Graft failure: II. Ocular surface complications
- First Online:
- Cite this article as:
- Al-Swailem, S.A. Int Ophthalmol (2008) 28: 175. doi:10.1007/s10792-007-9127-9
- 104 Views
Risk factors for corneal transplantation failure include both immunologic factors, such as graft rejection, corneal neovascularization, and peripheral anterior synechiae, as well as non-immunologic factors, such as ocular surface disorders (OSD) and glaucoma. This review highlights the necessity of having healthy ocular surface epithelia, tears, and eyelids. It presents different types of OSD, their underlying pathology, and their impact on native cornea and corneal grafts. In addition, a range of proposed donor and surgical factors influencing surface integrity following corneal transplant are addressed. Current medical and surgical research, both pre- and post-operative that promise to further improve the outcome of corneal grafts in the context of OSD are discussed.
KeywordsPenetrating keratoplastyOcular surface disordersLimbal stem cell deficiency
The primary function of the ocular surface is to provide clear vision when the eye is open, as the anterior surface of the cornea contributes more than two-thirds of the total refractive power of the eye . Furthermore, the mechanism by which ocular surface health is ensured is built into the intimate relationship between ocular surface epithelia, tear and the eyelids [1, 2]. Stratified, squamous, non-keratinizing epithelia with their basement membranes cover the ocular surfaces (cornea, conjunctiva, and limbus). The conjunctiva is the only directly exposed mucosal membrane and is the source of lymphatic tissue, which protects against infection . Functionally, all the three epithelial regions support tearing, which in turn protects against pathogen entrance and removes harmful substances, such as inflammatory cytokines, allergens, debris, and foreign body (ocular surface defense) [1, 3, 4].
To avoid desiccation, the conjunctival epithelium must be kept continuously moist. The corneal epithelium must remain avascular, transparent, dehydrated, optically pure and be protected by the lids and tears to avoid desiccation. Thus, surface epithelia cannot survive without tears [5, 6]. The tear film consists of a superficial lipid layer and a middle hydrated mucus gel that supply soluble mucins, electrolytes, immunoglobulins, antimicrobial enzymes (e.g., lysozyme, lactoferrin, and beta-lysin), vitamin A, and growth factors (e.g., epidermal growth factor (EGF), transforming growth factor alpha, and hepatocyte growth factor) [4, 5, 7–11]. The mucin secreted by conjunctival goblet cells is likely to facilitate the spreading of hydrophilic tears over the hydrophobic corneal epithelium. Also, it is likely that some of above-mentioned growth factors function to maintain the proliferation and differentiation of the corneal epithelium, whereas others regulate wound healing after ocular surface trauma as required following transplant .
Corneal stem cell theory
The corneal stem cell theory gave new insights into severe ocular surface abnormalities. It suggested that corneal and conjunctival epithelia were presumably supplied by their own stem cells (SC) [12–14]. Dua provocatively speculated that “some might even question the use of the term stem cell and prefer the term ‘committed progenitors’ instead” . The defining characteristics of SC are asymmetric self-renewal (ability to simultaneously replicate themselves and produce progenitor cells with high fidelity), potency (capacity to differentiate into different cell lineages), and niche (protective microenvironment, which is a specific location within the organ where SC reside and maintain their stemness) [15–17]. Limbal SC act as a “barrier” to conjunctival epithelial cells and normally prevent them from migrating onto the corneal surface . Again, these cells (but not conjunctival SC) inhibit fibroblast-stimulated angiogenesis and may therefore play a role in the maintenance of corneal avascularity . In 1983, Thoft and Friend proposed the X + Y = Z hypothesis where X represents the centripetal migration of corneal epithelial cells from the limbus, Y the anteroposterior movement of cells, and Z the cell loss at the surface. For the corneal epithelia to maintain a constant cell mass, the sum of X and Y must equal Z . Their renewal site or “niche” was proposed to be in the limbal palisades of Vogt and interpalisade rete ridges [15, 16, 18]. However, the true stem cell-specific biomarker was lacking . The location of the conjunctival renewal source (SC) was proposed to be either in the forniceal zone or the mucocutaneous junction [21, 22].
Stulting and colleagues showed that the absence of the corneal epithelium did not affect graft rejection, and the overall non-rejection-related failure rate in his series was higher in the group with the epithelium off than the group with epithelium on, underscoring the importance of a healthy epithelium . Thus, the integrity of the corneal epithelium after penetrating keratoplasty (PKP) is an important factor in wound healing and graft clarity. The critical period for stabilization of most of its surface problems is in the first 3 months [24, 25]. The superficial corneal epithelial cells are constantly being sloughed into the tear pool and renewed every three to 10 days under normal circumstances [13–15, 18, 26–28]. Terminal differentiations of cells coupled with cell death by apoptosis, prompts the cell loss via desquamation (exfoliation), a process aided by the shearing forces exerted through eye lid blinking . The forces exerted by the upper lid at the graft center are the greatest, in comparison to those by the lower lids and at the limbus. Exfoliation may be the driving force in epithelial renewal and be the cause of epithelial cell instability. In areas of less lid force, as with depressions near graft sutures, areas of epithelial cell stability are created . Therefore, exacerbation of epithelial instability can be caused by the pathologic process in the lids. In addition, epithelial regeneration on a graft is more complicated than that in a native cornea. Donor corneal epithelium itself, stored for days in medium, may not be amenable to instant resurfacing even under ideal conditions. Even if the donor epithelium remains on the cornea during surgery, it may not be viable and epithelial abnormalities may occur in the early postoperative period [24, 29]. Punctuate epithelial keratitis (PEK), hurricane (vortex) keratopathy, filamentary keratitis, and epithelial defects (ED) along with other types of epithelial abnormalities, are common after PKP, especially in the early postoperative period [1, 30, 31]. For example, the prevalence of PEK, ED and hurricane keratopathy after PKP are 63%, 75% and 30 to 70%, respectively [24, 30, 32, 33]. More importantly, the epithelial status in the immediate postoperative period is not truly predictive of short- or long-term epithelial outcome, and its central morphology will resemble that of normal patients 2 years or more after transplant [1, 25].
Ocular surface disorders
Ocular surface disorders that affect outcome of corneal transplantation
Type 1. Tear film disorders
• Aqueous tear deficiency
○ Sjogren’s syndrome
○ Non-Sjo gren’s syndrome
• Lipid tear deficiency
○ Meibomian gland dysfunction
• Mucus tear deficiency
○ Chemical injury
○ Vitamin A deficiency
Type 2. Limbal stem cell deficiency with or without conjunctival stem deficiency
○ Ectodermal dysplasia
○ Keratitis-ichthyosis-deafness syndrome
○ Chemical/thermal injury
○ Contact lens induced
○ Ocular cicatricial pemphigoid
○ Linear immunoglobulin A disease
○ Stevens-Johnson syndrome
○ Toxic epidermal necrolysis syndrome
• Graft versus host disease
○ Multiple limbal surgeries
○ Medication toxicity
• Lid abnormalities
○ Floppy eyelid syndrome
• Neurotrophic/neuroparalytic keratitis
Under normal circumstances, an epithelial cell-proliferation scheme for the normal cornea as proposed by Schermer et al covers limbal basal cells (SC) to basal corneal epithelium to transient amplifying cells (TAC) to suprabasal corneal epithelium “terminally differentiated cells” . That means during homeostasis and following injury to the corneal epithelium, the limbal stem cells divide to produce daughter TAC that proliferate, migrate and differentiate to replace lost cells . Unlike native corneas or limbal grafts that contain adequate limbal SC, penetrating and lamellar corneal grafts contain only TAC and terminally differential cells . The TAC have a limited life span (3–6 months), limited proliferative potential, and fail to provide a long-term stable epithelial surface to these grafts [24, 39]. Therefore, following PKP, the corneal epithelial repopulation process (or replacement) is dependent on the presence of healthy recipient limbal SC. A complete time for replacing the donor epithelium by recipient epithelium is unknown [21, 24]. However, both animal studies of epithelial rejection and sex chromatin and clinical observation indicated that donor epithelium may persist for as long as 12–13 months [19, 23, 40]. Even after epithelial repair by mitosis, migration, and transformation of the adequate host limbal SC population, firm adhesion of the newly reconstituted epithelium to the underlying tissue requires production of new basal lamina and proper hemidesmosomal attachment . The inability of recipient limbal SC to perform adequate replacement may be secondary to a loss in the number or to an abnormality in the function. In the majority of cases with LSCD, the true etiology is a combination of the two.
Factors influencing ocular surface function in conjunction with corneal transplantation
Donor-related factors that may affect ocular surface function, especially in the early postoperative period after corneal transplantation include those related to the storage media, preservation temperature, diabetes in the donor, death-to-enucleation time, and preservation-to-surgical time.
The prevalence of ED on the first day after PKP varies with different storage medium, ranging from as high as 80% with long-term organ culture medium and 70% with McCarey–Kaufman medium, to as low as 31% with intermediate-term Optisol [29, 32]. Lindstrom and Kim demonstrated that Optisol medium had no epithelial toxicity in vivo studies using rabbit and human eyes, respectively [41, 42]. Optisol-GS, which differs from Optisol by containing streptomycin in addition to gentamycin as an antibiotic, is one of the most commonly used commercially available corneal storage media to date [41–43].
Chou et al. reported diabetes in the donor to be a significant risk factor for a postoperative epithelial defect . That finding is consistent with a well-recognized entity of diabetic corneal epitheliopathy, which manifests as persistent ED or recurrent erosions, especially after pars plana vitrectomy . Thoft et al. demonstrated that lowering the temperature of the epithelium reduces its metabolic requirements and preserves its glycogen stores. Hence this is one of the rationales for placing ice packs on the eyelids from the time of death to enucleation .
There are conflicting data regarding the correlation of death-to-preservation time and its impact on the early ocular surface after corneal tranplantation [29, 32]. Chou et al. reported it to be a significant risk factor independent of other variables. That study concluded that the more important factor in determining epithelial integrity and viability is the postmortem time before enucleation rather than the preservation interval before surgery .
There is also conflicting data regarding the correlation between preservation-to-surgery time and an increased incidence of postoperative ocular surface complications. Machado’s data, in accordance with Mannis et al., Chou et al., Meyer et al. and Feiz et al., failed to establish such a correlation, whereas Wagoner et al., Kim et al., de Ocampo et al., and Khodadoust et al. reported a positive correlation [24, 25, 29, 32, 41, 43, 46–48]. The point to highlight is that the presence of a postoperative persistent ED (for more than 7 and 14 days) in association with longer Optisol-GS preservation to surgery time (7.0 to 14.5 days) did not have a statistically significant adverse impact on graft survival or visual outcome .
Intraoperatively, the donor epithelium should not be removed even if it does not appear viable, and all the measures should be taken to avoid its desiccation . Clinical studies described the beneficial effect on the epithelial status at 1 week postoperatively, after placing sodium hyaluronate on the corneal button during suturing instead of frequently irrigating with saline . In higher risk cases special personal biomicroscopic evaluation of donor epithelium presence and health by the surgeon, regardless of the eye bank assessment, are also recommended based on its reported positive effect on complete epithelialization postoperatively . In contrast, another study failed to find such correlation probably because of different storage media used . Other important issues are short surgical time and proper suturing technique. The latter is crucial to obtain a flush approximation of the donor to the recipient tissue and to facilitate adequate tear film dispersion (distribution) and epithelial migration over the donor cornea. Studies found that irregular graft surfaces in areas around sutures were associated with epithelial healing problems [25, 32]. Removal of the sutures caused disappearance of both central elongated superficial epithelial cells (in a vortex pattern) and palisade arrangement of epithelial cells around sutures [30, 49, 50]. These findings suggest that central and peripheral corneal epithelia are influenced by sutures because under normal conditions newer cells are relatively small and subsequently enlarge as they remain on the surface. This effect may be caused by tear-component alterations or by the inhibition of centripetal movement of the corneal epithelium . In the setting of OSD, the likelihood of the development or progression of corneal neovascularization (NV) is thought to be high due to associated inflammation, previous NV, or persistent epithelial defects. In 41% of patients with absent risk factors, NV can still developed 6–9 months after PKP [51, 52]. Therefore, it is relevant to keep in mind that the placement of suture knots in the host, large recipient beds, and active blepharitis can serve as additional significant risk factors for corneal NV . An interrupted suturing may be indicated especially with asymmetrical NV to allow for selective suture removal for two reasons: to avoid aggravation of NV and to allow for loosened suture removal due to accelerated healing in portions with maximum NV. In contrary, the collaborative corneal transplantation studies (CCTS) found that the risk of rejection was higher in the interrupted technique in comparison to running technique; probably due to selection bias in higher-risk cases. Their “high-risk corneal graft” was defined as patients having at least two quadrants of deep stromal NV and/or a history of previous graft rejection .
Some believe that the status of the recipient’s surface is far more important to efficient re-epithelialization of the graft than is the status of the donor epithelium on day one. An intact epithelium provides a “jump start” for surface maintenance, but it does not determine the subsequent course.
The graft is re-innervated slowly, incompletely, and with an abnormal pattern. Depletion of neurotrophic factors (e.g., acetylcholine, substance P(SP), insulin growth factor (IGF-1), and nerve growth factor (NGF)) secreted by corneal nerves and present in tear film cause two problems. First, altered cellular turnover because these factors normally direct epithelial cell division and maturation [1, 7, 54]. Second, diminished blink rate, thereby delaying the tear production and epithelial mucin expression [3, 55]. These problems will be compounded in the setting of pre-existing OSD leading to neurotrophic keratopathy (NK) such as chemical burns, viral keratitis, fifth nerve palsy after removal of acoustic neuroma or radiation etc., .
There are conflicting data regarding the correlation of advanced age with long-term epitheliopathy, which represent an abnormality in epithelial barrier function. Machado et al. failed to establish such correlation, whereas Feiz et al. and Mannis et al. reported a positive correlation [24, 25, 46]. Feiz data correlated well with one study, which documented more reduction in epithelial barrier function (i.e., increased permeability) among old patients following PKP in comparison to native corneas . In contrast, another report failed to find any change following PKP probably because they included young patients . Mannis et al. hypothesized that this is attributable to a functional diminution of tear production with aging, a less efficient lid function, a greater incidence of blepharitis in this age group, or perhaps an age-induced decrease in corneal stem cell activity . On the other hand, the young recipient was considered one of the risk factors for PKP failure in “high-risk grafts” at CCTS .
Dry eye syndrome “DES” or keratoconjunctivitis sicca “KCS” can be secondary to aqueous tear deficiency (ATD) or lipid tear deficiency (LTD). It ranges widely in severity from ocular fatigue associated with mild ocular surface inflammation to the destructive metaplastic changes as in chronic cicatrizing conjunctivitis (CCC). Defects in basal aqueous or mucin secretions and lipid layer are the common causes of ATD and LTD, respectively . The absence of even reflex tears makes CCC and Sjogren’s syndrome (SS) contraindications to PKP . Insufficient concentration of Vitamin A in tears leads to ocular surface xerosis . The antimicrobial enzymes and Ig A in the tear can provide protection against infection. There is decreased resistance of the external eyes with DES, especially for Staphylococcus aureus. Additionally, in severe cases, DES ultimately leads to corneal pannus, sterile ulceration of graft corneal epithelium, epithelial keratinization, and further damage of limbal SC and conjunctiva [5, 25, 59, 60]. Shimazaki et al. found that eyes with either Schirmer’s test (with & without nasal stimulation) value >10 mm and/or tear function index (TFI) >34 had a greater chance of successful ocular surface reconstruction than those with poor results in these tests .
Classification of the ocular surface disorders in the second type of ocular surface failure, based on amount of lost limbal stem cells and presence or absence of conjunctival inflammation 
Limbal stem cells lost (%)
Normal conjunctiva (stage a)
Previously inflamed conjunctiva (stage b)
Inflamed conjunctiva (stage c)
<50% (stage 1)
Iatrogenic, CIN, contact lens (stage 1a)
History of chemical or thermal injury (stage 1b)
Mild SJS, OCP, recent chemical injuries (stage 1c)
>50% (stage 2)
Aniridia, severe contact lens, and iatrogenic (stage 2b)
History of severe chemical or thermal injury (stage 2b)
Severe SJS, OCP, recent chemical or thermal injuries (stage 2c)
Conjunctival fibrosis obliterating the fornix, especially at the 12 o’clock and 6 o’clock positions, can destabilize the tear film by disrupting the formation of an adequate tear meniscus. Fornix foreshortening and symblepharon can interfere with eyelid blinking by causing lagophthalmos or entropion and can cause restrictive diplopia [71, 72]. Furthermore, misdirected eyelashes, involutional or cicatricial related-eyelid malposition, and keratinized lid margin can cause blink-related microtrauma [73, 74]. If the reserve of limbal SC is healthy, chronic microtrauma to ocular surface may lead to hyperproliferation-induced squamous metaplasia, generating the corneal pannus similar to superior limbic keratoconjunctivitis (SLK). However, if the reserve is poor, chronic microtrauma may facilitate their population attrition, leading to LSCD [70, 74]. The extent of corneal complications was most significantly correlated with lid margin/tarsal scarring in Stevens–Johnson syndrome (SJS) and toxic epidermal necrolysis syndrome (TENS). Whereas, others have shown negative correlation, including floppy eyelid, trichiasis, symblepharon, and aqueous tear deficiency, the positive relationship may be purely correlative and not causative, because inflammation and ulceration can attack both the lid margin/tarsus and the cornea simultaneously at the acute stage .
Several reports have shown association between topical drugs and corneal toxicity. Examples include topical 5-fluorouracil (FU), mitomycin C (MMC), generic form of nonsteroidal anti-inflammatory drugs (NSAID), aminoglycosides, and trimethoprim sulfate [24, 75, 76]. In some of these reports this may be related to either preservatives or to a patient selection bias, because patients with epithelial problems may have been more likely to receive prolonged topical drugs. The NSAID have been reported to be associated with corneal complications such as keratitis, ulceration, and perforation . Preservatives have been shown to disrupt the tear film and damage the epithelial surface. Even with low concentration, their prolonged presence on an already-compromised ocular surface, can cause worsening of the OSD. That is why some preservatives such as thiomersol have been abandoned .
At least two reports have shown the association between immunosuppressive medications and development of LSCD. Once toxicity is suspected, with regard to LSCD, correction can be achieved by switching to non-toxic alternatives and starting intensive supportive measures for the remaining SC [79, 80].
Stepwise approach for ocular surface rehabilitation in conjunction with corneal transplantation
Corneal or limbal SC grafts placed into an inhospitable environment generally will not survive. A stepwise approach for rehabilitation is the most realistic if one is to have some reasonable hope of improving the graft prognosis in patients with OSD.
Diagnosis and initial management of ocular surface disease
Meticulous preoperative evaluation of ocular surface is critical to a successful postoperative course. After obtaining a medical history, the simplest and the most important examination of the ocular surface involves staining with Lissamine green and fluorescein, which can reveal otherwise subtle corneal or conjunctival epithelial problems. Cytological or histopathologic confirmation of conjunctivalization as a critical sign of LSCD involves detection of goblet cells on an impression cytology taken from the corneal surface or a biopsy specimen of fibrovascular pannus covering the cornea, respectively [16, 34, 81].
Conventional pre- and post-PKP therapeutic options include intensive non-preserved tear supplements, hyaluronate drops, punctal occlusion, spectacle side panels, moist insert, treatment of meibomianitis, and topical vitamin A [13, 61, 82–84]. Frequent flushing with non-preserved tears is necessary for managing accumulated inflammatory cytokines after punctal occlusion. Hyaluronate alone has been reported to be effective in improving the corneal epithelial barrier function, in promoting wound healing of corneal epithelium, and in reducing the ocular surface damage and inflammation in DES [85–88]. Meibomianitis can be treated with lid hygiene, short course of topical antibiotic/steroid combination, and oral tetracycline or doxycycline for several months to control the associated belpharitis [89, 90]. Although topical vitamin A use in OSD is controversial, it is capable of increasing goblet cells as well as in reversing squamous metaplasia and keratinization in very severe DES, typically those caused by CCC or graft-versus-host disease (GVHD) [91, 92]. For advanced cases, additional topical treatment include autologous serum, cyclosporine and steroid. The latter two options may be effective because inflammation plays a role in DES by suppressing lacrimal secretion and by damaging the ocular surface. Autologous serum drops (20%) have been recommended in unresponsive cases with persistent ED, recurrent ED, SLK, neurotrophic keratopathy severe DES, and post-ocular surface reconstruction with a maximum objective response varying between 2 and 4 weeks [5, 15, 73, 93–102]. In three randomized trials, it has been found beneficial in improving the following: subjective comfort, ocular surface vital staining, impression cytology, and tear stability [73, 97, 98]. It is superior to tear supplements, because it harbors various growth factors, which are stable for 1 month and 3 months when refrigerated or frozen, respectively. These include vitamin A, EGF, transforming growth factor-β, and neurotrophic factors which are found in serum [73, 96, 98]. In vitro work has confirmed that tear supplements in contrast to serum eye drops do not maintain intracellular ATP levels and epithelial cell membrane integrity. It has to be continued after PKP because of reported epitheliopathy after its discontinuation in two cases undergoing PKP for persistent ED . The only side effects reported to date are immunoglobulin deposition in the form of ring infiltrate or ulceration (one case each), bacterial conjunctivitis (five cases), and microbial keratitis (two cases) secondary to contamination [97, 99, 103, 104]. Additionally, topical cyclosporine A (0.05%) was investigated for unresponsive moderate to severe DES cases in a controlled trial, and received FDA approval in December 2002. It was found to decrease inflammatory markers in the conjunctiva and to improve both subjective comfort and ocular surface vital staining [105, 106]. Following short-term efficacy with topical non-preserved corticosteroids, their use can be restricted for acute treatment of dry eye exacerbations with significant inflammation . A mucolytic agent such as n-acetylcysteine 10% has been useful in treating filamentary keratitis. Its major drawbacks include refrigeration and limited stability—features that precluded its commercialization . Alternatively, topical NSAID, which has not been tested for DES, has been useful in filamentary keratitis affecting SS patients .
Suppress ocular inflammation
Results indicated that control of underlying pathology is crucial . For instance, inflammatory cytokines, such as interferon gamma, up-regulates HLA class 2 antigen and encourages the conjunctival epithelium to undergo apoptosis in acute chemical burns . Also, any surgical manipulation often reactivates quiescent or triggers further conjunctival inflammation in eyes with ocular cicatricial pemphigoid (OCP) with resultant additional scarring [59, 82, 112]. Some advocate having a minimum of 1-year remission period by immunosuppressive therapy before attempting surgical reconstruction in eyes with severe OCP and Mooren’s ulcer. In contrast to the progressive nature of OCP, however, the typical form of SJS and TENS is one of chronic remission after an acute episode . Nevertheless, PKP should be delayed until the inflammatory activity associated with ocular inflammatory disease is controlled for a minimum of 6 months. Topical steroids and topical cyclosporine A are usually effective for the control, but systemic or peribulbar steroids and amniotic membrane transplantation (AMT) may be required for more severe diseases . During acute stage of chemical burn, SJS and TENS, AMT, in addition, can be used as a graft (or patch) to facilitate surface epithelialization, and prevent NV and scarring at the chronic stage [96, 113, 114]. Prophylactic acyclovir (400 mg BID) is recommended for at least 1 year after PKP to prevent recurrence, graft rejection and failure in patients with herpes keratitis [115, 116].
Reduce corneal neovascularization
Several suppressing modalities have been beneficial in reducing NV for a short period in human trials, and may be repeated. These include topical angiostatic steroids, topical NSAID, topical cyclosporine A, Argon laser photocoagulation, fine-needle diathermy, and AMT [52, 117–122]. Topical neutralization of vascular endothelial growth factor (VEGF), a potent angiogenic cytokine, has been shown to prolong corneal allograft survival in animal models and was associated with reduced corneal NV and inflammatory cells into the grafts .
Perform reparative conjunctival surgery
Excision of symblepharon is indicated if they cause lid margin abnormalities or restrictive diplopia. Narrow bands can be lengthened with a conjunctival Z-plasty. Broad symblepharon requires the excision of scar tissue, application of tissue graft and procedures to prevent re-adhesion, such as insertion of conformer or symblepharon ring, or subconjunctival injection of mitomycin C [71, 124–126]. Tissue graft may include conjunctiva, mucous membrane, and amniotic membrane (AM) [2, 71, 72, 127]. Reported techniques using the AM graft have shown its usefulness in reconstructing the conjunctival fornices after symblepharon release [2, 71, 72]. When an area of fibrovascular proliferation is noted after reconstruction with AM, triamcinolone can be injected postoperatively . A successful result has been reported with combined use of MMC in eyes with previous failed reconstruction. The effectiveness of this combined treatment requires further investigation with controls and longer-term follow-up .
Correct eyelid abnormalities
Gas-permeable scleral (not soft) contact lenses are advocated for the prevention of blink-related microtrauma prior to surgical correction of lid margin/tarsal pathologies and promotion of epithelial healing [128, 129]. Several options can address these abnormalities. They include electrocautery, cryocoagulation, argon laser ablation, or complete surgical extirpation for correcting misdirected eyelashes; bilamellar tarsal rotation for correcting trachomatous trichiasis ; and temporary or permanent tarsorrhaphy for the treatment of exposure keratopathy, floppy eyelid syndrome, poor blinking, severe DES, and to promote surface healing in combination with PKP . Temporary techniques include cyanoacrylate glue (lasts for few days), or levator paralysis using botulinum A toxin injections (last for an average of 16 days), or absorbable sutures (last for 4–6 weeks). Permanent techniques include excising opposing lid margins or tarsal pillar suturing; [132–134] debridement or excision, or overgrafting with AM or mucous membrane for the treatment of keratinized eyelid margin; lateral tarsal strip for correcting lower lids malposition secondary to involutional changes;  anterior lamellar recession or posterior lamellar surgery with grafting for correcting cicatricial entropion; [136, 137] and a skin Z-plasty or skin graft with or without horizontal shortening for correcting cicatricial ectropion .
Perform corneal transplantation with or without limbal stem cell transplantation
The laterality, extent of LSCD and degree of inflammation are all factors determining whether to perform corneal grafts with or without limbal stem cell transplantation (LSCT). For patients with severe OSD and normal endothelium, lamellar keratoplasty (LKP) are superior to PKP to eliminate the rejection-related failure frequently seen with PKP. The LSCT aims to restore ocular surface integrity by improving surface lubrication, replace conjunctival phenotypic epithelium with corneal epithelium, and promote the barrier function of the limbus. Improvement in corneal clarity and visual acuity as a result of LSCT may obviate the need for PKP. Delaying PKP for at least 3 months following LSCT appears to improve the outcome. To date, it appears that the long-term survival of PKP is uncertain, owing to increasing risks of graft rejection, non-rejection related failures and/or recurrent surface disorder, whenever PKP is performed with or without LSCT [139, 140]. Unfortunately, the existing clinical series that discuss corneal grafts and/or LSCT are, in general, small and have varying degrees of follow-up [15, 61, 102, 112, 141–155]. This limits our ability to draw any valid conclusions from these studies about which surgical technique of LSCT and immunosuppressive therapy are optimal.
For patients with pre-existing neurotrophic keratitis, there is no definitive cure. Current therapeutic options include bandage contact lens, tear supplements, hyaluronate drops, autologous serum, fibronectin, SP, IGF-a, and NGF [56, 87, 96, 156–158]. Reported surgical options include cyanoacrylate glue, conjunctival flap, tarsorrhaphy and AMT [159, 160].
The short- and long-term integrity of graft epithelium is dependent on the presence of healthy tears, eyelids, limbal SC, conjunctival SC and their normal interactions. In the setting of OSD, graft surface keratopathy is ubiquitous. While it may be transient, it can delay visual rehabilitation, and cause non-rejection-related failure. The following are the most essential preoperative measures to improve graft prognosis: (1) identify the ocular surface problems and their underlying causes, (2) eliminate the causes, and (3) ensure that donor corneas do not have risk factors influencing their epithelial viability. After transplantation, maintaining the preoperative measures are crucial. Finally, larger studies with considerably longer follow-up will be necessary before ideal corneal and/or limbal grafts and adjunct medical regimens can be determined.