Penetrating Keratoplasty Performed by Residents

Objective  To report the results of penetrating keratoplasty performed by residents.

Method  A retrospective medical record review of all patients undergoing penetratingkeratoplasty performed by residents at our institution from April 1998 toApril 2002.

Results  Forty penetrating keratoplasty procedures were performed by 8 residents.The most common indication was keratoconus (17 eyes [43%]), followed by cornealscarring (14 eyes [35%]). Mean preoperative best-corrected visual acuity was20/250. No intraoperative complications were reported. Mean follow-up timewas 15 months. Postoperatively, mean best-corrected visual acuity was 20/40,mean postoperative astigmatism was 3.4 ± 2.1 diopters, and graft survivalwas 92.5%. Postoperative complications included elevated intraocular pressure,wound dehiscence, and endophthalmitis.

Main Outcome Measures  Best-corrected visual acuity, postoperative astigmatism, graft survival,and intraoperative and postoperative complications.

Conclusion  Residents can be introduced to penetrating keratoplasty and achievesurgical success with intraoperative and postoperative complication ratessimilar to those previously published.

Results of resident surgery are important to analyze. Using this data,residency programs can determine the effectiveness of their surgical teachingstrategies and what modifications need to be made to improve outcomes. Theresults of cataract surgery performed by ophthalmology residents have beenwell described in the literature. Almost all of these publications have foundthat with proper supervision, residents can achieve results similar to experiencedsurgeons.^1-5 Thisimportant information provides both confidence to the residents and surgicalinstructors and reassurance to the patients who will be cared for.

There is a limited amount of data regarding results of residents andcornea fellows performing penetrating keratoplasty (PK). A study from DukeUniversity, Durham, NC, indicated that residents could be introduced to PKand achieve results comparable with those reported in the ophthalmic literature.⁶ Cornea fellows from the University of California,Davis, showed they could achieve rates of postoperative astigmatism followingPK similar to experienced corneal surgeons.⁷ Ourgoal is to enlarge this pool of data by reporting results of PK performedby residents at our institution.

The medical records of all patients undergoing PK from April 1998 toApril 2002 performed by residents at California Pacific Medical Center (CPMC),San Francisco, were reviewed for the following data: patient demographics,ocular diagnoses, preoperative clinical course, operative technique and procedures,intraoperative and postoperative complications, and postoperative clinicalcourse.

All patients were examined and approved for surgery by a senior residentand a member of the corneal faculty. The informed consent process was performedby the senior resident, often in the presence of the supervising faculty surgeon.This included a lengthy discussion of the postoperative diligence requiredby the patient and surgeon following PK. All surgical procedures were performedusing general anesthesia, sterile technique, standard trephination techniques,and 16 interrupted sutures.

Under the supervision of a corneal faculty member, residents performedall critical portions of the procedure including trephination of the donorand host tissue and suture placement including the first 4 cardinal sutures.One of the authors (A.L.S.) elected to place 4 noncardinal sutures for educationalpurposes.

All patients undergoing PK were followed up postoperatively in residentclinics attended by corneal faculty. Residents, with the guidance of the cornealfaculty, managed postoperative ocular medications, astigmatism, and complicationsduring these clinics. At CPMC, residents change rotations every 6 months.At that time, the residents managing these patients changed, while the supervisingcornea faculty remained the same.

Forty PK procedures were performed on 35 patients at CPMC from April1998 to April 2002. The mean patient age was 43 years (range, 4-86 years);23 (66%) were men, and 12 (34%) were women. The most common indication forPK was keratoconus (17 eyes [43%]) followed by corneal scarring (14 eyes [35%]),failed PK (3 eyes [7%]), pseudophakic corneal edema (2 eyes [5%]), infectiouskeratitis (active) (2 eyes [5%]), Fuch endothelial dystrophy (1 eye [2.5%]),and Peters anomaly (1 eye [2.5%]).

Eight senior residents performed the surgeries in this series. The averagenumber of PK procedures performed by each resident was 5 (range, 3-7). Themost common was PK alone (28 eyes [70%]), followed by PK with extracapsularcataract extraction and posterior chamber intraocular lens placement (6 eyes[15%]), PK with anterior vitrectomy and sulcus-sutured posterior chamber intraocularlens placement (3 eyes [7.5%]), PK with intraocular lens exchange (1 eye [2.5%]),and PK with anterior vitrectomy (1 eye [2.5%]). A single case of PK, extracapsularcataract extraction, and planned anterior vitrectomy was performed on a 4-year-oldboy with Peters anomaly. No intraoperative complications were reported.

Mean follow-up time was 15 months (range, 1 month to 3 years) with 24patients (68%) having current follow-up at the time of data collection. Onepatient died secondary to a myocardial infarction 1 month following surgery.

Mean best-corrected visual acuity (BCVA) was calculated as describedby Akpek et al.⁸ Mean preoperative BCVA was20/250 (range, 20/70-light perception). Postoperative BCVA was available for33 of 40 eyes and is summarized in Table1. Patients were grouped according to duration of follow-up. Seveneyes did not have a recording for postoperative BCVA. This was owing to death(1/7) or the absence of a manifest refraction at the time of data collection(6/7). Mean postoperative BCVA was 20/40 (range, 20/15-no light perception),with 22 eyes (67%) attaining a postoperative BCVA of 20/40 or better. Causesof postoperative BCVA < 20/200 (8 eyes) included graft failure (3), occlusionamblyopia (1), macular hole (1), persistent hyperplastic primary vitreousvariant (1), undiagnosed preexisting retinal detachment (1), and endophthalmitis(1).

The most common postoperative complications were elevated intraocularpressure (IOP) (10 eyes [25%]) and wound dehiscence (4 eyes [10%]). Elevatedintraocular pressure was defined as persistent IOP greater than 21 mm Hg requiringtreatment.⁹ Elevated IOP was controlled withtopical pressure-lowering agents in all eyes. Wound dehiscence was both traumatic(2/4) and spontaneous in the immediate postoperative period secondary to suturedehiscence (2/4). All cases were repaired without complication.

Allograft rejection (controlled) occurred in 3 eyes (7.5%) and was managedwith topical steroids. Graft failure was defined as a totally opacified cornealgraft secondary to corneal edema. This occurred in 3 (7.5%) of 40 eyes andin 3 (13%) of 23 eyes with more than 1 year of follow-up. Thus, when consideringall grafts with at least 1 year of follow-up (n = 23), 20 (87%) were clear.

Untolerated astigmatism occurred in 2 eyes (5%), and infectious sutureabscess, shallow anterior chamber (grade 1), and choroidal detachment occurredin 1 eye (2.5%) each.

The most significant postoperative complication was acute postoperativeendophthalmitis. After uncomplicated PK with extracapsular cataract extractionand posterior chamber intraocular lens placement, on postoperative day 4,the patient had a complete hypopyon, 30% graft dehiscence (secondary to tissuenecrosis), and visual acuity of no light perception. Despite surgical repairof the dehiscence, vitrectomy, and intravitreal antibiotic injection, thevision did not improve. Vitreous cultures grew Pseudomonasaeruginosa. The patient eventually required enucleation for oculardiscomfort.

Postoperative astigmatism measurements were available for 29 of 40 eyes.There were 11 eyes that did not have a recording for postoperative astigmatism.This was due to enucleation (1/11), graft failure (3/11), or lack of measurementin the immediate postoperative period (7/11). Two thirds of these measurementswere recorded using manual keratometry and the other third, using manifestrefraction. When both measurements were available, we used the larger of the2 (always manual keratometry) for our data analysis. Mean ± SD postoperativeastigmatism for all eyes was 3.4 ± 2.1 diopters (D) (range, 0.5-8.0D); for eyes with 6 to 12 months' follow-up, 3.2 ± 2.4 D (range, 0.5-8.0D); and for eyes with more than 1 year of follow-up, 3.5 ± 2.0 D (range,1.0-7.0 D).

Two cases of significant astigmatism required keratorefractive proceduresfor correction. The above quantification of astigmatism reflects the cornealmeasurements prior to these keratorefractive procedures.

Residents in our program have excellent exposure to PK, with each residentperforming an average of 5 PK procedures during their training. A typicalCPMC resident first performs PK during the second year of training, only afterapproximately 30 extracapsular cataract procedures, didactics regarding cornealsurgery, assisting faculty and senior residents performing PK, and extensivepractice with and without faculty supervision in a microsurgical wet lab.

The most common indication for PK in our series was keratoconus (17patients [43%]). This is consistent with some series in the literature wherekeratoconus has surpassed pseudophakic corneal edema as the leading indicationfor PK.^10,11 The number of patientsrequiring PK for pseudophakic corneal edema was low in our series.

Surgical success was evaluated by reviewing postoperative BCVA, postoperativeastigmatism, and graft survival. The prognosis for surgical success dependson the preoperative indication for PK. While some diagnoses carry a favorableprognosis (keratoconus), others are quite the opposite (advanced surface diseasewith loss of limbal stem cells).¹² The surgicalsuccess in any PK series is partially dependent on the percentage of differentindications present in the series.

Mean postoperative BCVA was 20/40, with 22 eyes (67%) attaining a postoperativeBCVA of 20/40 or better. This compares favorably with a range of 20/25 to20/60 published by experienced corneal surgeons.^13-17 Thegroup of patients with more than 1 year of follow-up had better mean BCVA(20/33) than those in the 6- to 12-month follow-up group (20/53), which isconsistent with previously published reports of BCVA stabilizing across timeafter PK.¹³

Mean ± SD postoperative cylinder was 3.4 ± 2.1 D. Thisalso compares well with previously published data looking at postoperativeastigmatism by cornea fellows (3.27-D change from preoperative astigmatism)⁷ and corneal specialists (2.0-5.4 D).^15,16,18-22 Weused manifest refraction for final postoperative astigmatism if manual keratometrywas not performed (approximately one third of patients). Because these valuestend to be lower than manual keratometry, our results may be lower than ifmanual keratometric values had been used in all patients. Postoperative astigmatismwas similar between the group of patients with more than 1 year of follow-upand those with 6 to 12 months of follow-up.

When considering all grafts with at least 1 year of follow-up (n = 23),20 (87%) were clear. Series with a similar length of follow-up have demonstratedgraft clarity ranging from 70% to 95%.^6,15 However,it is difficult to compare graft survival across a wide spectrum of preoperativediagnoses.

Our results for postoperative BCVA, postoperative astigmatism, and graftsurvival are comparable with those of the Duke University residents. However,while our most common indication was keratoconus, the Duke University residentscited previously failed grafts as their most common indication for PK. Manyof their patients had fully vascularized corneas. The prognosis for surgicalsuccess is markedly different between these 2 groups.¹²

No intraoperative complications were reported. The most common postoperativecomplication was elevated IOP (25%). This compares favorably with an incidenceof 18% published by Foulks.²³ All patientswere managed successfully with IOP-lowering agents (and/or reduction of topicalsteroids). Risk factors for post-PK elevated IOP are preoperative glaucoma,aphakia, anterior segment inflammation, vitrectomy, intraocular lens removal,and preoperative trauma.^9,24-26 Morethan half (7/10) of our patients with postoperative elevated IOP had 1 ofthese risk factors. Surgeons should be particularly watchful for post-PK elevatedIOP in patients with these risk factors.

The second most common complication was wound dehiscence occurring in4 eyes (10%). This was both traumatic and spontaneous in the immediate postoperativeperiod related to unstable sutures. One author (D.F.G.) buried all the resident-placedsuture knots evaluating the tension and stability. The resident replaced suturesthat were of inappropriate tension or that dehisced during this process. Thistechnique seems to have eliminated postoperative suture instability and mayhelp with postoperative astigmatism.

Weaknesses of this study are those that are inherent to a retrospectiveanalysis. Also, it would have been useful to have BCVA and manifest refraction,manual, and topographic keratometry values for all 40 eyes. However, in thefirst few postoperative months, the emphasis for evaluating patients post-PKin our cornea clinic includes IOP, the health of the corneal graft, graft-hostapproximation as an origin of large amounts of irregular astigmatism, anddetecting posterior segment pathologic features. Values for BCVA, manifestrefraction, corneal topography, and manual keratometry are obtained for some,but not all, patients in the early postoperative period.

In summary, results of this series indicate that with excellent supervisionresidents can be safely introduced to PK and achieve surgical success andintraoperative and postoperative complication rates similar to those previouslypublished. We attribute the postoperative results of PK by CPMC residentsto appropriate selection of patients, proper preoperative patient counseling,resident and patient diligence, and excellent teaching and supervision bythe corneal faculty at CPMC.

Correspondence: Ethan M. Kutzscher, MD, 1850 Sullivan Ave 540, DalyCity, CA 94015 (ethankutzscher@yahoo.com).

Submitted for publication June 24, 2003; final revision received October13, 2004; accepted January 6, 2004.

We thank the cornea faculty at California Pacific Medical Center, SanFrancisco, for their dedication to resident teaching and patient care: DavidDeMartini, MD; Jonathan Diamont, MD; Bernd Kutzscher, MD; Karen Oxford, MD;Niraj Patel, MD; Lee Schwartz, MD; Michael Turan, MD; David Vastine, MD; andRobert Webster, MD.

Dr Kutzscher had full access to all the data in the study and takesresponsibility for the integrity of the data and the accuracy of the dataanalysis.