Lamellar Keratoplasty in Keratoconus


  • Lamellar keratoplasty is often required for advanced keratoconus when contact lens wear is no longer comfortable or an ideal fit cannot be achieved.

  • Regional pachymetry and anterior segment optical coherence tomography (OCT) assessment preoperatively are essential to help decide on surgical technique.

  • Cannula-assisted, sequential air injection technique for big bubble (BB) bubble deep anterior lamellar keratoplasty (DALK) appears to be safe and effective, if scars do not involve the Descemet membrane (DM). The viscobubble technique is an alternate option for primary surgery or rescue technique for failed air injection.

  • Manual techniques should ensure a smooth residual bed with thickness less than 80 microns for optimal visual outcome.

  • Femtosecond laser and intraoperative anterior segment OCT can provide further assistance to surgical techniques and help improve outcomes.

  • Early recognition of complications and appropriate management can eliminate the need for conversion to penetrating keratoplasty (PKP).


Advanced cases of keratoconus often require surgical intervention to restore corneal anatomy and improve vision. Although there is no precise definition for advanced disease, most specialists would agree that a keratoconus patient is eligible for corneal transplant when spectacle or contact lens correction is insufficient, continued contact lens wear is intolerable, and visual acuity has reduced to unacceptable levels.

Traditionally, penetrating keratoplasty (PK) has been performed for advanced keratoconus with a higher success rate of graft survival than for other indications. Since 2002, lamellar keratoplasty (LK), especially deep anterior lamellar keratoplasty (DALK), has emerged as an alternative procedure to PK. , DALK allows surgical replacement of the recipient’s corneal stroma, preserving healthy Descemet membrane (DM) and endothelium, and is the ideal surgery for eyes affected by keratoconus that need a transplant. This technique offers substantial advantages compared with PK, primarily the avoidance of endothelial rejection and longer graft survival, which is particularly important in young patients with a long life expectancy. In a study of DALK performed for different preoperative diagnoses in over 660 eyes, with a mean follow-up of 4.5 years (range 0.5–10 years), the average graft survival rate has been reported as 99.3% (range 98.5%–100%). The Cornea Donor Study Group has demonstrated that, even in the absence of rejection, substantial endothelial cell loss occurs following PK, with a 76% endothelial cell loss from baseline after 10 years. On the contrary, the endothelial cell density (ECD) after DALK suffers from a small loss only during the first 6 months postoperatively but then remains stable. , Visual outcomes of DALK are comparable to PK if the surgical technique provides a smooth interface with minimal host residual stromal thickness. ,

Moreover, the DALK procedure is extraocular, carries a lower risk of endophthalmitis and expulsive hemorrhage, and allows earlier tapering of steroids with a decreased risk of secondary glaucoma and cataract. DALK also offers better wound integrity, which is very useful in patients with intellectual disability in which PK has a higher incidence of postoperative complications. ,


The goal of LK is to replace the affected stroma, preserving healthy host endothelium. Various LK surgical techniques have been described in the literature, including manual dissection, microkeratome-assisted, excimer laser, and femtosecond laser-assisted LK ( Box 31.1 ). The majority of anterior lamellar keratoplasty (ALK) techniques remove only 70% to 80% of host stroma, leaving behind >100 microns of residual host stroma. Wide variations in these techniques, along with suboptimal long-term visual outcomes compared with PK, did not make them popular worldwide. Currently, DALK is the most frequently performed surgery for advanced keratoconus. In this technique 85% to 90% of host stroma is removed by lamellar dissection, leaving behind a smooth interface, minimal residual stroma, and healthy DM and endothelium, and is replaced by full-thickness donor corneal tissue devoid of DM. Various DALK techniques have been described to create a good optical graft-host interface, including layer-by-layer manual dissection and stromal injection of air/fluid/viscoelastic material.

BOX 31.1

Classification of Anterior Lamellar Keratoplasty Procedures

DM , Descemet membrane; PDL , pre-Descemet layer.

1. Superficial Anterior Lamellar Keratoplasty (SALK)

  • Level: dissection <200 microns of anterior stroma

  • Technique: microkeratome-assisted or femtosecond laser–assisted

  • Apposition: sutureless

2. Anterior Lamellar Keratoplasty (ALK)/Midstromal ALK

  • Level: dissection 200–400 microns of stroma

  • Technique: manual/microkeratome-assisted or femtosecond laser–assisted

  • Apposition: sutures required

3. Deep Anterior Lamellar Keratoplasty (DALK)

  • Level: dissection >400 microns of stroma

  • pdDALK—pre-Descemetic DALK (minimal residual stroma with PDL and DM)

  • dDALK—Descemetic DALK (DM along with PDL)

  • Technique: stromal air/viscoelastic injection/manual near DM dissection

  • Apposition: sutures required.

Although agreement is not yet unanimous, DALK is broadly classified in the literature as descemetic DALK (dDALK), in which the dissection is achieved up to the DM, and as pre-descemetic DALK (pdDALK), in which a small amount of posterior stroma is left intact along with the DM. The majority of articles in the literature show that visual recovery after pdDALK is slower (2–5 years of follow-up) but comparable with dDALK when the residual recipient bed thickness is less than 80 microns and is homogeneous in its thickness. However, dDALK techniques are usually preferred because they are largely considered faster and more reliable, making the surgeon confident to have performed optimal stromal removal with a good visual prognosis (see Box 31.1 ).

The recent description of Dua’s layer, also called pre-Descemetic layer (PDL), has demonstrated that what was often thought to be a DM-endothelium surgical exposure in fact also includes a very thin layer of stroma. Taking this finding into consideration, a new nomenclature for DALK has been proposed recently.

DALK Techniques

Several techniques have been employed to achieve deep stromal dissections. , However, the most common of these techniques are air injection—big bubble (BB), viscodissection, , manual layer by layer, and stromal peeling.


The BB technique was originally described by Anwar and Teichmann in 2002, and it is probably the most popular DALK technique. This technique involves injection of air into deep corneal stroma to create a plane of separation between the stroma and DM, surgically seen as formation of a circular air-pocket that is referred to as a “big bubble.”

There are three types of bubble formation ( Fig. 31.1 ). Type 1 BB is well circumscribed, with white margins and a diameter measuring up to 8.5 mm. It starts in the center and enlarges circumferentially toward the periphery, and it is quite resistant ( Fig. 31.2 ). It is the most common type of bubble, and its posterior wall is formed by endothelium, DM, and PDL. Type 2 BB has clear margins and its diameter can measure up to 10.5 mm. It is usually eccentric in location, starting in the periphery and enlarging centrally (see Fig. 31.2 ). This rare type of bubble cleaves off the DM from the stroma and is very fragile. Type 3 BB consists of mixed types of bubble: BB type 1 and one or more smaller type 2 bubbles.

Fig. 31.1

Schematic diagram of different types of big bubble.

Fig. 31.2

Intraoperative photographs of type 1 and type 2 big bubble (BB) . Yellow arrow marks the edge of the BB.

With the original BB technique, a trephine is used to perform a partial thickness corneal trephination at about 60% to 80% depth. A 27- or 30-gauge needle, attached to an air-filled syringe, is then inserted deep into the paracentral stroma through the bottom of the trephination groove and is advanced, with the bevel facing downward and parallel to the DM, for 3 to 4 mm. Air is injected with formation of BB ( Fig. 31.3 ). A type 1 BB formation occurs in 60% to 70% of cases. Paracentesis is performed to normalize the intraocular pressure (IOP). Subsequently, anterior keratectomy is performed followed by deflation of BB by creating a small opening at the center of the anterior wall using a sharp blade. The remaining corneal stromal layers are lifted with an iris spatula, severed with a blade, and excised using scissors. The residual stroma is divided into four quadrants to facilitate easy removal. Any residual peripheral adhesions between the stroma and PDL are separated using a blunt dissector prior to excision. Scissors designed for a stromal excision, in which the lower blade is slightly longer than the upper blade, allow removal of tissue in a safe manner. Donor corneal tissue, usually of the same diameter as host trephine, is punched out, followed by removal of DM. The donor tissue is secured into position using an interrupted, continuous, or combined technique with 10-0 nylon ( Fig. 31.4 ). Postoperative medications include topical corticosteroids, antibiotics, and lubricants. Topical steroids are continued with a tapering dosage for 4 to 6 months, along with lubricants for an extended duration of time. The corneal graft clears rapidly with healthy endothelial cell counts ( Fig. 31.5 ).

Fig. 31.3

Schematic diagram of Anwar’s big bubble technique using stromal air injection. A) partial depth trephination, B) Anterior lamellar keratectomy , C) introduction of air injection cannula, D) Air injection to create big bubble

Fig. 31.4

Intraoperative photographs of big bubble (BB) deep anterior lamellar keratoplasty. (A and B) Deep trephination; (C) anterior lamellar keratectomy; (D) stromal track using pointed dissector; (E) placement of bottom port, blunt-tipped cannula; (F) stromal air injection and type 1 BB formation; (G) paracentesis to lower intraocular pressure; (H and I) BB deflation; (J–L) residual stromal removal; (M) Descemet membrane peel from donor button; (N) placement of donor on host bed; and (O) donor secured using 10-0 nylon sutures.

Fig. 31.5

Postoperative slit-lamp photograph showing a clear corneal graft following deep anterior lamellar keratoplasty, with good endothelial cell counts on specular microscopy.

One should be extremely careful with type 2 BB, given the high risk of spontaneous DM rupture, and even consider not opening the bubble and instead performing a manual dissection technique to near DM. The type 2 bubble spontaneously resolves in the early postoperative period. ,

The deeper the air is injected, the higher are the chances of creating a BB. Therefore several modifications of the original technique have been described that attempt to increase the bubble success rate.

Cannula Big Bubble

Bottom port blunt-tipped cannulas for air injection have been described by several surgeons; these reduce the risk of intraocular perforation, compared with a sharp needle. Anterior lamellar keratectomy to debulk 40% to 50% of stroma facilitates deeper placement of the cannula and also allows faster deflation of the BB. A pointed dissector (Fogla Pointed Dissector, Storz Ophthalmics, MO) is used to make an initial track in the peripheral stroma at the base of trephination. A 27-gauge blunt-tipped, bottom port, air injection cannula (Fogla 27G Air Injection Cannula, Storz Ophthalmics, MO) is attached to a 5 mL air-filled syringe. The tip of this cannula is introduced into the peripheral stromal track and moved toward the central/paracentral cornea for 3 to 4 mm using a wiggling motion. Air is then injected using a firm continuous pressure until the formation of BB is noted. Paracentesis is performed to normalize the IOP and reduce counter resistance from the aqueous in a closed anterior chamber (AC). Air is gradually injected to increase the bubble size. Once increasing resistance is felt again, air injection is discontinued. Using this modified technique, Fogla was able to increase the success rate of BB from 69% to over 95%. , The advantages of using a blunt-tipped bottom port cannula have been confirmed by several studies. ,


Intraoperative corneal thickness measurement to create a pachymetry-guided intrastromal air injection to increase the rate of BB formation has been described. After an initial partial trephination to approximately 60% to 70% of corneal thickness, intraoperative corneal thickness measurements using ultrasound pachymetry are taken 0.8-mm internally from the trephination groove in the 11 to 1 o’clock position. In this area, a 2-mm incision is made, parallel to the groove, with a micrometer diamond knife, calibrated to 90% depth of the thinnest measurement. The incision is then opened with toothed forceps and widened superficially with a 15-degree blade. The deep stroma is exposed, and an initial stromal track is made with an iris spatula, close to the DM. A bottom port blunt-tipped cannula attached to an air-filled syringe is inserted into the deep stromal track and advanced toward the central or paracentral cornea, and it is used to create a BB. Subsequent surgical steps are similar.

Bubble Size Optimization

Once the type 1 BB is initiated, its expansion is limited by counterpressure from aqueous humor in a closed AC. This is felt as increased resistance to air injection for BB expansion following its initiation. Because of this, the BB often fails to reach the margin of trephination, and if air injection is continued forcefully for further expansion, there is a risk of rupture of the BB. For a BB smaller than the trephination diameter, manual dissection of the residual stroma between the outer edge of the BB and trephination margin is required for optimal graft-host apposition, which can sometimes lead to unwanted microperforations. To avoid this, a sequential air injection technique can be performed, in which a paracentesis is performed after initiation of the BB, to lower the IOP and reduce counterpressure to bubble expansion. A repeat air injection now allows BB expansion with lower resistance, helping achieve an ideal size reaching up to or beyond the trephination margin.

Small Bubble Test

Stromal emphysema can occur following air injection and may hinder visualization of the BB. A simple test has been described, in which a small air bubble is injected into the AC via a limbal paracentesis. , If the small air bubble remains in the peripheral AC (usually with a sausage configuration), it confirms that a type 1 BB has been successfully accomplished ( Fig. 31.6 ). Alternatively, if the small bubble is located centrally beneath the opaque corneal stroma, this would suggest that a type 1 BB has not been obtained. Sometimes corneal emphysema, especially after multiple air injection attempts, may make it difficult to visualize the small bubble test. The presence of a BB can be ascertained by releasing aqueous via paracentesis to lower IOP. If a BB has formed, the cornea would have a protruding central dome shape, besides having a firm consistency to touch compared with the softer, flatter peripheral cornea.

Fig. 31.6

Intraoperative photographs of small bubble test. (A) Small bubble remains in periphery when type 1 big bubble (BB) is present. (B) Small bubble moves to the center following deflation of BB.

Bubble Deflation

The technique to open the BB described by Anwar and Teichmann is associated with a known risk of DM perforation. Goshe et al. refined the BB opening technique, advising to coat the stroma overlying the BB with a cohesive viscoelastic prior to entering the BB. A 1.0- to 1.5-mm incision is then created with a 1.0-mm diamond blade/sharp knife, using only the tip of the blade in a “lifting” motion. This limits the escape of air, preventing a sudden collapse of the bubble while entering with a sharp blade. Viscoelastic can be injected into the bubble to maintain space and facilitate easy removal of the overlying residual stromal tissue. To further limit the escape of air from the BB during its opening, a 2.2-mm keratotome can be applied in a horizontal plane for entry, instead of the lifting-motion cut using a sharp knife.


In 1999 Melles et al. described a technique that employed the injection of an ophthalmic viscoelastic device (OVD) to create a viscobubble and separate DM from the corneal stroma. This bubble mimics the behavior of the type 1 BB seen with air injection. In this technique, a 27-gauge needle or blunt-tipped bottom port cannula, attached to a viscoelastic-filled syringe, is inserted into the corneal stroma as close to the DM as possible. To visualize the depth of corneal incision and lamellar dissection during surgery, Melles proposed the creation of an air-to-endothelium interface by filling the AC with air, which then behaves as a convex mirror. A dark nonreflective band can be seen between the tip of the blade and its reflection, representing the nonincised corneal tissue between the blade and the air-to-endothelium interface. This dark band becomes thinner as the blade is advanced into deeper stromal layers. The needle/cannula is advanced in this plane, and when the tip appears to touch its reflection, the OVD can be injected to create the viscobubble. The formation of the viscobubble is outlined by a typical golden ring reflex. After the corneal pocket filled with the OVD has been created, a suction trephine can be centered over the anterior corneal surface. Trephination is performed until viscoelastic is seen to escape from the pocket through the trephine incision. The overlying stroma is excised and the recipient bed is thoroughly irrigated to remove all the OVD.

Some surgeons prefer to debulk the cornea before attempting the viscobubble, to enhance the chances of injecting the OVD as close as possible to the DM. The use of a cohesive OVD is preferred, because it is easier to remove and reduces the risk of a postoperative double AC. The viscobubble has also been suggested as a rescue bubble technique, after failed air injection: air-viscobubble (AVB). When air dissection does not result in the formation of a BB, superficial keratectomy is performed with a crescent blade and a new deeper tunnel is created into the stroma by using a blunt spatula/pointed dissector. Viscobubble though a blunt-tipped bottom port cannula is tried as a second strategy to separate DM from the corneal stroma. In a case series of 507 eyes affected by keratoconus, this combined technique (AVB) increased the percentage of bubble formation by 12%, bringing the total cases of successful bubble formation to 94%. When the BB fails, the cornea is generally pneumatized and offers many pathways of less resistance to air compared with the pre-Descemet space (i.e., leakage of air through the stroma, the trabecular meshwork, or the trephination groove). Because of its high viscosity, an OVD does not easily escape, creating a much higher intrastromal pressure, increasing the chances of bubble formation and probably also increasing the pressure inside the small stromal bubbles (from the failed BB attempt) that spontaneously merge to form a large DM detachment.


Despite being technically more challenging, manual dissection techniques are still valid options. They are mainly adopted in cases in which air- or viscodissection fails or is not indicated because of previous hydrops or deep stromal scars involving DM.

Peeling Technique

Malbran and Stefani described this easy and rapid manual technique in 1972, and it still represents a useful option, especially in eyes with keratoconus. After performing initial trephination, the inner stromal edges are raised using a Paufique knife and are then grasped securely with forceps, which are used to pull the stromal tissue to peel it from underlying deeper layers. Once the proper plane of 85% to 90% depth is reached, peeling off the stroma without applying excessive force is relatively easy, especially in the area of the cone of keratoconus in which the deeper lamellae have the lowest adhesion. This technique usually creates a recipient bed that is very deep, smooth, and regular in its thickness.

Some modifications of the original techniques have been proposed. Performing a partial debulking and using dedicated blunt-tipped forceps (Fogla stromal pocket forceps, Storz Ophthalmics, MO) or a blunt-tipped spatula (Sarnicola DALK spatula, ASICO, IL) to try to find an appropriate deep plane in the peripheral cornea all around the trephination groove, may facilitate the peeling of the stroma. The peeling of the stroma may by aided by a semisharp dissector (Fogla lamellar dissector, Storz Ophthalmics, MO), which may be particularly helpful in the presence of an intraoperative DM rupture, to limit the force of the peeling and the risk of enlarging the rupture.

Layer-by-Layer Manual Dissection

Dry manual dissection is one of the oldest described DALK techniques, which gained attention again when Tsubota et al. applied the cataract surgery principle of “divide-and-conquer” to corneal transplantation. After the initial trephination, to facilitate lamellar dissection, the recipient cornea is divided into four quadrants at approximately 70% of corneal depth. This division is then continued until a deep, smooth, and regular plane is exposed. “Layer-by-layer” manual dissection usually results in a pdDALK, as baring the DM is challenging.

This technique has a high risk of intraoperative perforation, especially while performing deeper dissections close to DM. Reducing the IOP by evacuating some aqueous through a peripheral paracentesis and wetting the stroma with some balanced salt solution to avoid dessication might help to reduce this risk. Injecting air or fluid into the stroma to create emphysema can be helpful to determine the depth of the stromal dissection, as described by Archila in 1985 and by Sugita and Kondo in 1997. , New technological tools such as intraoperative anterior segment optical coherence tomography (AS-OCT) or a handheld pachymeter may be helpful to assess the residual stromal bed and the need for further dissection. An irregular stromal bed with thickness >80 microns is likely to impact the visual outcome of DALK with the manual dissection technique.

Pachymetry-Assisted Manual DALK

Pachymetry-guided use of a diamond knife to initiate the plane of manual dissection for stromal removal has been successful in performing a pdDALK. , After a circular mark is made at the desired trephine diameter, intraoperative pachymetry is performed at the site of planned incision using a diamond knife. The diamond knife is set at a depth of 30 microns less than the pachymetry reading and used to make a 2.0-mm incision just internal to the trephination mark, at 11 to 12 o’clock position. The dissection plane is initiated at the base of the cut, followed by using curved medium-sized fine blade scissors to extend the incision on either side circumferentially. An open centripetal lamellar dissection is performed using lamellar dissectors.


Microkeratome-assisted ALK, also known as automated therapeutic lamellar keratoplasty (ATLK), involves the use of an automated microkeratome to perform anterior lamellar dissection of both the donor and recipient cornea. Being automated, the resultant bed is usually very smooth. However, this technique is usually not successful in performing an ideal lamellar dissection in thin corneas with areas of elevation, such as in advanced keratoconus. Semiautomated LK, a procedure that combines manual recipient bed lamellar dissection with automated donor preparation using a microkeratome, seems to be a safe and effective technique. It combines the benefits of smooth microkeratome lamellar dissection of the donor with customized lenticule thickness and diameter together with a manual lamellar dissection technique for the recipient, providing encouraging visual outcomes that show continuing improvement with time.

Busin et al. have also reported successful outcomes for DALK using a microkeratome-prepared large-diameter (9 mm) donor, along with baring of PDL only in the central 6 mm of the recipient bed.


The use of femtosecond lasers in DALK allows corneal-shaped wound creation in a precise manner instead of manual trephination and, in addition, the ability to create a plane for air injection at a desired depth in the corneal stroma. , Stromal removal is performed using the BB technique or near-DM manual dissection techniques. The shaped wound configuration has the advantage of better donor host apposition, with increased surface area contact, resulting in faster wound healing, greater tectonic stability, earlier suture removal, and possibly reduced astigmatism ( Fig. 31.7 ). The predefined stromal track facilitates air injection close to DM with improved rates of BB formation. Use of zigzag or mushroom-shaped edge configuration has been described for DALK cases with successful outcomes.

Oct 30, 2022 | Posted by in OPHTHALMOLOGY | Comments Off on Lamellar Keratoplasty in Keratoconus

Full access? Get Clinical Tree

Get Clinical Tree app for offline access