Purpose
To evaluate the outcome of pars plana vitrectomy, subretinal tissue plasminogen activator (t-PA) infusion and intraocular gas tamponade with and without postsurgical antivascular endothelial growth factor (VEGF) injection for thick submacular hemorrhage due to exudative age-related macular degeneration (AMD).
Design
Retrospective, comparative, interventional case series.
Methods
setting : 2 retina referral centers. The patient population included 101 eyes of 101 patients with neovascular AMD and thick submacular hemorrhage who underwent surgical displacement of the hemorrhage with or without postoperative anti-VEGF injections. Main outcome measures included degree of blood displacement, best and final postoperative visual acuity (VA), and adverse events. Snellen acuity was converted to logMAR for statistical analysis.
Results
All patients were followed for a minimum of 3 months (mean, 15.3 months, range, 3-70 months). In 83 (82%) of 101 eyes, the procedure resulted in complete hemorrhage displacement from the fovea. Mean preoperative VA was 20/2255 (2.05 logMAR). The acuity significantly improved to 20/893 (1.65 logMAR) at month 1 ( P < 0.001) at month 1; 20/678 (1.53 logMAR) at month 3 ( P < 0.001), and 20/1150 (1.76 logMAR) at month 12 ( P = 0.002). Best postoperative visual acuity improved by at least 1 line in 83 (82%) of 101 eyes, and 19.6% of eyes gained 3 lines or more at month 3. The visual acuity of the group of eyes that received postoperative anti-VEGF injection (n = 39) showed greater visual acuity improvement 6 months postoperatively compared to the group of eyes that did not receive postoperative anti-VEGF. Postoperative complications included vitreous hemorrhage in 2 eyes, rhegmatogenous retinal detachment in 4 eyes, and recurrent thick subretinal hemorrhage in 6 eyes.
Conclusions
Vitrectomy with subretinal t-PA injection and gas tamponade was found to be relatively effective for displacement of thick submacular hemorrhage with a significant improvement in visual acuity. There is a loss of acuity over time; the addition of postoperative anti-VEGF therapy may help maintain the visual acuity gains.
Thick submacular hemorrhage can cause sudden and severe vision loss in patients with age-related macular degeneration (AMD) and often results in visual acuity of 20/200 or worse. Thick submacular hemorrhage interferes with retinal function through several mechanisms, including shearing of photoreceptors by fibrin clots and physical separation of photoreceptors from the retinal pigment epithelium (RPE) and choroid, and through direct toxic effects of iron.
Because of the poor prognoses these patients commonly experience, there have been various attempts to remove the blood from the submacular space. There are several reports of pars plana vitrectomy (PPV) with active clot removal, either with or without tissue plasminogen activator (t-PA). Thus far, the results of these approaches have been disappointing. Others have used intravitreal gas with or without t-PA as a less invasive and less costly approach. (Heriot WJ. Intravitreal gas and t-PA: an outpatient procedure for submacular hemorrhage. Paper presented at AAO Annual Vitreoretinal Update, 1996; Chicago, IL).
Although t-PA may help to liquefy the clot and aid in hemorrhage displacement, there is experimental evidence to suggest that t-PA may not cross the neurosensory retina. To enable t-PA to more effectively enter the subretinal space, Haupert and associates described a technique of injecting t-PA into the subretinal space through a transretinal small-gauge cannula. A larger series was published a few years later; although encouraging, these results were limited by small size and relatively short follow-up. Also, these studies were performed in the era before intravitreal antivascular endothelial growth factor (VEGF) use was available to control the underlying exudative AMD. The purpose of this study was to evaluate the outcome of PPV, subretinal infusion of t-PA, and intraocular gas tamponade with and without postoperative intravitreal anti-VEGF injection for the treatment of thick submacular hemorrhage caused by AMD.
Methods
After Wills Eye Hospital Institutional Review Board approval was obtained, the medical records of 101 consecutive patients from 2 centers (Wills Eye Hospital and Midwest Eye Institute) with AMD and thick submacular hemorrhage who underwent PPV with subretinal t-PA injection and pneumatic displacement between April 1, 2001, and September 31, 2009, were reviewed. A thick submacular hemorrhage was defined as blood under the fovea causing an obvious elevation of the retina, with obscuration of the RPE on biomicroscopic examination. Patients with thick submacular hemorrhages treated with anti-VEGF therapy alone were not offered the procedure and were excluded from the study. Patients who were thought clinically to have extensive hemorrhage under the RPE, as well as patients with massive subretinal hemorrhage (defined as hemorrhage extending to the equator) were excluded. Although the duration of the hemorrhage was not an exclusion criterion, patients with chronic, khaki-colored hemorrhage generally were not offered the procedure. Optical coherence tomography (OCT), fluorescein and/or indocyanine green angiography, B-scan ultrasonography, and color fundus photography were performed at the discretion of the examining surgeon. However, as the hemorrhage often limited image quality, preoperative imaging was inconsistently performed. All operations were performed at the Wills Eye Hospital (Philadelphia, PA) or at the Midwest Eye Institute (Indianapolis, IN). After August 2005, some patients received postoperative anti-VEGF injections at various intervals at the discretion of the treating physician. The patients who underwent PPV with subretinal t-PA are defined as Group A. Those patients who also received postoperative anti-VEGF injection are defined as Group B.
The surgical procedure was as follows. After informed consent was obtained, a standard 3-port PPV was performed and a posterior vitreous detachment was created if not already present. Using a 41-gauge flexible cannula, t-PA (25 μg/0.1 mL) (Wedgewood Pharmacy, Swedenboro, NJ) was slowly injected into the subretinal space, and the injection was stopped once the retina was completely separated from the underlying hemorrhage (approximately 0.25 mL to 0.5 mL of t-PA). Between 1 and 4 separate infusion sites were created in order to detach the retina from the underlying hemorrhage. A subtotal air/fluid exchange (approximately 75%) was performed. The eye was left with either air, 18%-20% SF6 or 12%-14% C3F8 gas fill at the surgeon’s discretion. All patients were instructed to tilt their face forward at 45 to 60 degrees for approximately 1-5 days postoperatively.
The medical records were reviewed for the following data: age, gender, best corrected Snellen visual acuity using either manifest refraction or spectacle correction with pinhole preoperatively as well as at 1, 3, 6, and 12 months postoperatively, duration of hemorrhage, gas used, results of fluorescein angiography before submacular hemorrhage, postoperative fluorescein angiography, use of photodynamic therapy (PDT) and postoperative use of intravitreal anti-VEGF medications. Because OCT image quality is limited in cases with extensive hemorrhage, OCT was obtained sporadically and was not included in the analysis. Visual acuities were analyzed on a logarithm of minimal angle of resolution (logMAR) scale. For this purpose, counting fingers at 6 feet was converted to 48/1600; 3 feet to 24/1600; 1 foot to 20/8000; hand motions to 20/16,000; and light perception to 20/32,000. The degree of blood displacement was determined ophthalmoscopically or with the aid of fundus photographs and was graded as complete, partial or no displacement. Complete displacement was defined as no blood or only a scant amount of blood within 1 disc diameter of the foveal center after completion of prone positioning; partial displacement was defined as blood under the fovea that obscured the retinal pigment epithelium but that did not cause clinically visible elevation of the retina. The use of postoperative anti-VEGF medications was left to the discretion of the treating physician. The Wilcoxon signed-rank test was used for comparison of preoperative and postoperative visual acuities, and the chi-square test or the Fisher exact test was used for subgroup analysis. SPSS v 18.0 for Windows (IBM Corporation, Armonk, New York, USA) was used for statistical analysis. All tests were 2-tailed, and significance was defined as P ≤ 0.05.
Results
Included in the study were 101 eyes of 101 patients (61 women, 40 men). The mean age was 80 years (range, 60-98 years). All patients were followed for a minimum of 3 months (mean, 15.3 months; range, 3-70 months). Of 101 patients 95 (94%) had at least 3 months of follow-up; 73/101 (72%) had 6 months of follow-up; 63/101 (62%) had 12 months of follow-up, and 17/101 (16%) had 24 months of follow-up. At baseline, 47 eyes were phakic, 43 were pseudophakic, 2 were aphakic, and 9 did not have phakic status recorded. Of the 101 patients, 51 were taking some form of anticoagulant (aspirin or other platelet inhibitor and/or warfarin). The average duration of submacular hemorrhage before surgery was 16 days (range, 1-60 days). Mean preoperative visual acuity was 20/2255 (range, 20/100 to hand motion). The hemorrhage was completely displaced from the fovea in 83 eyes (82%) ( Figure 1 ). There was partial displacement in 17 eyes (17%) ( Figure 2 ). There was no displacement in 1 eye (1%) that had hemorrhage onset 2 days before surgery. As OCT, B-scan ultrasonography, and indocyanine green angiography were performed in fewer than 20 eyes, imaging analysis was not included in further analysis.
Mean preoperative baseline visual acuity was 20/2255 (2.05 logMAR). The visual acuity was 20/893 (1.65 logMAR) ( P < 0.001) at month 1; 20/678 (1.53 logMAR) at month 3 ( P < 0.001); 20/914 (1.66 logMAR) at month 6 ( P < 0.001); and 20/1150 (1.76 logMAR) at month 12 ( P = 0.002) ( Figure 3 ). Best postoperative visual acuity improved by at least 1 line in 83 of the 101 (82%) eyes, and 19.6% of eyes gained 3 lines or more at month 3. Postoperatively, 12 eyes showed no improvement, and 6 eyes had at lost at least 1 line of vision. Visual acuity at last follow-up improved by at least 1 line in 62 eyes (61%); remained unchanged in 19 (19%) eyes; and worsened by 1 line or more in 20 (20%) eyes ( P < 0.001, Figure 4 ). The visual acuities of the 45 patients who were pseudophakic at baseline were 20/4477 (logMAR 2.35) preoperatively; 20/957 (logMAR 1.68, n = 45) at month 3; 20/1074 (logMAR 1.73, n = 36) at month 6; and 20/1663 (logMAR 1.92, n = 28) at month 12. Linear regression analyses identified no statistically significant correlation between hemorrhage duration and postoperative visual acuity at any time point (final visual acuity, best visual acuity; P = 0.999, P = 0.375, respectively). Duration of hemorrhage of <2 weeks, 2 to 4 weeks or longer than 4 weeks also did not appear to correlate with visual outcome (baseline vs final logMAR P = 0.571, baseline vs best logMAR P = 0.981, respectively.) The longest duration of hemorrhage associated with visual improvement was 41 days in this series.
Prior to 2005, 8 eyes had undergone postoperative PDT. Of the 8 eyes, 7 had had complete anatomic displacement of the hemorrhage prior to PDT. The mean preoperative vision in this subgroup was 20/618 (logMAR 1.49). The best mean postoperative and post-PDT vision was 20/152 (logMAR .88), and the final visual acuity was 20/276 (logMAR 1.14).
In August 2005, intravitreal injection of the anti-VEGF agent bevacizumab (Avastin; Genentech, South San Francisco, California, USA) and shortly thereafter, ranibizumab (Lucentis, Genentech, South San Francisco, California, USA) became available. Prior to occurrence of submacular hemorrhage, 24 patients had been treated with either ranibizumab or bevacizumab using several different treatment regimens, and 21 of these patients continued to receive postoperative anti-VEGF injections. An additional 18 patients who had never received anti-VEGF therapy preoperatively began anti-VEGF therapy postoperatively. Of 101 patients, 59 never received anti-VEGF injections at any point, either because they were treatment-naive or because they presented prior to the routine use of anti-VEGF agents. Peak visual acuity improvement was reached at month 3 in the no postoperative anti-VEGF group (group A, n = 62). The patients who received postoperative anti-VEGF (Group B, n = 39) reached their peak visual acuity at month 6. Baseline visual acuity was better in group A (20/1664 [logMAR 1.92]) than in group B (20/3639 [logMAR 2.26], P = 0.018). By month 3, group A patients had improved to 20/815 (logMAR 1.55, P = 0.001 vs baseline), but by month 6, they had regressed to 20/1100 (logMAR 1.74, P = 0.206 vs baseline) ( Figure 5 ). Eyes in Group B received intravitreal anti-VEGF injection an average of 2.2 months (range, 0.5 to 12.6) after surgery and received a mean of 5.6 (range, 1-12) postoperative injections. By month 3, Group B’s visual acuity had improved to 20/710 (1.55 logMAR) and at month 6 was 20/693 (1.54 logMAR = 0.001 vs baseline).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
