VISUAL FIELD DEFECTS AFTER RADIAL OPTIC NEUROTOMY FOR CENTRAL RETINAL VEIN OCCLUSION

Transcription

1 VISUAL FIELD DEFECTS AFTER RADIAL OPTIC NEUROTOMY FOR CENTRAL RETINAL VEIN OCCLUSION ADIEL BARAK, MD, ANAT KESLER, MD, DANIEL GOLD, MD, ANAT LOEWENSTEIN, MD Purpose: Surgical decompression of the vein in central retinal vein occlusion (CRVO) by radial optic neurotomy (RON) was recently proposed as being surgically feasible, safe, and beneficial. The effect of RON on the visual field has not been systematically reported, although possible visual field defects are expected due to iatrogenic cutting of the optic nerve fibers. The authors report the results of visual field testing in patients who underwent RON at the Tel Aviv Medical Center. Methods: Twelve consecutive patients (8 men, mean age of all patients 68 years) with nonperfused or indeterminate CRVO whose initial visual acuity (VA) was 20/400 underwent RON. Nine of these patients were able to perform visual field tests at 6 months post-ron and their visual field results are presented. Results: VA improved by 3 lines in 5 patients (42%). Three (25%) patients had a final VA of 20/200 and another had a final VA 20/50. Fundus and VA improvement were relatively slow. Two patients had clearing of the intraretinal blood, resolution of the venous dilation, and improved VA at the 2-month follow-up visit. Temporal visual field defects consisting of temporal ones that could be correlated to the site of the RON incision were detected in five out of the nine patients who were able to perform visual field tests. No temporal visual field defects were found in the remaining four patients, and three other patients were unable to perform visual field tests due to inability to identify the largest target size. Conclusion: VA improvement in 5 of 12 patients with ischemic or indeterminate CRVO following RON may be better than the natural history of CRVO. The risk of visual field defects may, however, be heightened by possibly cutting off blood supply to the optic nerve head and possible damage to nerve fibers in the optic nerve head, both inherent to the surgical procedure and both likely to produce visual loss. RETINA 26: , 2006 From the Department of Ophthalmology, Tel Aviv Sourasky Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. Reprint requests: Adiel Barak, MD, Department of Ophthalmology, Tel Aviv Sourasky Medical Center, 6 Weitzman Street, Tel Aviv 64239, Israel; adielbarak@yahoo.com Central retinal vein occlusion (CRVO) is a common vision-impairing disease whose management remains highly controversial. CRVO consists of three distinct types, perfused, nonperfused, and indeterminate, each with very different prognoses, complications, visual outcomes, and management. 1 The nonperfused type of CRVO has been shown to be associated with poor prognosis for the recovery of vision. One prospective study found that the final visual acuity (VA) was 20/400 or worse in 87% of nonperfused patients, whereas it was 20/30 or better in 57% of perfused cases. 2 In a retrospective study, 3 93% of eyes with nonperfused CRVO had a final VA of less than 20/200, in contrast with 50% of perfused 549

2 550 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2006 VOLUME 26 NUMBER 5 Visual Field (Last Visit) Follow-up, mo Final VA VA Table 1. Patient Characteristics CRVO Type NVG/ NVI Duration, mo DM Glaucoma HBP Sex No defect 12 20/100 10/200 N 3 M 67 1 Temporal defect 11 20/80 FC N 3 F 42 2 Temporal defect 11 20/50 10/200 N 1 M 69 3 NA 11 HM HM I 4 M 47 4 NA 10 5/200 HM I 9 M 72 5 No defect 9 10/200 HM N 6 F 76 6 No defect 8 10/200 FC I 9 F 77 7 NA 8 HM HM I 16 M 75 8 No defect 7 5/200 5/200 N 4 M 62 9 Temporal defect 7 15/200 FC N 3 M Temporal defect 6 20/200 20/200 N 3 F Temporal defect 6 20/200 10/200 N 5 M Age, yr Patient No. HBP high blood pressure; DM diabetes mellitus; NVG neovascular glaucoma; CRVO central retinal vein occlusion; VA visual acuity; N nonperfused; FC finger counting; I indeterminate; HM hand movement; NA not applicable. eyes. As part of the Central Vein Occlusion Study Group (CVOS) investigation, a subgroup of 78 eyes with CRVO of varying duration, perfusion status (i.e., almost 21% of the group had type I [indeterminate] or N [nonperfused]), and baseline acuity level (between 20/50 and 5/200, a range of 12 acuity lines) were observed without intervention, and changes in their VA were measured. 3 Spontaneous improvement in VA tended to be infrequent and of small magnitude: there were only 4 eyes (6%) that gained at least three lines of vision compared with baseline acuity, and only two recovered at least four lines within the first year of follow-up (only three eyes reached this level of visual improvement after 3 years of follow-up). Although CVRO is a common disease, there is no consensus on therapeutic approach, and its management remains highly controversial. Opremcak et al 4 recently described a procedure for surgical decompression of the vein by making a radial cut from the vitreous side in the optic nerve head, extending all the way down to the lamina cribrosa and the adjacent sclera and cutting the arterial circle of Zinn and Haller, a procedure they called radial optic neurotomy (RON). They stated that RON is a surgically feasible and safe procedure and one that is beneficial in cases of CRVO. This surgery was performed in a series of 11 consecutive patients with nonperfused and indeterminate CRVO whose VAs were 20/400 or less. All the patients had clinical improvement as determined by fundus examination, photography, and fluorescein angiography. The postoperative VAs were the same or improved in 9 of their 11 (82%) patients. Eight of the 11 patients (73%) had rapid improvement of their VAs, with an average gain of five lines of vision. There were no complications noted with this procedure, although visual field examinations were not performed pre- and post-ron. 5 The lack of visual field examination raised some concern regarding the safety of the procedure. 6 We performed RON in 12 consecutive patients with nonperfused and indeterminate CRVO, and conducted visual field examinations on all of them 6 months after the procedure. We now report our surgical results and the results of our patients visual field examinations. Patient Selection Criteria Materials and Methods Patients were selected on a nonrandomized, caseby-case basis (Table 1). They were eligible for RON if they had either nonperfused or indeterminate CRVO with an initial VA of 20/400 or worse. There was a total of 12 patients and each one had intraretinal hemorrhage and macular edema. Eight were classified as nonperfused and four as indeterminate, the latter due to the presence of extensive retinal hemorrhage or vitreous hemorrhage that precluded accurate interpretation of fluorescein angiograms. Four patients had anterior segment neovascularization including iris neovascularization and two had neovascular glaucoma. Natural history study data were retrieved and informed consent was obtained from each patient. No patient was excluded on the basis of the duration of their CRVO. All the enrolled study patients underwent a comprehensive ophthalmologic examination, including significant past medical and ocular history, best-corrected Snellen VA, anterior segment biomicroscopy and gonioscopy, measurement of intraocular pressure (IOP) and dilated fundus examination with contact lens biomicroscopy, and indirect ophthalmoscopy. The date of CRVO onset was obtained from the his-

3 VISUAL FIELD DEFECTS AFTER RADIAL OPTIC NEUROTOMY FOR CRVO BARAK ET AL 551 torical information and a review of records from referring physicians. Twelve patients elected to undergo RON and underwent initial fundus examination, photography, and fluorescein angiography during the 2 weeks preceding surgery. The fluorescein angiographies were classified according to CVOS criteria 7 as perfused, nonperfused, or indeterminate. They were followed up postoperatively with serial best-corrected Snellen VA, IOP measurements, slit-lamp examination, and detailed biomicroscopic examination at 1 day, 1 week, 1, 2, 3, and 6 months, and at the final visit. The VA of patients who had a score of less than 20/400 Snellen (i.e., inability to see the largest projected Snellen optotype) was measured using a 20/200 E card held at varying premeasured distances until the patient reported being able to see it (the optotype was always held directly in front of the patient, but eccentric gaze was permitted). Vision was judged to be finger counting or hand motions if the hand-held optotype could not be visualized at 1 m but the patient could still perceive finger counting or hand motions. Finger counting, hand motions, and light perception acuity were assigned logmar scores that were 0.1 unit (or one line of acuity) higher than the logmar score corresponding to the lowest acuity measured with optotypes (5/200 Snellen). Surgical Procedure The surgical procedure was performed as described by Opremcak et al, 4 modified by the addition of laser photocoagulation in patients with anterior segment neovascularization. A standard three-port vitrectomy was performed, and the IOP was raised to minimize any potential bleeding. The RON procedure was always performed on the nasal side of the disk, and the incision was radial to the optic disk and parallel to the nerve fiber pattern. The tip of the microvitreoretinal (MVR) blade (Mani ophthalmic knife, MVR-Lance 20 G, Mani Inc., Utsunomiya Tochigi, Japan) was placed at the edge of the optic disk. The blade was then directed posteriorly into the optic nerve, ideally cutting an equal portion of the cribriform plate and adjacent sclera. The depth of the incision into the optic nerve placed the MVR blade just beyond the widest portion of the diamond-shaped tip. Laser photocoagulation was performed in all the patients with anterior segment neovascularization. Visual Field Examinations Visual field examinations using Goldmann kinetic perimetry (Haag-Streit AG Ophthalmologic Instruments, Bern, Switzerland) were performed on all 12 Fig. 1. Visual acuity (LogMAR) before and 6 months after RON. Improvements of three or more lines of VA were detected in 42% of the patients. patients at the 6-month follow-up visit. The largest/ brightest target (V-4e) stimulus with an average diameter of 1.72 degrees (64 mm 2 at 30 cm) was used for all patients. By following this protocol, even patients with only hand movement vision (or any vision better than light perception) can see the target. 8 Results Twelve of the suitable patients elected to undergo RON (Table 1). Their average age was 69 years (range years). Eight of them (67%) were men. CRVO affected the left eye in eight patients. The average duration of the CRVO before RON was 7 months (range 1 16). Seven of the 12 patients (58%) had CRVO symptoms for more than 3 months. Six patients (50%) had anterior segment neovascularization. RON was successfully performed in 12 consecutive patients and by a single surgeon (A.B.). The average follow-up after surgery was 6.5 months (range 6 12 months). There was an improvement of VA of three lines or more in 5 of the 12 patients (42%) (Figure 1). Five of the 12 (25%) patients had a final VA of 20/200 or better and 1/12 had a final VA of 20/50. Fundus and VA improvement was relatively slow. Two of the 12 patients had clearing of the intraretinal blood, resolution of the venous dilation, and improved VA at the 2-month follow-up visit. A comparison of the data revealed that patients with a shorter duration of CRVO had a better chance of regaining vision than patients with long-standing CRVO. Three out of five patients with CRVO of 3 months or less had improvement in their VA compared to two of seven patients with a longer duration. The most striking improvement was noted in the one patient who was operated on within 1 month of the initiation of symptoms of CRVO. No retinal or vitreous hemorrhages were detected during or directly after the procedure. A late vitreous hemorrhage developed in two patients at 2 and 4

4 552 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2006 VOLUME 26 NUMBER 5 our series were classified as being nonperfused before surgery, and none had glaucoma or iris neovascularization. Three of the five patients with visual field defects had a temporal visual field scotoma which interfered with their performance of daily activities, such as reading. Discussion Fig. 2. Patient 4, Table 1. A, Fundus photography 16 months after radial optic neurotomy. A nasal stab incision can be seen in the optic nerve head, with no significant macular hemorrhage. B, Goldmann visual field showing a temporal visual field defect correlating to the nasal optic nerve incision. months post-ron, and all of them cleared up without further intervention. We believe the vitreous hemorrhage had no direct relation to the RON, and that it was probably related to the natural history of the nonperfused CRVO. One patient (no. 2, Table 1) developed a retinal detachment due to peripheral break and was reoperated with scleral buckle and gas tamponade 3 months after undergoing RON. Temporal visual field defects corresponding to the site of the RON incision nasally were detected in five out of the nine patients (Table 1 and Figure 2) who were able to perform visual field tests. No temporal visual field defects were found in the remaining four patients who performed visual field tests, whereas the remaining three patients were unable to identify the largest/brightest (V-4e) target stimulus in the Goldmann kinetic perimeter due to their low VA. Visual field defects were detected in four out of five patients with final VA of 20/200 or better. There was no correlation between initial VA and subsequent visual field defects. All patients with visual field defects in The results of the current study showed an improvement of VA in 5 out of 12 patients with ischemic or indeterminate CRVO following RON, an outcome that may be better than the natural history of CRVO. As part of the CVOS investigation, a subgroup of 78 eyes with CRVO of varying duration, perfusion status (i.e., almost 21% of the group had type I or N angiograms), and baseline acuity level (between 20/50 and 5/200, a range of 12 acuity lines) were observed without intervention, and changes in their VA were measured. 3 Spontaneous improvement in VA tended to be infrequent and of small magnitude: there were only 4 eyes (6%) that gained at least three lines of vision compared with baseline acuity, and only two recovered at least four lines within the first year of follow-up (only three eyes reached this level of visual improvement after 3 years of follow-up). A direct comparison of this group can be made with the 12 eyes in our study, in which VA improved three lines or more in 5 of them (41%). Visual acuity of 20/200 or better was achieved in 5 of our 12 patients (41%) at 6 months following surgery. While our results are better than the natural history reports, they are poorer than those of published data. Opremcak et al 9 reported that RON was successfully performed in 100 patients with CRVO who had a VA of 20/200 or worse. Eleven patients had well-controlled intraoperative bleeding and four patients had postoperative vitreous hemorrhages. There was clinical and angiographic improvement in 95% of their patients, and VA improved by an average of four lines in 71%. 9 Weizer et al 10 reported a small series of five patients (four with CRVO and one with hemiretinal vein occlusion) who underwent RON. Four patients (80%) had improvement in VA and in 1 (20%) the VA worsened. Visual field defects after RON deserve special attention. During the procedure, a radial cut is made through the optic nerve head all the way to the lamina cribrosa and adjacent sclera, possibly cutting the arterial circle of Zinn and Haller as well. The circle of Zinn and Haller is one of the sources of blood supply to the optic nerve head. 11 Therefore, this procedure can be expected to cut off blood supply to the optic nerve head, resulting in possible acute ischemia of the optic nerve head and visual

5 VISUAL FIELD DEFECTS AFTER RADIAL OPTIC NEUROTOMY FOR CRVO BARAK ET AL 553 loss. RON also involves possible injury of nerve fibers in the optic nerve head by the radial cut itself and by hemorrhage in the cut margins, both features that are highly likely to produce visual loss. It must be stated that all our assumptions are speculative, and the exact amount of damage to either the arterial circle of Zinn or optic nerve fibers will remain unknown until histopathologic examinations of human subjects who underwent RON have been performed. An experimental RON performed in the rabbit optic nerve showed no damage to the central retinal vein and artery, and enhanced expression of glial fibrillary acid protein and vimentin indicating the activation of astrocytes. 12 Visual field defects were detected in five of our patients: the typical picture included temporal in addition to constricted ones. Although we did not perform presurgery visual field examinations due to the low VA of our patients, we believe that these defects are surgery related. Hayreh et al 13 performed visual field testing in 620 CRVO patients who were both nonischemic (81%) and ischemic (19%). They used three isopters (I-2e, I-4e, and V-4e) of the Goldmann perimeter in nonischemic CRVO, and the peripheral visual fields were entirely normal with V-4e and I-4e. They were also either normal with the I-2e target, or the I-2e could still be seen in spite of the presence of peripheral visual field defects. Patients with ischemic CRVO, however, either could not see whatsoever or they could not see the V-4e target, some could see the I-4e target, and most were unable to see the I-2e target at all. Thus, in their study, all three targets were seen in 71% of the eyes in nonischemic CRVO, and I-4e and V-4e were visibile in the remaining 29%, with no eye being unable to see any target or only V-4e. In contrast, only 8% of the eyes with ischemic CRVO could see all three targets, 63% saw I-4e and V-4e, and 18% only V-4e, while 10% could not see any target whatsoever. We, like Hayreh, 14 chose to use the Goldmann kinetic perimeter test for visual field testing because it was easy for the patients to perform and we could get more accurate information compared to the 24-2 or central 10 test (Humphrey Instruments, San Leandro, CA), especially in patients with low vision. The largest/brightest target (V-4e) stimulus in Goldmann perimeter that we used in these patients had an average diameter of 1.72 degrees (64 mm 2 at 30 cm). Therefore, patients who are able to see hand movement can see it. This stimulus size is not in standard use because it is larger than some of the scotomas and slightly smaller than the diameter of the blind spot. The stimulus can, however, be detected by low vision patients, unlike the standard stimulus size. 8 Although visual field defects were detected among our patients who had good visual acuity, we were unable to find the same defects in patients with poor vision, perhaps due to testing limitations and the lack of patients with poor vision who were able to do the test. We found that the temporal field s defects matched the nerve fiber field s defects. In recent publications, outcomes measured were best-corrected VA, macular thickness using OCT, and even multifocal electroretinography. 17 To our knowledge visual field defects post-ron have been mentioned only briefly before. 18 Specifically visual field loss has been reported as a complication after PPV, especially after macular hole surgery, as well as following epiretinal membrane removal 25 and removal of subretinal neovascular membrane. 26 The common characteristic among patients with postoperative visual field complications was air-fluid exchange, thus the emerging hypothesis was that trauma to or around the optic nerve during this maneuver may be responsible for the pathogenesis of field defects. 27 Most of the detected visual field defects were temporal scotomas, and the incidence and location of the postoperative ones was affected only by changing the location of the infusion cannula. 27 In recent years, visual field defects were prevented by passing air used for fluid-air exchange through water 28,29 or by lowering the air pressure during air-fluid exchange. 30 No air-fluid exchange was performed on our patients, thus our presumed explanation for the occurrence of the defects is direct damage to the optic nerve head. The introduction of the use of indocyanine green during internal limiting membrane peeling was also found to cause optic nerve head atrophy and visual field defects, but we did not use this substance on our patients. There are several limitations to our investigation. We did not perform presurgery visual field examinations that might have revealed pre-existing visual field defects that could have accounted for some of the defects seen postoperatively. This is a nonrandomized study, and we did not include a nontreated control group, thus an actual improvement over nontreated patients is only speculative. We did not obtain OCT data on optic nerve anatomy after RON and its possible correlation to visual field defects. Future evaluations of RON must include preoperative visual field evaluation, and patients who undergo RON must be informed about visual field defects and their occurring as a side effect of surgery.

6 554 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES 2006 VOLUME 26 NUMBER 5 Key words: central retinal vein occlusion, radial optic neurotomy, visual field defect. References 1. Hayreh SS. Occlusion of the central retinal vessels. Br J Ophthalmol 1965;49: Zegarra H, Gutman FA, Conforto J. The natural course of central retinal vein occlusion. Ophthalmology 1979;86: Evaluation of grid pattern photocoagulation for macular edema in central vein occlusion. The Central Vein Occlusion Study Group M report. Ophthalmology 1995;102: Opremcak EM, Bruce RA, Lomeo MD, Ridenour CD, Letson AD, Rehmar AJ. Radial optic neurotomy for central retinal vein occlusion: a retrospective pilot study of 11 consecutive cases. Retina 2001;21: Bynoe LA, Opremcak EM, Bruce RA, et al. Radial optic neurotomy for central retinal vein obstruction. Retina 2002; 22: ; author reply Hayreh SS. Radial optic neurotomy for central retinal vein occlusion. Retina 2002;22: Central Vein Occlusion Study of Photocoagulation. Manual of operations. Central Vein Occlusion Study Group. Online J Curr Clin Trials 1993; doc. no Anderson D. Perimetry With and Without Automation, 2nd ed. St Louis: Mosby Company; Opremcak EM. Radial optic neurotomy for central retinal vein occlusion. Presented at The Vitreous Society 19th annual meeting; Weizer JS, Stinnett SS, Fekrat S. Radial optic neurotomy as treatment for central retinal vein occlusion. Am J Ophthalmol 2003;136: Hayreh SS. The blood supply of the optic nerve head and the evaluation of it: myth and reality. Prog Retin Eye Res 2001; 20: Nozaki MSE, Okabe K, Ogura Y. Histopathological evaluation after radial optic neurotomy in rabbit. Presented at the ARVO meeting; Fort Lauderdale, FL; 2003; Program # Hayreh SS, Zimmerman MB, Podhajsky P. Incidence of various types of retinal vein occlusion and their recurrence and demographic characteristics. Am J Ophthalmol 1994; 117: Hayreh SS. Central retinal vein occlusion. Available at: Zambarakji HJ, Ghazi-Nouri S, Schadt M, Bunce C, Hykin PG, Charteris DG. Vitrectomy and radial optic neurotomy for central retinal vein occlusion: effects on visual acuity and macular anatomy. Graefes Arch Clin Exp Ophthalmol 2005; 243: Nagpal M, Nagpal K, Bhatt C, Nagpal PN. Role of early radial optic neurotomy in central retinal vein occlusion. Ind J Ophthalmol 2005;53: Martinez-Jardon CS, Meza-deRegil A, Dalma-Weiszhausz J, et al. Radial optic neurotomy for ischaemic central vein occlusion. Br J Ophthalmol 2005;89: Williamson TH, Poon W, Whitefield L, Strothidis N, Jaycock P. A pilot study of pars plana vitrectomy, intraocular gas, and radial neurotomy in ischaemic central retinal vein occlusion. Br J Ophthalmol 2003;87: Melberg NS, Thomas MA. Visual field loss after pars plana vitrectomy with air/fluid exchange. Am J Ophthalmol 1995; 120: Boldt HC, Munden PM, Folk JC, Mehaffey MG. Visual field defects after macular hole surgery. Am J Ophthalmol 1996; 122: Ezra E, Arden GB, Riordan-Eva P, Aylward GW, Gregor ZJ. Visual field loss following vitrectomy for stage 2 and 3 macular holes. Br J Ophthalmol 1996;80: Hutton WL, Fuller DG, Snyder WB, Fellman RL, Swanson WH. Visual field defects after macular hole surgery. A new finding. Ophthalmology 1996;103: ; discussion Paques M, Massin P, Santiago PY, Spielmann AC, Gaudric A. Visual field loss after vitrectomy for full-thickness macular holes. Am J Ophthalmol 1997;124: Arima T, Uemura A, Otsuka S, Doi N, Nakao K. Macular hole surgery-associated peripheral visual field loss. Jpn J Ophthalmol 1998;42: Haas A, Korner F. [Perimetry findings before and after surgical removal of epimacular membranes.] Klin Monatsbl Augenheilkd 1992;200: Yan H, Dhurjon L, Chow DR, Williams D, Chen JC. Visual field defect after pars plana vitrectomy. Ophthalmology 1998; 105: Welch JC. Dehydration injury as a possible cause of visual field defect after pars plana vitrectomy for macular hole. Am J Ophthalmol 1997;124: Welch JC. Prevention of visual field defect after macular hole surgery by passing air used for fluid-air exchange through water. Am J Ophthalmol 1999;128: Ohji M, Nao IN, Saito Y, Hayashi A, Tano Y. Prevention of visual field defect after macular hole surgery by passing air used for fluid-air exchange through water. Am J Ophthalmol 1999;127: Gass CA, Haritoglou C, Messmer EM, Schaumberger M, Kampik A. Peripheral visual field defects after macular hole surgery: a complication with decreasing incidence. Br J Ophthalmol 2001;85: Uemura A, Kanda S, Sakamoto Y, Kita H. Visual field defects after uneventful vitrectomy for epiretinal membrane with indocyanine green-assisted internal limiting membrane peeling. Am J Ophthalmol 2003;136: Ando F, Sasano K, Suzuki F, Ohba N. Indocyanine greenassisted ILM peeling in macular hole surgery revisited. Am J Ophthalmol 2004;138: Ando F, Yasui O, Hirose H, Ohba N. Optic nerve atrophy after vitrectomy with indocyanine green-assisted internal limiting membrane peeling in diffuse diabetic macular edema. Adverse effect of ICG-assisted ILM peeling. Graefes Arch Clin Exp Ophthalmol 2004;242: Kanda S, Uemura A, Yamashita T, Kita H, Yamakiri K, Sakamoto T. Visual field defects after intravitreous administration of indocyanine green in macular hole surgery. Arch Ophthalmol 2004;122: