Radial Optic Neurotomy in the Porcine Eye Without Retinal Vein Occlusion

Objective  To demonstrate the histopathologic changes in the porcine eye withoutretinal vein occlusion after radial optic neurotomy (RON).

Methods  A RON was performed in 14 normal eyes of 12 Yorkshire Cross pigs. Oneradial stab incision at the edge of the nasal optic nerve head was made usinga 20-gauge microvitreoretinal blade (Visitec) while the intraocular pressurewas elevated. Surgery was concluded when hemostasis was achieved. Weekly ophthalmoscopicexaminations were performed. Group 1 eyes (4 eyes of 2 pigs) were enucleatedat the end of surgery. Group 2 eyes (4 eyes of 4 pigs) were enucleated 1 weekpostoperatively, and group 3 eyes (4 eyes of 4 pigs) were enucleated 3 weekspostoperatively. In group 4 (2 eyes of 2 pigs), animals underwent vitrectomyand RON, and eyes were enucleated 3 weeks postoperatively.

Results  Ophthalmoscopic examination demonstrated engorged blood vessels at theRON site up to 3 weeks after surgery with minimal or no hemorrhage. Histologicalexamination of the optic nerve demonstrated foci of hemorrhage, interstitialedema, reactive gliosis, and rare inflammatory cells. At 3 weeks, there wascomplete axonal nerve fiber loss distal to the neurotomy site.

Conclusions  After RON, marked gliosis and complete axonal nerve fiber loss occurat the neurotomy site. Although bleeding was rare intraoperatively in thisporcine model, hemorrhage and interstitial edema were present within the opticnerve at the neurotomy site histologically.

Clinical Relevance  Radial optic neurotomy remains a controversial method of treatment forcentral retinal vein occlusion. To our knowledge, this is the first studyin the literature describing the histopathologic findings after RON.

Central retinal vein occlusion (CRVO) is one of the most frequent vascularcauses of visual loss.^1-3 Theloss of visual acuity in eyes with CRVO may be due to extensive intraretinalhemorrhage, edema, and retinal ischemia.

Radial optic neurotomy (RON) has been suggested as potentially efficaciousin the treatment of CRVO in some eyes.⁴ A compartmentsyndrome may exist at the level of the lamina cribrosa within the optic nervehead in eyes with CRVO. The scleral outlet contains the cribriform plate (laminacribrosa), central retinal artery and vein, and fibers of the optic nerveand is surrounded by the scleral ring. The anatomic tissue architecture atthe scleral outlet resembles a "bottleneck-like" configuration and may resultin neurovascular compression in susceptible eyes, similar to other compartmentsyndromes elsewhere in the body.⁴ Radial opticneurotomy may decompress the scleral outlet via an internal, vitreoretinalapproach, and the results appear promising in small, uncontrolled pilot studies.^4-6 An alternative mechanismof action is that RON may facilitate the development of chorioretinal shunts,and thus help bypass the obstructed central vein. To our knowledge, therehave been no animal or human studies examining the optic nerve after radialneurotomy.

We chose the pig as an experimental model because of the anatomic similarityof the pig eye to the human eye. The histological appearance of the neuroglialcomponent of the optic nerve in the pig is similar to that in the human. Thelamina cribrosa of the pig is very strong and highly pigmented and has a thicknessof 0.4 to 0.6 mm. It embeds the retinal blood vessels and surrounds thesewith considerable pigment. In one section, 5 large blood vessels and manysmall ones were seen in the lamina cribrosa.⁷ Thedeposition of the pigment is not only localized to the lamina cribrosa, butappears to involve all fibrovascular septa of the nerve. The pig retinal bloodvessels pass to the retina through the papilla centrally or near its borders.There are 4 main arterial branches, consisting of a nasal and a temporal vesselin both the superior and inferior segments of the fundus. The veins followa similar pattern, but the branches are less numerous.⁷ Ourobjective was to perform a RON in pigs and to examine histologically the opticnerve and adjacent tissues.

Twelve Yorkshire Cross pigs (14 eyes), aged 3 to 4 weeks, were treatedin accordance with the Association for Research in Vision and OphthalmologyStatement for the Use of Animals in Ophthalmic and Vision Research. All eyeswere included in the final analysis. Twelve eyes underwent RON and were randomlyassigned to 3 study groups. In group 1 (4 eyes of 2 animals), eyes were enucleatedat the end of surgery. In group 2 (4 eyes of 4 animals), eyes were enucleated1 week postoperatively. In group 3 (4 eyes of 4 animals), eyes were enucleated3 weeks postoperatively. In group 4 (2 eyes of 2 animals), eyes underwentpars plana vitrectomy and neurotomy and were enucleated 3 weeks postoperatively.

Animals underwent premedication with 0.1 mL of 1% atropine sulfate (PhoenixPharmaceutical Inc, St Joseph, Mo) intramuscularly and then anesthetized with10 to 40 mg/kg of intramuscular ketamine hydrochloride (Fort Dodge Laboratories,Fort Dodge, Iowa) and 2 to 5 mg/kg of intramuscular xylazine hydrochloride(Phoenix Pharmaceutical Inc). If additional anesthesia was needed during theprocedure, 10 to 40 mg/kg of intramuscular ketamine hydrochloride was used.For pupillary dilation, 0.25% scopolamine hydrobromide (Alcon LaboratoriesInc, Fort Worth, Tex), 1% atropine sulfate (Alcon Laboratories Inc), and 2.5%phenylephrine hydrochloride (Bausch & Lomb Pharmaceuticals Inc, Tampa,Fla) were used. Proparacaine hydrochloride (Bausch & Lomb PharmaceuticalsInc) was the topical anesthetic. The conjunctival sac was irrigated with 5%povidone-iodine solution. All animals were initially examined with indirectophthalmoscopy to exclude any preexisting vitreoretinal abnormalities. Oneor both eyes of each animal were subjected to proptosis, and 3 sclerotomieswere created using a 19-gauge microvitreoretinal (MVR) blade for the introductionof the infusion cannula, a fiberoptic light probe, and a 20-gauge MVR blade.Infusion of lactated Ringer solution maintained intraocular pressure via theinfusion line. The 20-gauge MVR blade was used to perform a radial incisionat the nasal edge of the optic nerve head while the intraocular pressure waselevated (Figure 1). In group 4,eyes underwent complete vitrectomy with creation of a posterior vitreous detachment,followed by RON as described. Surgery was concluded when hemostasis was achieved,and the sclerotomies and conjunctiva were sutured with 7/0 polyglactin 910(Vicryl; Ethicon Inc, a division of Johnson & Johnson, Somerville, NJ)and 6/0 plain gut, respectively. In groups 2, 3, and 4, 20 mg of gentamicinsulfate (Abbott Laboratories, North Abbott Park, Ill) was injected into thesubconjunctival space.

The 4 eyes of the 2 animals killed on the day of surgery (group 1) wereenucleated at that time and sent for histological evaluation. Eight of the10 remaining animals were followed up for 1 week (group 2) or 3 weeks (groups3 and 4), with examinations at 1, 2, and 3 weeks after surgery. At each time,the retina was examined by indirect ophthalmoscopy, and fundus photographswere obtained. After the final examination, animals were killed with intracardiacpentobarbital sodium, and the eye was enucleated for histological examination.

After enucleation, the eyes underwent immediate sharp-razor penetrationclose to the pars plana to ensure rapid penetration of fixative. The eyesremained immersed in 4% paraformaldehyde for at least 24 hours at 4°C.All tissues were embedded in paraffin, and 5-µm sections were stainedwith hematoxylin-eosin. Masson trichrome histochemical staining was performed.

Sections from the paraffin block were cut at 4 to 5 µm, placedon positively charged glass slides, deparaffinized in organic solvents, treatedwith methanolic hydrogen peroxide to quench endogenous peroxidase activity,and rehydrated. Sections were reacted with glial fibrillary acidic proteinand neurofilament protein monoclonal antibodies (dilutions, 1:4000 and 1:100,respectively; DAKO, Carpinteria, Calif). Nonimmune rabbit immunoglobulins(DAKO) were used as negative control specimens, and appropriate positive tissuecontrols were also tested. Phosphate-buffered saline was used throughout theprocedure. Incubation time for the primary reaction was 45 minutes at 40°C,and 20 minutes at 40°C for the secondary and tertiary reactions. The unlabeled,bound primary antibody was linked with biotinylated goat anti-rabbit IgG (1:300;Vector Laboratories, Inc, Burlingame, Calif) and detected with horseradishperoxidase–labeled streptavidin (1:800; Jackson ImmunoResearch Laboratories,Inc, West Grove, Pa). Immunoreactivity was visualized using diaminobenzidineas the chromogen, with Harris modified hematoxylin (Fisher Scientific Co,Pittsburgh, Pa) as the counterstain. In addition, the width of the MVR stabincision within the lamina cribrosa was measured using the 100-µm baron the hematoxylin-eosin sections in all 4 group 1 eyes. The width was measuredat the top, in the middle, and at the bottom of the lamina cribrosa.

No significant complications were encountered during surgery in anyeye. The optic nerve remained unchanged postoperatively in all but 6 eyesin which engorged blood vessels developed near the neurotomy site immediatelyafter RON and were present as long as 3 weeks postoperatively. In some cases,a small hemorrhage from the neurotomy site was observed at the time of surgery;however, no hemorrhage was present at 2 weeks postoperatively. The local lossof the nerve fibers adjacent to the neurotomy site was present for less than1 hour (Figure 2). The loss of nervefibers was most apparent 3 weeks after surgery.

Hematoxylin-eosin–stained sections of the optic nerve and posteriorretina confirmed the presence of the neurotomy site. The optic nerve was characterizedby foci of hemorrhage, interstitial edema, reactive gliosis, and rare inflammatorycells. Sections taken at the peripheral portion of the neurotomy site in proximityto retina demonstrated retinal tissue incarcerated in the sclera at the neurotomysite (Figure 3). Results of Massontrichrome histochemical staining showed early scarring at the neurotomy sitewithin 1 week and confirmed the presence of a healed scar at the neurotomysite at 3 weeks (Figure 4). Immunohistochemicalstaining for glial fibrillary acidic protein demonstrated early reactive gliosisin group 1 eyes that were immediately enucleated, and the reactive gliosiswas sustained and diffuse throughout the entire segment of optic nerve at3 weeks (Figure 5). Neurofilamentprotein showed weak staining of the axons in the immediate vicinity of theneurotomy site at the time of immediate enucleation, and progressive absenceof axons in the vicinity of the neurotomy site at 1 week. At 3 weeks, therewas complete axonal nerve fiber loss distal to the neurotomy site along theentire segment of optic nerve that was sharply demarcated from adjacent intactnerve fibers (Figure 6). Light microscopicexamination demonstrated no neovascularization or chorioretinal shunt formationat the neurotomy site at any time. The neurotomy wound had a mean width of105.5 µm within the lamina cribrosa; however, the gap was filled withfibrosis at 1 week (Figure 7). Therewere no differences in the histological findings of the optic nerve in thevitrectomized and nonvitrectomized eyes.

Radial optic neurotomy remains a controversial treatment for eyes withCRVO. No treatment has proven successful in treatment of this disease thusfar.^5,6 Intravenous streptokinaseand recombinant tissue plasminogen activator have been used to treat eyeswith CRVO, but despite encouraging results, possible severe systemic adverseeffects, including death, have prevented their widespread use.^8,9 Todecrease the risk of the systemic adverse effects of recombinant tissue plasminogenactivator, Lahey and colleagues¹⁰ demonstratedthat intravitreal administration did not lead to catastrophic hemorrhagicevents in their nonrandomized study, and some eyes appeared to benefit fromthe therapy. Weiss¹¹ demonstrated that it waspossible to deliver recombinant tissue plasminogen activator directly intothe retinal vein using a transvitreal approach, and some patients may benefitfrom this treatment. McAllister and colleagues¹² demonstratedthat the creation of a chorioretinal-venous anastomosis using a high-intensitylaser was beneficial in some eyes with CRVO. However, the potential for seriouscomplications exists, such as hemorrhage, preretinal fibrosis, traction retinaldetachment, vitreous hemorrhage, choroidal neovascularization, and choroidovitrealneovascularization.

Histological studies demonstrate that most CRVO cases are associatedwith a thrombus in the central retinal vein at the level of or just posteriorto the lamina cribrosa.¹³ Studies by Vasco-Posada¹⁴ and Arciniegas¹⁵ suggestedcutting the sclera around the optic nerve via an external approach. Despitepromising results, this technique was not further developed. Opremcak andcolleagues⁴ hypothesized that CRVO is a neurovascularcompression syndrome resulting from increased pressure within the confinedspace of the scleral outlet. In their nonrandomized study, a RON was performedtransvitreally with a 20-gauge MVR blade in an attempt to decompress the opticnerve.⁴ Eight (73%) of 11 patients with CRVOshowed an improvement of visual acuity with a mean gain of 5 lines of vision.⁴ In cadaver eyes, an MVR blade could cut the cribriformplate, scleral rim, and adjacent sclera without ocular perforation.⁴ To our knowledge, no study in the literature describesthe histological effects of RON in an animal eye. In the present study, nosignificant complications during or after neurotomy were observed; however,results of histological examination suggested that RON was a traumatic procedure.The neurotomy site was characterized by foci of hemorrhage, interstitial edema,and rare inflammatory cells. Fibrosis and a glial scar were evident within1 week. Immunohistochemical staining of the optic nerve demonstrated diffusereactive gliosis and complete axonal nerve fiber loss distal to the neurotomysite (Figure 5).

Because the mechanical constriction of the central retinal vein at thelevel of the lamina cribrosa is a possible pathoetiologic mechanism of CRVO,a surgical approach (such as RON) to relieve this compartment syndrome maybe beneficial. In our histological study, RON resulted in a mean wound widthof 105.5 µm; however, the neurotomy site was filled with fibrous tissueat 1 week. Therefore, it is not clear how long the RON can relieve any mechanicalforces on the retinal veins (Figure 7).Further animal studies such as the measurement of blood flow in the retinalveins after RON may provide some answers.

In a surgical technique study of cadaver human and porcine eyes, Litand colleagues¹⁶ demonstrated that punctureof the lamina cribrosa with a specially designed lancet tip was possible withoutserious injury to the optic nerve. These authors speculated that their techniquecould be another approach to decompress the central retinal vein via an intravitrealapproach. Although the authors performed additional in vivo experiments inrabbits, there was no information about the effectiveness of this procedurein humans. There are no data as to whether this technique is less traumaticthan RON.

Another possible explanation of how RON improves the course of CRVOis that RON leads to chorioretinal shunt formation after neurotomy.^17,18 In our study, we did not observeany new chorioretinal shunt formation on ophthalmoscopic or histological examinations.The lack of shunt development may be multifactorial. In our study, RON wasperformed in healthy, nonischemic pig eyes, the follow-up period was relativelyshort (3 weeks), and no imaging techniques that might be helpful in recognizingsuch shunts early, such as fluorescein angiography, were performed.

Because RON is a traumatic procedure, it remains a controversial methodof treatment for CRVO. Although preliminary clinical reports of RON are encouraging,further study is necessary to establish RON as a standard treatment for CRVO.

Correspondence: Sharon Fekrat, MD, Department of Ophthalmology, DukeUniversity Medical Center, Box 3802, Durham, NC 27710.

Submitted for publication May 20, 2003; final revision received October1, 2003; accepted December 3, 2003.

This study was supported by the Kosciuszko Foundation, New York, NY.