Eye (1990) 4, 693-697

Intraocular Surgery Following Penetrating Keratoplasty: The Risks and Advantages

L. A. FICKER, C. M. KIRKNESS, A. D. McG STEELE, N. S. C. RICE, A. M. E. GILVARRY London

Summary Graft survival has been evaluated for patients who underwent subsequent intra­ ocular surgery (extra-capsular surgery or trabeculectomy) between 1983 and 1989. The patients were different from the majority of keratoplasty patients as evidenced by the indications for keratoplasty; corneal perforation was the indication in 24% of cases. Perforated and inflamed were treated aggressively at the time of the acute event, including emergency keratoplasty and intensive topical steroids. Visco-elastic fluids were routinely used during secondary surgery and topical ste­ roids were administered intensively post-operatively. The incidence of post-oper­ ative graft rejection was low (less than 14%). Rejection episodes were diagnosed early, prior to the appearance of a Khodadoust line, and were treated aggressively with intensive topical steroids. which was not controlled by topical ther­ apy was surgically managed by trabeculectomy in the first instance. If this failed, tube drainage was performed and long-term topical steroids were administered. The only risk factor identified was uncontrolled glaucoma, P=O.1. The probability of graft survival (at five years) was 0.83 after cataract surgery and 0.62 after trabecu­ lectomy, but wide confidence limits indicate the difference is not significant.

The impact of intraocular surgery upon a pre­ For patients in whom corneal disease co­ existing penetrating keratoplasty (PK) is per­ exists with early cataract there are advantages tinent to various sub-groups of patients with in deferring cataract surgery in patients corneal disease. The most frequent proce­ undergoing PK. Stable keratometry following dures to be performed subsequently are PK should allow greater accuracy in calculat­ cataract extraction aud trabeculectomy. En­ ing IOL power. Furthermore, when PK is per­ dothelial cell loss is known to be accelerated formed as an emergency procedure for by intra-ocular surgery. 1 Expected graft sur­ perforation or progressive microbial , vival may therefore be compromised by accel­ deferred cataract surgery (in the absence of an erated endothelial cell loss following such intumescent ) may reduce post-operative surgery. Glaucoma may also threaten graft inflammation, hence reducing the risk of early survival. It may be a complication of ker­ graft rejection.4 Knowledge of the risks and atoplasty, or previously controlled glaucoma advantages of combined compared with two­ may become uncontrolled after further intra­ stage procedures would be useful in planning ocular surgery. 2 The introduction of visco­ the management of complicated cases such as elastic fluids, topical beta-blockade and tube these. drainage procedures3 represent therapeutic Glaucoma which is uncontrolled by topical advances with respect to these complications. therapy requires surgery. Trabeculectomy

Correspondence to: Linda Ficker FRCS, FCOphth, Moorfields Hospital, City Road, London EC1 2CV. 694 L. A. FICKER ET AL. and tube drainage surgery have been per­ between 4mm diameter and 9x12 mm. The formed in patients with a pre-existing corneal post-operative steroid regime was initially i-2 graft and it would be useful to identify any risk hourly topical dexamethasone 0.1% and was factor for graft survival which may be related to reduced according to the clinical course. Com­ glaucoma surgery. plications, graft survival, refractive and visual The effects upon graft survival of modem outcomes were analysed. microsurgery for cataract and glaucoma have been reported infrequently. We therefore Trabeculectomy reviewed our experience in the Corneal Clinic Trabeculectomy was performed aftef a mean at Moorfields Eye Hospital. Since 1983, ECCE post-PK interval of 7.8 months (n=29), ten of (mostly with lens implantation) has been the which were within one month. The most fre­ routine surgical technique for cataract surgery quent indications for grafting were keratitis and and tube drainage was adopted in the manage­ trauma as shown in Table I. Trabeculectomy ment of failed trabeculectomy, in preference to alone was performed in 20 patients; nine others cyclocryo-ablation.5 In addition, visco-elastic underwent additional surgery; tube drainage fluids were routinely used by this time for intra­ (5), ECCE (2), cyclo-cryotherapy (1) or com­ ocular surgery in the presence of a graft. bined ECCE + IOL (1). PK had been com­ bined with intra-capsular cataract extraction (ICCE) in eight cases; overall, 19 patients were Patients and Methods aphakic or pseudophakic, 13 following ICCE Patients who underwent intraocular surgery and six following ECCE. Most PKs (28) were after prior penetrating keratoplasty, between central, mean diameter 8.1 min, and only one 1983 and 1989, were identified from the Cor­ 9 mm PK was eccentric. Twenty-five were first neal Clinic database. Twenty-five patients and four were second PKs. Post-operative ste­ underwent extracapsular cataract extraction, 17 roids were given initially 1-2 hourly dexame­ with additional lens implantation. Two further thasone 0.1%, as for cataract surgery, and were patients underwent secondary lens implan­ reduced as indicated by the clinical course. Ste­ tation following combined PK and ECCE. roids were maintained indefinitely following Twenty-nine patients underwent trabeculec­ tube drainage to avoid the complication of tomy after PK and five of these underwent rejection.5 Complications, graft survival and tube drainage procedures following failed visual outcomes were analysed. trabeculectomy. Rejection episodes were diagnosed in the presence of anterior chamber cells with graft

Extracapsular Cataract Extraction Table I Indications for PK prior to further intraocular The mean post- PK interval was 73.3 months surgery overall (n=25) and 33.4 months for the sub­ group which had IOL implantation (n=17). Total number patients (%)

The interval was shorter (9 months) for ECCE (± IOL) Trabeculectomy inflamed eyes; two intumescent lenses were 25 (100%) 29 (100%) removed within one month. The two secondary 6 (24%) implants were inserted at a mean interval of 91 1 (3%) Interstitial keratitis 6 (24%) 3 (10%) months. Removal of graft sutures had pre­ Corneal perforation 6 (24%) 7 (24%) viously been performed in about half the cases. Acanthamoeba 3 (12%) 1 (3%) The most frequent indications for PK were ker­ Corneal dystrophies 2 (8%) atoconus, interstitial keratitis and corneal per­ Trauma 1 (4%) 4 (14%) HSK 1 (4%) 4 (14%) foration (Table I). There were 20 first, five ABK 3 (10%) second and two subsequent PKs. Trabeculec­ PBK 2 (7%) tomy had previously been performed in four Glaucoma cases. Most (22) were central grafts, mean complications 4 (14%)

7.6 mm diameter. Three were eccentric, ABK: aphakic bullous keratopathy adjacent to the limbus and measuring PBK: pseudo-phakic bullous keratopathy. INTRAOCULAR SURGERY FOLLOWING PENETRATING KERATOPLASTY 695 endothelial keratic precipitates; a Khoda­ pneumoniae and P. aeruginosa). Two cases doust line was not required for the diagnosis. were successfully treated; one with a suture Graft failure was defined as irrecoverable loss abscess (S. viridans) and one with endoph­ of graft clarity. thalmitis following tube drainage (S. epidermidis) . Results Eight grafts failed; three due to uncontrol­ Extracapsular Cataract Extraction led intra-ocular pressure (lOP), two due to Mean follow-up was 98 months post-PK and irreversible rejection, two due to recurrent 32 months post-ECCE (minimum six (HSK) and one months). Kaplan-Meier survival curves6 were gradual decompensation after tube insertion. constructed to demonstrate graft survival. Trabeculectomy alone controlled lOP in The probability of graft survival after ECCE, ten cases, additional topical therapy was with or without lens implantation, was 0.83 required in 11 (timolol 0.25% and/or pil­ after 66 months. The incidence of rejection ocarpine 4%) with acetazolamide in three was 12% (three cases), all of which occurred cases. Of the five cases which required tube during the immediate post-operative period. drainage, four were controlled. Visual out­ None resulted in graft failure. One patient come for this group was poor; 6/12 or better with dry eyes developed a suture abscess was achieved in only 35% of patients with (Streptococcus pneumoniae) which was suc­ clear grafts. Patients with vision less than cessfully treated. One patient required resu­ 6/36 (n=6) were noted to have glaucomatous turing of the PK after traumatic wound disc cupping (4), choroidal haemorrhage (1) dehissence. Four patients grafted for keratitis and poor ocular surface (1) which probably developed glaucoma; two after Acantha­ influenced visual potential. mo�ba keratitis (one required trabeculec­ tomy and a subsequent tube). There were Discussion three failures; due to intractable secondary Graft survival has been considered separ­ glaucoma (1), gradual decompensation (1) ately for patients who underwent subsequent and vascularisation of a 4 mm paralimbal cataract extraction or trabeculectomy. patch graft (1). Corneal disease more commonly co-exists Corrected visual acuity was 6/12 or better with cataract and is frequently managed by in 65% of cases with clear grafts. Poor vision combined surgery. Where corneal disease was attributable to graft failure in two cases. requires PK, a decision should be made with The remainder were due to maculopathy (3), respect to the mangement of any lenticular glaucoma (2), retinal vein occlusion (1) and opacity. Combined surgery has been advo­ (1). The spherical equivalent for cated in preference to two-stage surgery to patients with IOL was within 3.00 dioptres of reduce the risk of late endothelial decompen­ emmetropia in 65% of cases. sation. If no significant difference can be established for graft survival between com­ Trabeculectomy bined and two-stage surgery, the option of Mean follow-up was 46 months post-PK and deferring cataract surgery can be justified 38.5 months post-trabeculectomy. The prob­ with the advantage of a more predictable ability of graft survival after trabeculectomy, refractive outcome. including cases which subsequently under­ Grllft survival for 66 triple procedures went tube drainage procedures (5) or cyclo­ (PK, ECCE and lens implantation) was cryoablation (1), was 0.62 after 45 months. reported by Crawford et al (1986) to be Rejections occurred in 14% (4) cases, during 90%;7 mean post-operative follow-up was the post-trabeculectomy period (mean post­ 15.8 months. Kirkness reported no graft fail­ operative interval = 5 months) and resulted ures after combined surgery (n=28), with in two failures; neither had undergone trabe­ mean follow-up of 28.4 months.8 Without culectomy within 6 months of PK. Five cases detailed knowledge of survival and censor­ were complicated by infection; three were in ship for these patients, statistical comparison failed grafts which developed keratitis (S. with two-stage surgery cannot be performed. 696 L. A. FICKER ET AL.

The probability of PK survival for patients in trabeculectomy after PK is less than 50% this series (post-cataract surgery) was, how­ after two years.2 Since raised lOP results in ever, 0.83 (follow-up was 32 months) and endothelial dysfunction, patients in whom appears to compare favourably. PK survival, trabeculectomy fails should be offered tube however, is only a reflection of endothelial drainage to protect both and function. Endothelial cell counts were not graft endothelium. Combined PK and trabeculectomy is per­ done prospectively although donor material formed less frequently than combined PK was assessed clinically prior to keratoplasty. and cataract extraction, hence graft survival Quantitative endothelial assessment of the for combined and two-stage surgery cannot donor both pre-operatively and post-oper­ readily be compared for this group. Insler et atively would be useful further studies. al reported one failure in seven cases (mean Multiple procedures might be expected to be follow-up 16 months) which underwent com­ an adverse prognostic factor in PK survival. bined PK and trabeculectomy.9 Foulks Of the 12 patients who underwent two or reported similar results for PK survival after more procedures, however, only three failed trabeculectomy for post-keratoplasty glau­ (a similar proportion to the overall failure coma.lO Of five patients who had trabeculec­ rate ). tomy, one graft failed, but of 17 who had Post-operative refraction after combined cyclocryo-ablation, seven grafts failed. Kirk­ surgery was found by Crawford et al to be ness reported a 0.68 probability of success for within 2 dioptres of emmetropia in 62% of controlling intraocular pressure and main­ 5 cases and by Kirkness et at to be within 3 taining graft status after tube drainage. This dioptres of emmetropia in 50% of cases. Graft was statistically better than for cyclocryo­ ablation (P = 0.05). The choice of post-PK sutures had been selectively removed in glaucoma surgery is therefore important with these patients to reduce , but had respect to the risk of failure. Although the not been totally removed. It was suggested survival curves suggest better survival for by Crawford et al that the refractive accuracy PKs following subsequent cataract extraction of combined surgery (PK + ECCE + IOL) compared with trabeculectomy, the wide could be improved for individual surgeons by confidence limits (8.9%-114%) indicate this applying a formula derived from regression to be an imprecise estimate. analysis.7 The refractive results for patients Ocular pathology might be expected to reported in this study (65% of patients with play a role in PK survival. Various pathol­ IOL were within three dioptres of emmetro­ ogies were associated with failure (n=11), pia) were no better than for combined sur­ but none achieved statistical significance as a gery. Graft sutures had not been removed risk factor. This may in part be due to the routinely prior to cataract extraction, how­ small numbers involved. Coster4 found sur­ ever, hence final keratometry was not avail­ vival was reduced from 76% to 44% in eyes able when biometry was performed. This which had been inflamed. Management of should therefore be considered in the future inflamed eyes influences this risk as shown by 11 as a means of improving refractive accuracy. Ficker et at for grafting in HSK; inflam­ Visual outcome was significantly better for mation was a risk factor in an early cohort (P=O.OOl), but not in a later cohort (P=OA) the cataract group (P=0.025). Severe glauco­ managed by ECCE techniques, interrupted matous disc cupping occurred more fre­ suturing and more aggressive topical steroid quently in the trabeculectomy group (28%) therapy. Patients with corneal perforation or compared with the cataract group (8%) and uncontrolled inflammation, in this series, the difference in visual outcome was attribut­ were managed very aggressively during the able to glaucomatous field loss. acute event with early surgery and intensive Glaucoma was a risk factor for PK survival post-operative steroids to reduce the risks of (P=O.l) and occurred in both groups; 4% (1) peripheral anterior synechiae and early graft of ECCE patients and 14% (4) in the trabe­ rejection. This policy appears to have been culectomy group. The probability of main­ effective, with no PK failures in inflamed taining normal intraocular pressure (lOP) by eyes. INTRAOCULAR SURGERY FOLLOWING PENETRATING KERATOPLASTY 697

References Rejections were equally common in the 1 Abbott RL and Forster RK: Clinical specular micro­ cataract extraction and trabeculectomy scopy and intraocular surgery. Arch Ophthalmol groups and the risk of failure from graft 1979,97: 1476-9. rejection was similar for both groups 2 Gilvarry AME, Kirkness CM, Steele AD, et al.: The (P=O.5). Rejection episodes were managed management of post-keratoplasty glaucoma by early (in the presence of anterior chamber trabeculectomy. Eye 1989, 3: 713--18. 3 Schockett SS, Nirankari YS, Lakhanpal Y, et al.: cells, but not requiring a Khodadoust line) Anterior chamber tube shunt to an encircling with intensive topical steroids and this band in the treatment of neovascular glaucoma appears important for PK survival. 11 Infec­ and other refractory . Ophthalmology tions occurred in six cases of which 2/6 were 1985, 92: 553--62. 4 Coster DJ: Mechanisms of corneal graft failure: the suture abscesses. These may be avoidable by erosion of corneal privilege. Eye 1989, 3: 251--62. prompt removal of loose sutures. Suture 5 Kirkness CM, Ling Y, Rice NSC: The use of silicone infiltrates should be presumed to indicate drainage tubing to control post-keratoplasty glau­ infection. The suture should be removed for coma. Eye 1988, 2: 583-90. 6 culture in addition to a corneal scrape prior Kaplan E and Meier P: Non-parametric estimation from incomplete observations.l Am Statis Ass to treatment with intensive topical 1958, 53: 457-81. antibiotics. 7 Crawford GJ, Stulting RD, Waring GO: The triple We conclude that with modern micro-sur­ procedure. Analysis of outcome, refraction and gical techniques and the use of intra-ocular intraocular lens power calculation. Ophthal­ mology 1986,93: 817-24. visco-elastic fluids, intra-ocular surgery may 8 Kirkness CM, Cheong PY, Steele AD: Penetrating safely be performed after PK. Rejection keratoplasty and cataract surgery: the advantages caused PK failure in 2/54 cases overall. This of an extracapsular technique combined with pos­ compares with failures due to rejection in terior chamber intraocular implantation. Eye keratoconus and Fuchs' endothelial dys­ 1987, 1: 557--61. 9 Insler MS, Cooper HD, Kastl PR, et al.: Penetrating trophy. Uncontrolled glaucoma was also keratoplasty with trabeculectomy. Am J Ophthal­ identified as a risk factor, hence PK survival mol 1985, 100: 593-5. 10 may indeed depend on glaucoma surgery. Foulks GN: Glaucoma associated with penetrating keratoplasty. Ophthalmology 1987, 94: 871-4. Further prospective studies are required to 11 Ficker LA, Kirkness CM, Rice NSC, et al.: The determine whether, following elective graft changing management and improved prognosis suture removal, a more predictable refractive for corneal grafting in Herpes simplex keratitis. outcome may be achieved. Ophthalmology 1989, 96: 1587-96.