Postoperative endothelial cellloss of 293 homologous keratoplasties, and 29 normal-risk keratoplasties (keratoconus).
Endothelial cell loss of 7 autologousgrafts and 293 homologous grafts.
Comparison of endothelial cellloss between autologous and homologous transplants 6 to 12, 12 to 24, and24 to 36 months postoperatively.
A, Indications for 293 homologouskeratoplasties, 1996-2000. B, Indications for 7 autologous keratoplasties.
Bertelmann E, Hartmann C, Scherer M, Rieck P. Outcome of Rotational KeratoplastyComparison of Endothelial Cell Loss in Autografts vs Allografts. Arch Ophthalmol. 2004;122(10):1437-1440. doi:10.1001/archopht.122.10.1437
The nature of chronic endothelial cell loss in homologous corneal graftsis still unclear. Possible causes are cell migration to the recipient bedand chronic subclinical immune reaction.
To compare endothelial cell loss after autologous rotational keratoplastyand homologous keratoplasty and present the clinical outcome of patients afterrotational keratoplasty.
In this open prospective study, we included 7 consecutive patients whounderwent rotational keratoplasty between 1998 and 2000 in our hospital. Patientswere examined clinically every 3 months, and visual acuity, astigmatism, andendothelial cell density were evaluated. Endothelial cell densities were comparedwith endothelial cell counts of 293 homologous keratoplasties.
Mean follow-up for autologous grafts was 39 months. Mean increase invisual acuity was 3.5 lines. Mean astigmatism was 4.75 diopters in the autologousgraft group. Mean preoperative endothelial cell density was 2058 (637 cells/mm2). Mean endothelial cell density after 1 year was 1865 (639 cells/mm2), which represents a mean ± SD cell loss of 15% ± 7.19%.At the end of follow-up, endothelial cell number after autologous graftingwas 1630 ± 622 cells/mm2. Endothelial cell loss after 1year in homologous grafts was 40% ± 21.34%. There was 1 decompensationof autologous graft in the follow-up period.
Main Outcome Measure
Comparison of endothelial cell count at different postoperative timepoints using nonpaired t test.
Endothelial cell loss in autologous grafts is significantly lower thanin homologous grafts, which supports the hypothesis that chronic endothelialcell loss is due to chronic subclinical immune reactions in homologous grafts.Autologous keratoplasties can lead to good functional results and can be superiorto homologous corneal grafting in suitable situations.
The high rate of endothelial cell loss after penetrating keratoplastyrepresents an important unsolved problem: chronic endothelial cell loss probablylimits transplant survival time to a greater extent than do acute immune reactions.1- 4 The natureof cell number reduction is not completely understood. Factors possibly contributingto the endothelial cell loss are surgical trauma, cell exchange between donorand recipient,5 cell aging, acute immune reactions,6 and chronic subclinical immune reactions.7 In this context, we performed an open prospectivestudy to compare chronic endothelial cell loss in homologous and autologousrotational transplants.
Autologous keratoplasty was described very early in the history of cornealtransplantation.8 It has been performed using3 different approaches: (1) Bilateral autologous keratoplasty: the graft isobtained from the blind contralateral eye, and the 2 corneas are exchanged.(2) Ipsilateral autologous keratoplasty: 2 minigrafts (1 central and 1 peripheral)of the same cornea are exchanged. (3) Ipsilateral autologous rotational keratoplasty.
The first 2 approaches have particularly limited indications. The bilateralautologous keratoplasty requires a blind second eye with a healthy clear cornea.The altered cornea still has to be transplantable; a contraindication mightbe a painful bullous keratopathy. The ipsilateral keratoplasty with 2 minitransplantsrequires a small circumscribed central opacity and a sufficient area of clearhealthy peripheral cornea. The centrally located sutures can bias the visualresults by high astigmatism and subepithelial fibrosis.
The first description of ipsilateral autologous keratoplasty dates from1914.8 Since then, different modificationsof this technique have been proposed, for example, the rotation of an 8-shapedgraft.9 Rotational keratoplasty has frequentlybeen assessed as a (worse) replacement for homologous keratoplasty in situationsof graft shortage,10 mainly due to poor visualoutcome related to high or irregular astigmatism.11
For preoperative calculation of graft size and localization, geometricmodels have been proposed to alleviate preoperative planning. Required sizeand localization of the rotational transplant depend on the largest circleof clear cornea and on the distance of the circle arc of the optic center.12
In the present study, we compare endothelial cell loss rates of 7 autologousrotational grafts with those of homologous grafts.
The study included 7 patients (3 women, 4 men) who were consecutivelyoperated on by the same surgeon (C.H.). A signed informed-consent form wasobtained from all patients before enrollment into the study.
Preoperative endothelial cell density was examined by noncontact specularmicroscopy using the NonCon ROBO (Konan Medical Inc, Hyogo, Japan). A centralphotograph was obtained, as well as additional endothelial images in all 4quadrants of the periphery.
The rotation procedure was carried out following the technique describedby Bourne and Brubaker.12 In 4 patients, arotational keratoplasty was performed as a single procedure; in 2 patients,anterior synechiolysis was also performed. In 1 patient, the operation wascarried out as a triple procedure with extracapsular cataract extraction andposterior chamber lens implantation.
The trephine size ranged from 7 to 8.5 mm (mean, 7.64 mm). The trephinationwas performed using the Barron trephination system. Double-running sutureswere used in all 7 patients (10-0 and 11-0 nylon). Healon 5 (Pfizer Inc, NewYork, NY) was injected into the anterior chamber before trephination.
Postoperative treatment consisted of topical steroids 3 times a dayfor 2 weeks and topical lubricants 12 times a day. To avoid postoperativeglaucoma, all patients received systemic carboanhydrase inhibitors on theday of surgery.
Postoperative examinations were carried out every 2 weeks during thefirst 3 months and every 3 months thereafter. Visual acuity, astigmatism,anterior segment, intraocular pressure, and fundus examination were performedat each control. Endothelial cell density was measured by noncontact specularmicroscopy. Because of temporary postoperative swelling of the transplant,early postoperative endothelial cell counts could not be obtained. The firstpostoperative endothelial cell density was measured 3 months postoperatively.
Endothelial cell densities of the autologous grafts were compared withendothelial cell counts of 293 homologous corneal grafts performed between1996 and 2000 in our clinic. Indications for homologous keratoplasties includedbullous keratopathy, Fuchs endothelial dystrophy, keratoconus, stromal scars,stromal dystrophies, herpes keratitis, therapeutic keratoplasty, and repeatkeratoplasty. Indications are shown in Figure1. Endothelial cell counts of autologous grafts were compared withthose of the subgroup of patients undergoing normal-risk homologous keratoplastyperformed in cases of keratoconus (29 patients).
The corneal grafts used in homologous keratoplasties were organ culturedin the Cornea Bank, Berlin, following the guidelines of the European Eye BankAssociation. The culture medium was minimal essential medium–Earle's(Biochrom, Berlin, Germany) + 2% fetal calf serum (Gibco, Paisley, Scotland).The median culture time was 13 days.
Differences in endothelial cell loss between homologous and autologousgrafts after 12 months and 24 months were compared statistically using nonpaired t test. Preoperative and postoperative visual acuitieswere compared statistically using the Wilcoxon rank sum test.
Preoperative diagnoses of the 7 autograft patients included 4 cornealscars after perforating injury, 1 scar after keratitis due to contact lens–weardamage, 1 scar after corneal ulcer, and 1 scar after perforated corneal ulcer.The mean age of the patients was 48.2 years (range, 20-85 years). Preoperativevisual acuity ranged from less than 20/200 to 20/63.
Preoperative keratometric astigmatism ranged from 1.25 diopters to 13diopters (mean, 5.5 diopters); preoperative refractive astigmatism rangedfrom 1.25 diopters to 5.5 diopters (mean, 3.3 diopters). Mean ± SDpreoperative refractive power of the cornea was 40.6 diopters ± 3.67diopters (range, 34.75-44.5 diopters). Mean preoperative endothelial celldensity was 2063 cells/mm2; cell count ranged from 1200 cells/mm2 to 3000 cells/mm2.
Four corneas had a central scar and a circular clear periphery. Threepatients had complicated corneal scars with anterior synechiae. There were4 patients who were pseudophakic and had posterior chamber lenses. One patienthad secondary glaucoma after perforating injury. Intraocular pressure wasregulated preoperatively.
None of the 7 corneas showed neovascularizations. Ocular diagnoses otherthan the corneal scar were amblyopia in 2 patients (1 with perforating injuryas a child and 1 after keratitis with perforated ulcer as a child) and age-relatedmacular degeneration in 1 patient.
Visual acuity increased significantly in homologous and autologous transplantgroups (Wilcoxon rank sum test). Mean improvement in visual acuity in autologoustransplants was 3.5 lines. The best postoperative visual acuity after rotationalkeratoplasty was 20/25; the worst was less than 20/200. Mean postoperativeastigmatism was 4.75 diopters ± 1.5 diopters. Suture removal was carriedout 1 year postoperatively in cases of persisting astigmatism over 4 diopters.Mean keratometric astigmatism was 4.7 diopters before suture removal and 4.4diopters after suture removal in homologous keratoplasties. The differencesin autologous and homologous astigmatism were not statistically significant.
Complications occurred in 2 patients during follow-up after autologousgrafting. In 1 patient, a decompensation of the transplant occurred 6 weekspostoperatively, requiring a homologous transplant. Indication for rotationin this patient was a corneal scar due to a prior perforated ulcer. The rotationhad been carried out as triple procedure in this patient, and preoperativeendothelial cell density was the lowest of all 7 patients (1200 cells/mm2).
In another patient who had a perforating injury, a complicated retinaldetachment occurred several months after the keratoplasty, requiring vitrectomyand silicone oil tamponade. In this patient, several complicated operationshad been performed prior to the rotation, including lens extraction, secondaryposterior lens implantation, and posterior chamber lens reposition.
Mean endothelial cell loss in all autologous keratoplasty patients 1year after the operation was 15% ± 7.19%, compared with 35% ±15.16% in homologous normal-risk keratoplasties and 40% ± 21.34% inall 293 keratoplasties. Endothelial cell loss after 2 years was 44% ±16.5% in normal-risk keratoplasties and 49% ± 21.98% in all homologouskeratoplasties. In autologous grafts, endothelial cell loss after 24 monthswas 19.3% ± 7.32%.
Endothelial cell loss rates of 293 homologous keratoplasties are shownin Figure 1. Figure 1 also shows endothelial cell loss of 29 normal-risk keratoplastiesperformed in patients with keratoconus. Differences in endothelial cell loss(keratoconus compared with all homologous keratoplasties) were not statisticallysignificant. Figure 2 shows thecomparison of endothelial cell numbers after homologous and autologous grafting.
Differences in endothelial cell loss between homologous and autologousgrafts are statistically significant after 12, 24, and 36 months (P < .05) (autologous grafts vs homologous grafts [all indications]and autologous grafts vs homologous grafts [normal risk]) (Figure 3).
In this study, we demonstrate the outcome of a relatively small numberof patients after autologous rotational keratoplasty. Indications for thisspecial type of keratoplasty are rare; thus, studies have included comparablysmall numbers of patients.11,13- 15 Allpatients in our study had a complicated preoperative situation, as shown in Figure 4B.Visual outcome is comparable with results after homologous keratoplasty incomplicated situations, and postoperative astigmatism is in the same rangeas astigmatism after homologous keratoplasty. This is in contrast to the resultsof Jonas et al,11 who reported a significantlyhigher astigmatism in autologous transplants. The use of double-running suturesin our study, instead of interrupted sutures in all cases, may explain partlythis result.
Endothelial cell loss rate after autologous keratoplasty was significantlylower than endothelial cell loss rate after homologous keratoplasty. Lateendothelial failure is known as the most important reason for late-graft decompensationafter homologous keratoplasty.16
Comparison of preoperative endothelial cell densities between homologousand autologous transplants is difficult because homologous grafts undergoadditional endothelial cell stress by preparation and organ culture.17 However, it has been shown that there are no statisticallysignificant differences between endothelial cell loss rates in corneas storedin tissue culture medium (eg, Optisol; Chiron Ophthalmics, Irvine, Calif),the most widely used storage technique in the United States, and endothelialcell loss rates in organ-cultured corneas in the first 2 years.18
Furthermore, early postoperative endothelial cell counting is oftennot possible after homologous transplantation owing to early postoperativereduction of stromal transparency. However, despite these restrictions, differencesin endothelial cell densities between homologous and autologous transplantsincrease over at least 3 years (Figure 2),which cannot be explained by initial endothelial cell damage during organculture. There is only a minor additional endothelial cell loss in autologoustransplants after the first year.
Immune reactions naturally cannot occur in autologous transplants. Therefore,only short-term postoperative steroid treatment, if any, is necessary afterautologous grafting. For this reason as well, steroid-induced glaucoma cannotoccur.
The significantly reduced endothelial cell loss rate after autologouscorneal grafting supports the hypothesis that there is an immunogenic reasonfor chronic endothelial cell loss in homologous corneal grafts. Owing to thelimited indications, autologous keratoplasties will be restricted to raresituations. However, in suitable situations, autologous keratoplasty may besuperior to homologous grafting as far as long-term graft survival is concerned.
Submitted for publication March 25, 2003; final revision received February25, 2004; accepted February 25, 2004.
Correspondence: Eckart Bertelmann, MD, Augenklinik Charité CampusVirchow-Klinikum, Humboldt Universität Berlin, Augustenburger Platz 1,13353 Berlin, Germany (firstname.lastname@example.org).