Corneal Endothelial Replacement: Penetrating or Lamellar Keratoplasty

Restoration of corneal clarity by transplantation of healthy donor endothelium is the primary surgical goal in patients with oedematous corneas and decreased vision resulting from endothelial disorders. The most common causes of corneal endothelial decompensation in the adult population are cataract removal/IOL-related endothelial loss and Fuch's Dystrophy.

For over two decades, pharmacologic manipulation of endothelial cells has been investigated to increase endothelial density. I, as well as many other researchers, studied numerous growth factors in the hope of safely promoting endothelial cell proliferation. Despite these attempts and promising experimental studies using targeted gene therapy and endothelial cell cultures, transplantation of donor corneal tissues will remain the clinical strategy in the near future.

Indications for surgery and choice of procedures depend on the individual patient's visual requirements, visual handicap, extent of corneal oedema, ocular pathology causing corneal oedema, other ocular pathology also causing loss of vision, and other conditions that increase the risk of graft rejection or ocular trauma. The quality and availability of donor tissue, as well as surgical skill, technique and instrumentation are also critical determinants. All these factors must be considered prior to surgery. These factors also must be carefully considered when critically evaluating clinical studies and new procedures.

Penetrating keratoplasty with or without combined cataract extraction and intraocular lens insertion remains the gold standard for most of these surgical candidates who are not at high risk for graft rejection or ocular trauma.

Blanket statements claiming superiority of lamellar procedures (termed PLK, DLEK or ELK) for replacement of endothelial cells in all patients with Fuch's Dystrophy or pseudophakic bullous keratopathy are at best premature. Nonetheless, some current posterior lamellar techniques, when performed by experienced surgeons, may be beneficial in some patients, and further refinements may increase their indications.

Bourne has documented that endothelial cell loss from before preservation to a few months after penetrating keratoplasty performed by experienced surgeons is only 10%. However, there is an ongoing higher rate of loss of endothelial cells in transplanted corneas than in healthy corneas in unoperated eyes, and some of the causes are still unknown. To ensure maximum long term viability of the graft, it is essential to minimise surgical trauma, particularly manipulation of the donor endothelium.

Graft survival after penetrating keratoplasty in patients with Fuch's Dystrophy was documented this past year by Thompson to be 97% at five years and 90% at 10 years. Long term follow-up will be necessary to evaluate the survival of posterior lamellar grafts. Busin and Azar have documented the risks of epithelial ingrowth and flap melt, which can effect the survival of posterior lamellar grafts with microkeratome created flaps (anterior approach). Compressive forces from suturing of the donor posterior button can affect graft survival as well as deterioration of optical quality.

Until augmentation of donor or recipient endothelial cell density is possible, folding and rolling of donor endothelium should be evaluated very carefully. Preliminary results have shown additional cell loss from these manoeuvres. Will refinements in technique and instrumentation combined with increasing the diameter of the folded or rolled donor disc compensate for cell loss at the fold or from rolling and unrolling?

Does the Descemetorhexis remove full thickness Descemet's membrane with or without some stromal tissue? The electron microscopic study of Hirano, Sugita and Kobayashi (2002), demonstrated that, despite what can appear as separation between stroma and Descemet's membrane under a surgical microscope or by light microscopy, cleavage sometimes occurs within Descemet's membrane between the anterior banded and the posterior nonbanded layers. Also, depending on corneal pathology, stromal tissue may be more firmly attached to Descemet's membrane, and separation sometimes occurs within stromal tissue and not at the interface.

Despite increased technical complexity and cell loss, posterior lamellar keratoplasty via a small limbal incision – the Melles technique- could be indicated for faster visual rehabilitation in an elderly patient with pseudophakic bullous keratopathy, as long as donor tissue with a very high cell count is used.

It should be emphasised that no clinical study to date of posterior lamellar keratoplasty with transplantation of donor discs containing more than 150 microns of stromal tissue in addition to Descemet's membrane and endothelium performed, either via a limbal incision or microkeratome-assisted flap, has demonstrated better or more stable visual or refractive results than previously published controlled and randomised clinical studies of penetrating keratoplasty, even after suture removal. Ideal donor and recipient diameter disparities still need to be determined to best balance tight sealing wounds with minimal and stable inducement of sphere and cylinder. Corneal biomechanics are not yet adequately understood.

In 1996, Serdarevic, Renard and Pouliquen documented best spectacle-corrected visual acuities (BSCVA) of 20/40 or better at six months and 12 months postoperatively (before suture removal) and at six months after suture removal in all patients with Fuch's Dystrophy in their prospective clinical trial of triple procedures (penetrating keratoplasty, cataract extraction and IOL implantation). In 2003 Terry and Ousley reported BSCVAs of 20/40 or better in only 37% of Fuch's patients at six months after posterior lamellar keratoplasty (via a limbal incision either following or combined with cataract extraction and IOL implantation) and an average BSCVA of 20/48 at one year postoperatively.

Factors contributing to the inferior optical quality obtained to date with posterior lamellar grafts compared to full thickness grafts are not yet fully understood. Interface haze, orientation of recipient versus host lamellar fibres, and posterior surface irregularity have been suggested as important causes. Published results to date have documented better corneal regularity, as measured by computerised videokeratoscopy, after penetrating keratoplasty (Serdarevic 1995) both before running suture removal at six months postoperatively (SAI: 0.60 +/- 0.30) and after suture removal (SAI: 0.44 =/- 0.24) than after posterior lamellar keratoplasty (Terry 2003) at either six months (SAI: 0.77 +/- 0.51) or 12 months (SAI: 0.83 +/- 0.52) postoperatively.

Will the femtosecond laser improve dissection safety, accuracy and smoothness with improved optical clarity? Studies on LASIK flaps indicated that photodisruption with its secondary heat and shockwave effects required a minimal distance of about 150 microns to avoid endothelial damage. Soong and collaborators currently are investigating the laser's potential applicability in posterior lamellar keratoplasty. Power levels will need to be high enough to achieve adequate dissection but low enough to avoid inflammation.

Predictability of IOL power and resulting spherical equivalent outcomes after triple procedures depend on accuracy of predicted central corneal curvature. If only the central corneal curvature is measured in an oedematous, irregular cornea with bullous changes prior to posterior lamellar keratoplasty and then used for IOL power calculation, it is obvious that accuracy would not be improved compared with that of a triple procedure with penetrating keratoplasty, in which the surgeon's own average post keratoplasty keratometry value is used with or without modification based on pre-operative peripheral corneal powers measured 360º

Spherical equivalent refractive errors within +/- 2.0 D of the intended were achieved after penetrating keratoplasty, cataract removal and IOL insertion in 82% of eyes by Mattax and McCulley (1989) using the surgeon's average post keratoplasty K value and in 88% of eyes by Serdarevic, Renard and Pouliquen (1996) using the surgeons average post keratoplasty K value modified for a flat or steep peripheral cornea. In the clinical trial of Terry and Ousley (2003), in which the results of posterior lamellar keratoplasty procedures performed at the time of cataract removal were combined with those performed two weeks to 14 years after cataract extraction and IOL insertion, 83% of eyes attained SE refractions within +/- 2.0 D.

The range of predictive error was -2.54 D to + 1.22 D in the Serdarevic PK series and – 3.75 D to + 2.0 D in the Terry DLEK series. The deviation from intended central corneal power for correct IOL power was -1.5 D to +1.5 D in the Serdarevic study and -2.5 D to + 2.5 D in the Terry study.

Although predictability of triple procedures with lamellar keratoplasty do not yet match those with penetrating keratoplasty (unless the central K value prior to onset of corneal oedema is known), Melles has demonstrated a predictive error range of -1.75 D to + 0.50 D in a small series of PLK patients with pseudophakia.

Much work lies ahead to develop even better instrumentation and techniques for both penetrating and lamellar keratoplasty and to achieve refractive results after combined procedures in patients with Fuch's Dystrophy and cataracts that can match results after routine cataract surgery – before non surgical treatment for endothelial disorders is available.

Back to top..