Childhood corneal transplants have unique characteristics that make this a difficult field, in their practice and in their study. The difficulty begins by defining the pediatric age, variable according to the series. Although a conventional limit is often established at age 14, it is clear that there are at least three distinct periods: (a) that of breastfeeding and up to approximately 24 months, (b) that of childhood up to the moment that visual development is completed, between 2 and 8 years, and (c) that of pre-adolescence between 8 and 14 years. The distributions of the pathologies that can indicate a keratoplasty in a child are different not only from those of the adults but between these subgroups of age. In addition, the functional repercussion of a corneal problem will depend on the moment in which it has appeared and when it has been treated, in relation to visual development and amblyopia. Most of the experience in children's keratoplasty refers to penetrating (PK), with lamellar types relegated to certain special reconstructive cases. Although this is changing, in this chapter we will focus on the PK. Endothelial keratoplasty in children is dealt with in another chapter of this work (6.6.1).

HISTORICAL BACKGROUND

Until the mid-70s of the last century, PK was only recommended in children with severe bilateral corneal involvement or when they were older¹. This was because the results, quite poor, were believed to be due to the pediatric age itself. However, a recent Australian study on almost 15,000 pediatric corneal transplants² concludes that the problem is not age itself but the etiologies that indicate transplantation, often with a worse prognosis at younger ages – such as malformations of the anterior segment (Figure 1) –. In contrast, in older children there are causes of better prognosis such as trauma and keratoconus.

This has changed the concepts and attitudes little by little, and nowadays the pediatric cornea transplant is performed more and more frequently, around 2% of the total of keratoplasties. In a review of 966 PK’s performed at the Centro de Oftalmología Barraquer (COB) between 2001 and 2006, 24 patients (2.5%) were ≤14 years old. The specialists in this field promote early surgery, since the smaller the child, the greater the risk of amblyopia³ if it is not treated.

SPECIAL DIFFICULTIES OF PAEDIATRIC KERATOPLASTY

There are multiple causes why corneal transplants are more difficult or pose a greater risk. We can consider the following:

• Etiology. The diversity of causes, some with worse prognosis such as dysgenesis, especially due to associated problems such as glaucoma, etc. This also complicates the analysis of the results.

• Laterality. In bilateral cases it is not always easy to decide which eye to start with. The unilateral cases are nowadays operated if the visual acuity (VA) is <0.1 and there is potential for improvement.

• Visual development. In young children there is always a race against amblyopia, which can functionally fail a good anatomical result.

• Communication. Being limited with the patient, communication with parents or guardians is crucial, ensuring that they have realistic expectations regarding the possibilities of recovery, the risks and the need for multiple visits of control and anesthesia.

• Functional exploration. It is difficult to measure vision except with approximate methods such as Teller's, until they reach an age (around 3-4 years) in which it may be late if there is amblyopia. Distinguishing the cause of visual deficit – between corneal opacity and other possible problems – may not be easy. Obtaining information about the progress of vision in these patients is crucial in any case.

• Physical and complementary examination. It raises one of the most relevant practical difficulties, since in young children they must be performed often under anesthesia and there are many tests that cannot be performed.

• Surgical technique. It requires modifications due to the anatomical characteristics of the eye at these ages (see below).

• Postoperative treatment. It often becomes difficult to achieve compliance with treatment or even with frequent visits that may be needed. Success in rehabilitation and the fight against amblyopia depends above all on getting parents to understand their importance and get them involved properly.

• Immune reactivity. Children usually have a more intense immune response than adults. When an established corneal rejection occurs, it is irreversible in more than 70% of cases despite treatment⁴.

• Associated problems and procedures. Frequently, conditions can be found that overshadow the prognosis – optic atrophy, lesions of the posterior segment, glaucoma – that pose additional risks – infections due to obstruction of the lacrimal ducts –, or that require additional procedures – cataracts, glaucoma, iris alterations, strabismus, vitreous-retina, etc. –.

INDICATIONS OF THE PEDIATRIC KERATOPLASTY

Etiology

Pediatric corneal opacities have been classified into three groups according to their visual prognosis: congenital, traumatic and acquired nontraumatic^5,6. The congenital ones are divided into dystrophic (hereditary congenital endothelial dystrophy) (Figure 2a) and non-dystrophic, and the latter depending on whether they are associated with glaucoma (Peters’ anomaly and other anterior segment dysgenesis, congenital glaucoma; Figure 1) or not (obstetric trauma [Figure 2b], sclerocornea, dermoid cysts [Figures 2c and d], congenital aniridia, metabolic diseases, etc.). Non-traumatic acquired diseases include infectious keratitis and post-infectious scars, keratoconus and keratomalacia.

Figure 3 presents the distribution of diagnoses in a series of 203 PKs performed at the COB in patients up to 10 years of age during the period 1965-1995. The most important group was that of infectious post-keratitis leukomas (47.3%), followed by trauma (23.6%) and re-grafts (8.9%). Congenital leukomas were only 3.9% (plus 3.0% of dystrophies) and keratoconus 4.4%. Our impression is that in recent years there has been an increase in the number of cases of congenital opacities that are treated with PK.

When to perform the surgery?

In congenital opacities, it is important to assess whether or not there is glaucoma and associated ocular or systemic problems that may condition the prognosis³. In the unilateral ones it is recommended to operate between 5 and 6 weeks of life and a little later in the bilateral ones, but before 8 weeks, since this is a critical period of visual maturation. The deprivation in this period leads to an irreversible amblyopia. Oftentimes we do not wait many months to operate the second eye as in adults⁶. Indications for surgery include: opacities in visual axis > 3 mm, signs of deprivation (no fixation, nystagmus or resistance to occlusion), no visualization of fundus reflex or poor visualization of the fundus, preferential vision (Teller) < 1 cycle/grade, which is equivalent to AV < 0.03 decimal. If it is an active process – a keratitis or a recent trauma – it is advisable to wait at least 3 months of inactivity, if possible and depending on the age.

Which eye to operate first?

Until a few years ago in bilateral opacities, the most affected eye was operated first. However, today it is recommended to intervene the eye with the best prognosis, which is usually the one with the least opacity – although not always –. An attempt should be made to assess the visual potential, either with the vision tests (Teller, etc.) or the electrophysiological tests.

SPECIAL ASPECTS OF THE SURGICAL PROCESS

The surgical process of pediatric PK (video 3.4.1.1) presents a series of peculiarities:

• Preoperative assessment. It is important to perform an ophthalmological examination as complete as possible, to specify the diagnosis, rule out local or systemic associated problems – with the assistance of the pediatrician – and establish the prognosis and visual potential of each eye.

• Anesthesia. Given that general anesthesia will be used, it is necessary to assess its risks – such as the tendency to apnea, which decreases with age after one month. During the intervention, good ventilation is of great importance to avoid orbital and choroidal congestion.

• Preparation and position. Apart from the usual measures (povidone, eye massage), intravenous mannitol (1 mg/kg in 30 minutes) is useful to obtain adequate hypotonia. During the procedure, the anti-Trendelenburg position is recommended³.

• Donor tissue. It must be chosen with more stringent quality criteria – high cellularity, ideally youthful age – since the need for a functional transplant will be longer term than in adults and the postoperative risks are greater.

• Reduced scleral stiffness. It causes a tendency of the open globe to collapse, sometimes referred to as "positive vitreous pressure". To counteract it, it is advisable to fix a scleral ring (Flieringa), even in the presence of a crystalline lens.

• Diameter selection. Respecting 1.5-2.0 mm of peripheral cornea will be more important than in adults, due to the greater reactivity. This forces to have trephines of small diameters. The lack of stiffness of the tissue causes the cornea to retract within the suction trephines without obturator (such as the Hessburg-Barron) and the window becomes larger than the nominal diameter. In addition, we may be interested in the size of the graft being actually greater (0.5-1.0 mm) than the receiving window – not only the compensation for the trepanning from the endothelial side – to achieve a deeper anterior chamber and less risk of anterior synechiae. This has a limit because the excess of corneal curvature can cause surface problems⁷. A simple rule can be: to use an endothelial die 1 mm larger than the receptor trephine in infants < 4 months; 0.75 mm more in those from 4 to 24 months; 0.50 more in those from 2 to 6 years old and from there onwards as in adults.

• Difficulty in orientation. In the case of total sclerocornea (Figure 5), if there are no references to the position of the limbus, the position of the insertions of the middle and lateral recti can be used to center the trepanation (Ricardo Martínez’s technique).

• Associated intraocular procedures. It may often be necessary to perform intraocular reconstructive surgery of the iris or lens, either because of a congenital defect or after trauma (video 3.4.1.2).

• Increased reactivity. One must be especially careful not to traumatize the intraocular tissues in infants, as they often react by releasing fibrin, which causes greater postoperative inflammation and risk of synechiae. It may be useful to add sodium heparin to the irrigation solution.

• Sutures. Due to the increased reactivity, the sutures should always be independent, never continuous. The material should be the best tolerated (10-0 nylon or 11-0 polyester). It is preferable to bury the knots on the donor side.

POSTOPERATIVE TREATMENT

In the postoperative period, the two most important goals are for the graft to remain transparent and avoid amblyopia. The reviews should be frequent, which implies an important effort for both the ophthalmologist and the family. Most failures occur during the first year⁸. Topical treatment after surgery includes corticosteroids, antibiotics, lubricants without preservatives and mydriatics. Several authors propose the use of cyclosporine A, either topical⁹ or systemic¹⁰, although this is not free of side effects of greater repercussion at this age. It is recommended to start topical corticosteroids with high frequencies (every 1-2 h during the day and ointment at night), with slow reduction over two years or more. Systemic corticosteroids are used during the first weeks and in cases with increased inflammation or rejection.

Mydriatics can help, especially if the anterior chamber is narrow. In infants it is preferable to avoid atropine because of its toxicity (better use cyclopentolate). The oral cyclosporine A regimen used in the Galdakano Hospital starts with 5-7 mg/kg/day, seeking to maintain baseline levels (T0) between 100-120 ng/ml, with control of renal toxicity through indirect signs – decrease of glomerular filtration, increase of proteins in urine, increase of blood pressure – and with blood analysis initially every 2 weeks and then monthly. This regime is maintained for one year.

The healing is very fast, and removal of sutures can start within a few weeks (selectively to control astigmatism). A simple guideline is to start it, in children under 12 months, from weeks 4 to 6; between 12 and 24 months of age, from weeks 6 to 8; between 2 and 3 years, from weeks 8 to 12; between 4 and 6 years, from weeks 12 to 16; between 7 and 10 years, from 5-6 months; and in those from 10 to 20 years between 6 and 8 months¹¹. The formation of a firm scar can be assessed by the appearance of a whitish line (Figure 6). Any suture that loosens should be removed immediately, as it can become infected or induce rejection. It is fundamental to treat refractive errors from the beginning and vigorously prevent amblyopia, usually with penalties on the healthy eye.

RESULTS AND COMPLICATIONS OF PAEDIATRIC KERATOPLASTY

The results of pediatric PK are very variable according to the studies (table 1). This may reflect the heterogeneity of populations, ages and pathologies. The most relevant factor is the initial indication and not the age «per se»^2,12. The worst results appear associated with younger ages (<5 years, 50% of failures) because they are congenital opacities³. Within these, dystrophies, such as congenital hereditary endothelial dystrophy⁵, and Peters’ anomaly type I⁷ seem to have a better prognosis. The cases of keratoconus and inactive leukomas post-keratitis (older children) also have good results. The traumas have a medium prognosis and the chemical injuries, bad. After 5-12 years, the results are similar to the average of adults and after 12 they surpass it (as it happens in the juvenile group, see chapter 3.6).

In the 1965-1995 COB series, of 203 cases followed for an average of 60 months (range 1-240), opacification was observed in 72 (35.5%), with an average failure time of 13.7 months (range 1-68). Forty-eight of the failures (66.6%) occurred in the first 12 months, 14 (19.4%) in the 2^nd year and 5 (6.9%) in the 3^rd. This would mean a survival of 76.4% at 1 year, 69.5% at 2 years and 67.0% at 3 years. However, these values are not the result of an actuarial calculation and must be taken with caution. The frequency of opacification according to etiology is reflected in table 2.

The risk of rejection is clearly higher in children than in adults, due to their more active immune system. When rejection occurs, the risk of failure is also higher, between 22% and 50%¹³. The greatest number of rejections occurs during the first year, although they may occur several years after the transplant. The risk of infection is also higher in children than in adults, mainly due to Gram +⁶. The most commonly isolated organisms are Streptococcus pneumoniae and Staphylococcus aureus¹⁴.

Glaucoma after PK in children has been estimated at 5-9%⁶ and is an important cause of graft failure. It appears more in some of the congenital causes of PK, such as the dysgenesis of the anterior segment, but also with the use of corticosteroids. In case of finding high intraocular pressure (IOP) before surgery or during the postoperative period, the necessary treatment must be established. Remember that normal IOP values ​​in infants are lower than in adults and that general anesthetics have ocular hypotensive effect.

Traumas are not uncommon after corneal transplant in children. There are series with up to 19% of failures due to trauma¹⁴. That is why the continuous use of protective glasses is recommended. Other possible complications include: persistent epithelial defect, graft dehiscence, cataract, endophthalmitis (4-9%), expulsive hemorrhage (2-3%), retinal detachment (3-5%) and phthisis bulbi (4-13%)⁶. The causes of graft failure in the COB series were not always clear, so we assigned each case a more probable cause and a possible secondary cause (Table 3). We also analyzed the causes of low VA (<0.1) in the presence of a transparent graft, which reveals the great problem of amblyopia (Figure 7).

In case of graft opacification during the first years of life, re-intervention should be considered to try to develop a certain visual potential. Given that the prognosis of re-grafts darkens rapidly at these ages – at 4 years, 87% failure with the 2^nd and 100% with the 3^rd12 –, this would be a terrain where keratoprostheses can offer a hopeful option.

The idea of ​​performing keratoplasties of different shapes to the circular one goes back at least at the beginning of the 20^th century. In 1912 Magitot performed rectangular penetrating keratoplasties (PK) in humans and in 1921 Ebeling and Carrel presented their results¹. In 1923 Forster carved them with a cataract blade, in the form of an equilateral triangle, one of whose vertices exceeded the pupillary margin and the others were placed on the superior side². In 1931, Castroviejo began to practice his famous square PKs, cut by an instrument designed by himself with two parallel adjustable separation blades³ (see chapter 1.3). Paton also made them with a simple knife and scissors, with the help of square markers or others that only marked the four corners of the quadrilateral, which he drilled and then united with the knife⁴.

Although the PK evolved definitively towards the circular form, there are situations in which the preferable option – almost always with tectonic purpose – can be different from the usual central and with a diameter between 7 and 9 mm. These include small or "plugged" keratoplasties for peripheral corneal perforations, and keratoplasties of various shapes that adapt to more extensive peripheral lesions: semilunar or in crescent, fusiform or almond-shaped, and keratoplasties in crown sector, horseshoe and even complete crown^5,6.

PENETRATING KERATOPLASTY OF SMALL DIAMETER

Autoimmune pathologies such as ulcerative keratitis in rheumatoid arthritis and other systemic collagen diseases, often associated with severe lacrimal hyposecretion, and in some neurotrophic types, can lead to significant losses of substance in the peripheral cornea, with descemetocele and perforation. When these are of a size that does not allow to approach them with conservative measures, even being small (1-3 mm) and without affecting the visual axis, they can be solved by a small circular tectonic PK (Figure 1, video 3.4.2.1)⁷. This requires having trephines between 2.5 and 5 mm in diameter.

Surgical technique

Prior to the intervention, we induce meiosis with pilocarpine 1-2% (60 and 30 minutes before). We pass traction sutures through the upper and lower rectus muscles, to center the perforation in the operative field. We mark with a very sharp trephine and leave a margin of 0.5-1 mm around the hole. The most usual is between 3 and 4 mm in diameter. Trepanation is difficult in a hypotonic eye, with the anterior chamber (AC) flat or very narrow. We can reform and separate the incarcerated iris by injecting a cohesive viscoelastic – preferably of high molecular weight, e.g., Healon V (AMO) – through the perforation itself. Occasionally the tamponade of the iris allows some AC and a paracentesis can be performed; otherwise, we will do it after reforming the AC with the viscoelastic. Once the AC and the iris have stabilized, we continue with the trephine until it penetrates, and the button is resected with curved scissors. The hypotony causes that the window is with a certain bevel inward (smaller in the deep part), which helps to support the graft. Through the trepanation we perform an iridotomy as peripheral and far from the window as possible. If the incarcerated iris was necrotic or if it was very peripheral and is likely to form synechiae, it will be preferable to resect it in that area (Figure 2).

With the same trephine we obtain the graft of the donor cornea, either from its central part or from the periphery. We have not observed that this has an impact on astigmatism. The cylindrical profile of the graft is not a problem to adapt it to the somewhat trunk-conical window. In some cases, we have managed to perform two simultaneous small PKs in one eye with double perforated descemetocele, at the same time that the contralateral eye received a horseshoe PK due to a larger peripheral ulcer (Figure 3). The three grafts were obtained from the same donor eye (video 3.4.2.2).

We suture the graft with 10-0 nylon stitches, passed at the predescemetic level and bury the knots towards the recipient cornea. In smaller grafts 4 stiches may be sufficient, although above 3.5-4 mm 8 are usually necessary. They should avoid invading the visual axis. We perform an irrigation-aspiration of the viscoelastic and leave the reformed AC with balanced saline solution (BSS). If there is tendency of the iris to contact the window, an air bubble can be left to prevent the formation of synechiae and we apply a cycloplegic drop at the end.

SECTORIAL PERIPHERAL KERATOPLASTIES

In the peripheral lesions and up to approximately one quadrant of amplitude, we can perform peripheral sectorial reconstructive keratoplasty – be it penetrating or lamellar – that can have several morphologies. The choice of one or the other form will depend on the type of pathology, its nature – it may be perforations somewhat larger than the previous section, large slimming after peripheral keratitis, Mooren's ulcer or Terrien’s degeneration, after trauma, chemical injury or fistula –, its extension and degree of advance toward the center of the cornea, or if they should be removed “en bloc”, such as in tumors or epithelization of the AC^1,2. Sometimes these grafts will include a ring of scleral tissue (sclerokeratoplasty). In any case, the objective will be to encompass the existing pathology affecting the visual axis minimally. The different modalities include:

• Almond-shape or fusiform keratoplasty. This shape is defined by having two convex edges, but without being circular. It can be carved with the same trephine or two with different diameters (e.g. 8 and 10 mm), generally the largest, concentric with the cornea, marking in the limbus or slightly on the outside, and the smaller one centered on the outside the cornea to mark a peripheral convexity that will encompass the almond shaped lesion. It allows to treat pathologies that penetrate the cornea from the periphery, although the central convexity can affect the pupillary area (Figures 4 and 5, video 3.4.2.3).

• In crescent keratoplasty. It is a reverse configuration to the previous one. The outer edge is convex and the internal concave. It is carved with the same large diameter trephine applied on two different centers or with two trephines, the smaller for the external convexity, the larger diameter for the internal concavity. It tends to be used less because of the difficulty in matching, between donor and receiver, the tip-sharpened ends.

• Semilunar or semi-circular keratoplasty. Its outer edge is circular and marked with trephine, and the internal is linear and cut with scalpel and scissors. It is useful in lesions that advance more over the cornea such as tumors, cysts (Figure 6, video 3.2.4.4) or epithelization of the AC.

• Sectorial crown keratoplasty. It is a reduced version of the horseshoe keratoplasty, with two concentric trepanations of different diameter – 2 mm diameter difference per mm width (radius) of the crown – which are joined by radial cuts to the required amplitude (Figure 7).

Surgical technique

In general, all these techniques require a 2 or 3 mm peritomy wider than the altered corneal area and dissection of the conjunctiva to the fornix. Bipolar hemostatic diathermy and point of traction of the neighboring rectum muscle. If a large penetrating incision is to be made, a Flieringa ring should be applied. Marking of the receiving window with the trephines according to the chosen morphology. Almost always complete the cut with scalpel and scissors following the marks. In penetrating cases, one or multiple iridotomies or iridectomies must be done to avoid anterior synechiae, if the pathology does not require iris resection. Some of these lesions can be treated with lamellar versions of the same techniques, with non-penetrating dissection in the receiver (except perforation already present) and application of a graft of the same form but full thickness without endothelium.

PERIPERAL KERATOPLASTY IN HORSESHOE OR CROWN

In very extensive peripheral lesions, which affect two or more quadrants such as circumferential Terrien’s degenerations (Figure 8) and some Mooren’s ulcers, among others, the best reconstructive option without having to transplant the corneal center is the crown shape, be it sectorial, semi-circular (i.e., horseshoe) or even wider, until you reach the crown or complete ring. As in the previous section, sometimes these are corneoscleral grafts, since depending on the pathology it may be necessary to include a sclera ring in addition to the corneal tissue.

Surgical technique

Horseshoe or crown keratoplasty surgery is an expanded version of sectorial keratoplasty. It is marked with two concentric trephines. If the graft is to be 3 mm radial in width, the trephines must have 6 mm difference in diameter (e.g. 8 and 14 mm)^5,6. If intraocular structures are involved – e.g., cysts or tumors, AC epithelization (Figure 9, video 3.4.2.5) –, the root of the iris should be disinserted and resected in block. If this is not necessary, it is better to try to respect the angle, with a beveled dissection in the scleral part so that the penetration in the AC is made in front of the Schwalbe line. In any case, in penetrating surgery, iridotomies or peripheral iridectomies should be added. In the lamellar version, the marks with trephines are similar but the dissection only reaches a deep corneal plane, creating a horseshoe lamellar bed (Figure 10).

The graft is marked with the same trephines and is completed with scalpel and scissors, in the form of a complete crown or ring. The suture is done with radial and deep independent stitches. We begin with the peripheral or scleral circle and once partially fixed, we cut the donor ring according to the width of the bed (one quadrant, two, three, etc.) before completing the internal part of the suture. At the end the tightness is checked by filling the AC with BSS or an air bubble. The peripheral part is covered with the conjunctival flap. The latter is not done in Mooren's ulcers (Figure 11), where we leave about 2-3 mm of bare sclera for delayed epithelization, and it is protected with a large-diameter contact lens (video 3.4.2.6). This is intended to prevent conjunctival vascular tissue from replenishing the immune elements and reactivating the ulcer process.

In some patients, we have performed penetrating grafts in complete crown (Figure 8a), using the patient's central cornea as an autograft that is sutured to the ring of donor tissue⁵. The lamellar variant in full crown or ring is described in chapter 6.8.7, particularly for late circumferential ectasias in keratoplasty by keratoconus.

There are situations in which the transplantation of the central area of the cornea, which is usually understood as keratoplasty, cannot resolve the pathology, since this extends to the limbus or beyond. In these cases, one option is to resort to a penetrating keratoplasty (PK) of large diameter or even a sclerokeratoplasty. Although the prognosis of the PK clearly worsens with the diameter above 8.5 mm (see chapter 3.6), large diameter PKs generally have a reconstructive tectonic indication and, as we shall see, in some cases good optical and refractive results are obtained^1-3. On the other hand, if not achieved in a primary way, large diameter reconstructive PKs may be combined with a further procedure to obtain transparency and good vision, such as a second PK of conventional size or endothelial keratoplasty if the former has decompensated.

INDICATIONS

The most frequent indication of a large diameter PK – and also the one that has had the best results in our hands, has been the large diffuse corneal ectasia of keratoglobus type, with both central and peripheral thinning reaching almost to the limbus. It has also been used in completely unstructured corneas after keratitis – whether infectious or autoimmune – and other processes of diffuse keratolysis, sometimes of obscure cause. We will present three demonstrative cases.

PENETRANTATING KERATOPLASTY OF LARGE DIAMETER IN THE KERATOGLOBUS

The patient was a 31-year-old woman who presented with a large diffuse thinning and ectasia of both corneas, with globular morphology (keratoglobus) (Figures 1 a, b). The thickness was especially reduced in the corneal periphery in 360°, although more markedly in the lower sectors, with some superficial fine vessels at that level. The limbus was relatively normal and there was no alteration of the sclera and no systemic problem was found. Refraction with lenses achieved a visual acuity (AV) of 0.1 in both eyes (OU) with 50° -20 -20 in the right (RE) and 160° -18 -18 in the left (LE). He had worn rigid contact lenses for years but lately she developed intolerance to them.

Surgical technique

The RE was operated first and 12 years later the LE, with essentially the same technique (video 3.4.3.1). A PK of 11 mm in diameter was performed under general anesthesia. This graft was almost tangent to the limbus and left a thin margin of transparent cornea of ​​0.5 mm, although the donor cornea included a fine margin of scleral opaque tissue. The surgical technique was similar to the usual one, although it required a motorized trephine of the aforementioned diameter. A Flieringa ring had been fixed to avoid the collapse of the globe and to hold this during the trepanation. The trephine only carved a median groove in the receiving cornea, which was deepened with a Desmarres’ knife and penetrated into the anterior chamber (AC) with a diamond scalpel. The corneal button was cut with curved scissors, although in this case a certain bevel was left to improve coaptation. Four peripheral iridotomies were performed, one in each quadrant. The graft was obtained from a fresh whole eye globe with the same trephine and placed on the window after applying a drop of viscoelastic on the pupil and the iris. It was fixed provisionally with 4 cardinal sutures of 9-0 silk and a definitive continuous in the RE (in the LE were 32 independent stitches) of 10-0 nylon.

Outcome and evolution

At 7 days postoperatively (Figure 1c), the graft in the RE was in position, transparent and with deep AC, transparent lens and normal intraocular pressure (IOP). The treatment included topical and systemic corticosteroids (1 mg/kg/day). However, at 13 days there was a first episode of rejection (Figure 1d), with increased hyperemia and slight edema that was controlled with an increase in corticosteroids and oral azathioprine 2 mg/kg/day. This treatment was reduced in a few weeks and there were new episodes with blurred vision, endothelial precipitates, Tyndall + and sectorial edema at 4, 17 and 20 months postoperatively (Figures 2a, b). All of them responded to the previous treatment and the addition of topical 2 % cyclosporine A (CsA) (the patient quickly came when aware of any symptoms). After the last episode, she reached an VA = 0.8 with -7.5 D of astigmatism and -7 of sphere. The treatment was progressively reduced until it was completely finished in a few months, and since then the episodes of rejection have not been repeated. After 22 years, she developed a cataract that was intervened with an intraocular lens implant (IOL) in the bag. Three years later she reached VA = 0.3 without correction and 0.8 with 5 D of astigmatism and spherical equivalent of -0.25 D (Figure 2c). The endothelial count was 2,105 cells/mm² (Figure 2d).

Final comment

This case shows that it is possible to obtain good long-term results with large-diameter keratoplasty in keratoglobus. Although a deep anterior lamellar keratoplasty could be considered, it would be technically very difficult in these circumstances. Its evolution raises the question of how to establish the immunological tolerance of a corneal homograft that, after being recognized and subject to reaction from the early period and repeatedly in the first two years, remained well tolerated without immunosuppressant treatment for decades. Finally, the unusually high value of the endothelial population, for a graft of that age (26 years) and the aforementioned rejection history, makes us speculate about whether a transplant of such diameter could be providing a peripheral endothelial population with a certain regenerative capacity compared with that of central grafts.

We have intervened 10 cases as the present with PK of between 10 and 14 mm, with good reconstructive results in all of them and visuals in 7 of them. The other three ended up decompensating and required a central secondary PK of standard diameter^4,5. In any case, a close and prolonged control of these patients and their adequate education and collaboration to respond quickly to any incident in the postoperative period is crucial.

SCLEROKERATOPLASTY IN KERATOLYSIS ON OLD CORNEAL GRAFT

75-year-old woman, of Iranian origin. She had been repeatedly operated on with PK in OU, for a cause she did not remember (perhaps keratoconus), the first time in the 50s of the last century by Dr. Castroviejo and the other eye (she does not remember which) in the '70s. Then she was re-intervened 3 more times with PK by Dr. Khodadoust in the '80s. When we saw her for the first time in 2007, she showed in the RE a VA = light perception with poor localization and a semi-opaque graft with calcium deposits. The LE, with which she could defend herself in recent years, had VA = 0.03 with a descemetocele and diffuse keratolysis of the graft that also affected the peripheral cornea outside the graft towards the lower limbus (Figure 1). Lacrimal function was acceptable for age.

Surgical technique

After a few weeks of waiting with topical and systemic treatment against keratolysis (medroxyprogesterone, doxycycline, TCL, etc.) and initiating systemic CsA, we proceeded to a 12 mm tectonic sclerokeratoplasty, plus cataract extraction and IOL implant in the bag. The technique was similar to the previous one, under peribulbar anesthesia and sedation. An attempt was made to carve the receiving window with a greater bevel to avoid affecting the angle and it was sutured with 31 independent 11-0 polyester stitches.

Outcome and evolution

The initial evolution was favorable (Figures 4 a, b) and at one month she reached VA = 0.5 with 30° -2 -1, intraocular pressure (IOP) of 8 mmHg and an endothelial count of 2,617 cells/mm². The initial treatment included dexamethasone phosphate without preservatives 6 times/day, oral prednisone 1 mg/kg/day and oral CsA, apart from topical antibiotic and artificial tears. The corticosteroids were reduced after one month and the CsA was maintained. She came at the 5^th month follow-up appointment with VA with correction (VACC) = 0.2, pigmented endothelial precipitates (Figure 4c) without forming a line and Tyndall + with cells in AC. We restarted the systemic corticosteroids and increased the topic ones to 30 times/day. One week later she had recovered (VACC = 0.4), without inflammatory signs and IOP = 15 mmHg.

We slowly reduced the corticosteroids until the systemic ones were finished in 4 months (control of 9 months) and we left the topics in 4 times/day, plus the oral CsA, lubricants, etc. She returned at 13 months, "feeling worse for a long time", with VACC = 0.15 and again edema (not very marked), endothelial pigment and mild ciliary injection (Figures 4d and e). We restarted the aggressive corticoid treatment like the previous one and she slowly improved. At 3 weeks, a dendritic ulcer appeared in the upper periphery of the graft and we added acyclovir ointment 5 times/day and autologous serum. During the following months she remained with the eye calm, with minimal residual edema. At 21 months she had VACC = 0.1 and the posterior capsule was somewhat opaque. After capsulotomy, Nd:YAG she recovered VACC = 0.5 and remained well until month 28, when she fell back to VACC = 0.25 without large visible changes in the cornea. The patient lived alone in Tehran, had difficulty in following the treatment, was depressed in her fluctuating and often insufficient vision and wanted a solution. We decided to perform an endothelial lamellar keratoplasty (ELK/DSAEK) with which she maintained an acceptable vision (0.35-0.4) until the last visit two years later (Figure 5).

Final comment

This patient shows the value of sclerokeratoplasty as a reconstructive technique in losses of corneal substance that reach the limbus. The cause of keratolysis was obscure, with a possible neurotrophic component linked to the old graft and to old age. Although a large-diameter tectonic lamellar keratoplasty could also be considered, due to the predictable difficulty of dissection and limited visual results, we preferred PK. It also highlights the importance of long-term follow-up and the difficulty of the patient's distant residence, as well as the option of endothelial keratoplasty to treat graft decompensation once the graft has resolved its tectonic function.

SCLEROKERATOPLASTY IN ANTERIOR BILATERAL STAPHILOMA SECONDARY TO KERATITIS (POSSIBLE MOOREN’S ULCER)

This 24-year-old patient was visited during a campaign of our Foundation in southern Senegal at the beginning of 2013. After having good vision, she had developed inflammation in OU since a few months before and presented hyperemia and ulcers in the entire extension of both corneas, without secretions. When she arrived at our center one month later, the appearance was of anterior staphyloma in OU, with opaque and thinned corneas of irregular shape with embossed surface, although in general epithelialized, with zones of fibrosis and vascularization and others with descemetocele. The areas with a certain transparency showed a completely attached iris (Figure 6). The vision was of light perception in OU, with good localization in RE and poor in LE. The pressure to the touch was normal or somewhat high in RE and low in LE, although Seidel + was not observed. The ultrasound showed a normal background in OU. The ERG and PEV were altered but the RE seemed better.

The analysis was normal except for eosinophilia, with negativity for hepatitis C virus. The general medical study did not reveal relevant processes. Conjunctival cultures isolated S. aureus only in RE and those of cornea only nonspecific Corynebacterium. She was treated with antibiotic and topical corticosteroid eye drops 3 times/day and oral methotrexate 10 mg/week plus folic acid. We performed a conjunctival covering in the LE and after a few weeks a sclerokeratoplasty of 13 mm in diameter in the RE.

Surgical technique

Under general anesthesia and in anti-Trendelenburg position, after practicing a 360° peritomy, with resection of 1 mm of perilimbal conjunctiva, we sutured a Flieringa ring of 18 mm in diameter with 6 stitches of 7-0 vicryl, leaving the 4 verticals fixed to the operative field. We marked manually with a 13 mm trephine, fixing the eye by the sutured ring. We obtained the graft of a whole fresh globe, marking with the same trephine and cutting through the scleral groove with Katzin-Troutman corneal scissors. We separated the corneo-sclera from the uvea by gentle traction and with the help of a sponge. We left the graft protected with viscoelastic.

The dissection of the receiving window was continued with a piriform knife (Grieshaber) in a deep lamellar way from the edge of the scleral trephination to the AC around the perimeter until a transparent cornea is visualized. We penetrated with a Desmarres knife just in front of Schwalbe's line. Although the iris was attached, it could be separated with viscoelastic in the periphery of the AC and the opening was expanded first with the Desmarres and then with scissors. The iris was poorly attached to the cornea except in the area of ​​the pupillary ridge. We separated it gently with a sponge and fine tweezers. We applied coaxial bipolar diathermy of the few bleeding points.

With two iris separators (Visitec) we explored the lens, which presented a cortical and subcapsular opacity. We decided to extract it “open-sky” and implant a PMMA IOL of +23 D (the anterior-posterior axis was normal). We thought that leaving a pseudophakic eye would also help to deepen the AC. We reconstructed the pupil with two 10-0 polypropylene sutures and practiced 8 iridotomies distributed around the peripheries. We applied the graft on the window and fixed it with 26 radial sutures of 10-0 nylon. The conjunctiva was left fixed outside the edge of the graft, a little prominent because the scleral dissection of the receptor was lamellar (video 3.4.3.2).

Outcome and evolution

The early postoperative period ran without incidents and at two weeks she reached VA = 0.1 with 160° -5 + 7 (Figure 7a). Due to the circumstances of the patient and to obtain at least one tectonic reconstruction in OU, we decided to operate the LE one month later, with a similar technique. At 3 weeks, LE reached VACC = 0.2, with IOP = 18 mmHg with a hypotensive eye drop (Figure 7b). However, at that time the RE initiated a rejection that was treated with topical dexamethasone 30 times/day, oral prednisone 1 mg/kg/day and oral and topical CsA 2%, plus methotrexate (Figures 7c, d). The RE stabilized but developed a retro-corneal membrane that leaped from the iris, with fibrosis on the periphery of the iris and formation of extensive anterior synechiae – despite multiple iridotomies. One week later she started with edema in LE, which responded to the increase in topical corticosteroids (Figures 8a-c).

In spite of everything, at 3 months after the 1^st intervention, the VACC were RE = 0.2 and LE = 0.95, with IOP 12-16 mmHg (Figures 8d, e). After starting to reduce oral prednisone because she started with Cushing’s syndrome, the patient had a new episode of rejection in LE, which required restarting systemic corticosteroids and reinforcing the topic therapy. On the last visit before returning to her country, the VACC were of RE = 0.15 and LE = 0.4 and the IOP = 14 and 10 mmHg respectively. The retina examination and the Goldmann visual fields were normal in OU. She was prescribed treatment with topical dexamethasone 3 times/day in RE and 5 times/day in LE, and CsA 2% 3 times/day and Timolol 2 times/day in OU. The systemic treatment was changed to azathioprine 2 mg/kg/day.

Final comment

This case highlights the tectonic indication of sclerokeratoplasty, since the situation of this patient could hardly be resolved in any other way. Due to the incomplete history, bilateralism, rapidly progressive course and absence of secretions, the most probable diagnosis seems to be Mooren's ulcer, which is known to be more aggressive in bilateral cases in young people of African origin. The transplants performed achieved a good reconstructive and visual acceptable result. There was a significant fibrosis reaction and episodes of rejection despite preventive treatment. From an optical point of view, a keratoprosthesis could be proposed, but it would also be necessary to start with anatomical reconstruction and, due to her place of origin, it did not seem like a good option. If these corneas deteriorate in the future they could be the object of a smaller central PK or even an endothelial keratoplasty.

The risks of “open-sky” surgery and the postoperative complications inherent to PK explain the rise of the new lamellar surgery. However, the frequency of pathologies that involve all layers of the cornea and therefore require a PK justifies the search for total-thickness techniques safer than “open-sky” PK. In order not to pierce the recipient cornea during the transplant, we propose a new technique that we have called “pseudo-chamber protected keratoplasty” (PsPK), which we define as a full-thickness transplant, but which, at least initially, is not penetrating – although it can be transformed into a penetrating one according to the needs of the case.

Hallerman in 1959¹, was the first to use a full-thickness corneal button, with endothelium included, in a lamellar keratoplasty. Later, Lazar in 1991² and Loewenstein in 1993³ propose maintaining Descemet’s membrane (DM) and the recipient endothelium in transplants for edematous keratopathy (bullous). Their idea was not to pierce the cornea and provide security to the technique, but the donor endothelium, when coming into contact with the receptor stroma, did not work properly and the graft failed. The difference of our technique with these ones is the creation – and the maintenance if necessary – of a space between both corneas that we call pseudo-chamber (PsC).

SURGICAL TECHNIQUE

The PsPK surgery begins with a partial thickness trephination followed by a dissection of the stroma leaving a thin residual layer (Figure 1). Next, the donor cornea is placed, without removing the endothelium, on a viscoelastic mattress and suturing is carried out. The key of the procedure is the final introduction of viscoelastic and/or gas in the space between the residual bed and the donor cornea, so that the PsC is formed (Figure 2). Once created, we have several options: (a) proceed immediately to the cutting and removal of the stromal lamina and endothelium of the recipient cornea or endotheliectomy (Figure 3); (b) perform this maneuver in a second stage, days, weeks or months later – in both cases we transform the procedure into penetrating (Figure 4) –, and (c) maintain the PsC indefinitely, without removing the residual bed (Figure 5).

Dissection of the recipient bed and creation of the pseudo-chamber

As usually in transplants, we evaluate the initial situation of the recipient cornea and select the diameter of the trephination (Figure 6a). We proceed to partial trepanation until about 2/3 of its thickness (Figure 6b). We begin the lamellar dissection of the deep layers of the stroma, first with a semilunar knife (mini-crescent of 1.25 mm) and then with the spatulas of Melles (Figure 6c). The intention is to leave a residual stroma and endothelium sheet <100 μm thick, ideally 50-60 μm. We also create a peripheral pocket of 0.5 mm outside the limits of the trephination, to facilitate suturing the graft (Figure 6d). A faster way to reach the deeper planes of the stroma and save dissection time, is to use a femtosecond laser (FsL) to create a lamellar plane at a depth of 2/3 of the corneal thickness, from which it is continued with manual dissection until reaching the predescemetic plane.

After completing the predescemetic dissection, we place the donor cornea with its intact endothelium (Figure 6e) on a cohesive viscoelastic mattress and proceed to suture it with 16 independent nylon stitches (10-0). Finally, we inject viscoelastic and/or gas (SF₆) between both corneas with a 23 G cannula to create the PsC (Figure 6f).

Complementary intraocular surgery

If it is necessary to perform a surgical procedure in the anterior chamber, the most appropriate time is before placing the donor cornea. Through corneal, limbal or scleral incisions, the anterior segment can be reconstructed (synequiaetomies, repositioning or lens explant, etc.) or complete any other associated surgery. We recommend extending methyl cellulose over the residual stromal bed to obtain a better visualization. The new technique allows us to associate intraocular surgery without damaging the transplanted cornea.

Extraction of the residual bed: endotheliectomy

The extraction of the residual stroma-DM-endothelium (SDE) complex or endotheliectomy can be performed in the same surgical act of transplantation, deferred to a 2^nd stage according to our interest or need, or even leave the SDE indefinitely. We indicate intraoperative endotheliectomy (Figure 7) when all the problems of the anterior or posterior segments have been solved in surgery, so that a secondary action will not be necessary.

When we are interested in keeping the donor cornea isolated for a while, we defer postoperative endotheliectomy (Figure 8), usually between the 3^rd and the 6^th month, although it can be done at any time, days or years after the intervention, according to the evolution of the PsC or the needs of the case. The layer can be removed even when the suture is still in place. In fact, it is normal that on the 3^rd month we remove the SDE and 8 of the usual 16 sutures. Whether it is intraoperative or deferred, the extraction of the residual bed can be done with manual or FsL-assisted techniques.

• Manual technique. We practice two 20 G paracentesis in the corneal limbus at 10 and 2 o’clock. Through them, a 29 G cystotome is inserted with the tip bent upwards, below the receptor endothelium. We drill the SDE at several points that will serve as a gateway to a vitreous-retina scissor, to make a circular cut of the bed. Finally, we extract the trimmed SDE with vitreous-retinal forceps (Figure 9).

• Technique assisted by femtosecond laser. This greatly simplifies the procedure. We use the LenSx system (Alcon Labs) that incorporates an optical coherence tomography (OCT), which allows to evaluate in real time the structures of the anterior segment and focus the laser exactly on the SDE (Figure 10). The diameter of the cut usually ranges between 5 and 6 mm. Next, and as in the manual technique, we proceed to extract the SDE complex with curved vitreous-retinal forceps.

Maintenance of the residual bed: pseudo-chamber

In certain cases, we may be interested in maintaining the residual bed and therefore leaving the PsC temporarily or definitively, due to intraocular or peri-corneal problems. For example, in combined transplant and vitreous-retina surgery we can prevent the donor cornea from being exposed to the intraocular environment, such as the possibility of contact with silicone oil, until a specific issue is solved. In others, it is not interesting for the inside of the eyeball to be exposed to coexisting surface pathologies, such as active infections or their alteration in chemical injuries. In this case we must consider the natural evolution of the PsC and how to maintain it.

• Natural evolution of PsC. There are two possibilities: a) The residual SDE remains separated from the donor endothelium and the presumed PsC between the two is maintained (Figure 11). The donor endothelium functions normally and the cornea acquires good transparency. With the passage of time, the SDE becomes fibrous and remains separated forming a horizontal plane (Figure 12). b) The residual SDE tends to move towards the donor endothelium (Figure 13). When the viscoelastic is exchanged for aqueous humor, the space of the PsC decreases and the two corneas come closer. If they come to contact, the donor endothelium may find its viability compromised and the graft fail.

• Maneuvers to maintain the PsC. To prevent the collapse of the PsC we can do two things during the initial surgery (Figure 14): a) Perform a cross-shaped micro-perforation in the SDE with a 15° knife, to facilitate the flow of aqueous humor and balance the pressures between the inside of the eyeball and the PsC. b) Insert, in the peripheral corneal pocket made during the predescemetic dissection, a closed PMMA ring, diameter 0.5 mm greater than trepanation and 0.25 mm height. With this ring we increase the separation between both corneas, in addition to providing more stiffness to the scar. When the tendency of the two layers to come closer is perceived, we must introduce again viscoelastic or gas in the PsC to avoid losing it. This maneuver should be performed in the operating room with a 23 G cannula, taking advantage of the space between two sutures.

INDICATIONS

PsPK would be indicated in cases of high-risk PK and those with poor prognosis of lamellar keratoplasty (Table 1), whether deep anterior with endothelium of doubtful viability or endothelial in situations of limited prognosis such as aphakia (Figure 15).

Endothelial decompensating, secondary to intraocular surgery, is possibly the main indication for PsPK. In complex or complicated surgeries (intra- and postoperatively), both in crystalline⁴ and vitreous-retina⁵ or glaucoma⁶, this is where the technique becomes a preferable alternative to classical approaches. The combined surgeries of keratoplasty with other intraocular surgery⁷, herpetic keratitis with vascularized stroma and endothelial involvement⁸, corneal chemical injury with limbal insufficiency⁹, are also situations in which the poor prognosis of a PK has been demonstrated and where PsPK provides safety.

A typical situation is aphakia with intraocular silicone oil. The corneal endothelium in contact with it decompensates irreversibly, raising the option of a new PK of poor prognosis. In contrast, PsC protects the donor endothelium and keeps it viable for an indefinite period. Another option in these cases is keratoprosthesis¹⁰, but the complications related to them make us consider PsPK preferable.

ADVANTAGES AND DISADVANTAGES OF THE PsPK

Advantages during the initial surgery

During the transplant, the PsPK presents a series of advantages with respect to a PK: it avoids the possible complications of an open surgery, greater safety to work simultaneously inside the eyeball and less damage to the donor endothelium. The PsPK allows to perform any type of surgery combined with the transplant, from a lensectomy to a complex reconstruction of the anterior segment, a vitreous-retinal surgery or a valve implant for glaucoma. In all these cases the donor endothelium is protected by the PsC.

Advantages of maintaining the residual bed

• Protection of the donor endothelium. In the immediate postoperative period, the donor endothelium will be less exposed to the usual inflammation of a penetrating technique and possibly more immunologically isolated, so it will be less likely to fail. The use of corticosteroids is less necessary and therefore the chances of infection will be lower. All this results in a faster functional recovery. In the aphakic with intraocular silicone, the presence of PsC offers the best option to preserve the viability of the graft.

• Rapid healing. In the advanced postoperative period, PsPK allows the sutures to be removed early. The scar is formed more quickly as in an anterior lamellar keratoplasty and without standing the intraocular pressure. The residual bed behaves like a horizontal tensioner that facilitates the process. From the 3^rd month onwards, we begin to remove sutures and residual astigmatisms similar to those of the anterior lamellar technique are obtained.

Disadvantages of maintaining the residual bed

• Limited visual acuity. Leaving the SDE layer may limit visual acuity in the initial postoperative period. This drawback is relative in the context of astigmatism and/or corneal flattening by the sutures that also hinders vision after a PK until the stitches are removed – which on the other hand, is done earlier in a PsPK .

• PsC tendency to collapse. Histological studies of lamellar keratoplasties show that the donor endothelium is damaged if it is placed in contact with the receptor stroma¹¹. Therefore, close follow-up of the patient during the first postoperative month is important to correct the tendency to collapse of the PsC if it occurs.

• Damage to donor endothelium due to viscoelastics. It is not known if a long exposure of the endothelium to the viscoelastic in the PsC can be harmful, be it left at the end of the surgery or later injected to maintain it. In our experience, the cell count at 3 months does not differ from the values ​​obtained in an endothelial keratoplasty (DEK/DMEK). It is possible that the residual bed of <100 μm allows the passage of nutrients to the PsC and that the viscoelastic is replaced progressively by aqueous humor. To reduce the amount of viscoelastic, we also introduce gas (SF₆). The mixture maintains the PsC in the same way.

PsPK COMBINED WITH CATARACT SURGERY

The advantages of the PsPK are especially relevant in a combined surgery. For this reason, we have developed in a special section the aspect of PsPK combined with lensectomy and intraocular lens implantation, which means a new form of triple procedure. Due to lack of space, it has not been possible to include it here, but it will be available on the website of the Spanish Society of Ophthalmology (SEO).