Reasons for performing study
We wanted to investigate the visual outcome of horses presented with iris prolapse and treated with corneal transplantation.
We wanted to investigate the visual outcome of horses presented with iris prolapse and treated with corneal transplantation.
To evaluate the visual outcome of horses with iris prolapse treated with penetrating keratoplasty alone and penetrating keratoplasty in combination with overlying conjunctival or amniotic membrane grafting.
A retrospective medical records study of horses presented to the University of Florida Veterinary Medical Center for iris prolapse and treated with penetrating keratoplasty in the period of 1998-2010. Data collected from the medical records included signalment, clinical descriptions of ocular lesions, treatments, and therapeutic outcome.
Iris prolapses in this study were caused by corneal ulcers with keratomalacia (n = 37). All horses were treated medically for infection, hyperproteinase activity and iridocyclitis, and then surgically treated with either penetrating keratoplasty alone (n = 9) or penetrating keratoplasty with either a conjunctival pedicle flap (n = 22), amniotic membrane transplant (n = 5) or amnion membrane and conjunctival pedicle flap (n = 1). The eyes were visual postoperatively in a majority of the cases (n = 24; 64.9%). Limited vision was noted in 6 eyes (16.2%), 3 eyes became phthisical (8.1%) and 4 globes were enucleated (10.8%). Graft rejection manifested as some degree of donor corneal graft opacification in all cases. Anterior synechiae were present in 48.6% of the eyes. Wound dehiscence and aqueous humour leakage were also common as post operative problems.
Penetrating keratoplasty alone or in combination with an overlying graft of conjunctiva or amniotic membrane can achieve a successful visual outcome in a high percentage of horses with iris prolapse.
Iris prolapse is a serious and often visually catastrophic ophthalmic condition in horses associated with corneal perforation or rupture. An iris prolapse can occur secondary to corneal rupture or perforation due to rapid enzymatic degradation of stromal collagen in infectious and noninfectious corneal ulcers and following full-thickness corneal lacerations from acute, sharp corneal trauma . Aqueous humour leaks through the corneal perforation to cause anterior chamber narrowing or collapse, hyphaema, hypotony and anterior movement of the corpora nigra or iris into the perforation. The hypotony may induce forward movement of the vitreous and retina to result in retinal detachment. The globe with corneal perforation and iris prolapse is at high risk of infectious endophthalmitis. Obstruction of the aqueous humour leak with iris tissue can restore intraocular pressure to maintain globe stability and preserve vision. The globe with iris prolapse generally results in some degree of permanent anterior synechiation in the horse.
Veterinarians have been frustrated by a lack of viable treatment options for the horse with iris prolapse. A study in 1997  reviewed the visual outcomes and degree of globe survival in horses with iris prolapse treated with medical therapy and conjunctival flap or graft surgery. The standard treatment for iris prolapse at that time was to replace the protruding prolapsed iris into the anterior chamber if possible, amputate protruding necrotic iris and then ‘patch’ the affected cornea with a vascularised conjunctival flap grafted from the adjacent bulbar conjunctiva. The data in that study indicated that combined medical and conjunctival flap surgical treatment yielded a visual outcome of 33–40% depending on the cause of the iris prolapse .
Microsurgical techniques developed in the last 15 years allow surgeons to provide good visual outcomes after repair of deep cornea ulcerations, melting corneal ulcers, corneal lacerations and iris prolapses in horses [2-9]. Medical therapies to reduce bacterial and fungal infection, suppress iridocyclitis and reduce excessive proteinase activity have also improved [2-8]. Penetrating keratoplasty involves the removal of full-thickness cornea and replacement of the resultant defect with full-thickness donor cornea [2, 4, 5]. This technique can be combined with conjunctival and/or amniotic membrane grafts to reduce post operative complications and speed corneal healing. This paper compared the results obtained with the use of penetrating keratoplasty alone or when combined with an overlying membrane of conjunctiva and/or amnion for treating severe corneal diseases with iris prolapse in horses.
A corneal transplant medical records database and ophthalmic image gallery were reviewed to identify horses presented to the University of Florida Veterinary Medical Center Large Animal Hospital between 1998 and 2010 for evaluation and treatment of iris prolapse caused by corneal ulceration. The cases selected for this study required the presence of an iris prolapse and a history of corneal transplant surgery. Case details from each case included gender (stallion, mare, gelding), age (years) and breed. The cause of the iris prolapse, surgical method of repair, surgical complications, healing time and visual outcome were noted for each case. Healing time was defined as the time post operatively when graft epithelialisation was complete, graft incorporation into the recipient site had occurred and uveitis was absent. All medications were then discontinued. Due to reorganisation of the medical records data system, it was not possible to access complete microbiological culture, cytological or histopathological results or specific medical treatment records. As a result, these factors were not included in this study.
All horses in this study had a positive consensual pupillary light reflex and a positive dazzle response from the eye with the iris prolapse, indicative of some degree of retinal activity prior to surgery. Surgical treatments were divided into 4 different surgical procedures: penetrating keratoplasty alone or with conjunctival flap, amniotic membrane transplant or conjunctival flap and amnion membrane. Eyes with iris prolapse with large diameter (>6 mm) melting ulcers only underwent the latter 3 procedures (Figs 1-9). Amnion was not available during the first years of this study but once it became available there was a biased tendency for surgeons to utilise it as the layering material.
Visual outcomes were categorised as: 1) Visual (normal menace reflex, dazzle response, positive pupillary light reflexes, ability to examine fundus, normal behaviour), 2) Limited vision (menace, dazzle and pupillary light reflex responses intact but not from all directions) and 3) blind (phthisis bulbi; menace reflex, dazzle response and consensual pupillary light reflex absent) or enucleated (eyes removed due to endophthalmitis or other cause).
Preoperative medications generally included topical antimicrobials, atropine and serum and systemic antibiotics and flunixin meglumine as described in previous studies (1–4,6). All surgical procedures were performed under general anaesthesia and muscular paralysis (0.07 ± 0.01 mg/kg bwt). An operating microscope was utilised for each procedure.
A standard surgical approach for penetrating keratoplasty was utilised . The anterior chamber was reformed in each procedure by injecting a viscoelastic solution (hyaluronate sodium, 10 mg/ml, Hylartin V)1 into the anterior chamber. This solution moved the iris posteriorly breaking down any adhesions between the cornea and iris. Direct contact with the lens capsule was also avoided by reforming the anterior chamber with the viscoelastic solution that was not removed from the anterior chamber. The protruding iris tissue was amputated and any haemorrhage controlled with topical phenylephrine (2.5%) or epinephrine (1:1000), pressure and low voltage cautery. The size of the corneal lesion was determined with calipers. The recipient globe was stabilised with scleral fixation sutures of 5-0 nylon. A corneal trephine larger in diameter than the site of iris prolapse was centred over the diseased area of cornea and then rotated with minimal downward pressure to obtain a clear-cut, round incision with vertical sides. The trephine incision approached but did not penetrate Descemet's membrane. The remaining intact stromal tissue was vertically incised with a No. 65 Beaver blade to enter the anterior chamber, being careful to avoid the iris, corpora nigra and lens. The button of diseased host tissue was then removed with corneal section scissors.
A full-thickness button of cornea 1 mm larger in diameter than the recipient bed was then trephined from the endothelial to the epithelial side of the donor cornea. The donor button was grasped with fine-toothed forceps while paying particular attention to the orientation of the epithelium/endothelium, placed on a gauze swab and kept moistened with lactated Ringer's solution. The epithelium was not removed from the corneal donor button.
The donor cornea was placed in the recipient bed and 4 cardinal sutures of 8-0 Vicryl2 placed at the 12, 6, 9, and 3 o'clock positions. Simple interrupted sutures were placed to fill in the remaining sectors in each quadrant or, alternatively, a simple continuous suture was placed to hold the graft. Once the donor cornea was sutured into place, the viscoelastic solution was again injected via a limbal incision to completely reform the anterior chamber. The viscoelastic was not removed from the anterior chamber.
If signs of infection or severe keratomalacia were present, a conjunctival pedicle graft, amniotic membrane graft, or both, were placed over the keratoplasty site. Conjunctival pedicle grafts were created by separating a bulbar conjunctival graft from the underlying Tenon's capsule and amniotic membrane was used as a free-island graft. The edges of each graft were sutured to normal cornea surrounding the keratoplasty site with 8-0 vicryl in a simple interrupted pattern . A temporary lateral tarsorrhaphy was performed to minimise eyelid trauma to the keratoplasty site.
Donor cornea was harvested from fresh equine globes within 2–4 h of death. Globes were obtained from horses subjected to euthanasia for noninfectious diseases such as trauma, colic or laminitis. The corneas were hemisected and stored frozen at -20°C in either ophthalmic gentamicin, ophthalmic triple antibiotic (neomycin, polymyxin B, gramicidin), or corneal storage media (Optisol)3.
Equine amnion was aseptically harvested from caesarean section or term pregnancy. The allantoamnion was separated from the allantochorion and was rinsed with sterile saline containing penicillin (Spectrum)4 and streptomycin (Spectrum)4. The amnion was separated from the allantois by blunt dissection, sectioned and placed on a nitrocellulose membrane5 in Delbecco's modified Eagle's medium (DMEM)6 with glycerol5, penicillin (Spectrum)4, streptomycin (Spectrum)4, neomycin (Spectrum)4 and amphotericin B (Spectrum)4 and stored at -80° for use.
For surgical use of amniotic membrane, an individual section was thawed and rinsed with sterile saline for 30 min to remove any remaining glycerol in which it was stored. Amniotic membrane was left on the nitrocellulose membrane with the allantoic side against the paper until just prior to placement on the cornea.
Preoperative medications were continued post operatively and the frequency of application of topical autogenous serum was generally increased after penetrating keratoplasty. The horses were examined twice daily the first week and then weekly for corneal graft opacity and integrity, degree of uveitis and presence of vision. Once the corneas were healed, all medications were discontinued and a final vision assessment was performed.
This study compared differences in visual outcome (visual and limited vision) vs. blindness (including phthisis bulbi and enucleated) using penetrating keratoplasty alone vs. penetrating keratoplasty combined with either conjunctival flap or amnion membrane transplant and between the latter 2 procedures. Since visual outcomes are observations, descriptive statistical analysis was performed using the SAS statistical programme7. The Fisher's exact test was then used to assess the data in the contingency tables to identify any outcome differences between the same surgical procedures. P values <0.05 were considered to be statistically significant.
Thirty-seven horses with iris prolapse that underwent penetrating keratoplasty were identified. Of the 37 horses included in this study, 25/37 (67.5%) were mares, 7/37 (18.9%) geldings and 5/37 (13.5%) stallions. The mean age of the horses for this study was 6.5 ± 6.5 years. Thoroughbreds were the most common horse breed in this study with 16/37 cases (43.2%), followed by Quarter Horses with 10/37 cases (27.0%), Arabian and Miniature horse with 2/37 cases each (5.4% each) and one case each of an Andalusian, Appaloosa, Haflinger, Paint horse and Rocky Mountain horse (1/37 cases; each 2.7%). Two horses were of unknown breed.
The cause of iris prolapse in this study was corneal ulceration with keratomalacia. Surgical treatments utilised were divided into 4 different surgical techniques (Figs 1-9): (1) penetrating keratoplasty alone = 9/37 (24.3%), (2) penetrating keratoplasty and conjunctival flap = 22/37 (59.5%), (3) penetrating keratoplasty and amnion membrane transplant = 5/37 (13.5%), (4) penetrating keratoplasty, conjunctival flap and amnion membrane transplant = 1/37 (2.7%).
Based on final visual assessment and management of complications, 24/37 (64.9%) of the horses were visual, 6/37 (16.2%) had limited vision, 3/37 (8.1%) developed phthisis bulbi (Fig 10) and 4/37 (10.8%) had to be enucleated due to complications such as endophthalmitis (Fig 11) and glaucoma. The average healing time from surgery to cessation of medical treatment was 7 weeks.
Statistical comparisons with a Fisher's exact test of the visual outcome of the 3 surgical options resulted in no significant differences between penetrating keratoplasty alone vs. penetrating keratoplasty combined with either conjunctival flap (P = 0.3026) or amnion membrane transplant (P = 0.4895), or when penetrating keratoplasty and conjunctival flap was compared with penetrating keratoplasty and amnion membrane transplant (P = 0.4837).
Surgical complications included scarring of the surgical site in all horses, microleaks of aqueous humour with partial or complete suture dehiscence were identified with a positive Seidel's test  in 12/37 (32.4%) of the horses, severe posterior synechia causing the pupil to be completely constricted in 6/37 (16.2%), shallow anterior chamber in 5/37 (13.5%), endophthalmitis in 3/37 (8.1%) and glaucoma and retinal detachment in one horse each (2.7%).
The overall visual outcome of horses with iris prolapse from corneal ulcers treated with penetrating keratoplasty alone or in combination with a layering technique was 64.9% overall and is thus considered a clinically successful therapeutic approach for iris prolapse in horses. This contrasts with the results of a previous study that utilised conjunctival flap surgery alone for therapy of iris prolapse in horses and had a visual outcome of 40% in horses with iris prolapse due to an ulcer and 33% in horses with iris prolapse after a traumatic corneal laceration . There may be several explanations for why replacing absent cornea by penetrating keratoplasty rather than by covering the corneal lesion with a conjunctival flap alone in horses with iris prolapse resulted in an overall improvement in visual outcome despite the inevitable donor graft rejection and graft opacification . The layered collagen of the corneal graft tissue may increase the tectonic strength of the weakened cornea and could more effectively block aqueous humour leakage from the corneal perforation, thereby allowing the anterior chamber to restabilise and more quickly reduce an iris prolapse associated iridocyclitis than a conjunctival flap alone. Microleaks of aqueous humour at the incision or around the sutures are common and serious post operative complications of corneal surgery in horses as they can lead to sustained hypotony, endophthalmitis, repeat surgery and phthisis bulbi .
Improvements in medical therapies over the last 15 years that reduce bacterial and fungal infection, suppress iridocyclitis and reduce excessive proteinase activity also may have contributed to the improved visual outcome obtained with either penetrating keratoplasty alone or when combined with a conjunctival flap for horses with iris prolapse when compared to placement of a conjunctival graft alone .
Corneal transplantation in man is most commonly done for optical reasons in corneas with oedema but no vascularisation, infection or inflammation . Human corneas with infection or vascularisation do not tolerate penetrating keratoplasty well with the graft readily becoming rejected and opaque post operatively . The most common corneal environment present in horses that are candidates for penetrating keratoplasty or lamellar keratoplasties are corneas with severe infectious keratitis, extreme hyperproteinase activity and corneal vascularisation and are thus also considered high risk for graft rejection and subsequent opacity [4, 5, 10]. The ability to maintain a clear cornea in horses under these conditions is probably impossible but penetrating keratoplasty still serves an important function in the treatment of iris prolapse in horses.
There were no significant differences identified in the present study in visual outcome between horses that underwent penetrating keratoplasty alone vs. those in which this procedure was combined with an ‘overlaying’ biological membrane of vascularised conjunctiva or nonvascularised amnion. There are several possible reasons why covering the surgery site with either of these tissues may reduce the incidence of incision dehiscence and microleakage of aqueous humour. For example, these tissues may provide physical support to the graft and plasma-derived antiproteinases to reduce tear film and stromal proteinases that digest the corneal graft, surrounding cornea and sutures. The decision to combine a penetrating keratoplasy with either conjunctiva or amnion in the surgical treatment of iris prolapse in horses is the surgeon's preference but the results of this study strongly suggest that one of these procedures should be used. Both the amnion and conjunctival graft physically support the surgery site and provide antiprotease activity during the healing process but the conjunctival graft may cause a slightly higher degree of corneal fibrosis [8, 9].
The donor cornea used in a penetrating keratoplasty provides physical support, a source of layered collagen to replace the absent corneal tissue and possibly an improved barrier-effect against aqueous humour leakage and microbial invasion than would be obtained with conjunctival grafting alone. Penetrating keratoplasty alone or in combination with an ‘overlaying’ biological membrane (conjunctival graft or amniotic membrane) has a high level of positive visual outcome for horses with iris prolapse due to corneal ulceration.
No competing interests have been declared.
This work was not funded.
Michala de Linde Henriksen: data collection: 45%; study execution: 45%; data analysis and interpretation: 33.33%; preparation of the manuscript: 40%. Caryn Plummer: data collection: 10%; study execution: 10%; data analysis and interpretation: 0%; preparation of the manuscript: 10%. Brendan Mangan: data collection: 10%; study execution: 10%; data analysis and interpretation: 0%; preparation of the manuscript: 10%. Gil Ben-Shlomo: data collection: 5%; study execution: 5%; data analysis and interpretation: 0%; preparation of the manuscript: 5%. Hiroki Tsjuita: data collection: 5%; study execution: 5%; data analysis and interpretation: 0%; preparation of the manuscript: 5%. Shari Greenberg: data collection: 5%; study execution: 5%; data analysis and interpretation: 0%; preparation of the manuscript: 5%. Nils Toft: data collection: 0%; study execution: 0%; data analysis and interpretation: 33.33%; preparation of the manuscript: 5%. Dennis Brooks: data collection: 20%; study execution: 20%; data analysis and interpretation: 33.33%; preparation of the manuscript: 20%.
Pfizer Animal Health, New York, USA.
Ethicon, Somerville, New Jersey, USA.
Chiron Ophthalmics, Irvine, California, USA.
Gardena, California, USA.
Biorad Laboratories, Richmond, California, USA.
Fischer Scientific, Hampton, New Hampshire, USA.
Statistical Analysis System, Cary, North Carolina, USA.