Manual Suction versus Femtosecond Laser Trephination for Penetrating Keratoplasty: Intraocular Pressure, Endothelial Cell Damage, Incision Geometry, and Wound Healing Responses

Abstract

Purpose.: To measure real-time intraocular pressure (IOP) during trephination with a manual suction trephine (MST) and the femtosecond laser (FSL), and to assess endothelial cell damage, incision geometry, and wound healing response with these procedures.

Methods.: IOP was monitored with an intracameral sensor. Eight rabbits underwent manual suction trephination. Eight rabbits had FSL trephination (FSL-T). Slit lamp photography, confocal microscopy, and anterior segment optical coherence tomography (AS-OCT) were performed at baseline and postoperatively. Animals were sacrificed at 4 hours and 3 days. Tissue was examined with scanning electron microscopy (SEM) and immunohistochemistry for an array of wound-healing markers. Separately, 6 human corneas had MST (3) and FSL-T (3). Incision geometry was imaged with high resolution Optovue AS-OCT.

Results.: The average IOP during MST and FSL-T was similar (37 mm Hg). There was wider IOP fluctuation during the MST cutting phase (60 mm Hg maximum). There were 1–2 rows of endothelial loss on either side of the incision for FSL-T and 2–5 rows deep for MST. Immune cell responses at 4 hours (CD11b) were comparable, greater apoptosis with FSL-T (TUNEL) occurred at 4 hours, and there was increased keratocyte proliferation at 3 days (Ki67) with FSL-T. There was significantly greater undercutting of the cornea with MST (46.86 degrees versus 16.72 degrees).

Conclusions.: There is more IOP variation during MST. Average IOP is 37 mm Hg for both techniques. More endothelial damage and undercutting of the cornea occurs with MST. The wound healing response to FSL-T appears greater at 3 days.

Introduction

Despite the recent adoption of endothelial keratoplasty in many countries, penetrating keratoplasty remains the commonest form of corneal transplantation surgery worldwide. Corneal surgeons in the early 20th century rapidly realized the importance of accurate graft–host apposition for reducing postoperative complications, including wound leakage and postoperative astigmatism after penetrating keratoplasty (PK). Various trephines have been designed since that time to overcome the limitations, such as the difficulty of maintaining a perpendicular plane and correct graft–donor size match, with freehand trephination for PK. These have included mechanically driven trephines, punch devices, and vacuum trephines. ¹ Despite these innovations, mean postoperative astigmatism is reported to be between 2.5 to 5.0 dioptres in large keratoplasty series. ^2–6  

The geometry of the trephine incision is determined by a number of complex factors, including corneal curvature, intraocular pressure, corneal edema, and the biomechanical properties of the cornea. Thus, the desired perfectly perpendicular trephination incision is perhaps never achieved and probably always mismatched to some degree between graft and host. The requirement for typically oversizing the graft button in relation to the host trephination is a further testament to these factors. ^7,8 In addition to potential circumferential graft–host interface mismatch, the recent availability of anterior segment ocular computed tomography (AS-OCT) imaging has provided detailed information that suggests that posterior graft–host malposition may be a relatively overlooked common phenomenon. ⁹ Moreover, suture technique, suture tension, individual corneal wound healing, biomechanical properties, and surgeon factors are further complicating variables. 

The femtosecond laser (FSL), a 21st century laser scalpel, provides the potential for enhanced accuracy for the modern corneal surgeon and has been rapidly adopted as the instrument of choice for creating corneal flaps of predetermined depth during LASIK. With the use of proprietary software, FSLs are now capable of creating circular through-corneal trephinations for PK or multiplanar keratoplasty incisions, which potentially increase graft–host interface surface area and fit. These latter adaptations have been shown to create a better graft–host fit with less likelihood of wound leakage, greater wound strength, and potentially less postoperative astigmatism. ^10–14  

The purpose of this study was to critically evaluate the similarities and differences between fixating vacuum manual trephination and FSL through-corneal trephination in terms of real-time procedural IOP changes, corneal endothelial cell damage, wound healing responses, and through-corneal incision geometry. 

Methods

The aims of this research were to investigate procedural (IOP) and anatomic differences between manual suction trephination and FSL laser trephination of the cornea. To accomplish these aims, experiments were structured according to the schematic in Figure 1. 

Figure 1.

 

Schematic of experimental protocol. MT, manual trephine; FSLT, femtosecond laser trephination; CFM, confocal microscopy; SLP, slit-lamp photography; IHC, immunohistochemistry; EC, endothelial cells.

Animals

Sixteen male or female New Zealand White rabbits (2–2.5 kg) were procured from the animal holding unit of the National University of Singapore, Singapore. All procedures performed in this study complied with the ARVO Statement for the Use of Animals in Ophthalmology and Vision Research guidelines and approved by the Institutional Animal Care and Use Committee. 

Rabbits were anaesthetized with a combination of ketamine and xylazine (ketamine 40 mg/kg; xylazine 20 mg/kg, IM). Further proparacaine drops were used for topical anesthesia during surgery. Animals were sacrificed with an overdose intracardiac injection of pentobarbitol. 

Human Corneal Tissue

Six cadaveric human donor corneas of research grade were obtained from the Lions Eye Institute for Transplant & Research (Tampa, FL). Donor ages ranged from 24 to 77 years. Mean endothelial cell density was 2358 (SD 309). Mean death to enucleation time was 10.7 hours. 

Intraocular Pressure Measurements

Real-time IOP was measured in the anterior chamber using a previous published technique. ¹⁵ Briefly a 30-gauge winged infusion cannula connected to a pressure transducer was inserted through the limbus into the anterior chamber. The transducer was prepared according to the instructions of the manufacturer to assure a tight seal, and all air was flushed from the system. Calibration was performed before starting each trephination. To ensure the transducer was working accurately, IOP measurements were also taken with a Tono-pen (Reichert-Jung, Depew, NY) before commencement of the experiment. During each procedure, the IOP was recorded continuously from the time of application of suction through to the end of the trephination and cessation of suction. 

Rabbit Manual Trephination Procedure

Each rabbit was anaesthetized as described above. A lid speculum was inserted to separate the eyelids of the right eye of each rabbit. A 6/0 silk traction suture was placed through the nictitating membrane, which was retracted. A new 8-mm suction trephine (Ultrafit vacuum trephine; UK Network Medical, Ripon, UK) was used for each of the rabbits (Fig. 2). The trephine was centered on the cornea, and suction was applied using a 5-mL syringe. Partial trephinations were performed on each rabbit to a depth of roughly 180 to 200 μm by advancing the trephine blade by three turns (each turn advances appoximately 60 μm). This constituted roughly ⅓ (100 microns) of the full rabbit corneal thickness of 300 μm. Full-thickness trephination was avoided as this would have led to collapse of the anterior chamber and possibly prolapse of intraocular contents. A depot injection of subconjunctival cefuroxime was given (0.1 mL) after trephination, before a bandage contact lens (PureVision; Bausch & Lomb, Madison, NJ) was placed over the cornea.The lid speculum and traction suture was then removed, and the lids were closed with a suture tarsorrhaphy (6/0 silk). 

Figure 2.

 

Image showing manual suction trephination of the rabbit cornea.

Rabbits in the manual trephination group were subdivided for sacrifice at either 4 hours (N = 4) or 3 days (N = 4). To evaluate endothelial cell loss as a result of the trephination process, each of the rabbits in the 3-day group was anaesthetized again prior to sacrifice, and manual trephination was performed on the fellow eye (left) of each animal as described above. On this occasion, full-thickness trephination of each cornea was performed until the point of entry into the anterior chamber. Endothelial cell damage resulting from anterior chamber collapse was minimized by injecting a visco-elastic agent (Viscoat; Alcon, Inc., Hunenberg, Switzerland) into the anterior chamber of each animal using a 27-gauge needle prior to trephination.The trephine incision was not extended circumferentially with scissors as this would have led to endothelial cell loss from the scissors rather than the trephine alone. 

Each animal was immediately sacrificed by pentobarbital, and corneoscleral rims were excised for immunohistochemistry (right eyes, N = 4) and electron microscopy (left eyes, N = 4). 

Femto Second Laser Trephination Procedure

The 500 kHz VisuMax FSL (Carl Zeiss Meditec, Jena, Germany) was used to perform laser trephination. The VisuMax has a curved treatment cone, which conforms to the curvature of the human cornea and is likely to result in minimal corneal distortion on applanation. This contrasts with the Intralase FSL (Abbott Medical Optics Inc., Abbott Park, IL), which has a flat treatment cone that compresses the cornea on applanation.Briefly, each animal was prepared as described above. A small-sized curved docking cone was used in the right eye of each rabbit (N = 8). Following application of suction, the penetrating keratoplasty firing sequence was initiated using the following parameters: diameter 8 mm, energy 180 nJ, 2-μm spot and line separation, angle 90°. In the case of the FSL, the trephination incision advances from the posterior to the anterior cornea. Full-thickness trephination was possible with the FSL without collapse of the anterior chamber as small tissue bridges prevented immediate separation of the incision plane. No attempt was made to manually separate these bridges. Following completion of the laser sequence, each rabbit had subconjunctival cefuroxime, a bandage contact lens, and tarsorraphy performed as described above. 

The FSL trephination group was subdivided into two groups, with animals for sacrifice at 4 hours (N = 4) and 3 days (N = 4). Endothelial cell damage was assessed in the left eyes of rabbits in the 3-day group by repeating the FSL trephination procedure as described above, before sacrifice of the animals. Viscoelastic was not used as there was no chamber collapse. Corneoscleral rims were excised for immunohistochemistry (right eyes, N = 4) and electron microscopy (left eyes, N = 4). 

Assessment of Trephine Incision Geometry in Human Corneas

Human corneas (N = 6) were mounted and pressurized in a disposable artificial anterior chamber (AAC) (Coronet; UK Network Medical). Pressure was maintained during trephination at a physiological level. Trephination was performed using either an 8-mm-diameter suction trephine (N = 3) (Ultrafit vacuum trephine; UK Network Medical) or an 8-mm-diameter FSL trephination sequence (N = 3). Immediately following trephination, each cornea was imaged while still mounted in the AAC using the RTVue spectral domain anterior segment optical coherence tomography (SDASOCT) (Optovue, Freemont, CA) with a resolution of <5μ. ¹⁶ Cross-sections of the trephination incision were obtained at a minimum of 8 points circumferentially. The bevel angle of each trephine incision was assessed by measuring the angle between the epithelial and endothelial surfaces of the trephine incision using Photoshop software (Adobe Creative Suite 5, http://www.adobe.com/products/) that allowed a virtual perpendicular line to be drawn from the anterior limit of the incision to the posterior cornea (representing the ideal incision) and measurement of the angle of deviation from this line. 

Slit-Lamp Photography, Optical Coherence Tomography (OCT), and Confocal Microscopy of Rabbit Corneas

Slit-lamp photographs (Nikon FS-3V; Nikon, Tokyo, Japan) were captured before surgery, immediately after surgery, and 3 days postoperatively. Corneal cross-sectional visualization in the rabbits pre- and postoperatively was performed using a Visante ASOCT unit (Carl Zeiss Meditec). All rabbit corneas had confocal corneal microscopy (CM) with the Heidelberg retina tomography HRT3 (Heidelberg Engineering GmbH, Heidelberg, Germany). A carbomer gel (Vidisic; Mann Pharma, Berlin, Germany) was used as immersion fluid. All corneas were examined centrally with at least 3 z-axis scans through from epithelium to endothelium. In vivo confocal micrographs were analyzed with the Heidelberg Eye Explorer version 1.5.1 software (Heidelberg Engineering GmbH). Image measurements of trephine incision separation were made using the ImageJ software package (http://rsbweb.nih.gov/ij/). Animals were imaged with these devices at the time points shown in Figure 1. 

Scanning Electron Microscopy (SEM)

The rabbit cornea was excised from the whole globe. Specimens were fixed in 2.5% glutaraldehyde (Sigma, St. Louis, MO) and 0.1 M sodium cocodylate (pH 7.4) overnight at 4°C. They were then transferred and stored in sodium cocodylate buffer (EMS, Hatfield, PA). Before processing, the samples were washed twice in distilled water for 10 minutes before immersion in 1% osmium tetraoxide (FMB, Singapore) for 2 hours at room temperature. Following this, they were dehydrated in an ascending concentration of ethanol, with 95% and 100% concentrations performed twice. The samples were then dried in a critical point dryer (Bal-Tec, Balzers, Liechtenstein) and mounted on SEM stubs using carbon adhesive tabs. They were then sputter coated with a 10-nm-thick layer of gold (Bal-Tec) and examined with a scanning electron microscope (JSM-5600; JEOL, Tokyo, Japan) at 15 W. 

Tissue Fixation and Sectioning

For immunofluorescent staining, excised rabbit corneas were first embedded in optical cutting temperature (OCT) embedding compound (Leica Microsystems, Nussloch, Germany). Frozen tissue blocks were stored at −80°C until sectioning. Serial sagittal corneal 10-μm sections were cut using a Microm HM550 cryostat (Microm, Walldorf, Germany). Sections were placed on polylysine-coated glass slides, air-dried for 15 minutes, and then processed for immunohistochemistry. 

Immunohistochemistry

Sections were fixed with freshly prepared 4% paraformaldehyde (Sigma) for 20 minutes, washed with 1X PBS, blocked with 4% bovine serum albumin (Sigma) in 1X PBS, 0.15% Triton X-100 (Sigma) for 1 hour, and then incubated with either rat monoclonal antibody against CD11b (BD Pharmingen) diluted 1:100 in the blocking solution or with 1:100 rat monoclonal antibody against Ki-67 (Invitrogen, Carlsbad, CA) at 4°C overnight. After washing with 1X PBS, the sections were incubated with goat antirat Alexa Fluor 488-conjugated secondary antibody (Invitrogen) at room temperature for 1 hour. Slides were then mounted with Vectashield containing DAPI (Vector Labs, Burlingame, CA). For negative controls, nonimmune serum was used in place of the specific primary antibody. Sections were observed and imaged with a Zeiss Axioplan 2 fluorescence microscope (Zeiss, Oberkochen, Germany). Normal corneas were used as uninjured controls. 

TUNEL Assay

To detect apoptotic cells, a fluorescence-based TUNEL assay (In Situ Cell Death Detection Kit; Roche Applied Science, Indianapolis, IN) was used following the manufacturer's instructions. 

Statistical Analysis

Data were expressed as mean ± SD where appropriate. The P value was determined using ANOVA and t-test with the GraphPad Prism statistical package (GraphPad Software, Inc., La Jolla, CA). Findings were considered to be statistically significant when P < 0.05. 

Results

The trephine incision was clearly identifiable on slit-lamp photography (Figs. 3a, 3b) and confocal microscopy (Figs. 3c, 3d) immediately after the surgery. Manually trephined corneas had wider separation of the cut surface of the cornea (129.23 μm, SD 20.14) compared with FSL trephined incisions (36.48 μm, SD 7.28) (Figs. 3d and 3c, respectively). On confocal microscopy tissue debris was clearly evident in the depth of both the manual and FSL trephine incisions (Figs. 3c, 3d). ASOCT confirmed the incision depth (not shown). 

Figure 3.

 

Color slit-lamp photos of rabbit corneal trephination with the femtosecond laser trephination (FSLT) (a). Note the small air bubbles visible within the anterior chamber (arrow). The manual trephine incision (MT) (b) is clearly visible (arrow). Confocal micrographs show tissue debris within the incisions after FSLT (c) and MT (d) (arrows). Note that the separation of the incision is wider following MT (d) (mean 129 μm) compared with FSLT (c) (mean 36.45). Keratocytes can be seen in the adjacent stroma. Yellow scale bar is 50 μm.

IOP Variation during Manual and FSL Trephination

IOP measurements during different phases of trephination are summarized in Table 1. Although the overall average procedural IOP was similar in both groups (37.73 mm Hg SD 9.21 mm Hg manual and 37.79 mm Hg SD 5.14 mm Hg for FSL), there were variations at different stages of trephination (Fig. 4a).IOP was significantly lower during application of suction with the manual trephine (32.65 mm Hg SD 8.37 mm Hg vs. 40.61 mm hg SD 8.15 mm Hg for FSL, P < 0.001). IOP was only slightly higher during manual cutting (39.43 mm Hg compared with 38.11 mm Hg for FSL). Further analysis of intra-operative pressure variations revealed variability during the cutting phase during manual trephination (Fig. 4b) compared with the FSL (Fig. 4c). The duration of trephination was greater with the FSL, 26.04 seconds (SD 5.66) and 17.02 (SD 5.64) seconds with the manual trephine (Fig. 4d). Maximal IOP occurred with the FSL (63 mm Hg). 

Table 1.

 

Real-Time Intraocular Pressure at Key Stages of Trephination

Procedure	Baseline	SD	Suction on	SD	Cutting	SD	Suction off	SD	Avg. procedure	SD	Duration	SD
MT	8.31	3.22	32.65	8.37	39.43	10.32	18.14	4.54	37.73	9.21	17.02	5.64
FSLT	7.94	2.54	40.61	8.15	38.11	4.55	19.38	4.70	37.79	5.14	26.04	5.66

 Mean intra-operative real-time IOP measurement at baseline, suction-on, cutting, suction-off phases of trephination with the femtosecond laser and manual trephine. Mean overall procedural IOP and duration of trephination for each group are also given.

Figure 4.

 

(a) Mean IOP measurements during various phases of manual and femtosecond laser trephination (FSLT). (b) IOP variation during individual phases of manual suction trephination; note variation between suction and cutting phases. (c) IOP variation during FSLT; note constant IOP during suction and cutting phases. (d) Graph showing duration of mean IOP versus time for manual and FSLT.

Endothelial Cell Damage during Trephination

SEM images of endothelial cells were analyzed in the area adjacent to the trephination incision. The FSL laser created a clean cut with minimal collateral endothelial cell damage (Fig. 5a). Sporadic cell loss occurred with the FSL, maximally resulting in one row of cells on either side of the trephine incision being damaged. In contrast, damage to adjacent cells with the manual trephine was more widespread, involving three to four rows of cells either side of the trephination (Fig. 5b). By measuring the width of an individual endothelial cell (11.13 μm) and determining the circumferential distance of the 8-mm-diameter trephine incision (25,132 μm), the number of cells lost per row of endothelial cell damage on either side of the trephination incision was able to be determined (25,132/11.13 × 2 = 4516 cells). Thus, the FSL laser trephination, which caused cell damage to one row of cells on either side of the incision resulted in a total cell loss of 4516 cells compared with 13,548 to 18,064 cells with the MT (3–4 rows), P < 0.0001. 

Figure 5.

 

SEM images of endothelial cells at the edge of the trephine incision after (a) femtosecond laser trephination (FSLT) and (b) manual trephination. Note that only a row of cells is damaged after FSLT compared with up to four rows of cells following manual trephination.

Wound Healing Responses after Trephination

Corneal sections stained for cellular apoptosis using the TUNEL assay showed maximal keratocyte apoptosis at 4 hours after trephination in both groups. Apoptotic cell numbers were significantly greater after manual trephination at 4 hours (P < 0.001) but not at 3 days, when apoptotic cell numbers were similar between both groups (Fig. 6a). Although the distribution of apoptotic cells was limited to the area immediately adjacent to the trephination with the FSL (Fig. 7a), a separate additional cluster of apoptotic keratocytes was evident outside the trephination incision in the manual trephination group, possibly arising from where corneal suction may have damaged underlying cells during application of suction (Fig. 7b). 

Figure 6.

 

(a) TUNEL, (b) CD11b, and (c) Ki67 positive cell numbers following femtosecond laser (FSLT) trephination and manual trephination at 4 hours and 3 days.

Figure 7.

 

Representative images of cells staining positively for TUNEL positive (red) apoptotic cells at 4 hours after (a) FSL (femtosecond laser) and (b) manual trephination. Note the second larger cluster of apoptotic cells in the manual group probably as a result of corneal suction. CD11b inflammatory monocyte staining (arrows) at 4 hours after (c) FSL and (d) manual trephination. Ki67 positive cells (green and arrowed) at 3 days after (e) FSL and (f) manual trephination.

CD11b staining for monocytes was maximally positive at 4 hours after trephination and significantly greater in the FSL trephination group (P < 0.001). CD11b positive cell numbers were similar between both groups at 3 days postsurgery (Fig. 6b). The distribution of monocytes was typically adjacent to the site of the incision in both groups (Figs. 7c, 7d). 

Proliferating keratocytes staining positively for Ki67 were minimally raised at 4 hours in both groups and significantly raised at 3 days in the FSL trephination group (P < 0.001) (Fig. 6c). Proliferating cells were typically distributed adjacent to the incision site (Figs. 7e, 7f). Epithelial cell hyperplasia/plugging of the incision defect was evident at 3 days following trephination in both groups. 

Through Corneal Trephine Geometry in Human Corneas

The mean bevel angle was 16.72° (SD 6.09) sloping outward (internal diameter greater than external = undercutting) in the FSL group and 46.86° (SD 11.36) for the MT group (P < 0.0001) (Fig. 8). Stepped irregularities in the trephine incision were visible in the manual group but not the FSL group on spectral domain ASOCT (Fig. 8). 

Figure 8.

 

RTVue Fourier domain anterior segment optical coherence tomography images of the trephine incisions (arrowed) following (left) femtosecond laser and (right) manual trephination. Note that the bevels of both incisions have a wider internal than external diameter (undercutting). The manual incision profile appears serrated and irregular.

Discussion

Trephination of the cornea is beset by various confounding variables that lead to deviation from a perfect perpendicular incision match between graft and host tissue. This study reveals previously unreported real-time changes in IOP during trephination, and highlights important differences between a commonly used disposable fixating vacuum trephine and trephination with a modern FSL. 

It is reassuring in terms of retinal and optic nerve perfusion to find that the average IOP during manual vacuum trephination (MVT) and FSL is approximately 38 mm Hg. Thus IOP elevation during trephination with the VisuMax FSL is similar to manual trephination. This is particularly relevant in elderly patients, where the risk of retinal vascular occlusion may be greater at increased levels of IOP. Peaks of up to approximately 65 mm Hg did occur during both types of trephination. IOP elevation with the Visumax laser also appears to be significantly lower than with other FSLs, where IOPs up to >180 mm Hg have been reported. ^17,18 Despite starting at a lower IOP during application of manual suction (32.65 mm Hg), IOP levels varied significantly during the cutting phase of MVT. In contrast, there is no significant variation in IOP between the suction and cutting phases of FSL trephination. This is an important finding, as constant intraocular pressure is known to be an important factor in maintaining a constant trephine plane. ¹  

During manual trephination, the surgeon typically holds the trephine with his nondominant hand and incrementally advances the trephine blade by turning the blade forward. Variation in IOP during the manual cutting phase may arise from the need to turn, regrasp, and turn the trephine and the resultant inadvertent down force on the eye. In contrast, once the curved applanation cone of the VisuMax laser engages with the cornea and once suction is successfully achieved, no further movement occurs during FSL trephination. This is reflected in the absence of any significant variability in IOP between suction and cutting phases. 

Spikes in IOP cause increased central corneal curvature and pushes the cornea further into the central manual trephine aperture, resulting in undercutting of the cornea. These spikes and resultant deviation in trephine plane can be appreciated as subtle staggered ridges in the cut cornea on ASOCT imaging (Fig. 7C). A central obturator present in manual trephines such as the Hanna trephine, but absent in the Coronet trephine used in this study, may reduce central corneal tissue herniation and the resulting undercutting of the cornea (internal diameter > external diameter = bevel slopes outward). ¹⁹ Although the FSL laser maintains constant IOP during cutting, the curved applanator of the treatment pack also prevents upward herniation of tissue, hence helping to maintain a constant trephine plane. 

Analysis of ASOCT images of trephine incisions in this study revealed a significantly greater degree of undercutting with the manual trephine compared with the FSL (P < 0.001). A number of other factors apart from fluctuations in intraoperative IOP are likely to contribute to this finding. Importantly, the anatomically curved applanation cone of the VisuMax FSL treatment cone is less likely to cause tissue distortion than some other models of FSL, such as the Intralase (Abbott Medical Optics Inc.) or Ziemer FSLs (Ziemer Ophthalmic Systems, Port, Switzerland) that utilize a flat applanating cone.Apart from the logical benefits of an undistorted cornea, the absence of distortion may also facilitate more precise laser spot focusing and less laser attenuation. The FSL is a near infrared laser that causes precise photo-disruption of tissue with the generation of plasma cavitations. The confluence of these cavitations in a plane results in tissue dissection. Invariably small bridges of tissue remain between cavitations and require manual separation, typically with a Sinskey hook or scissors. In the case of trephination for PK, these tissue bridges help hold the incision together while preventing gaping of the incision as it progresses through the tissue, thereby helping to maintain a constant anteroposterior trephination plane. Less gaping is also likely as the incision proceeds from the posterior concave surface of the cornea to the anterior convex surface of the corneal surface (the opposite applies to the manual trephine, which cuts from convex to concave surfaces). In contrast, manual trephination results in immediate tissue gaping as a result of the elastic biomechanics of the cornea that results in contraction and upward herniation of the central cornea (within the trephine) and retraction of peripheral cornea (outside the trephine blade), with resulting undercutting of the cornea. 

An essential prerequisite of trephination for PK is to cause minimal damage to the endothelial cells. Kim et al. (using a 60 kHz, Intralase FSL; Abbott Medical Optics Inc.), as the authors did, found significantly less cell loss during FSL trephination. ²⁰ In this study, the authors found that endothelial cell loss was nearly four times greater along the circumference of the MT incision (range 13,548–18,064) (P < 0.0001). Previous studies have reported even higher levels of cell loss after manual trephination than have been reported here. ^20–22 The lower rate of cell loss in the manual group in this study compared with other published studies may be related to the use of viscoelastic before the trephination. In practice, manual trephination rarely results in a circumferentially complete incision, and scissors or a blade are needed for completion, possibly leading to further endothelial cell damage. This caveat may also apply to FSL incisions that require separation of tissue bridges using either a Sinskey hook or scissors. These rates of cell loss should be considered in the context of an overall endothelial cell population of more than 10⁶ in a healthy cornea and may not be clinically significant. 

Immunohistochemical analysis of stromal tissue injury and wound-healing responses in our study showed greater levels of inflammatory cell infiltration and keratocyte proliferation after FSL trephination. Interestingly, although apoptotic cells were generally restricted to the peri-incision tissue following FSL, a separate cluster of apoptotic keratocytes was observed away from the incision, with the manual trephine, possibly corresponding to the area of corneal suction. These two populations of apoptotic cells resulted in a greater number of total apoptotic cells in the manual trephine group at 4 hours. However, further analysis excluding the cluster of apoptotic cells caused by manual suction revealed less apoptotic cells in the immediate peri-incisional tissue following manual trephination compared with the FSL (six cells). The increased inflammatory response following FSL trephination may arise as a consequence of corneal stromal tissue interaction with the laser and the tissue sequelae of energy dissipation, heat generation, shockwave, and intrastromal pressure distribution from cavitation bubble expansion. ²³ In this model, the full-thickness FSL incisions (compared with only partial thickness manual trephine incisions) originating from the endothelial to the epithelial surface may also allow anterior chamber pro-inflammatory cytokines access into the stroma and thereby may stimulate a greater inflammatory cell response. Greater amounts of tissue inflammation and keratocyte proliferation may result in more scar formation and stronger donor–host healing with the FSL. The FSL is known to cause stronger flap adhesion compared with the microkeratome in LASIK. ^24,25 However, Malta et al. found only slightly increased wound strength with the Intralase FSL (Abbott Medical Optics Inc.) compared with the manual trephine for vertical incisions (did not reach statistical significance), but significantly greater wound strength with multiplanar mushroom and top-hat FSL incisions. ²⁶ The latter finding probably reflects the greater contact area afforded by multiplanar incisions. 

In conclusion, trephination with the FSL laser has a number of significant advantages over manual suction trephination for PK. The FSL laser results in a more precise trephine incision, with less graft undercutting, a constant IOP throughout the cutting phase, and significantly less endothelial cell damage. Surgeons with access to a FSL in their practice may consider the use of this device for trephination during PK based on the findings presented here. 

Ackowledgements

The authors thank Livia Yong and Jessica Yu (Carl Zeiss Meditec) for their technical assistance during the use of the Visumax laser as well as Lee Wing Sum for his assistance with the animal experiments. 

Footnotes

 Supported by grants from the National Research Foundation of Singapore-Funded Translational and Clinical Research Programme (NMRC/TCR/002-SERI/2008). Supported by a Pfizer Fellowship Award from the Royal College of Ophthalmologists (UK) (RA).

Footnotes

 Disclosure: R.I. Angunawela, None; A. Riau, None; S.S. Chaurasia, None; D.T. Tan, None; J.S. Mehta, None