Penetrating keratoplasty has been performed for more than a century and has evolved into a highly refined surgical practice. The evolution of keratoplasty surgery was closely linked to the broadening of anatomical-pathological knowledge and led to the further perfection of the technique applied. The introduction of motor and vacuum trephines opened up new perspectives; however, postoperative astigmatism remained the main source of patient dissatisfaction and has been a subject of pivotal interest to surgeons. Non-mechanical excimer trephination, invented by Naumann1 in 1989 in Erlangen, Germany, partially overcame this complication and has been proven a superior alternative to established methods of surgery in previous prospective clinical studies.2 The technique of non-contact excimer laser-assisted penetrating keratoplasty (ELAK) improves donor and recipient decentration and reduces vertical tilt and horizontal torsion of the graft, thus reducing postoperative keratometric astigmatism and providing higher surface regularity and better visual acuity compared to mechanical trephination.3
More than a decade later, femtosecond lasers have been acknowledged to be capable of creating penetrating corneal wounds for keratoplasty (femtosecond laser-assisted keratoplasty [FLAK]). Femtosecond laser trephination has been available for clinical practice since 2006.4 Buratto and Böhm5 reported seven cases with 3 months of follow-up and Price and Price6 published 1-year experiences of six cases in 2008 with a conclusion of a better donor corneal centration and a quicker visual recovery. Through the various cutting shapes, FLAK provides further advantages in wound sealing and inevitably the potential of creating lamellar corneal cuts.
In addition to corneal topographic and refractive results, in vivo confocal corneal microscopic studies have been reported on corneal cellular reactions after penetrating keratoplasty.7–11 Despite the improving surgical technology, increasing knowledge of the corneal immune system and developing imaging possibilities, it is still not fully understood how different techniques of penetrating keratoplasty affect the microstructure of the cornea.
The purpose of our study was to evaluate the corneal microstructure and cell densities by means of in vivo confocal corneal microscopy, comparing ELAK and FLAK in eyes with Fuchs’ dystrophy and keratoconus.
Patients and Methods
The study was performed in accordance with the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants, and the study protocol was reviewed and approved by an independent ethics committee of the institution. This was an observational study with analysis of epithelial, stromal, inflammatory, and Langerhans cell endothelium in patients with either ELAK or FLAK and in control subjects.
Fifty-seven eyes of 57 patients (age: 57.1 ± 29.2 years [range: 23 to 86 years]) were recruited to the study and assigned to four groups according to the corneal disease and the type of surgery performed. The four treatment groups were Fuchs’ dystrophy and ELAK; Fuchs’ dystrophy and FLAK; keratoconus and ELAK; and keratoconus and FLAK. The control group comprised individuals without any ophthalmic disease or ocular surgery in the medical history.
Penetrating keratoplasty was performed by one experienced surgeon (BS) under general anesthesia at the Department of Ophthalmology, Saarland University Medical Center, Homburg, Germany. The type of trephine was randomly selected and masked from other investigators. Control patients were randomly selected from healthy volunteers. Patients with previous surgeries, intraocular inflammation, dry eye, or allergic conditions were excluded.
Demographic data are summarized in Table 1. Different parameters were applied in all groups. The Zeiss-Meditec MEL 70 laser (Carl Zeiss Meditec, Jena, Germany) was used for ELAK. Donor tissue diameter was 7.6 mm for Fuchs’ dystrophy and 8.1 mm for keratoconus. Recipient cornea ELAK diameter was 7.5 mm for Fuchs’ dystrophy and 8.0 mm for keratoconus. Eight orientation teeth were used in all cases. The IntraLaseFS system (Abbot Medical Optics, Inc., Abbott Park, IL) was used for FLAK. A top hat configuration was used in Fuchs’ dystrophy and a mushroom-shaped configuration was used in keratoconus. The diameter of the donor was set to 8.6 mm in Fuchs’ dystrophy and 8.5 mm in keratoconus. A 7.5-mm diameter trephination was performed with FLAK in the recipient corneas. Grafts were fixated with double-running 10-0 nylon sutures as described before by Hoffmann.12
Demographic Data of Patients
All patients received topical antibiotics (ofloxacin 0.3%, Floxal EDO; Dr. Gerhard Mann, Chem.-pharm. Fabrik GmbH, Berlin, Germany) until epithelial closure and topical 10 mg/mL prednisolone acetate (Inflanefran forte; Allergan Pharmaceuticals, Ltd., Dublin, Ireland) five times daily after epithelial closure for 6 weeks, tapered gradually (every 6 weeks one drop less per day was applied). Artificial tears were also applied. Systemic therapy comprised oral glucocorticoid therapy: cortisone 150-150-125-125-100-100-80-80-40-40-20-20-10-10 mg tapered gradually daily. Oral carboanhydrase inhibitor and potassium were taken in the first three postoperative days.
In Vivo Confocal Corneal Microscopy
Ophthalmic examinations were performed by masked investigators. Besides routine ophthalmic (slit-lamp) examination, we performed in vivo confocal cornea microscopy using the Heidelberg Retina Tomograph with Rostock Cornea Module (HRT-RCM; Heidelberg Engineering, Inc., Heidelberg, Germany).
This confocal microscope uses a 670-nm wavelength diode laser source with a definition of 384 × 384 pixels over an area of 400 × 400 μm. Before each examination, a drop of proxymetacaine hydrochloride 0.5% (Propakain-POS 0.5%; Ursapharm GmbH, Saarbrücken, Germany) was instilled in the lower conjunctival fornix. Fixation light was used for the contralateral eye to maintain the stability of the eye examined. We used a disposable plastic cap (TomoCap; Heidelberg Engineering, Inc.) as a cover on the microscope head to keep the cornea away from the microscope head at a given distance. Carbomer gel (Vidisic; Dr. Gerhard Mann, Chem.-pharm. Fabrik GmbH) served as a coupling medium. The positioning of the eye examined was monitored by a digital camera. In contrast to the traditional histological cross-section of the cornea, this method allowed en face images to be obtained. Cornea was systemically scanned in the center from the anterior to the posterior surface. Our interest was focused on corneal epithelial layers, subepithelial nerves, Langerhans cells, and keratocytes in the anterior and posterior stroma and endothelium. The ten best-focused images of each layer and location were considered for the analysis in a masked fashion. Special attention was paid to the graft–host junction zone, according to morphology of cut edges, fibrotic tissue, nerve regeneration, inflammatory reaction, and sutures.
Corneal Image Analysis
Cell densities (cells/mm2) were calculated with the inbuilt semi-automatic system as described previously.13 Morphology of Langerhans cells was determined according to the established classification.14–16
Statistical analysis was performed using the STATISTICA version 11.0 (StatSoft, Tulsa, OK) software. Mann–Whitney tests were performed for the comparison of control and treatment groups. A P value of less than .05 was considered significant.
All corneas were transparent on slit-lamp examination; no epithelial defect or significant stromal opacity could be observed and no sign of rejection could be identified. In vivo confocal corneal microscopy could be performed without significant discomfort or any corneal injury. Statistical results are summarized in Table 2.
Comparison of Corneal Cell Densities (cell/mm2) After Femtosecond and Excimer Laser Corneal Transplantation (Mean ± SD)
Superficial epithelial cells had variable reflectivity in all patient groups. No significant difference could be observed between groups and none of the groups were different from the control group (P > .05) (Figure A, available in the online version of this article). Corneal basal epithelial cells in the center showed normal size, reflectivity, density, and structure in all patient groups irrespective of the trephination technique applied and were not different from that of the control group (P > .05) (Figure A).
In vivo confocal microscopic image of the corneal epithelium: (A–D) superficial and (E–H) basal cells after (A, E) excimer laser-assisted penetrating keratoplasty (ELAK) in Fuchs’ dystrophy and (B, F) femtosecond laser-assisted keratoplasty (FLAK) and after (C, G) ELAK and (D, H) FLAK in keratoconus. Original size of all images in all figures is 400 × 400 μm. Bar indicates 50 μm.
Keratocyte density was found to be lower in the anterior and posterior stroma in all patient groups compared to the control group (P < .05) (Figure B, available in the online version of this article).
In vivo confocal microscopic image of the (A–D) anterior and (E–H) posterior stromal keratocytes after (A, E) excimer laser-assisted penetrating keratoplasty (ELAK) in Fuchs’ dystrophy and (B, F) femtosecond laser-assisted keratoplasty (FLAK) and after (C, G) ELAK and (D, H) FLAK in keratoconus. Size of all images is 400 × 400 μm.
Endothelial cells were well structured, and neither significant polymagnetism nor pleomorphism could be observed. Cell density was reduced independent from trephination technique or corneal pathology (Figure C, available in the online version of this article).
In vivo confocal microscopic image of the endothelium in Fuchs’ dystrophy after (A) excimer laser-assisted penetrating keratoplasty (ELAK) and (B) femtosecond laser-assisted keratoplasty (FLAK) in keratoconus after (C) ELAK and (D) FLAK.
Subepithelial nerves could be detected in the center only in 2 eyes after penetrating keratoplasty (Figure 1). Stromal nerve fibers (Figures 1A, 1C, 1D) and subepithelial nerve plexus (Figure 1B) could be found in the graft–host junction zone, especially in the proximity of sutures (Figure 1A). Gracile subepithelial nerve fibers were noted (Figures 1E–1F).
In vivo confocal microscopic image of the (A, C, D) stromal and subepithelial (B, E–H) corneal nerves in Fuchs’ dystrophy after (A, E) excimer laser-assisted penetrating keratoplasty (ELAK) and (B, F) femtosecond laser-assisted keratoplasty (FLAK) in keratoconus after (C, G) ELAK and (D, H) FLAK. Stromal nerve fibers are indicated with white arrows (A, C, D) and subepithelial nerve plexi with empty arrows. Nerves could be found in the graft–host junction zone, especially in the neighborhood of sutures (A). Note the gracile subepithelial nerve fibers (E, F).
Graft–Host Junction Zone
Graft–host junction was characterized by highly reflective fibrous tissue along the cutting edges. In the fibrotic tissue at the interface, some keratocyte-like cells (suspected to be fibroblasts) and some inflammatory cells could be seen. No clear difference between FLAK and ELAK trephined corneas could be observed with confocal microscopy. Orientation teeth could be well visualized after ELAK trephination (Figure 2). No morphologic difference of the cut edge after FLAK and ELAK could be demonstrated.
In vivo confocal microscopic images of the corneal graft–host junction zone in Fuchs’ dystrophy after (A) excimer laser-assisted penetrating keratoplasty (ELAK) and (B) femtosecond laser-assisted keratoplasty (FLAK) in keratoconus after (C) ELAK and (D) FLAK. Orientation tooth after ELAK is indicated with a white arrow (A). Some fibroblasts (arrowhead) in the (B) fibrotic tissue, (C) gracile subepithelial nerve plexi, (D) and inflammatory cells in the interface and epithelial cells are visible.
Langerhans cells could be detected only in 5 cases in the center of the grafts (all in Fuchs’ dystrophy, 3 following FLAK, 2 after ELAK). Langerhans cells could be found in the periphery of the graft in all groups. Regarding Langerhans cell morphology, types 1 and 2 were observed according to our classification15,16 or Langerhans cells without dendrites and with small protrusions according to Zhivov et al.’s classification14 (Figure 3).
In vivo confocal microscopic signs of inflammatory reaction after penetrating keratoplasty. Some corneal Langerhans cells with elongated dendrites (A, arrow) in the subepithelial layer, 48 μm from the corneal surface after femtosecond laser-assisted keratoplasty (FLAK) in Fuchs’ dystrophy. (B) Empty arrow indicates corneal precipitate after excimer laser-assisted penetrating keratoplasty (ELAK) in Fuchs’ dystrophy.
It has been shown that non-contact trephination systems provide superior cutting edges in the donor and recipient cornea and less injury to the graft endothelium compared to mechanical trephination.2 In our study we compared the most advanced laser trephine systems (the excimer laser and the femtosecond laser devices) and used a confocal corneal microscopy as a tool to have better understanding of the postoperative structural and cellular changes of the cornea.
There are several comparative studies published on FLAK and manual penetrating keratoplasty. Shivanna et al.17 reviewed 15 studies comparing FLAK and conventional penetrating keratoplasty in advanced keratoconus. All suggested that FLAK is superior to manual penetrating keratoplasty in wound closure, visual recovery, visual acuity, astigmatism, and corneal healing. Although mechanical penetrating keratoplasty can only provide vertical wound surface, FLAK is capable of cutting different configurations, including the so-called top hat, mushroom, zig-zag, and Christmas tree configurations.18 Despite these advantages, there are drawbacks of this technique. Because the cornea must be stabilized during the cutting procedure by using a suction ring to dock the eye, corneal deformation can occur and postoperative astigmatism may subsequently worsen. This limits the applicability of the procedure, especially for eyes with unstable corneas. The other limitation of FLAK is that it can only be performed on corneas with preserved transparency. On the other hand, ELAK lacks these disadvantages and has been proven to be a reliable tool in performing vertical corneal cuts.
In vivo confocal microscopy is a powerful instrument to study the cornea and to explore the pathological alterations associated with certain corneal diseases.15,16,19 The horizontal resolution of the device is 1 to 2 μm, so this given magnification served us well to investigate corneal conditions at a cellular level after keratoplasty.
To the best of our knowledge, this is the first study to investigate the corneal epithelial, keratocyte, and endothelial densities and the morphological alterations of the cornea after ELAK by means of confocal microscopy in comparison with FLAK and a control group. Imre et al.7 described corneal re-innervation after mechanical penetrating keratoplasty and showed that only 2 of 7 corneas were re-innervated 1 year after penetrating keratoplasty, but all corneas were re-innervated 5 years later. Our study is in line with Imre et al.’s findings because we observed only minimal re-innervation and those nerves were gracile and irregular.
Bucher et al.20 demonstrated by confocal microscopy that Fuchs’ dystrophy itself can alter corneal nerve morphology. They noted significant inverse correlations for total nerve length, total nerve number, number of main nerve trunks, and number of nerve branches with increasing severity of Fuchs’ dystrophy.
The literature is controversial in terms of keratocyte density after penetrating keratoplasty. Feizi et al.21 described comparable keratocyte densities to the control group in their nonrandomized cross-sectional study of 19 patients, whereas Patel et al.8 found the keratocyte density to be decreased in penetrating clear grafts compared to normal corneas years after surgery. Niederer et al.9 came to the same conclusion in their cross-sectional study examining 42 patients after penetrating keratoplasty. Imre et al.7 also found similar patterns in both anterior and posterior keratocyte density.
We found decreased keratocyte densities in both the anterior and posterior stroma layers similar to what Imre et al. found in their study after penetrating keratoplasty. In our study, it ranged between 890 and 956 cells/mm2 in the anterior and 529 to 596 cells/mm2 in the posterior stroma. In Imre et al.’s report, further decrease of keratocyte density was reported with approximately 400 cell/mm2.7
The cause of the lower keratocyte densities found after perforating keratoplasties is not fully understood. According to some, programmed cell death (apoptosis) may play a key role in that and it may occur in corneas with minimal collateral damage or during development and wound healing of other causes. Longstanding use of steroid drops may have an effect not only on the activation, but also on the migration and programmed cell death of keratocytes.
It has been shown that either topical or oral steroid use interferes with Langerhans cell migration and maturation.15,22 It has also been proposed that Langerhans cells are using the subbasal nerve plexi as a skeleton for horizontal movement.
The proper sizing of the graft is essential for the integrity of the wound. Price and Price6 performed FLAK with identical donor and recipient diameters. In the early postoperative period, the posterior rim buckled with suture placement, but by month 3, apposition of the posterior surface recovered. If the graft was undersized relative to the recipient, the posterior donor tissue was well opposed after surgery, but a gap developed between the posterior rim of the donor and recipient with wound healing. No exact data are available yet to achieve the best visual and biomechanical outcome. In vivo confocal microscopy of our patients demonstrated exact alignment of the graft on the anterior and posterior surface with donor corneas larger than 8.1 mm.
Birnbaum et al.23 summarized their experience with 123 consecutive FLAK procedures and found the wound healing to be faster and more stable, which they considered the consequence of profiled trephinations and larger wound area. They presented anterior optical coherence tomography images around the graft–host junction zone, which were consistent in terms of wound structure with our confocal microscopic results.
Szentmáry et al.24 summarized the immunological aspects of excimer and femtosecond laser-assisted trephine penetrating keratoplasty. Immunological graft reactions occur, which may lead to endothelial cell loss and irreversible transplant rejection. They reported that ELAK did not have any immunological disadvantages, because 13.9% rejection occurred in eyes with keratoconus and 2.9% in eyes with Fuchs’ dystrophy after 3 years. In their results, the “mushroom-shaped” trephination in FLAK resulted in a higher rate of transplant rejection (21.8% in keratoconus after 14 months) in contrast to “top hat-shaped” penetrating keratoplasty leading to rejection in 6.6% of the patients in Fuchs’ dystrophy after 14 months. Our in vivo study showed minimal activity of the Langerhans cells, which could suggest that the T-cell–mediated immune response is suppressed. It has been postulated that Langerhans cells require subbasal corneal nerves for migration.
Our comparative study included only full-thickness corneal transplantation, whereas different techniques of lamellar keratoplasty (deep anterior lamellar keratoplasty,25,26 Descemet stripping endothelial keratoplasty, and Descemet membrane endothelial keratoplasty27) are more frequently applied in many situations, such as endothelial dystrophies. However, in many cases lamellar keratoplasty cannot be substituted for penetrating keratoplasty (eg, advanced bullous keratopathy and keratoconus with hydrops, scarring, and ulcerations). Therefore, a better understanding of the exact mechanisms of femtosecond and excimer laser-assisted penetrating keratoplasty is required to achieve good results.
The better wound integrity and faster donor–recipient junction wound healing could not be demonstrated and excimer laser trephination cut edges were not different from femtosecond cuts according to the confocal corneal microscopic results. Gaster et al.28 presented 56 cases of FLAK zig-zag–shaped incisions in keratoconic eyes with favorable results.
One of the limitations of the study is that the HRT RCM uses laser light and the reliability of data collected greatly depends on the transparency of the organ examined. A further limitation of our study is that only a single examination of patients is reported. A prospective follow-up of all parameters is planned to be reported soon with more reliable correlations.
Our data demonstrated that ELAK is not inferior to FLAK in performing corneal cuts, corneal integrity, and immune reaction. In vivo confocal microscopy provided evidence that good alignment of graft–host junction could be achieved with both trephination techniques, which provide the chance for sufficient epithelial regeneration and potential re-innervation. The low density of Langerhans cells revealed well-controlled cellular immunological response in all cases.