3
Corneal anatomy, physiology and response to wounding
Sandip Doshi
The mainstay of current laser refractive surgery centres on manipulation of the properties of the cornea to achieve the desired optical affect. As a result it is essential that the clinician has a strong understanding of the intricate architecture and physiological properties of the organ. Moreover, a firm understanding of the basic science of the cornea allows the clinician to plan treatments that result in a minimal disruption to its structure, and hence achieve a preferred visual outcome.
The cornea occupies approximately 7 per cent of the outer coat of the eye. It is a highly organized five-layered structure (Figure 3.1) that consists of:
•The epithelium;
•Bowman’s layer;
•The stroma;
•Descemet’s layer; and
•The endothelium.
With a standard slit-lamp biomicroscope and appropriate magnification and observation techniques only the epithelium, stroma and endothelium are visible. Each corneal layer is discussed separately in this chapter.
Corneal epithelium
The corneal epithelium is the outermost layer (Figure 3.2) and is an anatomical continuation of the conjunctival epithelium. It is thinnest centrally, where it is typically around 50–60μm in thickness, and thickens to around 70–80μm in the periphery. Thus, the central corneal epithelium constitutes approximately 10 per cent of the total corneal thickness.
The cornea is the major refractive component of the eye. The epithelium is arguably the most important layer for this property.
The reported refractive index of the epithelium varies for different researchers, but commonly is stated to range from 1.375 to 1.543. A uniform regularity and transparency of the epithelium is essential if the cornea is to be a perfect optical surface.
Being the outermost layer of the cornea, the epithelium functions as a barrier to protect the deeper layers from various insults. It also provides a barrier against fluids from the tear film. As any refractive surgeon knows, the epithelium is remarkably tough and resilient to significant trauma, but when damage does occur the epithelium has an excellent recovery rate (see the section on corneal wound healing later).
Epithelium microanatomy
The normal human corneal epithelium can be described as a non-keratinized, stratified, squamous epithelium. Typically, the corneal epithelium is five to six cell layers thick. Three distinct cell types can be identified in the epithelium: basal, wing (or umbrella) and squamous cells.
Figure 3.2
Corneal epithelium and
Bowman’s layer
Figure 3.1
Transverse section cornea
18 ■ Refractive surgery: a guide to assessment and management
Basal cells are the innermost cells of the epithelium and form a monolayer of large, columnar cells. They are relatively uniform in size and are typically 15–20μm in height. Basal cells produce the basement membrane of the corneal epithelium. The organization of the basement membrane allows the epithelium to attach to the remaining layers of the cornea.
Superior (and so more external) to the basal cell layer is the wing cell layer. This consists of two or three layers of cells characterized by a flattened dome-shaped appearance. The wing cell layer tends to be 20–25μm in height. Wing cells are thought to be more mature than basal cells. At the outermost aspect of the epithelium is the squamous cell layer. This superficial layer consists of one or two cell layers. Typically, this layer tends to be about 10–15μm. These cells are the most mature of the epithelium and complete their lifespan by sloughing off into the tear film.
Epithelium ultrastructure
Resident cells of the corneal epithelium
Basal cells are large, columnar cells with a slightly apically displaced nucleus that is spherical or slightly oval and contains dispersed chromatin. The cytoplasm contains many intermediate filaments (mostly grouped into bundles known as tonofilaments), free ribosomes, sparse mitochondria, little granular endoplasmic reticulum, glycogen granules and occasionally Golgi complexes.
Basal cells are attached to the basement membrane via a series of hemidesmosomes. Normally, the latter are so numerous that they occupy at least one-third of the area of the membrane. Interestingly, these bonds are readily broken and reformed – a feature that proves to be vital as part of the wound-healing process.1 Hemidesmosomes are macula junctions that are continuous with basal cell tonofilaments and consist of local thickenings of the plasma membrane (lamina densa) opposite a thickened zone of basement membrane. A lighter interval between the two thickened membranes (lamina lucida) contains numerous fine, connecting filaments, which in turn are attached posteriorly to fine, branched collagen fibrils, called attachment (or anchoring) plaques, that anchor the basal lamina and epithelium to Bowman’s layer.
Anchorage of the epithelium to the basal lamina occurs via a large number of unevenly distributed hemidesmosomes. This results in a strong adhesion of the epithelium to the underlying surface. The
attachment of the basement membrane is very stubborn and normally a blunt trauma does not remove it. However, in some varieties of refractive surgery, such as photorefractive keratectomy (PRK), the removal of the epithelium and the basement membrane (and Bowman’s layer) allows the anterior stroma to become involved in the healing process, which can lead to scarring. Moreover, the presence of a normal basal lamina is essential for reepithelialization.
Basal epithelial cells continue to lay down the basement membrane throughout life and, therefore, there is an increase in its thickness with age. Bergmanson suggested that this might be a reason for a weakening with age of the attachment between the corneal stroma and epithelium.2
Away from the innermost aspect, basal cell borders are characterized by shallow interlocking ridges that cover most of their surfaces, with little or no space between cells. These ridges are least frequent in apposed membranes of basal cells. Adjacent cells are joined by numerous desmosomes. This results in the epithelium being able to withstand a considerable amount of abuse. A relatively small number of gap junctions are also seen throughout the basal layer.
The wing cell layer is characterized by dome-shaped cells with central round nuclei. These cells flatten as they move anteriorly away from the basal layer. The number of cell layers varies from two to three in the central cornea, but may increase to four or five in the periphery. Cells of this layer are derived from the basal layer and represent more mature cells. Cell organelles are more sparse than in the basal layer, which suggests metabolic activity in this layer is slower than that in the basal layer. This is of some relevance when the corneal epithelium is examined in relation to stem cell theory (see later). Like all cells of the corneal epithelium, wing cells are tightly packed and there is very little intracellular space. Cells attach to their neighbours via numerous membrane interdigitations and desmosomes.
As cells continue to mature through the wing cell layer they become flatter and ultimately move more superficially to the squamous cell layer. This layer is characterized by the presence of long, thin, flattened cells that display an intensely stained, elongated nuclei. Squamous cells contain the fewest organelles of the epithelium, which indicates that they possess the lowest metabolic activity. Cells are joined to their neighbours via membrane interdigitations and desmosomes and gap junc-
tions on the inner and lateral aspects. Gap junctions contribute to intercellular adhesions and are probably communicating junctions that permit ionic exchange. Additionally, the surface cells display tight junctions (zonula occludentes) on their most superficial surface. Unlike the other junctions, these girdle the cells and form the closest of contacts without achieving complete fusion. This limits permeability and permits access to the intracellular space from the tear film or, in reverse, through discrete pores only.3
Unlike surface cells of the skin, those of the normal corneal epithelium are non-keratinized and retain some organelles, which indicates that even at this late stage of maturation their metabolic processes are still functioning. This is particularly relevant as part of the normal wound-healing process. Squamous cells display numerous microvilli (and rarely microplicae). In humans, these can be quite substantial, reaching up to 0.75μm in height. Pedler suggested that as cells slough from the surface, desmosomes must split to achieve detachment, and cytoplasmic extrusions at these points may be responsible for microvilli development.4 Microplicae represent fusion of adjacent microvilli. The number of microvilli on a cell surface is a good indicator of its age – the greater the number of microvilli on a cell surface, the older it is.
Microvilli provide an increased surface area for the attachment of a fine glycoprotein layer, the glycocalyx. This layer provides anchorage for the pre-ocular tear film. Damage to microvilli during the flapmaking process in laser in situ keratomileusis (LASIK) and the resultant lack of anchorage for the tear film components is thought to be one of the reasons for dry eye after the procedure.
Non-native cells of the corneal epithelium
A non-native cell in any epithelia can be described as one that does not form any junctional complexes with its neighbours. In humans, the normal corneal epithelium has a complement of non-native cells present at any time. Langerhans’ cells are found in the basal layer of the corneal epithelium. These dendritic, polymorphous cells are thought to play an important role in the ocular surface immune response.5 Although they perform a similar function to their counterparts in skin, ocular Langerhans’ cells are known to vary in that they lack the thymocyte antigen (T6).6 More recently, Doshi suggested that ocular Langerhans’ cells might also vary morphologically from those in the skin, in
that they lack the Birbeck granules characteristic in skin.7 It is normal to see occasional lone lymphocytes and macrophages within a normal corneal epithelium, but it appears that their presence is of no clinical consequence. When present, these cells tend to be confined to the peripheral cornea.
Corneal epithelium stem cell theory
The concept of corneal epithelial stem cells is now firmly established. Among many other classifications, this theory describes cells in relation to their proliferative capacity and their state of maturation. The cornea is unique in that its progenitor cell, the stem cell, is located away from the organ itself and is found in the basal layer of the limbal conjunctival epithelium.8,9 Stem cells are immature and slow to multiply under normal conditions. Division of a stem cell produces two offspring, one that is a replica of itself and a second, transient amplifying cell (TAC). The TAC is responsible, by relatively rapid division, for increasing cell volumes.
The sequestration of corneal epithelial stem cells to the limbus requires their offspring to migrate centripetally to reach the cornea. Evidence of this centripetal flow is seen in individuals with pigmented conjunctivae, who sometimes display pigment slide into the peripheral cornea.10 Further support for this observation was given by Thoft et al.,11 who suggested that corneal epithelial stem cells are located predominantly in the vertical meridian. Additionally, they suggested that the entire basal layer of the corneal epithelium represents TACs that had migrated centripetally from the limbus. Lauweryns et al. corroborated this observation and suggested that, consequently, centripetal movement of epithelial cells was limited to the vertical meridian.12
Stem cell theory and the subsequent centripetal migration that must exist for this theory to hold true are the focal points of corneal wound healing (see later).
Bowman’s layer
Bowman’s layer lies beneath the epithelial basement membrane and separates the epithelium from the stroma (Figure 3.2). It is acellular and uniform in thickness, which typically measures around 12μm. Its thickness remains constant in a healthy cornea. Bowman’s layer is present throughout the cornea and terminates at the periphery, which marks the beginning of the limbus.
Corneal anatomy, physiology and response to wounding ■ 19
Normally, epithelial damage occurs readily without the involvement of Bowman’s layer. This provides evidence of its relative toughness. However, if damage does occur to Bowman’s layer, fibrous scar tissue is laid down and results in an opacity, which tends to reduce in density with time. Bowman’s layer does not regenerate, and significant damage or surgical removal of this layer, such as in PRK, results in a permanent loss. However, in such cases the anterior stroma becomes more compact and loses its cellularity to form a pseudo Bowman’s layer.
Bowman’s ultrastructure
Electron microscopy reveals a fine, randomly orientated mesh of collagen fibrils. These fibrils are finer than those of the stroma. Bowman’s layer modifies anteriorly, where the anchoring filaments of the basement membrane insert. Over the whole of its area, Bowman’s layer is penetrated by fine unmyelinated nerve fibres that pass from the stroma to the epithelium. To maintain transparency, the nerve fibres loose their Schwann cell sheaths as they leave the stroma.
Corneal stroma
The stroma constitutes 90 per cent of the corneal thickness and gives the cornea its strength. Despite its apparent acellularity, the stroma is far from being passive. Stromal cells are important in the production and maintenance of corneal transparency. This property is further aided by the regularity of this layer and the absence of blood vessels within it.
Stromal microanatomy
The stroma consists of around 200 layers of lamellae of collagen fibrils. The fibres, which are buried in a matrix of proteoglycans, have a periodicity that is characteristic of collagen. The adult human cornea lacks elastic fibres.13 Collagen fibrils are of a regular size at any given depth of the cornea, and typically measure 34nm in humans.14
Cells occupy 2–10 per cent of the corneal volume. Keratocytes, nerve fibres and occasionally cells of a vascular origin lie between the lamellae. Keratocytes predominate and are responsible for secreting the proteoglycan matrix and procollagen. Blood-borne cells are relatively small in number. During inflammation, polymorphonuclear leukocytes (PMNs) produce localized opacities called infiltrates, which may contribute to the induction of stromal oedema.15 Interestingly, keratocytes
also have phagocytic capabilities and congregate towards sites of inflammation.
Stromal ultrastructure
The majority of the stromal lamellae have
asimilar thickness (1.5–2.5μm) and lie parallel to each other, except in the anterior third of the stroma where some of the lamellae run obliquely. Fibrils within a lamina are parallel, but the fibrillar orien-
tation in adjacent lamellae is angled, and tends to be arranged orthogonally.16 A bias of lamellar orientation has been claimed, but the direction is not agreed as the result differs according to the technique used.
Lamellae widths are difficult to measure; most are up to 250μm, but some appear to be in excess of 1mm. Although discreteness of adjacent lamellae prevails, at least in the posterior two-thirds of the stroma, occasionally slightly oblique branches connect one lamella to another. This arrangement explains the ease with which the stroma may be split parallel to the surface, as in the preparation of flaps for LASIK or lamellar grafts.
At the corneoscleral margin, the stromal lamellar undulate, branch and interweave. The fibrils of single lamellae remain parallel to each other, but their diameters increase significantly, up to tenfold.
The matrix of the stroma is composed largely of glycosaminoglycans (GAGs) covalently bound to protein, which constitutes proteoglycans. The two major types are keratan sulphate and chon-
droitin sulphate with a filamentous structure demonstrated by Hirsch et al.17 It is the filaments that attach through their core proteins to collagen fibrils, bridging the spaces between them. Details of this bottle organization are unclear, but Scott proposed a model that may explain the
manner in which the matrix influences the regular separation of collagen fibrils.18 During oedema that results from a com-
promised endothelium, GAGs are lost from the cornea.19
Keratocytes are positioned in the interface between adjacent lamellae. In a single interface, keratocyte cell bodies are spaced well apart across the cornea, but their thin, lengthy processes are so extensive that they may come into contact with processes from neighbouring cells, which gives the appearance of a fine, wide-mesh network. This is repeated at each lamellar interface. The nuclei of these cells are flat, oval and embedded in a sparse perikaryon. More than one nucleus may often be present. In
anormal eye there is little or no proliferative activity among keratocytes.20
20 ■ Refractive surgery: a guide to assessment and management
Descemet’s layer
Descemet’s layer (or membrane) is the basement membrane of the corneal endothelium and can be found in embryos as early as 8 weeks gestation (Figure 3.3). There is a two-part formation and structure to this layer: the anterior striated or banded portion is formed in utero and the non-banded section is laid down after birth. Descemet’s layer is around 5μm thick at the first post-natal year and increases by approximately 1.3μm each decade.21 Stromal thickness, by comparison, remains unchanged.
Figure 3.3
Corneal endothelium, Descemet’s membrane and posterior stroma
Descemet’s ultrastructure
Under the light microscope, Descemet’s layer appears void of any cells and lacks internal structure. However, the anterior part of this membrane displays a fine, regular organization under the electron microscope. In tangential section it has a two-dimensional lace network appearance with a repeating hexagonal unit in which seven dense nodes mark the angles; fine filaments of equal length connect these. The networks are stacked in depth register, as revealed by transverse sections. Dark bands are discernible perpendicular to the plane of the cornea and consist of dark granules, which are the nodes of the tangential section network.16 In contrast, the posterior part of this layer has the same fine, granular appearance in whichever plane it is sectioned and shows no sign of patterned organization.
The biochemical composition of Descemet’s layer is not understood well in humans. In other species it has been found to consist primarily of type IV collagen.22 There is evidence for the presence
of types I, II and V collagens in nonhuman models.23–25
Corneal endothelium
The corneal endothelium is a monolayer of hexagonal cells that lines the inside of the cornea. These cells assume a hexagonal array that varies with age, trauma and disease. This layer plays a pivotal role in the maintenance of corneal clarity because its function is to maintain stromal deturgescence (Figure 3.4).
Endothelial microanatomy
Seen in tangential section, cell borders are ill-defined because of the oblique cell interfaces and the interdigitation of the broad processes of adjacent cells. Cell nuclei are often oval or kidney-bean shaped and the cytoplasm appears grainy. The endothelial
mosaic is not always regular because of the variation in cell size or as a result of polymegathism. This increases with age and is exaggerated by some forms of contact lens wear.
Endothelial ultrastructure
Svedbergh and Bill reported that most primate endothelial cells averaged 20–25μm in diameter.26 Therefore, with a corneal surface of approximately 100mm2 to cover, there are about 400,000 endothelial cells in the typical cornea.
Endothelial cells are well stocked with organelles, especially mitochondria. Cells also display a prominent endoplasmic reticulum. Both indicate an extensive metabolic activity. Near the posterior border of this layer, the intercellular space is reduced to form a tight junction of width about 10nm that restricts movement in and out of the cornea between adjacent endothelial cells.
Endothelial replication, regeneration and healing after wounding
The importance of the corneal endothelium in the maintenance of stromal deturgescence and clarity means a considerable
amount of literature has been published on the response of this organ to wounding.
Spontaneous mitosis and migration occur in young rabbit endothelium after trauma. This is followed by mitotic activity to replenish the normal cellular density. Early specular microscopic studies suggested that there was no mitotic activity in human endothelium, and wound healing was accomplished by the spreading, enlargement and finally contact inhibition of adjacent endothelial cells. Over time an equilibration of endothelial cell size across a large area was seen to occur, even when the initial wound was limited to a small central area.
Treffers challenged this view on finding that tritated thymidine was incorporated both in vitro and in vivo in humans,27 which indicated that the corneal endothelium does have a proliferative capacity. Some cells typical of those seen in the M- phase of mitosis were observed by specular microscopy28, which corroborated the Treffer’s study27.
Human corneal endothelial cells grown in culture have responded favourably to the administration of epidermal and fibroblast growth factors.29 In fetal tissue, endothelial cells are seen to respond to
Figure 3.4
Tangential section of the corneal endothelium
similar agents.30 Cytofluorometric techniques have indicated that most human endothelium may be stable in the postmitotic G1 phase.31
Whether human corneal endothelial cells actually undergo mitosis and assist in wound healing or in the repair of natural endothelial cell loss remains unclear, with strong arguments for and against. Specular microscopic studies strongly suggest that if mitosis does occur, it does not appear to result in the formation of endothelial cells of normal size. It is likely that the response to endothelial wounding (e.g., after surgery) is an initial lag phase followed by migration and enlargement of endothelial cells that surround the defect until these cells re-establish contact with one another. At this stage, intercellular junctions are reformed and the eventual establishment of endothelial function by thinning of the overlying corneal stroma is observed.
Corneal innervation
The cornea is served by 70–80 small sensory nerves that issue from ciliary nerves which branch from the ophthalmic division of the trigeminal nerve. They enter the sclera from the uvea at the level of the ciliary body and pass anteriorly to enter the cornea radially and predominantly in the middle layers of the cornea. Other nerves from the same source enter the cornea more superficially. They enter the conjunctival epithelium from the subepithelial tissue at the limbus and pass directly into the corneal epithelium at basal level.32 A minority of the nerve fibres that enter the cornea possess a myelin sheath, but this is lost at the limbus or within 1mm of entering the cornea. Rarely, myelin persists a little further. The perineurium and the fibres and cells of the endoneurium also terminate at the limbus. Only the nerve fibre bundles advance into the cornea. Each bundle consists of several axons enclosed by a Schwann cell sheath.
Initially, the fibre bundles of each nerve are grouped together. These separate and spread, overlapping and running together with branches of neighbouring nerves to produce the plexiform arrangements seen in full thickness preparations of the cornea. The plexus is particularly dense beneath Bowman’s layer.
Axons separate and some divide at intervals and form fine terminal branches, some of which may lose their Schwann cell covering; these terminal axons follow a lengthy course between the stromal fibrils. They possess numerous small, bead-
Corneal anatomy, physiology and response to wounding ■ 21
like varicosities, with a final, often larger, one that marks the end of the axon.
Fibres from a single nerve bundle at the limbus may be distributed to as much as two-thirds the area of the cornea. Consequently, there is considerable overlap of nerve fibres from different bundles. This arrangement explains why sensitivity persists in all areas of the cornea subsequent to large surgical incisions, and is also the reason why the cornea localizes stimuli poorly.
The epithelium receives a prolific supply of terminal fibres that pass perpendicularly from the anterior stromal plexus and penetrate Bowman’s layer. The small nerve fibre bundles lose their Schwann cell covering before they enter the epithelium, where the fine, naked axons disperse and turn sharply to lie nearly parallel to Bowman’s layer. Varicosities similar to those in the stroma occur in the epithelium. Such axons may run a course up to 2mm long, and the fine beaded branches that issue from them are directed through successive layers of the epithelium to almost the surface of the cornea.
Matsuda observed two types of epithelial nerve terminal beads in rabbits and humans.33 One contained mitochondria and the other contained mitochondria and vesicles. He suggested that beads without vesicles serve a sensory function, while those with vesicles were probably motor. There is no reliable evidence of parasympathetic fibres in the cornea. There is, however, a strong body of evidence for some sympathetic innervation to the cornea.
Variety in their chemistry suggests that sensory fibres may consist of functionally distinct subgroups; some contain the neuropeptide substance P, and others of unknown chemical identity do not. Calcitonin gene-related protein (CGRP) also exists within nerves. It is thought to coexist with substance P in the same terminal.34
Corneal sensitivity
It is generally undisputed that the sensitivity of the normal cornea is unsurpassed by any other organ of the body. Aesthesiometry has furnished the clinician with essential data relating to the sensitivity of the corneal surface. From such data it is now accepted that the sensitivity of the cornea varies from a maximum apically to a minimum at the periphery, with a further considerable drop in sensitivity at the limbal conjunctiva.
Sensitivity varies with age. In a study of patients between 10 and 90 years of age Boberg-Ans found peak sensitivity to be up to three times greater in the younger indi-
viduals than that in the eldest.35 Most sen-
sitivity reduction occurs between the ages of 50 and 70 years.36,37
Sensitivity variations between the two eyes are normally minimal. Millodot and Lamont found significant reduction (almost half) in a sample of pre-menstru- al and menstrual women,38 which indicates the possibility of a variation, albeit for a limited time, between the sexes.
The cornea displays a diurnal variation in sensitivity, with about a third greater sensitivity as the day progresses from morning to evening.39 However, perhaps the most striking variation displayed is that between individuals with different iris colour. Blue-eyed individuals have a greater sensitivity than those with darkbrown irides. Non-white people with darkbrown irides have less sensitive corneas than Caucasians with a similar iris colour. Generally, non-white people have fourtimes less sensitive corneas than blue-eyed individuals and half as sensitive corneas as those of brown-eyed Caucasians.40
A normal nervous supply is essential to maintain regular corneal function. Without this several key features of the cornea are diminished or absent. Epithelial cell migration diminishes and epithelial cell turnover is hampered. It is widely accepted that both LASIK and PRK cause corneal hypoaesthesia, but there seems to be disagreement in the literature as to which procedure has a more profound and longer effect. The variety in results may reflect differences within surgical techniques. Moreover, variation in clinical techniques and corneal location in aesthesiometric results cannot be excluded.
Yang et al. reported that corneal sensitivity was more reduced in PRK than in LASIK in the early stages post-proce- dure.41 In eyes that had undergone PRK, they found it took 6 months to recover baseline (pre-operative) levels. By contrast, eyes that had undergone LASIK recovered to baseline levels within 1 month.
The majority of the literature, however, seems to corroborate the findings of Perez-Santonja et al.42 They found that corneal sensitivity was reduced for the first 3 months after LASIK and only recovered to pre-operative levels after 6 months. In PRK, corneal sensitivity recovered its preoperative values within 1 month, except at the central cornea, which took 3 months. In comparing both groups, they found that corneal sensitivity was more depressed after LASIK than after PRK during the first 3 months. No differences were found between the groups at 6 months. Matsui et al. reported recovery in corneal sensitivity as early as 1 week after PRK,
22 ■ Refractive surgery: a guide to assessment and management
with recovery to pre-operative values within 3 months.43 In agreement with the results of Perez-Santonja et al.,42 they found LASIK had a more profound effect on corneal sensitivity, with recovery beginning around 3 months post-procedure. However, over the short duration of this study (3 months) these workers found that sensitivity after LASIK failed to reached baseline levels.43
In vivo confocal microscopy has provided excellent information about the regeneration of corneal nerves after laser refractive surgery. Kauffmann et al. compared the regeneration of corneal nerves after PRK and LASIK.44 In PRK, recovery of subepithelial re-innervation started from the margin of the ablated zone towards the centre of the cornea. At 8 weeks post-oper- atively, rarefied subepithelial nerve fibres were visible at the edges, and after 3 months single non-branched nerve fibres were present at the centre of the ablation zone. By 6–8 months after PRK, subepithelial nerve regeneration seemed to be complete; however, abnormal branching and thin accessory nerve fibres were present without exception.
After LASIK, corneal nerve-fibre regeneration followed the same course as described for PRK, except that the regenerated subepithelial nerve fibres were barely visible in the central cornea after 6 months. Further changes in nerve structure were visible for up to 12 months postoperatively. These observations correlate well with clinical data obtained on the return of corneal sensitivity.
Corneal transparency
If each corneal layer has the same refractive index, then the transparency of the cornea is explained easily. However, this is not the case. The refractive index of the epithelium is quoted to range from 1.375 to 1.543,45 while that of the stroma is typically around 1.55 in the dry state. It is probably fair to say that the exact reasons for corneal transparency are still poorly understood.
Maurice has offered an explanation of the transparency of the stroma.46,47 His theory embraces light of all incidences and explains how transparency is lost in various circumstances. According to Maurice, six neighbouring fibrils surround each collagen fibril in a regular, hexagonal array or lattice. The fibrils are arranged so because they act as a series of diffraction gratings that permit transmission through the liquid ground substance, which has a lower refractive index of 1.34. As the fibrils in adjacent regions of the stroma have
very regular diameters and spacing, a duplication of the diffraction grating exists in any plane.
Spacing between adjacent fibrils is approximately one-tenth the wavelength of visible light.48 Normally, this spacing is quite regular. The incident light that impinges on the collagen fibrils either passes through or reflects off the fibril, such that light scatter cancels by destructive interference. As a result, the matrix can transmit visible light with an efficacy of 90–98 per cent. Supporting this hexagonal theory is that the collagen fibrils of the sclera have a larger diameter and are spaced more irregularly than those in the cornea.
When the cornea is oedematous its transparency is reduced. This may be explained in terms of the lattice theory in that the excessive fluid disturbs the regularity of the fibrillar spacing, so the efficacy of the fibrils as grating elements is lost. Alternatively, transparency loss with oedema may be explained by the formation of spaces within stroma. A similar explanation can be applied when loss of stromal transparency occurs as an adverse response to surgery.
Corneal wound healing
In the context of laser refractive surgery, corneal wound healing is best described in terms of epithelial and stromal healing. Neither is exclusive and now a strong body of evidence indicates that there is interaction between these organs as part of the normal wound-healing process. This is thought to occur via a number of chemotactic factors. Endothelial wound healing is a less important feature in the context of laser refractive surgery, but is discussed briefly here.
Epithelial wound healing
The pattern of epithelial wound healing is generally size dependent. Small, central wounds tend to recover more slowly than larger more peripheral ones. The rate of corneal wound healing is also dependent on the presence or absence of the epithelial basement membrane. When present, re-epithelialization takes a shorter period of time, typically 2–3 days, but when absent the same process can take longer, normally 5–7 days.49 Epithelial wound healing is described in four distinct stages: the latent phase, cell migration, cell proliferation and adhesion.
Latent phase
Extensive reorganization occurs at both cellular and subcellular levels as a result of wounding. Initially, PMNs from the tear
layer and limbus appear in the basal layer at the edge of the wound to remove dead cells and debris. After this, cells at the leading edge of the undamaged epithelium lose their surface microvilli and subsequently flatten and separate. These flattened cells develop surface ruffles and filopodia at their free edges.50 Concurrently, hemidesmosomes are broken between basal cells and the basal lamina, which allows the cells to slide. The ruffles and long, fine filopodia extend to form attachments to the basal lamina, which gives the impression of a capacity to draw cells forwards into the area of the defect.
Cell migration
Between 4 and 6 hours after wounding, epithelial cells migrate across the wounded area. This migration results, initially, in a monolayer of cells that plug the wound. Consequently, this accounts for the disappearance of symptoms 4–6 hours post-wounding. As a result of numerous chemical changes in the basal lamina, the formation of hemidesmosomes is suspended. Consequently, sliding cells are supported by actin filaments, located in the filopodia, which act as a cytoskeleton.
Cell migration occurs in a centripetal manner and as a continuous sheet. It is rare for cells to migrate independently. Individual cells generally maintain the same position within a sheet. Sheet migration occurs from several directions, which meet at a junction as the wound closes.
Cell proliferation
After cell coverage of the wound with a monolayer of epithelial cells, the corneal epithelium undergoes stratification. This is facilitated by mitosis of the corneal epithelial stem cells located in the basal layer of the limbal conjunctiva. In the normal ocular surface, stem cells are relatively quiescent and rarely undergo mitosis. However, in response to wounding these cells readily divide.
Division of a stem cell produces two offspring, one a stem cell and the other a TAC. The newly produced TACs migrate centripetally from the basal layer of the limbus, where they are generated originally, to the basal layer of the cornea. It is principally the rapid division of a TAC that results in stratification of the corneal epithelium. As these cells mature and become more differentiated they move away from the basal lamina to become post-mitotic cells (PMCs). Near the corneal surface, PMCs become fully differentiated into terminally differentiated cells that are eventually sloughed off into the tear film.
Adhesion
The final stage of epithelial healing involves reconstruction of the normal epithelium adhesion structures to Bowman’s layer. Intraepithelial attachments also form and can take up to 8 weeks to become complete. Prior to this the attachments are relatively weak. The speed of hemidesmosomal attachment to the basement membrane is dependent on whether the latter is intact. A more rapid regeneration occurs when the basal lamina is undisturbed.
Stromal wound healing
As epithelial wound healing begins, stromal keratocytes disappear. This is a rapid process and can begin as soon as 30 minutes after wounding. Within 15 hours of the initial injury almost 40% of the anterior stroma is void of keratocytes.51 Reduction of keratocyte numbers occurs by apoptosis. Helena et al. suggested that in refractive surgery this was because these cells became redundant.52 A benefit of apoptosis is that minimal damage occurs to the surrounding tissue, as a consequence of which corneal clarity is preserved. Another advantage of apoptosis is that a potential vehicle for infection is closed down.
As time passes stromal keratocytes undergo migration and proliferation to regenerate a normal stroma. Stromal keratocytes migrate from the posterior stroma to the surface and expand the cell numbers by undergoing mitosis. Keratocyte reproduction begins after the wound has been covered by new epithelial cells and, typically, reaches a peak 3–6 days later. The newly generated keratocytes synthesize collagens, glycoproteins and proteoglycans (Figure 3.5). After corneal injury, this effect is intense for the first 3 months and tapers off at 6–15 months.
Corneal anatomy, physiology and response to wounding ■ 23
Endothelial wound healing
Any significant decrease in endothelial cell density or a change to the cell mosaic results in corneal decompensation. Additionally, it is well established that human corneal endothelium has a low capacity for regeneration. Both of these observations have stimulated research into the effect of laser refractive surgery on this layer of the cornea. It appears that the majority of work has found no significant change in either property in PRK or LASIK. Stulting et al. noticed, however, an increase in central cell density and a correlating decrease in the periphery.53 This was attributed to recovery of the central area through a migration of some of the cell mass from the periphery after discontinuation of contact lens wear.
Confocal microscopy after refractive surgery
The advent of in vivo confocal microscopy has furnished the clinician with a way to improve imaging of the living cornea. This facility has been used to study the cornea at the cellular level, to describe normal morphology, keratitis and other corneal pathologies, and lately the effects of refractive surgery. Moreover, confocal microscopy has furnished the clinician with highly accurate morphometric data.
In normal human corneas, Patel et al. demonstrated that the full thickness density of corneal keratocytes was 20,522 ± 2981 cells/mm3.54 This study investigated 70 subjects who did not wear contact lenses and had normal corneas, with ages that ranged from 12 to 80 years. These workers found that keratocyte density was highest in the anterior 10% of the stroma and that density decreased with age at a rate of 0.45% per year. Using a continu-
Figure 3.5
Corneal fibroblasts (keratocytes) viewed topographically
ous through-focus method (CTFM), they found that the normal central corneal thickness was 563.0 ± 31.1μm and the central epithelial thickness was 48.6 ± 5.1μm.
Erie et al. investigated the affect of LASIK on epithelial and stromal thickness.55 They found a similar epithelial
thickness (46 ± 5μm) to Patel et al. (48.6
± 5.1μm) before treatment.54,55 After surgery, epithelial thickness had increased 22% by 1 month. Thereafter, epithelial thickness did not change, but remained thicker at 12 months after LASIK (54 ± 8μm) than before. Post-operative epithelial hyperplasia has also been linked to refractive regression after myopic LASIK.56 Spadea et al. demonstrated that epithelial thickness increased as early as 1 week after LASIK, reached maximum thickness between 1 and 3 months, and then remained stable for up to 1 year.57
An increase in epithelial thickness has also been demonstrated after myopic PRK treatments.58,59 However, unlike in LASIK, the epithelium continues to thicken for up to 1 year after surgery. Normally, LASIK does not disrupt the corneal epithelium or Bowman’s layer and, consequently, it should affect anterior corneal homeostasis less than PRK. Erie et al. suggested, however, that although preservation of the anterior corneal layers does not prevent thickening, it does seem to allow the earlier establishment of a stable epithelium compared with that in PRK.55
The cause of epithelial remodelling after LASIK is unclear. Epithelial hyperplasia is noted frequently in corneal diseases associated with stromal loss as the epithelium attempts to fill and restore a smooth corneal surface. Dierick and Missotten suggested a tension model in which the epithelium attempts to restore the original curvature of the cornea.60 Reinstein et al. used high-frequency ultrasound to demonstrate that the epithelium varies in thickness after LASIK and appears to possess the ability to remodel itself to compensate for underlying stromal surface anomalies.61 The origin of this property is thought to be the semirigid concave tarsus of the upper eyelid, which polishes and remodels the epithelial surface during blinks.
Erie et al. found no significant change in the thickness of the total stroma, flap stroma or base stroma between 1 and 12 months after LASIK.55 However, rather than being relatively quiescent, as this observation would suggest, Vesaluoma et al. demonstrated marked cellular activity near the ablation zone.62 They found activated keratocytes near the interface. The
24 ■ Refractive surgery: a guide to assessment and management
initial response to LASIK is the creation of a thin keratocyte-free zone on both sides of the lamellar cut. Apoptosis is thought to be the mechanism that underlies the disappearance of keratocytes.
Activated keratocytes have been indicated as part of the healing process after PRK,63 and have been implicated in the stromal thickening shown to occur after this procedure.64 The duration of activated keratocytes in LASIK was shown to diminish after 1–2 weeks by Vesaluoma’s group,62 whereas it typically peaks 3 weeks to 4 months after PRK.65
Keratocyte activation after PRK is thought to be caused by an epithelial–stro- mal interaction mediated by the release of cytokines during epithelial removal. However, as the epithelium and Bowman’s layer are left intact on LASIK, it is possible that keratocyte activation and subsequent stromal regeneration are less than that observed in PRK, which may in part explain the lack of stromal thickening found after LASIK. Studies in rabbits have shown that the wound-healing process occurs only in the periphery of the corneal flap and in relative proximity to the epithelium,66 which further supports the theory that epithelial–stromal interaction mediates keratocyte activation.
A surprising and novel finding after LASIK is the apparent loss of cells in the most anterior keratocyte layers, beginning at 6 months after surgery.62 The exact reason for this is poorly understood. It is now thought that there is a direct innervation of keratocytes by stromal nerve fibres.67 During LASIK, most of the stromal nerve trunks are cut – only those at the hinge are spared. Consequently, most of the keratocytes in the flap zone probably lose their neural input. Lack of communication with the sensory nerves may be the reason for the loss of the anterior-most keratocytes. However, it is important to remember that inconsistencies in this theory exist. Keratocyte loss is not observed until after 6 months, and innervation in LASIK is on the whole restored after 6 months.
References
1Gipson IK, Spurr-Michaud S, Tisdale A and Keough M. (1989). Reassembly of the anchoring structures of the corneal epithelium during wound repair in the rabbit. Invest Ophthalmol Vis Sci. 30, 425–434.
2Bergmanson JPG (1989). CCC or continuously changing cornea. Contact Lens J. 17, 10–14.
In addition to morphometric data, confocal microscopy allows the clinician to view the morphological changes induced in the cornea as a result of refractive surgery. A common feature after LASIK, present in almost every eye, are microfolds.62 These typically appear in two forms, as a wavy unevenness in Bowman’s layer or as more prominent folds that extend into the anterior stroma. The latter variety might affect topography and result in irregular astigmatism.
Flap particles are readily visible by confocal microscopy. The potential origin of the material in the flap interface includes metal from the microkeratome blade, fibres from swabs, lipids or inflammatory cells from the tear film, or epithelial particles carried into the interface by the microkeratome.
Post-interface keratocyte morphology has been shown to be profoundly different between LASIK and PRK. Variations occur in extent and onset.64 In LASIK, the first change in keratocyte morphology is the presence of oval, brightly reflective keratocyte nuclei and thick cell processes behind the flap interface by 3 days. These changes are still present at 1–2 weeks, although the processes became thinner with time. Similar cells are present in the mid-stroma 1 month after PRK.64 Vesaluoma et al. suggested these might represent activated keratocytes.62
Abnormal reflective bodies were reported in all layers of the stroma after PRK.68 These are confined to ablated areas only. Two distinct forms are found, neither of which are visually significant. The first, the so-called rods, are long (≥50μm), slender (2–8μm in diameter) and dimly reflective. Rods sometimes contain bright, punctate, crystal-like inclusions, arranged linearly and at irregular intervals. The second variety, needles, are shorter (<25μm) and more slender (<1μm in diameter), but are highly reflective. Needles are composed of crystal-like granules in linear array, with an individual appearance similar to the bright punctate inclusions seen in rods, but more densely packed.
3Gumbiner BM (1993). Breaking through the tight junction barrier. J Biol. 123, 1631–1633.
4Pedler C (1962). The fine structure of the corneal epithelium. Exp Eye Res. 1, 286–289.
5Gillette TE, Chandler JW and Greiner JV (1982). Langerhans’ cells of the ocular surface. Ophthalmology 89, 700–705.
6Seto SK, Gillette TE and Chandler JW (1987). HLA-DR+/T6– Langerhans’ cells
The incidence of needles and rods does not correlate to either the volume of tissue ablated or the length of post-operative interval. Although they are present throughout the stroma, there is a predominance of these entities in the anterior layers. Interestingly, in contact lens wearers, highly reflective granules, reminiscent of those from which needles are composed, are found scattered as isolated entities throughout the entire depth of the corneal stroma, but rods and needles are not encountered.
Rods and needles are probably manifest some months after surgery and are not thought to be present in the early postoperative phase. Böhnke et al. suggested that rods represented the processes of keratocytes that have undergone chronic change during wound healing, whereby their light-scattering properties are enhanced.68 Such a modification could be attributed to an augmented synthetic activity, which indicates that stromal wound healing continues for a long time after the initial ablation (up to 3 years in Böhnke’s et al.’s study68).
The highly reflective, crystal-like granules encountered sporadically within rods may represent lipofuscin granules. During the course of keratocyte degeneration and necrosis, the processes that are thought to constitute the rods may shrivel, shrink backwards to the perikaryon and eventually disappear, during which sequence of events the granules may be shunted against one another and thereby condense into a shorter length. This could account for the existence and morphological features of the needles. As part of this theory, rods and needles could be indicative of apoptotic activity.
Another possibility is that rods could represent pathological collagen synthesized in response to corneal inflammation after surgery. Alternatively, rods and needles may represent accumulations of reflective material deposited in the corneal matrix, alongside the surfaces of collagen fibres.
in the human cornea. Invest Ophthalmol Vis Sci. 28, 1719–1728.
7Doshi S (1998). The Limbal Palisades of Vogt. PhD Thesis. (London: City University).
8Schermer A, Galvin S and Sun T-T (1986). Differentiation-related expression of a major 64K corneal keratin in vivo and in culture suggests a limbal location of corneal epithelial stem cells. J Cell Biol. 103, 49–62.
9Cotsarelis G, Cheng SZ, Dong G, Sun T-T and Lavker RM (1989). Existence of slowcycling limbal epithelial basal cells that can be preferentially stimulated to proliferate – implications on epithelial stem cells. Cell 57, 201–209.
10 Davanger M and Evensen A (1971). Role of the pericorneal papillary structure in the renewal of the corneal epithelium. Nature 229, 560–561.
11 Thoft RA, Wiley LA and Sundraj N
(1989). The multipotential of cells at the limbus. Eye 3, 109–113.
12 Lauweryns B, Vandenoord JJ and Missotten L (1993). A new epithelial cell type in the human cornea. Invest Ophthalmol Vis Sci. 34, 1983–1990.
13 Alexander RA and Garner A (1983). Elastic and precursor fibres in the normal human cornea. Exp Eye Res. 36, 305–315.
14 Jakus MA (1961). The fine structure of the human cornea. In: The Structure of the Eye, Ed. Smelser GK (New York: Academic Press).
15 Chusid MJ, Nelson DB and Meyer LA
(1986). The role of the polymorphonuclear leukocyte in the induction of corneal edema. Invest Ophthalmol Vis Sci. 66, 192–198.
16 Jakus MA (1964). Ocular fine structure. Selected electron micrographs. In Institute of Biology and Medical Science Monographs and Conferences, Vol. 1, Ed. Retina Foundation (London: Churchill).
17Hirsch M, Nicolas G and Pouliquen Y (1989). Interfibrillary structures in fastfrozen deep-etched and rotary-shadowed extracellular matrix of rabbit corneal stroma. Exp Eye Res. 49, 311–315.
18Scott JE (1992). Morphometry of cupromeronic blue-stained proteoglycan in animal corneas, versus that of purified proteoglycans stained in vitro, implies that tertiary structures contribute to corneal ultrastructure. J Anat. 180, 155–164.
19Kangas TA, Edelhauser HF, Twining SS and O’Brien WJ (1990). Loss of stromal glycosaminoglycans during corneal edema. Invest Ophthalmol Vis Sci. 31, 1994–2002.
20Hanna C and O’Brien JE (1961). Thymidine-tritium labelling of the cellular elements of the corneal stroma. Arch Ophthalmol. 31, 29–33
21Murphy C, Alvarado J and Juster R (1984). Prenatal and postnatal growth of the human Descemet’s membrane. Invest Ophthalmol Vis Sci. 25, 1402–1415.
22Tseng S, Smuckler D and Stern R (1982). Comparison of collagen types in the adult and fetal bovine corneas. J Biol Chem.
257, 2627–2633.
23Fitch J, Gibney E, Sanderson R and Lisenmayer T (1982). Domain and basement membrane specificity of a monoclonal antibody against chick type IV collagen. J Cell Biol. 95, 641–647.
24Von der Mark K, Von der Mark A, Timpl R and Trelstad R (1977). Immunofluorescent localisation of collagen type I, II and III in the embryonic chick eye. Dev Biol. 59, 75–85
25Lisenmayer T, Fitch J and Mayne R (1984). Extracellular matrices in the developing avian eye. Type V collagen in corneal and non-corneal tissues. Invest Ophthalmol Vis Sci. 25, 41–47.
Corneal anatomy, physiology and response to wounding ■ 25
26 Svedbergh B and Bill A (1972). Scanning electron microscopic studies of the corneal endothelium in man and monkeys. Acta Ophthalmol. 50, 321–335.
27 Treffers W (1982). Corneal Endothelial Wound Healing (Nijmegen: Janssen Print).
28 Laing R, Neubauer L, Leibowitz H and Oak S (1983). Coalescence of endothelial cells in the traumatised cornea II. Clinical observations. Arch Ophthalmol. 101, 1712–1715.
29 Yue B, Sugar J, Gilboy J and Elvart J (1989). Growth of human corneal endothelial cells in culture. Invest Ophthalmol Vis Sci. 30, 248–253.
30 Hyldahl L (1986). Control of proliferation in the human embryonic cornea: An autoradiographic analysis of growth factors on DNA synthesis in endothelial and stromal cells in organ culture after explantation in vitro. J Cell Sci. 83, 1–21.
31Ikebe H, Takamatsu T, Itol M and Fujita S (1984). Cytofluorometric nuclear DNA determination on human corneal endothelial cells. Exp Eye Res. 39, 497–504.
32Lim CH and Ruskell GL (1978). Corneal nerve access in monkey. Graefe’s Klin Exp Arch Ophthalmol. 208, 15–23.
33Matsuda H (1968). Electron microscopic study of the corneal nerve with special reference to its endings. J Physiol. 122, 367–391.
34Stone RA and McGlinn AM (1988). Calcitonin gene-related peptide immunoreactive nerves in human and rhesus monkey eyes. Invest Ophthalmol Vis Sci. 29, 857–863.
35Boberg-Ans J (1956). On the corneal sensitivity. Acta Ophthalmol. 35, 149–162.
36Jalavisto E, Orma E and Tawast M (1951). Ageing and relation between stimulus intensity and duration in corneal sensibility. Acta Physiol Scand. 23, 224–233.
37Sédan J, Farnarier G and Ferrand G (1958). Contribution à l’ etude de la keraesthésie. Ann Oculist. 191, 736–751.
38Millodot M and Lamont A (1974). Influence of menstruation on corneal sensitivity. Br J Ophthalmol. 58, 752–756.
39Millodot M (1972). Diurnal variation of corneal sensitivity. Br J Ophthalmol. 56, 844–877.
40Millodot M (1975). Do blue-eyed people have more sensitive corneas than browneyed people? Nature 255, 151–152.
41Yang B, Chen J and Wang Z (1998). Changes in corneal sensitivity after excimer laser corneal refractive surgeries.
Chung Hua Yen Ko Tsa Chih 34, 50–52.
42Perez-Santonja JJ, Sakla HF, Cardona C, Chipont E and Alio JL (1999). Corneal sensitivity after photorefractive keratectomy and laser in situ keratomileusis for low myopia. Am J Ophthalmol. 127, 497–504.
43Matsui H, Kumano Y, Zushi I, Yamada T, Matsui T and Nishida T (2001). Corneal sensation after correction of myopia by photorefractive keratectomy and laser in situ keratomileusis. J Cataract Refract Surg.
27, 370–373.
44Kauffmann T, Bodanowitz S, Hesse L and Kroll P (1996). Corneal reinnervation after photorefractive keratectomy and laser in situ keratomileusis: an in vivo study with a confocal videomicroscope. Ger J Ophthalmol. 5. 508–512.
45 Clark BAJ and Carney LG (1971). Refractive index and reflectance of the anterior surface of the cornea. Am J Optom. 48, 333–338.
46 Maurice DM (1957). The structure and transparency of the cornea. J Physiol. 136, 263–286.
47 Maurice DM (1962). Clinical physiology of the cornea. Int Ophthalmol Clin. 2, 561–572.
48 Gyi TJ, Meek KM and Elliott GF (1988). Collagen interfibrillar distances in corneal stroma using synchrotron X-ray diffraction: A species study. Int J Biol Macromol. 10, 265–269.
49 Dayhaw-Barker P (1995). Corneal wound healing: II. The process. Int Contact Lens Clin. 22, 110–116.
50 Pfister RR (1975). The healing of corneal epithelial abrasions in the rabbit: A scanning electron microscopic study.
Invest Ophthalmol Vis Sci. 14, 648–641. 51 Nassaralla BA, Szerenyi K and Pinheiro
MN (1995). Prevention of keratocyte loss after corneal de-epithelialization in rabbits. Arch Ophthalmol. 113, 506–511.
52 Helena MC, Baerveldt F, Kim WJ and Wilson SE (1998). Keratocyte apoptosis after corneal surgery. Invest Ophthalmol Vis Sci. 39, 276–283.
53 Stulting RD, Thompson KP, Waring GO and Lynn M (1996). The effect of photorefractive keratectomy on the corneal endothelium. Ophthalmology 103, 1357–1365.
54 Patel SV, McLaren JW, Hodge DO and Bourne WM (2001). Normal human keratocyte density and corneal thickness measurement by using confocal microscopy in vivo. Invest Ophthalmol Vis Sci. 42, 333–339.
55 Erie JC, Patel SV, McLaren JW, et al. (2002). Effect of myopic laser in situ keratomileusis on epithelial and stromal thickness: A confocal microscopic study. Ophthalmology 109, 1447–1452.
56 Lohmann CP and Güell JL (1998). Regression after Lasik for the treatment of myopia: The role of the corneal epithelium. Semin Ophthalmol. 13, 79–82.
57 Spadea L, Fasciana R, Necozione S and Balestrazzi E (2000). Role of the corneal epithelium in refractive changes following laser in situ keratomileusis for high myopia. J Refract Surg. 16, 133–139.
58 Gartry DS, Kerr-Muir MG and Marshall J (1992). Excimer laser photokeratectomy. 18-month follow-up. J Cataract Refract Surg. 99, 1209–1219.
59 Gauthier CA, Epstein D and Holden BA
(1995). Epithelial alterations following photorefractive keratectomy for myopia. J Refract Surg. 11, 113–118.
60 Dierick HG and Missotten L (1992). Is corneal contour influenced by a tension of the superficial epithelial cells? A new hypothesis. Refract Corneal Surg. 8, 54–59.
61 Reinstein DZ, Silverman RH, Sutton HFS and Coleman DJ (1999). Very high frequency ultrasound corneal analysis identifies anatomic correlates of optical complications of lamellar refractive surgery. Anatomic diagnosis in lamellar surgery. Ophthalmology 106, 474–482.
62 Vesaluoma M, Perez-Santonja J, Petroll WM, Linna T, Alió J and Trevo T (2000). Corneal stromal changes induced by myopic Lasik. Invest Ophthalmol Vis Sci. 41, 1447–1452.
26 ■ Refractive surgery: a guide to assessment and management
63Del Pero RA, Gigstad JE and Roberts AD (1990). A refractive and histopathologic study of excimer laser keratectomy in primates. Am J Ophthalmol. 109, 419–429.
64Møller-Pedersen T, Cavanagh HD, Petroll WM and Jester JV (2000). Stromal healing explains refractive instability after photorefractive keratectomy: A 1-year confocal microscopic study. Ophthalmology 107, 1235–1245.
65 Frueh BE, Cadez R and Böhnke M (1998). In vivo confocal microscopy after photorefractive surgery in humans. A prospective long-term study. Arch Ophthalmol. 116, 1425–1431.
66Perez-Santonja JJ, Linna TU, Trevo KM, Sakla HF, Alio y Sanz JL, Trevo TM, et al. (1998). Corneal wound healing after laser in situ keratomileusis. J Refract Surg. 14, 602–609.
67 Müller L, Pels L and Vrensen GFJM (1996). Ultrastuctural organisation of human corneal nerves. Invest Ophthalmol Vis Sci.
37, 476–488.
68 Böhnke M, Thaer A and Schipper I (1998). Confocal microscopy reveals persisting stromal changes after nyopic photorefractive keratectomy in zero-haze corneas. Br J Ophthalmol. 82, 1393–1400.