VH ⫽ vitreous hemorrhage; FVPED ⫽ fibrovascular pigment epithelial detachment; DD ⫽ disc diameter.

Complications attributed to the procedure included breakthrough vitreous hemorrhage in three eyes, one of which was mild and resolved over 2 weeks. One of the two dense vitreous hemorrhages occurred in the eye with a retinal arterial macroaneurysm (patient 8) and was treated with pars plana vitrectomy. The other dense nonclearing vitreous hemorrhage occurred in the eye with a massive subretinal hemorrhage (patient 11), and the patient declined further intervention. The only other complication observed in this series was coagulase-negative staphylococcal endophthalmitis in one eye that responded to intravitreous antibiotic injection and was judged not to have affected the final visual outcome (limited because of macular atrophy and advanced glaucoma). We observed no pigmentary changes outside areas of previous subretinal hemorrhage to suggest retinal toxic reactions to the tPA solution in any eye. Although intraocular pressure elevations inevitably occurred coincident with the tPA and gas injections, no patient in our series had intraocular pressure elevations during subsequent follow-up that prompted the use of ocularhypotensive medication. Finally, no new crystalline lens opacities were noted in the medical record of any patient during follow-up.

patients, with only a small proportion (16%) achieving postoperative visual acuities of 20/80 or better.13–19 The safety and efficacy of this procedure are undergoing further study in one of the Submacular Surgery Trials, a set of multicenter, randomized, clinical trials sponsored by the National Eye Institute. Our retrospective study confirms the experience of Heriot and demonstrates that thick blood under the neurosensory retina can routinely be displaced away from the center of the macula with a minimally invasive procedure consisting of intravitreous injection of tPA and gas followed by a relatively brief period of prone positioning. Buhl and coworkers recently reported similar results with the same procedure (Invest Ophthalmol Vis Sci 1998;39 [Suppl]: 390). Based on this combined experience, it appears that in many or most patients, the goal of removing fresh thick

Discussion Thick submacular hemorrhage, particularly in patients with ARMD, is generally associated with a poor visual outcome.1–3 Several mechanisms, including shearing of photoreceptors by fibrin clots, physical separation of photoreceptors from the retinal pigment epithelium, and toxic effects of iron, have been suggested as explanations for the retinal damage caused by thick subretinal blood.5–7,20,21 These natural history and experimental data have prompted the search for a safe and effective method of removing thick blood from under the macula to speed visual recovery and prevent irreversible blood-induced damage to the outer retina. To date, the most commonly used approach has been that of pars plana vitrectomy and aspiration of submacular blood, with or without intraoperative clot dissolution by subretinal tPA. Published results of pilot studies using this procedure suggest improved visual outcome in selected

1902

Figure 1. Final postoperative visual acuity plotted against initial preoperative acuity in 15 eyes. Eyes experiencing visual improvement fall above the oblique line. HM ⫽ hand motion.

Hassan et al 䡠 Intravitreous tPA/Pneumatic Displacement of Submacular Hemorrhage

Figure 2. (Patient 5) A 74-year-old man with submacular hemorrhage of 1-week duration. A, preoperative appearance of left macula. Visual acuity ⫽ 20/200. B, 7 days after treatment with tissue plasminogen activator (tPA) and pneumatic displacement, there is complete displacement of blood out of the macula. C, recurrent submacular hemorrhage 1 month after treatment. D, 2 weeks after repeat treatment with tPA and pneumatic displacement, the blood is again displaced. Final visual acuity improved to 20/25.

blood from under the macula can be accomplished simply and without the higher costs associated with a more invasive vitrectomy procedure. Because most of the patients in our series had hemorrhages that were moderately large (2–5 disc

diameters in longest dimension), it is unknown whether this procedure would be effective for massive hemorrhages. The only patient in our series with a massive subretinal hemorrhage developed a dense vitreous hemorrhage after surgery,

Figure 3. (Patient 2) A 13-year-old boy with thick submacular hemorrhage after blunt trauma. A, preoperative appearance of left fundus. Visual acuity ⫽ 20/200. B, 2 months after treatment with tissue plasminogen activator and pneumatic displacement, the visual acuity is 20/30. Displaced and degenerating blood is seen outside the macula along with an extrafoveal choroidal rupture.

1903

Ophthalmology Volume 106, Number 10, October 1999 but ultrasound examination suggested that thick blood had been displaced out of the macula. The applicability of this procedure to eyes with chronic hemorrhages (⬎1 month old) and to those with a significant subretinal pigment epithelial component is unknown. Most patients in our series (93%) experienced visual improvement coincident with or soon after displacement of blood out of the macula. Some patients, particularly those with limited vision in the fellow eye, reported that this prompt recovery of central vision significantly enhanced their ability to perform tasks of daily living. In four patients, this benefit was only temporary, because subsequent neovascular or atrophic progression led to a decline in visual acuity back to within two lines of pretreatment acuity. By final follow-up, 67% of eyes had a visual acuity that was improved by two or more Snellen lines over pretreatment acuity. Although several authors have suggested that hemorrhage size and duration may serve as predictors of visual outcome after surgical evacuation,13,17 we could identify no preoperative or intraoperative variables of prognostic importance. The small sample size and relative homogeneity of our series (with respect to hematoma size and duration) gave us limited power to detect differences in outcome based on preoperative variables. In the absence of a control group, we cannot conclude that this procedure reduces permanent blood-induced macular damage and improves final visual outcome over observation alone. However, comparison of our data with published retrospective natural history studies suggests the possibility of benefit. Of the patients in our series with ARMD, 38% (5 of 13) achieved a final visual acuity of 20/80 or better. In contrast, natural history studies that report raw patient data show that among patients with ARMD with submacular hemorrhages similar in size to those in our series (thick and large, ⬎2 disc diameters) only 13% (2 of 15 eyes) had final acuities of 20/80 or better.1,2 Review of published pilot studies on surgical evacuation of submacular hemorrhage complicating ARMD reveals that only 16% (22 of 138) of reported patients achieved a final visual outcome of 20/80 or better,13–19 suggesting that pneumatic displacement may result in similar or possibly even superior visual outcomes compared with those of submacular surgery. However, these comparisons with other published reports probably are weak, given the small number of cases involved. Only a randomized, prospective, clinical trial can conclusively compare the efficacy and safety of different management approaches to this difficult problem. For most patients with submacular hemorrhage, the final visual outcome appears to depend in large part on the natural history of the underlying macular pathology. We speculate that previously asymptomatic patients who lose vision abruptly because of thick submacular hemorrhage commonly harbor otherwise inactive (nonleaking) fibrovascular pigment epithelial detachments. Our data show that after blood displacement, some of these lesions remain relatively quiescent, permitting visual recovery and stability over prolonged follow-up. In addition to simplicity and lower cost, greater safety is an anticipated advantage of pneumatic displacement over surgical evacuation of submacular hemorrhage. Although

1904

both procedures carry the risks of cataract formation, retinal breaks and detachment, macular pucker, proliferative vitreoretinopathy, endophthalmitis, and submacular and vitreous hemorrhage, these complications might be expected to occur more frequently with vitrectomy procedures. However, only a randomized, clinical trial could confirm greater safety of pneumatic displacement compared to surgical evacuation of submacular hemorrhage. Apart from a single case of coagulase-negative staphylococcal endophthalmitis, the only complication noted in our series was vitreous hemorrhage (three eyes). In one case, this was mild and assumed to represent transretinal migration of blood. Dense vitreous hemorrhage occurred in one eye with massive subretinal hemorrhage and in one eye later found to harbor a retinal arterial macroaneurysm. Although new bleeding consequent to tPA injection could not be excluded in these cases, this was thought to be unlikely since no new subretinal blood was seen in either case. Based on our data and other anecdotal experience (Robert L. Avery, MD, personal communication), we urge caution in the use of intravitreous tPA in patients with arterial macroaneurysm or massive subretinal hemorrhage because of a possible increased risk of vitreous hemorrhage. We did not observe new subretinal hemorrhage attributable to intravitreous tPA injection in any of our patients. Nevertheless, injecting tPA into eyes with recent (⬍72 hours) experimental vitreous hemorrhage from sectioned retinal blood vessels has been shown to cause recurrent bleeding.22 Based on this theoretical concern, it may be advisable to avoid intravitreous tPA injections within 3 days of an acute submacular bleed. The ideal time period between injecting intravitreous tPA and commencing prone positioning is unknown. Albumin, a protein with a molecular weight similar to that of tPA, appears in the subretinal space within 1 hour after intravitreous injection in rabbit eyes.23 We believe a delay of 2 to 3 hours after tPA injection should be sufficient to allow for its migration across the retina. We saw no evidence of retinal or other intraocular toxic reactions to tPA solution in the doses used in this study (25–100 ␮g in 0.1 ml). Specifically, we observed no retinal pigmentary changes outside areas of previous subretinal hemorrhage, no unexplained visual acuity or visual field loss, and no association between tPA dosing and visual acuity outcomes. Nevertheless, the potential for retinal toxicity from the commercially available tPA solution remains a serious concern. Johnson and coworkers24 reported dosedependent retinal toxicity in rabbit eyes with intravitreous injections of 50 ␮g/0.1 ml or greater, including severe retinal necrosis at higher concentrations. The toxic reaction was attributed to the arginine-based vehicle of the commercial tPA solution. Higher concentrations subsequently have been injected into human eyes, including most eyes in this series, on the assumption that the larger vitreous volume, greater vitreous liquefaction, and vascularized retina might raise the threshold for toxicity. More recently, however, we found retinal toxic reactions from intravitreous tPA injections of 50 ␮g/0.1 ml or greater in the cat, an animal model with vitreous volume and retinal vascularity similar to human eyes (Hrach CJ, et al. Invest Ophthalmol Vis Sci

Hassan et al 䡠 Intravitreous tPA/Pneumatic Displacement of Submacular Hemorrhage 1998;39 [Suppl]:883). Additionally, other investigators have presented observations suggesting severe toxicity in a human eye after intravitreous tPA injection of 100 ␮g/0.1 ml (Howard Gilbert, MD, unpublished data, presented at Vitreous Society annual meeting, New Orleans, Louisiana, September 1997) and pigmentary changes after injected tPA concentrations as low as 33 ␮g/0.1 ml (Ruiz-Lapuente CJ, et al. Invest Ophthalmol Vis Sci 1998;39 [Suppl]:988). Factors affecting the sensitivity of a given eye to toxic reactions by commercial tPA solution are unknown but may include vitreous volume, extent of vitreous liquefaction, position of needle tip relative to the retina, and degree of fundus pigmentation.25 Because recent evidence suggests that formed vitreous appears to restrict free circulation of tPA solution (Hrach CJ, et al. Invest Ophthalmol Vis Sci 1998;39 [Suppl]:883; Egana BC, et al. Invest Ophthalmol Vis Sci 1997;38 [Suppl]:212), we now believe it is unsafe to assume complete dilution of injected tPA solution in the vitreous cavity. Based on the animal studies and preliminary human experience cited above, we currently recommend avoiding intravitreous injections of tPA in concentrations greater than 25 ␮g/0.1 ml. In our small series, the efficacy of intravitreous tPA and gas in displacing blood from under the macula did not vary with tPA dose over the range used (25–100 ␮g). This fact, coupled with the recent observation that intravitreous gas without tPA effectively displaces submacular hemorrhage in some cases,26 raises a question about the role for intravitreous tPA in this procedure. Several lines of evidence suggest that intravitreous tPA is capable of traversing the retina in at least some species. The molecular weight of tPA (70 kD) is similar to that of albumin (68 kD), a protein that has been shown to diffuse across intact rabbit retina from the vitreous cavity.23 Coll et al10 recently demonstrated that intravitreously injected tPA (50 ␮g) routinely crosses the rabbit retina and induces complete lysis of 1-day-old subretinal blood clots. In a pig model, Boone and coworkers9 found partial lysis of large subretinal clots treated after 24 hours with intravitreous tPA (25 ␮g). Finally, Kimura and colleagues27 reported complete liquefaction of acute (2– 4day-old) subretinal hemorrhages in six patients treated with intravitreous tPA (6 ␮g) 12 to 36 hours before surgical evacuation of the blood. Only a prospective, randomized, clinical trial will establish which, if any, eyes require intravitreous tPA adjunctive therapy to facilitate complete pneumatic displacement of submacular blood. However, we believe it is unlikely that intravitreous gas without tPA will produce complete displacement in all eyes, particularly those with freshly clotted blood. We currently are pursuing a treatment strategy in which tPA is injected only when 24 hours of face-down positioning with intravitreous gas alone fails to produce adequate displacement of the submacular clot. Weaknesses of our study include its retrospective design, lack of control group, variable follow-up, and nonstandardized refraction and visual acuity testing. We acknowledge that our visual acuity data must be interpreted with caution in light of these limitations. Although the lack of a standard protocol with respect to tPA dosing is a relative weakness, this allowed us retrospectively to assess possible differences

in efficacy and toxicity with varying doses. Finally, our inclusion criteria used the same definition of thick submacular hemorrhage that was found in the natural history studies used as historical “controls.”1,2 However, this definition allows for a range of hemorrhage thickness. Future prospective studies of submacular hemorrhage would therefore benefit from a standardized method of measuring and grading hemorrhage thickness, such as echography, standardized stereophotography, or possibly optical coherence tomography. We conclude that intravitreous injection of tPA and gas, followed by a brief period of prone positioning, is effective in displacing thick blood from beneath the center of the macula. The procedure is technically simple, and the rate of serious complications appears to be low. Most patients enjoy visual improvement coincident with blood displacement. Although the final visual outcome is often limited by progression of ARMD, significant and stable visual recovery over extended follow-up is possible in some cases. Acknowledgment. The authors thank David C. Musch, PhD, for assistance in statistical analysis.

References 1. Bennett SR, Folk JC, Blodi CF, Klugman M. Factors prognostic of visual outcome in patients with subretinal hemorrhage. Am J Ophthalmol 1990;109:33–7. 2. Berrocal MH, Lewis ML, Flynn HW Jr. Variations in the clinical course of submacular hemorrhage. Am J Ophthalmol 1996;122:486 –93. 3. Avery RL, Fekrat S, Hawkins BS, Bressler NM. Natural history of subfoveal subretinal hemorrhage in age-related macular degeneration. Retina 1996;16:183–9. 4. Hochman MA, Seery CM, Zarbin MA. Pathophysiology and management of subretinal hemorrhage. Surv Ophthalmol 1997;42:195–213. 5. Glatt H, Machemer R. Experimental subretinal hemorrhage in rabbits. Am J Ophthalmol 1982;94:762–73. 6. Toth CA, Morse LS, Hjelmeland LM, Landers MB III. Fibrin directs early retinal damage after experimental subretinal hemorrhage. Arch Ophthalmol 1991;109:723–9. 7. Johnson MW, Olsen KR, Hernandez E. Tissue plasminogen activator treatment of experimental subretinal hemorrhage. Retina 1991;11:250 – 8. 8. Benner JD, Morse LS, Toth CA, et al. Evaluation of a commercial recombinant tissue-type plasminogen activator preparation in the subretinal space of the cat. Arch Ophthalmol 1991;109:1731– 6. 9. Boone DE, Boldt HC, Ross RD, et al. The use of intravitreal tissue plasminogen activator in the treatment of experimental subretinal hemorrhage in the pig model. Retina 1996;16:518 – 24. 10. Coll GE, Sparrow JR, Marinovic A, Chang S. Effect of intravitreal tissue plasminogen activator on experimental subretinal hemorrhage. Retina 1995;15:319 –26. 11. Johnson MW, Olsen KR, Hernandez E. Tissue plasminogen activator thrombolysis during surgical evacuation of experimental subretinal hemorrhage. Ophthalmology 1992;99:515– 21. 12. Lewis H, Resnick SC, Flannery JG, Straatsma BR. Tissue plasminogen activator treatment of experimental subretinal hemorrhage. Am J Ophthalmol 1991;111:197–204.

1905

Ophthalmology Volume 106, Number 10, October 1999 13. Lewis H. Intraoperative fibrinolysis of submacular hemorrhage with tissue plasminogen activator and surgical drainage. Am J Ophthalmol 1994;118:559 – 68. 14. Ibanez HE, Williams DF, Thomas MA, et al. Surgical management of submacular hemorrhage. A series of 47 consecutive cases. Arch Ophthalmol 1995;113:62–9. 15. Lim JI, Drews–Botsch C, Sternberg P Jr, et al. Submacular hemorrhage removal. Ophthalmology 1995;102:1393–9. 16. Moriarty AP, McAllister IL, Constable IJ. Initial clinical experience with tissue plasminogen activator (tPA) assisted removal of submacular haemorrhage. Eye 1995;9:582– 8. 17. Kamei M, Tano Y, Maeno T, et al. Surgical removal of submacular hemorrhage using tissue plasminogen activator and perfluorocarbon liquid. Am J Ophthalmol 1996;121:267– 75. 18. Hesse L, Meitinger D, Schmidt J. Little effect of tissue plasminogen activator in subretinal surgery for acute hemorrhage in age-related macular degeneration. Ger J Ophthalmol 1997; 5:479 – 83. 19. Claes C, Zivojnovic R. Efficacy of tissue plasminogen activator (t-PA) in subretinal hemorrhage removal. Bull Soc Belge Ophtalmol 1996;261:115– 8. 20. el Baba F, Jarrett WH II, Harbin TS Jr, et al. Massive hemorrhage complicating age-related macular degeneration. Clin-

21.

22. 23. 24. 25. 26. 27.

icopathologic correlation and role of anticoagulants. Ophthalmology 1986;93:1581–92. Koshibu A, Kaga N, Ohkuma H, et al. Ultrastructural studies on the absorption of experimentally produced subretinal hemorrhage [in Japanese] [Eng. abstr]. Folia Ophthalmologica Japonica 1978;29:20 –7. Sternberg P Jr, Aguilar HE, Drews C, Aaberg TM. The effect of tissue plasminogen activator on retinal bleeding. Arch Ophthalmol 1990;108:720 –2. Takeuchi A, Kricorian G, Yao XY, et al. The rate and source of albumin entry into saline-filled experimental retinal detachments. Invest Ophthalmol Vis Sci 1994;35:3792– 8. Johnson MW, Olsen KR, Hernandez E, et al. Retinal toxicity of recombinant tissue plasminogen activator in the rabbit. Arch Ophthalmol 1990;108:259 – 63. Zemel E, Loewenstein A, Lei B, et al. Ocular pigmentation protects the rabbit retina from gentamicin-induced toxicity. Invest Ophthalmol Vis Sci 1995;36:1875– 84. Ohji M, Saito Y, Hayashi A, et al. Pneumatic displacement of subretinal hemorrhage without tissue plasminogen activator. Arch Ophthalmol 1998;116:1326 –32. Kimura AE, Reddy CV, Folk JC, Farmer SG. Removal of subretinal hemorrhage facilitated by preoperative intravitreal tissue plasminogen activator [letter]. Retina 1994;14:83– 4.

Discussion by Wilson J. Heriot, FRACO, FRACS The authors have presented their experience treating thick submacular hemorrhages by intravitreous injection of tissue plasminogen activator (tPA) and an expanding gas bubble. The data are presented in a retrospective, noncomparative case series collected from multiple (5) participating centers. Fifteen patients with a mean follow-up of 9.7 months were included. One potentially serious complication (endophthalmitis) occurred. Blood was displaced from under the fovea in all 15 treated cases. Breakthrough vitreous hemorrhage occurred in three eyes, and in four eyes recurrent subretinal hemorrhage occurred. When the research proposal was prepared, the most important unanswered question was whether a sufficient clot could be displaced pneumatically without having to surgically drain the blood directly. Now that that concept is established, the subsequent critical issues are as follows: Is tPA needed in all cases, or do subretinal clots liquefy (at least in part) over 1 to 2 weeks (e.g., as for choroidal expulsive hemorrhages)? What is the optimal dose of intravitreous tPA in humans? Is more tPA needed for larger clots and less if it is smaller? What causes toxic changes? Is tPA itself or part of the carrier toxic? Is the tPA rapidly bound to fibrin to activate plasminogen such that the larger the clot (and hence the more fibrin), the less “free” tPA is available to cause toxicity? Although I support the authors’ caution regarding tPA dosage and applaud their thorough review of experimental data on retinal toxicity of tPA, extrapolation of data from animal eyes (rabbit, cat,

From the Vitreo-Retinal Unit, The Royal Eye and Ear Hospital, E. Melbourne, Australia. Address correspondence to W. J. Heriot, FRACO FRACS, 1/766 Elizabeth Street, Melbourne, 3000 Australia. E-mail: [email protected].

1906

and pig) with or without intraocular clot may be misleading. Clumping of the retinal pigment epithelium (RPE) in the inferior quadrants was observed in 2 of the 35 cases reported in my series.1 No changes occurred in the macula and vision was not affected (visual acuity nor measurable field). No toxicity was observed in the six patients reported here who were injected with 100␮g nor the six patients with 75 ␮g. Similarly, there were no cases of toxicity in a series of 12 cases injected with 100 ␮g tPA (M. S. Jacobson, MD, unpublished data, presented at Vitreous Society annual meeting, 1998) and none of the 15 eyes injected with 50 ␮g reported by Buhl et al (Invest Ophthalmol Vis Sci 1998;39 [Suppl]:390). In my series, the RPE clumping occurred in the inferior quadrants, consistent with either a toxic reaction or an RPE response to clearing the clot. If due to a toxin, the inferior distribution could be (1) because of a specific gravity greater than vitreous (liquefied or gel); (2) the localization inferiorly in formed vitreous if a posterior detachment is present; or (3) because the solution was displaced by the gas bubble. The observation by Ruiz–Lapuento et al (Invest Ophthalmol Vis Sci 1998;39 [Suppl]: 883) that RPE toxicity correlated strongly with dilution of the tPA in the manufacturer’s carrier rather than balanced salt solution (used by the authors in this series) is consistent with the findings of Irvine et al,2 Benner et al,3 and Kamp et al (Invest Ophthalmol Vis Sci 1991;32 [Suppl]:2729) that the carrier, and perhaps more specifically the arginine phosphate within it, is the toxic agent, not the tPA itself. Whatever the toxic agent in commercially available tPA, the observation by Kawai et al (Invest Ophthalmol Vis Sci 1998;39 [Suppl]:390) that submacular clot can be displaced by a gas bubble alone in some cases suggests that perhaps the better approach to thick submacular clot would be to inject a gas bubble and position the patient for 24 hours. If the clot is not displaced, a subsequent intravitreous injection of tPA could liquefy the clot for displacement by the pre-existing bubble. We should aim to use the lowest possible dose; the authors are now assessing the effect of 25 ␮g of tPA. I look forward to hearing about their subsequent experience.

Hassan et al 䡠 Intravitreous tPA/Pneumatic Displacement of Submacular Hemorrhage The optimal timing after bleed, the dose of tPA, and the safest method of tPA preparation are yet to be determined. No retinal tears or detachments occurred in this series, which may reflect at least some of the authors’ extensive experience with pneumatic retinopexy. One tear and one detachment occurred early in my series1 and retinal tears should be searched for carefully after surgery. In summary, the technique of intravitreous injection of tPA and pneumatic displacement of submacular blood has improved the vision of many patients. To date, there have been few cases of toxicity, even at the initial dosage of 100 ␮g. The procedure is inexpensive and is extremely well-tolerated by elderly patients. Conceptually, it forces us to recognize that the neurosensory retina is not an impermeable barrier but a matrix of cells with an extracellular water compartment through which molecules move physiologically, accumulate in excess in disease (lipoproteins from, for example, diabetic microangiopathy or Coat’s disease), and can be used to manipulate the subretinal space. The blood– retinal barrier is highly selective in regulating the retinal environ-

ment, but we can over-ride this barrier by direct intravitreous injection. Further research to clarify the potentially toxic component of commercial tPA and establish the optimal dose and timing of intravitreous tPA injection in human eyes is needed. Unfortunately, the majority of patients with submacular hemorrhage do not regain normal vision because of the causal macular pathology rather than complications of this treatment. References 1. Heriot WJ. Intravitreal gas and tPA: an outpatient procedure for submacular hemorrhage. Aust N Z J Ophthalmol (in press). 2. Irvine WD, Johnson MW, Hernandez E, Olsen KR. Retinal toxicity of human tissue plasminogen activator in vitrectomized rabbit eyes. Arch Ophthalmol 1991;109:718 –22. 3. Benner JD, Morse LS, Toth CA, et al. Evaluation of a commercial recombinant tissue-type plasminogen activator preparation in the subretinal space of the cat. Arch Ophthalmol 1991;109:1731– 6.

1907