- Email: [email protected]
Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion
Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion One-Year Results of a Pilot Study Nicolas Feltgen, MD, Bernd Junker, MD, Hansjuergen Agostini, MD, Lutz L. Hansen, MD Purpose: Retinal endovascular lysis is a new therapeutic option for patients with central retinal vein occlusion (CRVO). In this procedure, a fibrinolytic agent is injected directly into a cannulated retinal vein after pars plana vitrectomy. Design: Prospective interventional case series. Participants: Thirteen strictly defined patients with ischemic CRVO. Methods: Patients with a decimal visual acuity (VA) of 0.2 or worse were scheduled for surgery within the first 5 months after onset of CRVO. A full ocular examination, determination of VA (Early Treatment Diabetic Retinopathy Study charts), and fluorescein angiography were done preoperatively and 6, 12, 26, and 52 weeks postoperatively. Main Outcome Measure: Visual acuity 1 year after retinal endovascular lysis. Secondary study end points were (1) correlation of VA and successful recombinant tissue plasminogen activator injection into a retinal vein, (2) complication rate, and (3) number of additional surgical procedures within the first year after retinal endovascular lysis. Results: All patients had an ischemic CRVO and completed the 1-year follow-up visit. Preoperative decimal VA was 0.063 ⫹0.025/⫺0.018 (VA range, light perception [LP]– 0.2); 6-week postoperative VA, 0.049 ⫹0.024/ ⫺0.016 (LP– 0.4); 3-month postoperative VA, 0.043 ⫹0.019/⫺0.014 (LP– 0.3); 6-month postoperative VA, 0.035 ⫹0.022/⫺0.013 (blindness– 0.4); and 12-month postoperative VA, 0.04 ⫹0.026/⫺0.016 (blindness– 0.4). Visual acuity changed 1 year after retinal endovascular lysis by ⫺1.923⫾1.619 lines (⫹6 to ⫺16 lines; P ⫽ 0.258). We considered the retinal endovascular lysis procedure to have been technically successful in 10 eyes. Visual changes did not depend on successful lysis. Six eyes developed neovascular glaucoma, of which 2 globes ended up with painful phthisis and had to be removed. Retinal detachment was found in 3 eyes and cataract in 4. Together, the 13 eyes needed 22 additional surgical procedures. Preoperative and postoperative angiographic examinations showed no significant changes. Conclusion: Ischemic CRVO patients did not profit from retinal endovascular lysis in this pilot study. Visual results and the risk of developing iris neovascularization and neovascular glaucoma took the natural course. Although these results may be due to the overall bad prognosis of these particular ischemic eyes, the number of postoperative complications is unacceptably high. Ophthalmology 2007;114:716 –723 © 2007 by the American Academy of Ophthalmology.
Retinal vein occlusion is one of the most common retinal vascular disorders encountered in clinical practice. Despite its frequency, the etiology of central retinal vein occlusion (CRVO) remains controversial, and certainly involves more than one factor.1–3 The ischemic CRVO type is particularly associated with a poor visual and ocular prognosis. To date, Originally received: May 2, 2006. Accepted: June 27, 2006. Manuscript no. 2006-500. From the Department of Ophthalmology, University of Freiburg, Freiburg, Germany. Correspondence to Nicolas Feltgen, MD, Department of Ophthalmology, University of Freiburg, Killianstrasse 5, 79106 Freiburg, Germany. E-mail: [email protected].
716
© 2007 by the American Academy of Ophthalmology Published by Elsevier Inc.
therapeutic results disappoint. Although a slightly better visual outcome after isovolemic hemodilution has been demonstrated by several randomized studies,4 – 6 that approach generally is not accepted.2,3,7 Over the last few years, new surgical therapies have been described.8 –10 Green et al11 could demonstrate that a thrombus had formed in all postmortem eyes with a history of CRVO. Thus, drug-assisted thrombolysis seems a reasonable approach for treating CRVO. But what are the best mode and best time to deliver a fibrinolytic drug to the thrombus? Systemic use,12–14 intraarterial approach,15 and injection of fibrinolytic agent into the vitreous cavity16 –18 were tried, but had serious complications12,13 or were not efficient. ISSN 0161-6420/07/$–see front matter doi:10.1016/j.ophtha.2006.06.064
Feltgen et al 䡠 Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion Retinal endovascular lysis using recombinant tissue plasminogen activator was first described by Weiss and Bynoe.8,19,20 In the retinal endovascular lysis procedure, the surgeon must pierce a large and dammed peripapillary branch of the central retinal vein using a special glass cannula. After successful cannulation, the retinal vein is flushed with recombinant tissue plasminogen activator. Findings from a 28-patient case series revealed positive results: 54% of all CRVO patients recovered ⬎2 lines independent of the perfusion.19 Only minor complications were reported. Similarly favorable results were corroborated by Bynoe et al.20 Unfortunately, the evidence from those studies was limited by the inhomogeneity of the types of CRVO and quality of the postoperative controls. We therefore decided to test retinal endovascular lysis in a prospective case series to define surgical outcome based on visual results, technical feasibility, and the complication rate in a strictly defined group of patients with severely ischemic CRVO.
Patients and Methods Between 2003 and 2005, patients with CRVO were screened for eligibility (see below) and consecutively included in the study.
Inclusion Criteria ● Clinically and angiographically diagnosed ischemic CRVO21–23 between 6 and 20 weeks after CRVO onset ● Optimally corrected visual acuity (VA) of 0.7 (logarithm of the minimum angle of resolution [logMAR]) or more (decimal VA ⱕ 0.2) ● Age over 18 years ● Ability to give informed consent
Exclusion Criteria ● Patients with retinal or disc neovascularization needing photocoagulation at first presentation
● Other eye diseases that reduced VA, except cataract (e.g., glaucoma with visual field [VF] loss in the other eye; diabetic retinopathy; macular degeneration; uveitis; vitreous opacity; history of retinal detachment [RD] with visual impairment, of retinal vein or artery occlusion, and of neuro-ophthalmological diseases with VF defects; patients with amblyopia in the affected eye) ● Inability to give informed consent
Study End Points The study end point was change in VA 1 year after retinal endovascular lysis as compared with the initial VA. Secondary study end points were (1) correlation of VA and successful recombinant tissue plasminogen activator injection into a retinal vein, (2) complication rate, and (3) number of additional surgical procedures within the first year after retinal endovascular lysis.
Preoperative Examination The following data were registered preoperatively: duration of CRVO before surgery, ophthalmological and medical histories, patient age and gender, best-corrected VA (Early Treatment Diabetic Retinopathy Study charts), and full ocular examination with swinging flashlight test for estimation of the relative afferent pupillary defect (recorded in log units of the neutral density filters).24 We also documented retinal changes by color fundus photographs (FF450, Zeiss, Oberkochen, Germany) and fluorescein angiography with a scanning laser ophthalmoscope (Heidelberg Retina Angiograph, Heidelberg Engineering, Heidelberg, Germany). Those examinations were made preoperatively and 6, 12, 26, and 52 weeks after surgery. For angiographic assessment, we assumed that the change in arteriovenous passage (AP) between preoperative and postoperative examinations of the affected eye was the relevant parameter and, thus, defined 2 different APs: (1) early AP, the interval between first appearance of the fluorescein in the central artery and that in the corresponding vein, and (2) maximum filling AP, the interval between maximum fluorescein filling in the central artery and that in the corresponding vein. To measure the foveal avascular zone (FAZ), we marked the inner capillary arcade surround-
Table 1. Patient Characteristics VA†
No.
Age (yrs)
Onset of CRVO (Preoperative Weeks)
Preoperative Treatment (IHD)*
1 2 3 4 5 6 7 8 9 10 11 12 13
71 83 69 54 76 76 75 59 75 56 67 59 57
16 18 18 12 9 17 13 12 7 8 11 11 18
Complete Complete Complete No IHD No IHD Incomplete Complete No IHD Incomplete Incomplete Incomplete Incomplete Complete
Preoperative
6 Weeks Postoperative
1 Year Postoperative
Change in VA After 1 Year (Lines)
0.2 0.1 0.2 0.2 0.014 (CF) 0.08 0.04 0.1 0.05 0.1 0.1 0.0025 (LP) 0.05
0.4 0.04 0.2 0.05 0.04 0.1 0.03 0.1 0.005 (HM) 0.04 0.2 0.0025 (LP) 0.05
0.32 0.0025 (LP) 0.08 0.1 0.014 (CF) 0.08 0.08 0.1 — 0.05 0.32 — 0.1
⫹2 ⫺16 ⫺4 ⫺3 0 0 ⫹3 0 ⫺10 ⫺3 ⫹6 ⫺3 ⫹3
Retinal Endovascular Lysis Successful
rt-PA Used (g)
No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
200 300 300 100 200 100 200 300 200 100 200 200 180
CF ⫽ counting fingers; CRVO ⫽ central retinal vein occlusion; HM ⫽ hand movements; IHD ⫽ isovolemic hemodilution; LP ⫽ light perception; rt-PA ⫽ recombinant tissue plasminogen activator; VA ⫽ visual acuity. — ⫽ eye was removed/enucleated. *Complete, 6 wks; incomplete, ⬍6 wks. † Decimal.
717
Ophthalmology Volume 114, Number 4, April 2007 blood reflux within the cannula showed intraluminal placement. Recombinant tissue plasminogen activator solution (Actilyse, Boehringer Ingelheim, Ingelheim, Germany) at 200 g/ml was injected using a volume between 0.5 and 1.5 ml corresponding to 100 to 300 g/ml. No standard gas tamponade was required. The retinal endovascular lysis procedure was considered successful if the recombinant tissue plasminogen activator solution could be injected into the retinal vein (venous whitening), as demonstrated in Figures 2 and 3(the latter available at http://aaojournal.org). The retinal endovascular lysis procedure was considered unsuccessful if venous whitening was seen secondary to papillary paleness (injection into the optic nerve head tissue) or retinal bleb formation occurred (intraretinal or subretinal injection). No panretinal photocoagulation was carried out during the retinal endovascular lysis surgery. Figure 1. Visual acuity (VA) follow-up (VA ⫾ standard error of the mean). Decimal values for comparison (n ⫽ 13). logMAR ⫽ logarithm of the minimum angle of resolution; postop ⫽ postoperatively.
ing the fovea interactively. The morphologic data were corrected automatically for refractive error using Heidelberg Eye Explorer software (Heidelberg Engineering). Two masked examiners quantified early AP, maximum filling AP, and FAZ using offline digital image analysis, and the mean of both values was calculated.
Isovolemic Hemodilution Isovolemic hemodilution was carried out on an outpatient basis by patients’ general practitioners. It was performed over 6 weeks (depending on the hematocrit) if the ischemic CRVO was not over 8 weeks old.4,5 If that was so, the therapy was designated as complete. If the isovolemic hemodilution period was under 6 weeks (e.g., due to technical problems or incombatibility), it was designated as incomplete. Patients whose CRVO was older than 8 weeks underwent no isovolemic hemodilution therapy (Table 1). To ensure that the isovolemic hemodilution did not interfere with the retinal endovascular lysis results, we included in our study only patients who had not profited from the isovolemic hemodilution therapy (Fig 1).
Surgical Procedure Patients underwent a standard 3-port pars plana vitrectomy, with detachment of the posterior hyaloid if necessary. The macular region’s internal limiting membrane was not removed. To penetrate the vessel wall, it was necessary to free the vessel from the surrounding nerve fibers adjacent to the vein. Thereafter, the bevelled bent glass cannula (Micron Ophthalmic Inc., Margate, FL) was placed parallel to the lumen of the peripapillary branch retinal vein chosen for cannulation, as described by Weiss and Bynoe.19 The retinal vein was pierced manually with the glass cannula. In case of vessel perforation, the device was redrawn until
Study Design and Ethics This nonrandomized, prospective, interventional case series was reviewed by University Hospital Freiburg’s ethics committee and performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. All patients gave their informed consent before their inclusion in the study.
Statistics All statistical evaluations were done on the approximately normally distributed log (VA) ⫽ ⫺logMAR scale. For intuitive display, all results were transformed to the decimal VA scale [decimal VA ⫽ 10log(VA)]. According to Holladay and our study group’s results, blindness was set at 0.00125/2.9 (decimal/logMAR), light perception (LP) at 0.0025/2.6, hand movements at 0.005/2.3, and counting fingers at 0.014/1.85.25,26 Preoperative versus postoperative values were tested with analysis of variance (Bonferroni/ Dunn). Values are given as mean ⫾ standard error of the mean and VA range. The null hypothesis was that the difference in mean VA measurement between preoperative and postoperative VAs was zero.
Results From August 2002 to March 2005, we prospectively enrolled 13 patients in the study. So far, all patients have completed the 1-year control. Patient characteristics are depicted in detail in Table 1.
Visual Acuity At the patients’ first presentation in our hospital, VA was 0.097 ⫹0.044/⫺0.031 (range, LP– 0.4). Immediately before surgery, preoperative VA was 0.063 ⫹0.025/⫺0.018 (range, LP– 0.2). Despite isovolemic hemodilution therapy, no patient showed an increase in VA between first presentation and preoperative measurement. The
™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 Figure 2. Successful recombinant tissue plasminogen activator injection towards the periphery (blue arrow, vein; black arrow, artery). Left, Artery and vein before injection. Middle, Artery and vein at start of injection. Right, Artery and vein during injection. Figure 4. Decimal visual acuity (VA) results 1 year postoperatively. Diagonal small lines, 2-line difference in VA; Œ, lysis not successful. Figure 5. Retinal endovascular lysis. Injection towards optic nerve head (arrow). Left, Preoperative early fluorescein angiography. A ⫽ branch of retinal artery (white); V ⫽ branch of retinal vein (dark). Colored pictures, During the injection, the optic nerve head and all central vessels turned pale. Figure 6. Bleeding from the cannulated area (circle) started when reflux from the lower branch of the central retinal vein (black arrow) was completed. Yellow arrow, peripheral retinal vein.
718
719
Ophthalmology Volume 114, Number 4, April 2007 Table 2. Patient Surgical Characteristics No.
Attempts of Venous Puncture
Intraoperative Complications
1 2
1 5
— Serous retinal detachment
— Proliferative retinopathy Glaucoma
1 4
3
3
rt-PA intraretinal Vitreous hemorrhage
4
4
3
Retinal detachment Cataract Glaucoma Luxation of lens Vitreous hemorrhage Cataract Retinal detachment
5 6 7 8 9
1 2 1 1 3
Cataract
1 0 0 0 4
10 11
2 2
12
1
13
2
—
Vitreous hemorrhage — — Vitreous hemorrhage —
— — Vitreous hemorrhage —
Postoperative Complications
— Glaucoma — Vitreous hemorrhage Glaucoma Painful phthisis — Retinal detachment Redetachment Glaucoma Painful phthisis —
No. of Reoperations
4
0 2 2 0
Reoperations Cataract surgery CPC and peripheral retinal cryocoagulation Vitrectomy, laser, Vitrectomy, laser, and oil filling Oil removal Encircling band, vitrectomy, laser Cataract surgery Gas removal Repositioning of IOL Vitrectomy, laser, and oil filling Oil removal, membrane peeling, laser Cataract surgery, encircling band, oil filling Oil removal, retinotomy, laser, and gas filling Cataract surgery — — — Vitrectomy and laser Peripheral retinal cryocoagulation, CPC Vitrectomy and laser Enucleation — Encircling band, gas filling Vitrectomy, laser, and gas filling CPC, peripheral retinal cryocoagulation Enucleation —
CPC ⫽ cyclophotocoagulation; IOL ⫽ intraocular lens; rt-PA ⫽ recombinant tissue plasminogen activator.
interval between the 2 measurements was 4.2⫾2.7 weeks (range, 0 – 8). Six-week postoperative VA was 0.049 ⫹0.024/⫺0.016 (range, LP– 0.4); 3-month postoperative VA, 0.043 ⫹0.019/ ⫺0.014 (range, LP– 0.3); 6-month postoperative VA, 0.035 ⫹0.022/⫺0.013 (range, blindness– 0.4); and 12-month postoperative VA, 0.04 ⫹0.026/⫺0.016 (range, blindness– 0.4) (Fig 1, Table 1). None of these differences was significant. Visual acuity worsened 1 year after retinal endovascular lysis by about 1.923⫾1.619 lines (6 to ⫺16 lines). Again, this difference was not significant (P ⫽ 0.26). In evaluating the patients individually, VA improved in 3 eyes by ⬎2 lines, 4 eyes remained stable within ⫾2 lines, and 6 eyes worsened by ⬎2 lines (Fig 4). Although 2 globes had to be enucleated, we added the final VA (blindness/0.00125) to present the realistic situation 1 year after initial surgery. All patients had ischemic CRVO according to the 1995 Central Vein Occlusion Study’s definition, with a mean of 52.7⫾11.2 disc areas of nonperfusion,23 although the VA of 3 eyes was 0.2. The mean relative afferent pupillary defect was 1.12⫾0.15 log units.
Surgical Procedure Posterior vitreous detachment and venous piercing were achieved in all eyes. Multiple attempts to pierce a vein were necessary in 8 eyes, as summarized in Table 2. Injection was regarded successful in 10 eyes, as demonstrated in Figure 2 (primary venous whitening). Despite high-quality videotaping (Videos 1, 2 [available at http://aaojournal.org]), we could not determine retrospectively for certain whether venous whitening actually was caused by successful intraluminal recombinant tissue plasminogen activator injection in all cases. It also may have been caused by increased
720
periluminal pressure due to paravasal injection into the optic nerve head tissue. Those eyes (n ⫽ 3) were assessed as having unsuccessful lysis. Nevertheless, we noticed the optic nerve head turning pale most frequently during the injection. To explain these findings, the most likely explanation is that the thrombus is neither dissolved nor flushed away, but rather the central retinal vein is bloated by the injection of recombinant tissue plasminogen activator. The bloated vein may have compressed the central artery (which usually courses very close to the vein) and hampered the arterial inflow, mimicking a compartment syndrome. Hence, the optic nerve became pale, and the vessels collapsed (Fig 5). Upon removal of the cannula, the piercing site did not start bleeding immediately, but blood was observed when the cannulated vein filled up completely due to central blood reflux. We could not identify bleeding from the periphery, indicating an unchanged venous outflow (Fig 6). However, visual improvement did not depend on successful lysis (Fig 4). Visual changes were also independent of the duration of ischemic CRVO within the group of patients with successfully pierced retinal veins.
Problems and Complications Intraoperative Complications. Retinal endovascular lysis was feasible, but we often needed more than one piercing attempt for successful injection. After touching the selected vein, some vessels became constricted and did not allow further piercing. We occasionally perforated the vein and mistakenly injected recombinant tissue plasminogen activator into the retinal tissue or the subretinal
Feltgen et al 䡠 Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion space. In such a case, one can pull back the cannula into the lumen of the vein. Although it is not difficult to hold the cannula within the vein, infusion pressure frequently causes the recombinant tissue plasminogen activator solution to leak into the surrounding tissue. When extracting the glass pipette from the vein, the wound closes within a few seconds to minutes, allowing only small amounts of blood to escape. Postoperative Complications. Postoperative complications (Table 2) within the first year after initial surgery were frequent. Only one patient suffered from heavy vitreous bleeding during the first postoperative day. Despite panretinal laser coagulation, 6 eyes developed neovascular glaucoma, of which 2 globes ended with painful phthisis. Retinal detachment was found in 3 eyes and cataract in 4. Together, the 13 eyes required 22 additional surgical procedures, including cataract surgery (4 eyes); transscleral cyclophotocoagulation and peripheral cryocoagulation (3 eyes); vitrectomy, laser coagulation, and oil filling (4 eyes); encircling band and vitrectomy for RD (3 eyes [1 eye with redetachment]); and enucleation of the globe (2 eyes).
Angiographic Findings Postoperative control was possible in 12 patients at 3 months. Later controls could not be performed in all patients because of corneal edema, anterior chamber or vitreous hemorrhage, RD, cataract, or patient unwillingness. The mean early AP was 5.03⫾0.4 seconds before surgery and 8.1⫾7.0 seconds postoperatively. This increase in AP time was not significant (P ⫽ 0.39). The maximum filling AP increased from 16.5⫾1.2 seconds before surgery to 20.1⫾7.0 seconds postoperatively, which was also not significant (P ⫽ 0.73). We measured the FAZ to assess pole ischemia, but in 10 eyes the macula was partly covered with blood. The mean FAZ was 0.87⫾0.56 mm2.
Discussion In the management of acute major vessel occlusion, the use of fibrinolytic agents such as recombinant tissue plasminogen activator is widely accepted. Today, the spectrum of indications for thrombolytic drugs comprises acute myocardial infarction, lung embolism, ischemic stroke, deep vein thrombosis, and acute arterial occlusions of the lower limbs. Based on the histopathological and clinical features of retinal vessel occlusion, fibrinolysis aiming at early restoration of blood flow appears to be a promising therapeutic approach. However, there are serious concerns regarding the systemic use of lytic agents in retinal vein occlusion because of considerable risk of cerebral or gastrointestinal hemorrhage.13,14 To avoid life-threatening complications, the local endovascular application of lytic agents directly was developed.8 Weiss and Bynoe reported the results of this novel surgical technique with various series of CRVO patients. Of those, up to 54% recovered ⬎2 lines of VA within 3 months, whereas 50% reached this goal even after a mean follow-up of 12 months.19,20,27 This compares favorably to the natural-course studies with improvement rates of only 10% to 14%.4 – 6,28 In those studies, patients with poor initial VA (⬍0.1) had an 80% risk of having VA⬍0.1 at final visit, whether perfused or not perfused initially.23 The data from those initial studies are quite inhomogeneous. Central retinal vein occlusion duration varied be-
tween 0.25 and 30 months, both ischemic and nonischemic patients were included, and different volumes of recombinant tissue plasminogen activator solution were injected (0.6 –7.5 ml). To validate those initial results, we performed this pilot study on a narrowly defined group of patients with severe ischemic CRVO, including all preoperative and postoperative controls at our study center over a 1-year period. As opposed to the initial studies, our data did not rely on the referring ophthalmologists. In short, our visual results after vitrectomy and retinal endovascular lysis are clearly disappointing. Three eyes improved by ⬎2 lines, and only 1 patient had a final VA of ⬎0.1—results that closely resemble the natural course of the ischemic type of CRVO. Additionally, our angiographic data showed an increase in AP after surgery that is worse than the preoperative data. Although these changes were not significant, one would have expected a reduction in AP time by flushing the veins. It is very unlikely that isovolemic hemodilution, which is performed routinely in our hospital and by the local ophthalmologists, negated our retinal endovascular lysis results. Previous studies have shown that if visual improvement occurs, it will do so within the first 6 weeks of therapy.4,5 Yet none of our 13 patients presented improved vision within the preoperative period (Fig 1). What might explain the failure of retinal endovascular lysis in our case series? 1. Did we operate too late? Low venous perfusion with subsequent thrombus formation and organization11,29 causes CRVO and determines the prognosis. The duration of CRVO in our patients was between 7 and 18 weeks; hence, thrombi were rather old and possibly less amenable to easy dissolution. Thrombus flushing was the explanation30 for the good results obtained in the Weiss and Bynoe studies. We found it possible to flush in a retrograde manner, but this was not always feasible towards the optic nerve head. Furthermore, it is difficult to understand how a well-organized thrombus can be loosened from the wall and flushed away, as thrombus organization includes capillary growth and ultimately fibrous scarring.30 We therefore find it improbable that simply flushing it with thrombolytic agents can dissolve an already organized thrombus. On the other hand, even ischemic CRVO is seldom caused by a complete obstruction of the central retinal vein.31 It therefore may be possible to improve overall venous flow by flushing the remaining venous lumen, hence reducing the macular edema. Unfortunately, we have no means of telling how old a thrombus is in CRVO eyes. For ethical reasons, we decided to operate on patients with older CRVOs so as not to mask spontaneous recovery. Theoretically, retinal endovascular lysis performed within the first hours or days should be more effective. The most likely explanation based on our intraoperative findings is that the thrombus itself remained unchanged. However, the already dammed central ret-
721
Ophthalmology Volume 114, Number 4, April 2007 inal vein was bloated further by the injection of additional fluid, thus compressing the central artery and hampering arterial inflow. 2. Were our patients’ eyes too ischemic for any therapy to be visually effective? Our poor results also may be due to our negatively selected group of patients in which we treated only eyes with severe ischemia, which in turn have a higher risk of neovascular complications such as glaucoma or vitreous bleeding.32 In our study, 6 of 13 (46.2%) patients developed neovascularization, which corresponds closely to the high rate of iris neovascularization (35%) in ischemic CRVO eyes.23 The high complication rate and high number of additional surgical procedures required in our patient series are remarkable. Five of 12 phakic eyes needed cataract surgery. Other severe complications were RD, repeated vitreous hemorrhage, and phthisis. Despite intensive treatment, neovascular glaucoma could not be avoided in all cases, and 2 patients even required enucleation of the globe. This negative selection may be a drawback of the present study. But we identified 10 ischemic CRVO patients in the Weiss and Bynoe study as well,19 and it is astonishing that their visual improvement and complication rate were independent of both initial VA and retinal perfusion (perfused, nonperfused, or intermediate, according to the 1993 Central Vein Occlusion Study32). 3. Does the surgeon need to be more experienced in retinal endovascular lysis? One also must consider the learning curve in this sophisticated surgical method. Puncture and flushing of the vein were certainly feasible in most patients. But we did not demonstrate a direct measurable improvement in the flow as indicated by angiographic data. Perhaps the short exposure time of about 30 to 60 seconds was not sufficient to dissolve a thrombus 7 to 18 weeks after formation. To summarize, retinal endovascular lysis is feasible but remains a sophisticated surgical method with many complications. Although we anticipated postoperative complications in eyes with severe retinal ischemia, numbers of complications and necessary additional surgical procedures were unacceptably high. As we have not been able to reproduce the good results of others in our pilot study, we can neither recommend nor give up on this new surgical method unless it is tested on patients suffering from fresh CRVO. However, without demonstrating a distinct improvement in VA, it is hard to justify the increased risk due to surgical complications.
References 1. Hayreh SS. Retinal vein occlusion. Indian J Ophthalmol 1994; 42:109 –32. 2. Hayreh SS. Prevalent misconceptions about acute retinal vascular occlusive disorders. Prog Retin Eye Res 2005;24:493– 519.
722
3. Williamson TH. Central retinal vein occlusion: what’s the story? Br J Ophthalmol 1997;81:698 –704. 4. Hansen LL, Danisevskis P, Arntz HR, et al. A randomised prospective study on treatment of central retinal vein occlusion by isovolaemic haemodilution and photocoagulation. Br J Ophthalmol 1985;69:108 –16. 5. Hansen LL, Wiek J, Schade M, et al. Effect and compatibility of isovolaemic haemodilution in the treatment of ischaemic and non-ischaemic central retinal vein occlusion. Ophthalmologica 1989;199:90 –9. 6. Wolf S, Arend O, Bertram B, et al. Hemodilution therapy in central retinal vein occlusion: one-year results of a prospective randomized study. Graefes Arch Clin Exp Ophthalmol 1994; 232:33–9. 7. Hayreh SS. Management of central retinal vein occlusion. Ophthalmologica 2003;217:167– 88. 8. Weiss JN. Treatment of central retinal vein occlusion by injection of tissue plasminogen activator into a retinal vein. Am J Ophthalmol 1998;126:142– 4. 9. Opremcak EM, Bruce RA, Lomeo MD, et al. Radial optic neurotomy for central retinal vein occlusion: a retrospective pilot study of 11 consecutive cases. Retina 2001;21:408 –15. 10. Hattenbach LO, Steinkamp G, Scharrer I, Ohrloff C. Fibrinolytic therapy with low-dose recombinant tissue plasminogen activator in retinal vein occlusion. Ophthalmologica 1998; 212:394 – 8. 11. Green WR, Chan CC, Hutchins GM, Terry JM. Central retinal vein occlusion: a prospective histopathologic study of 29 eyes in 28 cases. Retina 1981;1:27–55. 12. Kohner EM, Pettit JE, Hamilton AM, et al. Streptokinase in central retinal vein occlusion: a controlled clinical trial. Br Med J 1976;1:550 –3. 13. Elman MJ. Thrombolytic therapy for central retinal vein occlusion: results of a pilot study. Trans Am Ophthalmol Soc 1996;94:471–504. 14. Hattenbach LO, Wellermann G, Steinkamp GW, et al. Visual outcome after treatment with low-dose recombinant tissue plasminogen activator or hemodilution in ischemic central retinal vein occlusion. Ophthalmologica 1999;213:360 – 6. 15. Paques M, Vallee JN, Herbreteau D, et al. Superselective ophthalmic artery fibrinolytic therapy for the treatment of central retinal vein occlusion. Br J Ophthalmol 2000;84: 1387–91. 16. Lahey JM, Fong DS, Kearney J. Intravitreal tissue plasminogen activator for acute central retinal vein occlusion. Ophthalmic Surg Lasers 1999;30:427–34. 17. Glacet-Bernard A, Kuhn D, Vine AK, et al. Treatment of recent onset central retinal vein occlusion with intravitreal tissue plasminogen activator: a pilot study. Br J Ophthalmol 2000;84:609 –13. 18. Elman MJ, Raden RZ, Carrigan A. Intravitreal injection of tissue plasminogen activator for central retinal vein occlusion. Trans Am Ophthalmol Soc 2001;108:219 –21, discussion 222–3. 19. Weiss JN, Bynoe LA. Injection of tissue plasminogen activator into a branch retinal vein in eyes with central retinal vein occlusion. Ophthalmology 2001;108:2249 –57. 20. Bynoe LA, Hutchins RK, Lazarus HS, Friedberg MA. Retinal endovascular surgery for central retinal vein occlusion: initial experience of four surgeons. Retina 2005;25:625–32. 21. Magargal LE, Donoso LA, Sanborn GE. Retinal ischemia and risk of neovascularization following central retinal vein obstruction. Ophthalmology 1982;89:1241–5. 22. Hayreh SS, Klugman MR, Beri M, et al. Differentiation of ischemic from non-ischemic central retinal vein occlusion
Feltgen et al 䡠 Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion
23. 24. 25. 26.
27.
during the early acute phase. Graefes Arch Clin Exp Ophthalmol 1990;228:201–17. Central Vein Occlusion Study Group. Natural history and clinical management of central retinal vein occlusion. Arch Ophthalmol 1997;115:486 –91. Servais GE, Thompson HS, Hayreh SS. Relative afferent pupillary defect in central retinal vein occlusion. Ophthalmology 1986;93:301–3. Holladay JT. Proper method for calculating average visual acuity. J Refract Surg 1997;13:388 –91. Schulze-Bonsel K, Feltgen N, Burau H, et al. Visual acuities “hand motion” and “counting fingers” can be quantified with the Freiburger visual acuity test. Invest Ophthalmol Vis Sci 2006;47:1236 – 40. Bynoe LA, Weiss JN. Retinal endovascular surgery and intra-
28. 29. 30. 31. 32.
vitreal triamcinolone acetonide for central vein occlusion in young adults. Am J Ophthalmol 2003;135:382– 4. Quinlan PM, Elman MJ, Bhatt AK, et al. The natural course of central retinal vein occlusion. Am J Ophthalmol 1990;110: 118 –23. Cotran RS, Pober JS. Effects of cytokines on vascular endothelium: their role in vascular and immune injury. Kidney Int 1989;35:969 –75. Hayreh MS. t-PA in CRVO. Ophthalmology 2002;109:1758 – 61. Seitz R. The retinal blood vessels. Comparative ophthalmoscopic and histological studies on the healthy and pathologically changed eye [in German]. Buch Augenarzt 1962;40:1–175. Baseline and early natural history report: the Central Vein Occlusion Study. Arch Ophthalmol 1993;111:1087–95.
723
Ophthalmology Volume 114, Number 4, April 2007
Figure 3. Successful retinal endovascular lysis. The most important parts of Video 2 are shown. The thrombus (T) remains unchanged during the injection. The vein (V) is bloated by the injection of recombinant tissue plasminogen activator. The arterial (A) inflow is hampered and the vessels collapse.
723.e1
Retinal vein occlusion is one of the most common retinal vascular disorders encountered in clinical practice. Despite its frequency, the etiology of central retinal vein occlusion (CRVO) remains controversial, and certainly involves more than one factor.1–3 The ischemic CRVO type is particularly associated with a poor visual and ocular prognosis. To date, Originally received: May 2, 2006. Accepted: June 27, 2006. Manuscript no. 2006-500. From the Department of Ophthalmology, University of Freiburg, Freiburg, Germany. Correspondence to Nicolas Feltgen, MD, Department of Ophthalmology, University of Freiburg, Killianstrasse 5, 79106 Freiburg, Germany. E-mail: [email protected].
716
© 2007 by the American Academy of Ophthalmology Published by Elsevier Inc.
therapeutic results disappoint. Although a slightly better visual outcome after isovolemic hemodilution has been demonstrated by several randomized studies,4 – 6 that approach generally is not accepted.2,3,7 Over the last few years, new surgical therapies have been described.8 –10 Green et al11 could demonstrate that a thrombus had formed in all postmortem eyes with a history of CRVO. Thus, drug-assisted thrombolysis seems a reasonable approach for treating CRVO. But what are the best mode and best time to deliver a fibrinolytic drug to the thrombus? Systemic use,12–14 intraarterial approach,15 and injection of fibrinolytic agent into the vitreous cavity16 –18 were tried, but had serious complications12,13 or were not efficient. ISSN 0161-6420/07/$–see front matter doi:10.1016/j.ophtha.2006.06.064
Feltgen et al 䡠 Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion Retinal endovascular lysis using recombinant tissue plasminogen activator was first described by Weiss and Bynoe.8,19,20 In the retinal endovascular lysis procedure, the surgeon must pierce a large and dammed peripapillary branch of the central retinal vein using a special glass cannula. After successful cannulation, the retinal vein is flushed with recombinant tissue plasminogen activator. Findings from a 28-patient case series revealed positive results: 54% of all CRVO patients recovered ⬎2 lines independent of the perfusion.19 Only minor complications were reported. Similarly favorable results were corroborated by Bynoe et al.20 Unfortunately, the evidence from those studies was limited by the inhomogeneity of the types of CRVO and quality of the postoperative controls. We therefore decided to test retinal endovascular lysis in a prospective case series to define surgical outcome based on visual results, technical feasibility, and the complication rate in a strictly defined group of patients with severely ischemic CRVO.
Patients and Methods Between 2003 and 2005, patients with CRVO were screened for eligibility (see below) and consecutively included in the study.
Inclusion Criteria ● Clinically and angiographically diagnosed ischemic CRVO21–23 between 6 and 20 weeks after CRVO onset ● Optimally corrected visual acuity (VA) of 0.7 (logarithm of the minimum angle of resolution [logMAR]) or more (decimal VA ⱕ 0.2) ● Age over 18 years ● Ability to give informed consent
Exclusion Criteria ● Patients with retinal or disc neovascularization needing photocoagulation at first presentation
● Other eye diseases that reduced VA, except cataract (e.g., glaucoma with visual field [VF] loss in the other eye; diabetic retinopathy; macular degeneration; uveitis; vitreous opacity; history of retinal detachment [RD] with visual impairment, of retinal vein or artery occlusion, and of neuro-ophthalmological diseases with VF defects; patients with amblyopia in the affected eye) ● Inability to give informed consent
Study End Points The study end point was change in VA 1 year after retinal endovascular lysis as compared with the initial VA. Secondary study end points were (1) correlation of VA and successful recombinant tissue plasminogen activator injection into a retinal vein, (2) complication rate, and (3) number of additional surgical procedures within the first year after retinal endovascular lysis.
Preoperative Examination The following data were registered preoperatively: duration of CRVO before surgery, ophthalmological and medical histories, patient age and gender, best-corrected VA (Early Treatment Diabetic Retinopathy Study charts), and full ocular examination with swinging flashlight test for estimation of the relative afferent pupillary defect (recorded in log units of the neutral density filters).24 We also documented retinal changes by color fundus photographs (FF450, Zeiss, Oberkochen, Germany) and fluorescein angiography with a scanning laser ophthalmoscope (Heidelberg Retina Angiograph, Heidelberg Engineering, Heidelberg, Germany). Those examinations were made preoperatively and 6, 12, 26, and 52 weeks after surgery. For angiographic assessment, we assumed that the change in arteriovenous passage (AP) between preoperative and postoperative examinations of the affected eye was the relevant parameter and, thus, defined 2 different APs: (1) early AP, the interval between first appearance of the fluorescein in the central artery and that in the corresponding vein, and (2) maximum filling AP, the interval between maximum fluorescein filling in the central artery and that in the corresponding vein. To measure the foveal avascular zone (FAZ), we marked the inner capillary arcade surround-
Table 1. Patient Characteristics VA†
No.
Age (yrs)
Onset of CRVO (Preoperative Weeks)
Preoperative Treatment (IHD)*
1 2 3 4 5 6 7 8 9 10 11 12 13
71 83 69 54 76 76 75 59 75 56 67 59 57
16 18 18 12 9 17 13 12 7 8 11 11 18
Complete Complete Complete No IHD No IHD Incomplete Complete No IHD Incomplete Incomplete Incomplete Incomplete Complete
Preoperative
6 Weeks Postoperative
1 Year Postoperative
Change in VA After 1 Year (Lines)
0.2 0.1 0.2 0.2 0.014 (CF) 0.08 0.04 0.1 0.05 0.1 0.1 0.0025 (LP) 0.05
0.4 0.04 0.2 0.05 0.04 0.1 0.03 0.1 0.005 (HM) 0.04 0.2 0.0025 (LP) 0.05
0.32 0.0025 (LP) 0.08 0.1 0.014 (CF) 0.08 0.08 0.1 — 0.05 0.32 — 0.1
⫹2 ⫺16 ⫺4 ⫺3 0 0 ⫹3 0 ⫺10 ⫺3 ⫹6 ⫺3 ⫹3
Retinal Endovascular Lysis Successful
rt-PA Used (g)
No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
200 300 300 100 200 100 200 300 200 100 200 200 180
CF ⫽ counting fingers; CRVO ⫽ central retinal vein occlusion; HM ⫽ hand movements; IHD ⫽ isovolemic hemodilution; LP ⫽ light perception; rt-PA ⫽ recombinant tissue plasminogen activator; VA ⫽ visual acuity. — ⫽ eye was removed/enucleated. *Complete, 6 wks; incomplete, ⬍6 wks. † Decimal.
717
Ophthalmology Volume 114, Number 4, April 2007 blood reflux within the cannula showed intraluminal placement. Recombinant tissue plasminogen activator solution (Actilyse, Boehringer Ingelheim, Ingelheim, Germany) at 200 g/ml was injected using a volume between 0.5 and 1.5 ml corresponding to 100 to 300 g/ml. No standard gas tamponade was required. The retinal endovascular lysis procedure was considered successful if the recombinant tissue plasminogen activator solution could be injected into the retinal vein (venous whitening), as demonstrated in Figures 2 and 3(the latter available at http://aaojournal.org). The retinal endovascular lysis procedure was considered unsuccessful if venous whitening was seen secondary to papillary paleness (injection into the optic nerve head tissue) or retinal bleb formation occurred (intraretinal or subretinal injection). No panretinal photocoagulation was carried out during the retinal endovascular lysis surgery. Figure 1. Visual acuity (VA) follow-up (VA ⫾ standard error of the mean). Decimal values for comparison (n ⫽ 13). logMAR ⫽ logarithm of the minimum angle of resolution; postop ⫽ postoperatively.
ing the fovea interactively. The morphologic data were corrected automatically for refractive error using Heidelberg Eye Explorer software (Heidelberg Engineering). Two masked examiners quantified early AP, maximum filling AP, and FAZ using offline digital image analysis, and the mean of both values was calculated.
Isovolemic Hemodilution Isovolemic hemodilution was carried out on an outpatient basis by patients’ general practitioners. It was performed over 6 weeks (depending on the hematocrit) if the ischemic CRVO was not over 8 weeks old.4,5 If that was so, the therapy was designated as complete. If the isovolemic hemodilution period was under 6 weeks (e.g., due to technical problems or incombatibility), it was designated as incomplete. Patients whose CRVO was older than 8 weeks underwent no isovolemic hemodilution therapy (Table 1). To ensure that the isovolemic hemodilution did not interfere with the retinal endovascular lysis results, we included in our study only patients who had not profited from the isovolemic hemodilution therapy (Fig 1).
Surgical Procedure Patients underwent a standard 3-port pars plana vitrectomy, with detachment of the posterior hyaloid if necessary. The macular region’s internal limiting membrane was not removed. To penetrate the vessel wall, it was necessary to free the vessel from the surrounding nerve fibers adjacent to the vein. Thereafter, the bevelled bent glass cannula (Micron Ophthalmic Inc., Margate, FL) was placed parallel to the lumen of the peripapillary branch retinal vein chosen for cannulation, as described by Weiss and Bynoe.19 The retinal vein was pierced manually with the glass cannula. In case of vessel perforation, the device was redrawn until
Study Design and Ethics This nonrandomized, prospective, interventional case series was reviewed by University Hospital Freiburg’s ethics committee and performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. All patients gave their informed consent before their inclusion in the study.
Statistics All statistical evaluations were done on the approximately normally distributed log (VA) ⫽ ⫺logMAR scale. For intuitive display, all results were transformed to the decimal VA scale [decimal VA ⫽ 10log(VA)]. According to Holladay and our study group’s results, blindness was set at 0.00125/2.9 (decimal/logMAR), light perception (LP) at 0.0025/2.6, hand movements at 0.005/2.3, and counting fingers at 0.014/1.85.25,26 Preoperative versus postoperative values were tested with analysis of variance (Bonferroni/ Dunn). Values are given as mean ⫾ standard error of the mean and VA range. The null hypothesis was that the difference in mean VA measurement between preoperative and postoperative VAs was zero.
Results From August 2002 to March 2005, we prospectively enrolled 13 patients in the study. So far, all patients have completed the 1-year control. Patient characteristics are depicted in detail in Table 1.
Visual Acuity At the patients’ first presentation in our hospital, VA was 0.097 ⫹0.044/⫺0.031 (range, LP– 0.4). Immediately before surgery, preoperative VA was 0.063 ⫹0.025/⫺0.018 (range, LP– 0.2). Despite isovolemic hemodilution therapy, no patient showed an increase in VA between first presentation and preoperative measurement. The
™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 Figure 2. Successful recombinant tissue plasminogen activator injection towards the periphery (blue arrow, vein; black arrow, artery). Left, Artery and vein before injection. Middle, Artery and vein at start of injection. Right, Artery and vein during injection. Figure 4. Decimal visual acuity (VA) results 1 year postoperatively. Diagonal small lines, 2-line difference in VA; Œ, lysis not successful. Figure 5. Retinal endovascular lysis. Injection towards optic nerve head (arrow). Left, Preoperative early fluorescein angiography. A ⫽ branch of retinal artery (white); V ⫽ branch of retinal vein (dark). Colored pictures, During the injection, the optic nerve head and all central vessels turned pale. Figure 6. Bleeding from the cannulated area (circle) started when reflux from the lower branch of the central retinal vein (black arrow) was completed. Yellow arrow, peripheral retinal vein.
718
719
Ophthalmology Volume 114, Number 4, April 2007 Table 2. Patient Surgical Characteristics No.
Attempts of Venous Puncture
Intraoperative Complications
1 2
1 5
— Serous retinal detachment
— Proliferative retinopathy Glaucoma
1 4
3
3
rt-PA intraretinal Vitreous hemorrhage
4
4
3
Retinal detachment Cataract Glaucoma Luxation of lens Vitreous hemorrhage Cataract Retinal detachment
5 6 7 8 9
1 2 1 1 3
Cataract
1 0 0 0 4
10 11
2 2
12
1
13
2
—
Vitreous hemorrhage — — Vitreous hemorrhage —
— — Vitreous hemorrhage —
Postoperative Complications
— Glaucoma — Vitreous hemorrhage Glaucoma Painful phthisis — Retinal detachment Redetachment Glaucoma Painful phthisis —
No. of Reoperations
4
0 2 2 0
Reoperations Cataract surgery CPC and peripheral retinal cryocoagulation Vitrectomy, laser, Vitrectomy, laser, and oil filling Oil removal Encircling band, vitrectomy, laser Cataract surgery Gas removal Repositioning of IOL Vitrectomy, laser, and oil filling Oil removal, membrane peeling, laser Cataract surgery, encircling band, oil filling Oil removal, retinotomy, laser, and gas filling Cataract surgery — — — Vitrectomy and laser Peripheral retinal cryocoagulation, CPC Vitrectomy and laser Enucleation — Encircling band, gas filling Vitrectomy, laser, and gas filling CPC, peripheral retinal cryocoagulation Enucleation —
CPC ⫽ cyclophotocoagulation; IOL ⫽ intraocular lens; rt-PA ⫽ recombinant tissue plasminogen activator.
interval between the 2 measurements was 4.2⫾2.7 weeks (range, 0 – 8). Six-week postoperative VA was 0.049 ⫹0.024/⫺0.016 (range, LP– 0.4); 3-month postoperative VA, 0.043 ⫹0.019/ ⫺0.014 (range, LP– 0.3); 6-month postoperative VA, 0.035 ⫹0.022/⫺0.013 (range, blindness– 0.4); and 12-month postoperative VA, 0.04 ⫹0.026/⫺0.016 (range, blindness– 0.4) (Fig 1, Table 1). None of these differences was significant. Visual acuity worsened 1 year after retinal endovascular lysis by about 1.923⫾1.619 lines (6 to ⫺16 lines). Again, this difference was not significant (P ⫽ 0.26). In evaluating the patients individually, VA improved in 3 eyes by ⬎2 lines, 4 eyes remained stable within ⫾2 lines, and 6 eyes worsened by ⬎2 lines (Fig 4). Although 2 globes had to be enucleated, we added the final VA (blindness/0.00125) to present the realistic situation 1 year after initial surgery. All patients had ischemic CRVO according to the 1995 Central Vein Occlusion Study’s definition, with a mean of 52.7⫾11.2 disc areas of nonperfusion,23 although the VA of 3 eyes was 0.2. The mean relative afferent pupillary defect was 1.12⫾0.15 log units.
Surgical Procedure Posterior vitreous detachment and venous piercing were achieved in all eyes. Multiple attempts to pierce a vein were necessary in 8 eyes, as summarized in Table 2. Injection was regarded successful in 10 eyes, as demonstrated in Figure 2 (primary venous whitening). Despite high-quality videotaping (Videos 1, 2 [available at http://aaojournal.org]), we could not determine retrospectively for certain whether venous whitening actually was caused by successful intraluminal recombinant tissue plasminogen activator injection in all cases. It also may have been caused by increased
720
periluminal pressure due to paravasal injection into the optic nerve head tissue. Those eyes (n ⫽ 3) were assessed as having unsuccessful lysis. Nevertheless, we noticed the optic nerve head turning pale most frequently during the injection. To explain these findings, the most likely explanation is that the thrombus is neither dissolved nor flushed away, but rather the central retinal vein is bloated by the injection of recombinant tissue plasminogen activator. The bloated vein may have compressed the central artery (which usually courses very close to the vein) and hampered the arterial inflow, mimicking a compartment syndrome. Hence, the optic nerve became pale, and the vessels collapsed (Fig 5). Upon removal of the cannula, the piercing site did not start bleeding immediately, but blood was observed when the cannulated vein filled up completely due to central blood reflux. We could not identify bleeding from the periphery, indicating an unchanged venous outflow (Fig 6). However, visual improvement did not depend on successful lysis (Fig 4). Visual changes were also independent of the duration of ischemic CRVO within the group of patients with successfully pierced retinal veins.
Problems and Complications Intraoperative Complications. Retinal endovascular lysis was feasible, but we often needed more than one piercing attempt for successful injection. After touching the selected vein, some vessels became constricted and did not allow further piercing. We occasionally perforated the vein and mistakenly injected recombinant tissue plasminogen activator into the retinal tissue or the subretinal
Feltgen et al 䡠 Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion space. In such a case, one can pull back the cannula into the lumen of the vein. Although it is not difficult to hold the cannula within the vein, infusion pressure frequently causes the recombinant tissue plasminogen activator solution to leak into the surrounding tissue. When extracting the glass pipette from the vein, the wound closes within a few seconds to minutes, allowing only small amounts of blood to escape. Postoperative Complications. Postoperative complications (Table 2) within the first year after initial surgery were frequent. Only one patient suffered from heavy vitreous bleeding during the first postoperative day. Despite panretinal laser coagulation, 6 eyes developed neovascular glaucoma, of which 2 globes ended with painful phthisis. Retinal detachment was found in 3 eyes and cataract in 4. Together, the 13 eyes required 22 additional surgical procedures, including cataract surgery (4 eyes); transscleral cyclophotocoagulation and peripheral cryocoagulation (3 eyes); vitrectomy, laser coagulation, and oil filling (4 eyes); encircling band and vitrectomy for RD (3 eyes [1 eye with redetachment]); and enucleation of the globe (2 eyes).
Angiographic Findings Postoperative control was possible in 12 patients at 3 months. Later controls could not be performed in all patients because of corneal edema, anterior chamber or vitreous hemorrhage, RD, cataract, or patient unwillingness. The mean early AP was 5.03⫾0.4 seconds before surgery and 8.1⫾7.0 seconds postoperatively. This increase in AP time was not significant (P ⫽ 0.39). The maximum filling AP increased from 16.5⫾1.2 seconds before surgery to 20.1⫾7.0 seconds postoperatively, which was also not significant (P ⫽ 0.73). We measured the FAZ to assess pole ischemia, but in 10 eyes the macula was partly covered with blood. The mean FAZ was 0.87⫾0.56 mm2.
Discussion In the management of acute major vessel occlusion, the use of fibrinolytic agents such as recombinant tissue plasminogen activator is widely accepted. Today, the spectrum of indications for thrombolytic drugs comprises acute myocardial infarction, lung embolism, ischemic stroke, deep vein thrombosis, and acute arterial occlusions of the lower limbs. Based on the histopathological and clinical features of retinal vessel occlusion, fibrinolysis aiming at early restoration of blood flow appears to be a promising therapeutic approach. However, there are serious concerns regarding the systemic use of lytic agents in retinal vein occlusion because of considerable risk of cerebral or gastrointestinal hemorrhage.13,14 To avoid life-threatening complications, the local endovascular application of lytic agents directly was developed.8 Weiss and Bynoe reported the results of this novel surgical technique with various series of CRVO patients. Of those, up to 54% recovered ⬎2 lines of VA within 3 months, whereas 50% reached this goal even after a mean follow-up of 12 months.19,20,27 This compares favorably to the natural-course studies with improvement rates of only 10% to 14%.4 – 6,28 In those studies, patients with poor initial VA (⬍0.1) had an 80% risk of having VA⬍0.1 at final visit, whether perfused or not perfused initially.23 The data from those initial studies are quite inhomogeneous. Central retinal vein occlusion duration varied be-
tween 0.25 and 30 months, both ischemic and nonischemic patients were included, and different volumes of recombinant tissue plasminogen activator solution were injected (0.6 –7.5 ml). To validate those initial results, we performed this pilot study on a narrowly defined group of patients with severe ischemic CRVO, including all preoperative and postoperative controls at our study center over a 1-year period. As opposed to the initial studies, our data did not rely on the referring ophthalmologists. In short, our visual results after vitrectomy and retinal endovascular lysis are clearly disappointing. Three eyes improved by ⬎2 lines, and only 1 patient had a final VA of ⬎0.1—results that closely resemble the natural course of the ischemic type of CRVO. Additionally, our angiographic data showed an increase in AP after surgery that is worse than the preoperative data. Although these changes were not significant, one would have expected a reduction in AP time by flushing the veins. It is very unlikely that isovolemic hemodilution, which is performed routinely in our hospital and by the local ophthalmologists, negated our retinal endovascular lysis results. Previous studies have shown that if visual improvement occurs, it will do so within the first 6 weeks of therapy.4,5 Yet none of our 13 patients presented improved vision within the preoperative period (Fig 1). What might explain the failure of retinal endovascular lysis in our case series? 1. Did we operate too late? Low venous perfusion with subsequent thrombus formation and organization11,29 causes CRVO and determines the prognosis. The duration of CRVO in our patients was between 7 and 18 weeks; hence, thrombi were rather old and possibly less amenable to easy dissolution. Thrombus flushing was the explanation30 for the good results obtained in the Weiss and Bynoe studies. We found it possible to flush in a retrograde manner, but this was not always feasible towards the optic nerve head. Furthermore, it is difficult to understand how a well-organized thrombus can be loosened from the wall and flushed away, as thrombus organization includes capillary growth and ultimately fibrous scarring.30 We therefore find it improbable that simply flushing it with thrombolytic agents can dissolve an already organized thrombus. On the other hand, even ischemic CRVO is seldom caused by a complete obstruction of the central retinal vein.31 It therefore may be possible to improve overall venous flow by flushing the remaining venous lumen, hence reducing the macular edema. Unfortunately, we have no means of telling how old a thrombus is in CRVO eyes. For ethical reasons, we decided to operate on patients with older CRVOs so as not to mask spontaneous recovery. Theoretically, retinal endovascular lysis performed within the first hours or days should be more effective. The most likely explanation based on our intraoperative findings is that the thrombus itself remained unchanged. However, the already dammed central ret-
721
Ophthalmology Volume 114, Number 4, April 2007 inal vein was bloated further by the injection of additional fluid, thus compressing the central artery and hampering arterial inflow. 2. Were our patients’ eyes too ischemic for any therapy to be visually effective? Our poor results also may be due to our negatively selected group of patients in which we treated only eyes with severe ischemia, which in turn have a higher risk of neovascular complications such as glaucoma or vitreous bleeding.32 In our study, 6 of 13 (46.2%) patients developed neovascularization, which corresponds closely to the high rate of iris neovascularization (35%) in ischemic CRVO eyes.23 The high complication rate and high number of additional surgical procedures required in our patient series are remarkable. Five of 12 phakic eyes needed cataract surgery. Other severe complications were RD, repeated vitreous hemorrhage, and phthisis. Despite intensive treatment, neovascular glaucoma could not be avoided in all cases, and 2 patients even required enucleation of the globe. This negative selection may be a drawback of the present study. But we identified 10 ischemic CRVO patients in the Weiss and Bynoe study as well,19 and it is astonishing that their visual improvement and complication rate were independent of both initial VA and retinal perfusion (perfused, nonperfused, or intermediate, according to the 1993 Central Vein Occlusion Study32). 3. Does the surgeon need to be more experienced in retinal endovascular lysis? One also must consider the learning curve in this sophisticated surgical method. Puncture and flushing of the vein were certainly feasible in most patients. But we did not demonstrate a direct measurable improvement in the flow as indicated by angiographic data. Perhaps the short exposure time of about 30 to 60 seconds was not sufficient to dissolve a thrombus 7 to 18 weeks after formation. To summarize, retinal endovascular lysis is feasible but remains a sophisticated surgical method with many complications. Although we anticipated postoperative complications in eyes with severe retinal ischemia, numbers of complications and necessary additional surgical procedures were unacceptably high. As we have not been able to reproduce the good results of others in our pilot study, we can neither recommend nor give up on this new surgical method unless it is tested on patients suffering from fresh CRVO. However, without demonstrating a distinct improvement in VA, it is hard to justify the increased risk due to surgical complications.
References 1. Hayreh SS. Retinal vein occlusion. Indian J Ophthalmol 1994; 42:109 –32. 2. Hayreh SS. Prevalent misconceptions about acute retinal vascular occlusive disorders. Prog Retin Eye Res 2005;24:493– 519.
722
3. Williamson TH. Central retinal vein occlusion: what’s the story? Br J Ophthalmol 1997;81:698 –704. 4. Hansen LL, Danisevskis P, Arntz HR, et al. A randomised prospective study on treatment of central retinal vein occlusion by isovolaemic haemodilution and photocoagulation. Br J Ophthalmol 1985;69:108 –16. 5. Hansen LL, Wiek J, Schade M, et al. Effect and compatibility of isovolaemic haemodilution in the treatment of ischaemic and non-ischaemic central retinal vein occlusion. Ophthalmologica 1989;199:90 –9. 6. Wolf S, Arend O, Bertram B, et al. Hemodilution therapy in central retinal vein occlusion: one-year results of a prospective randomized study. Graefes Arch Clin Exp Ophthalmol 1994; 232:33–9. 7. Hayreh SS. Management of central retinal vein occlusion. Ophthalmologica 2003;217:167– 88. 8. Weiss JN. Treatment of central retinal vein occlusion by injection of tissue plasminogen activator into a retinal vein. Am J Ophthalmol 1998;126:142– 4. 9. Opremcak EM, Bruce RA, Lomeo MD, et al. Radial optic neurotomy for central retinal vein occlusion: a retrospective pilot study of 11 consecutive cases. Retina 2001;21:408 –15. 10. Hattenbach LO, Steinkamp G, Scharrer I, Ohrloff C. Fibrinolytic therapy with low-dose recombinant tissue plasminogen activator in retinal vein occlusion. Ophthalmologica 1998; 212:394 – 8. 11. Green WR, Chan CC, Hutchins GM, Terry JM. Central retinal vein occlusion: a prospective histopathologic study of 29 eyes in 28 cases. Retina 1981;1:27–55. 12. Kohner EM, Pettit JE, Hamilton AM, et al. Streptokinase in central retinal vein occlusion: a controlled clinical trial. Br Med J 1976;1:550 –3. 13. Elman MJ. Thrombolytic therapy for central retinal vein occlusion: results of a pilot study. Trans Am Ophthalmol Soc 1996;94:471–504. 14. Hattenbach LO, Wellermann G, Steinkamp GW, et al. Visual outcome after treatment with low-dose recombinant tissue plasminogen activator or hemodilution in ischemic central retinal vein occlusion. Ophthalmologica 1999;213:360 – 6. 15. Paques M, Vallee JN, Herbreteau D, et al. Superselective ophthalmic artery fibrinolytic therapy for the treatment of central retinal vein occlusion. Br J Ophthalmol 2000;84: 1387–91. 16. Lahey JM, Fong DS, Kearney J. Intravitreal tissue plasminogen activator for acute central retinal vein occlusion. Ophthalmic Surg Lasers 1999;30:427–34. 17. Glacet-Bernard A, Kuhn D, Vine AK, et al. Treatment of recent onset central retinal vein occlusion with intravitreal tissue plasminogen activator: a pilot study. Br J Ophthalmol 2000;84:609 –13. 18. Elman MJ, Raden RZ, Carrigan A. Intravitreal injection of tissue plasminogen activator for central retinal vein occlusion. Trans Am Ophthalmol Soc 2001;108:219 –21, discussion 222–3. 19. Weiss JN, Bynoe LA. Injection of tissue plasminogen activator into a branch retinal vein in eyes with central retinal vein occlusion. Ophthalmology 2001;108:2249 –57. 20. Bynoe LA, Hutchins RK, Lazarus HS, Friedberg MA. Retinal endovascular surgery for central retinal vein occlusion: initial experience of four surgeons. Retina 2005;25:625–32. 21. Magargal LE, Donoso LA, Sanborn GE. Retinal ischemia and risk of neovascularization following central retinal vein obstruction. Ophthalmology 1982;89:1241–5. 22. Hayreh SS, Klugman MR, Beri M, et al. Differentiation of ischemic from non-ischemic central retinal vein occlusion
Feltgen et al 䡠 Retinal Endovascular Lysis in Ischemic Central Retinal Vein Occlusion
23. 24. 25. 26.
27.
during the early acute phase. Graefes Arch Clin Exp Ophthalmol 1990;228:201–17. Central Vein Occlusion Study Group. Natural history and clinical management of central retinal vein occlusion. Arch Ophthalmol 1997;115:486 –91. Servais GE, Thompson HS, Hayreh SS. Relative afferent pupillary defect in central retinal vein occlusion. Ophthalmology 1986;93:301–3. Holladay JT. Proper method for calculating average visual acuity. J Refract Surg 1997;13:388 –91. Schulze-Bonsel K, Feltgen N, Burau H, et al. Visual acuities “hand motion” and “counting fingers” can be quantified with the Freiburger visual acuity test. Invest Ophthalmol Vis Sci 2006;47:1236 – 40. Bynoe LA, Weiss JN. Retinal endovascular surgery and intra-
28. 29. 30. 31. 32.
vitreal triamcinolone acetonide for central vein occlusion in young adults. Am J Ophthalmol 2003;135:382– 4. Quinlan PM, Elman MJ, Bhatt AK, et al. The natural course of central retinal vein occlusion. Am J Ophthalmol 1990;110: 118 –23. Cotran RS, Pober JS. Effects of cytokines on vascular endothelium: their role in vascular and immune injury. Kidney Int 1989;35:969 –75. Hayreh MS. t-PA in CRVO. Ophthalmology 2002;109:1758 – 61. Seitz R. The retinal blood vessels. Comparative ophthalmoscopic and histological studies on the healthy and pathologically changed eye [in German]. Buch Augenarzt 1962;40:1–175. Baseline and early natural history report: the Central Vein Occlusion Study. Arch Ophthalmol 1993;111:1087–95.
723
Ophthalmology Volume 114, Number 4, April 2007
Figure 3. Successful retinal endovascular lysis. The most important parts of Video 2 are shown. The thrombus (T) remains unchanged during the injection. The vein (V) is bloated by the injection of recombinant tissue plasminogen activator. The arterial (A) inflow is hampered and the vessels collapse.
723.e1