Fig. 23.1
Clinical picture of a keratoconic eye after penetrating keratoplasty with a double continuous 10/0 Nylon suture
Interrupted sutures, CCIS, and a single continuous suture (SCS) have shown comparable postoperative astigmatism [27]. In addition, a comparison of astigmatism in keratoconus patients utilizing a single continuous versus a double continuous suture showed that after suture removal, astigmatism was comparable (DCS −4.6 D, SCS −5.2 D) between the two groups [28]. Therefore, it is apparent that all methods of suture closure can work well. The ultimate choice rests with the surgeon.
Regardless of the preferred method, it is very important to have clear concepts of each suture technique. As a basic idea for standard graft suturing, the needle is passed 90 % depth through the donor cornea and then through the host cornea. The ideal bite is as close to Descemet’s membrane as possible, and there should be an equal amount of tissue purchased in the donor and host cornea in order to approximate Bowman’s layer in both the donor and host. Discrepancies frequently exist in the thickness of the donor and host cornea if donor corneas are thick due to the hyperosmolar glycosaminoglycans in the preservation medium, or fresh donor tissue is used in patients with severe corneal edema. In keratoconic eyes this scenario is frequent, where the graft is sutured to a relatively thin host cornea. Closing Bowman’s layer to Bowman’s layer should always be attempted to avoid steps in the graft–host junction and subsequent exposed sutures, so in areas where the recipient cornea presents thin (assessed preoperatively by slit lamp examination) partial thickness bites (50–70 % depth) in the donor tissue should be in relation with deep bites (95 % depth) in the host thin stroma (Fig. 23.2).
Fig. 23.2
Normal appearance of the graft–host junction with correct aligning of Bowman’s layer of the donor and host corneas, with needle passed at a 90 % depth in both sides (a). If care is not taken in cases of a thin recipient cornea, steps will remain at the graft–host junction, leaving an irregular astigmatism and exposed sutures that need to be replaced (b). To avoid this, a partial thickness bite (50–70 % depth) should be performed at the donor side (c)
The postoperative astigmatism management and elective suture adjustment/removal for PK in cases of previous keratoconus does not differ from other PK indications , with a complete suture removal generally recommended after 12–15 months.
23.3.3 Outcomes
Penetrating keratoplasty offers good long-term visual rehabilitation for keratoconus patients, and compared with other indications for PK there is a relatively low rate of graft failure and long mean graft survival. Rejection rate has been reported to be 5.8–41 % with a long-term follow-up, where most rejections occurred in the first 2 years [29–33]. Larger host trephine size, male donor gender, and nonwhite donor race have been associated with increased rejection hazard [29]. Despite this observed rejection rate, only a 4–6.3 % graft failure rate has been reported with a mean follow-up of 15 years, with an estimated 20-year probability of 12 % [29, 30, 34]. Fukoka et al. reported a cumulative probability of graft survival at 10, 20, and 25 years after PK of 98.8 %, 97.0 %, and 93.2 %, respectively, while Pramanik et al. estimated a graft survival rate of 85.4 % at 25 years after initial transplantation [30, 34]. Summarizing, the existing evidence shows that the graft survival rate gradually decreases after 20 years post-PK.
An average best-corrected visual acuity (BSCVA ) in logarithm of the minimum angle of resolution (LogMAR ) at preoperation, 10, 20, and 25 years after surgery of 1.54 ± 0.68, 0.06 ± 0.22, 0.03 ± 0.17, and 0.14 ± 0.42, respectively, has been reported [30]. Best spectacle-corrected visual acuity (BSCVA ) of 0.14 ± 0.11 LogMAR has been reported with a mean period of 33.5 months, while a BSCVA of 20/40 or better with a mean follow-up of 14 years was observed in 73.2 % of patients [33, 34].
An open angle glaucoma rate of 5.4 % with a mean follow-up of 14 years has been reported [34].
Claesson et al. reported a poorer survival and worse visual outcome of regrafts compared with first grafts in patients where the original indication was keratoconus: the failure rate was three times higher with regrafts and the observed visual acuity with preferred correction was ≥ 0.5 in 69 % of first grafts, while only 55 % of regrafts achieved that level [35].
23.4 Deep Lamellar Anterior Keratoplasty in Keratoconus
The goal of deep lamellar anterior keratoplasty in keratoconus is to achieve a depth of dissection as close as possible to the Descemet Membrane (DM ). There are various ways to create a plane of separation between DM and the deep stromal layers, mainly variations of the two basic strategies: the Anwar big bubble method and the Melles manual dissection.
23.4.1 Surgical Techniques
23.4.1.1 The Big Bubble Method
Anwar based the big bubble method on a discovery in 1998 that intrastromal injection of balanced salt solution (BSS) was often effective at establishing cleavage plane just above the DM [36], taking advantage of the loose adhesion between DM and the posterior stroma. Anwar and Teichman described the current big bubble procedure in 2002 using air instead of BSS [9].
After a partial trephination of 70–80 % of the corneal stroma, pneumatic pressure is used to detach DM by injecting air into the deep stroma with a 30G needle. The air injected into the stroma produces a dome-shaped detachment of the DM that is seen under the surgical microscope as a ring meaning that the big bubble has been formed. The stromal tissue above the DM plane is removed with spatula and scissors, making first sure to exchange the air in the supradescemetic plane with viscoelastic to avoid inadvertent puncture of the DM . When all of the stromal tissue is successfully removed, the DME membrane exposed is characteristically smooth (Fig. 23.3).
Fig. 23.3
DALK Big Bubble Technique: after a partial trephination of 70–80 % of the corneal stroma, pneumatic pressure is used to detach DM by injecting air into the deep stroma with a 27G needle (a). Once the air is injected it produces a dome-shaped detachment of the DM that is seen under the surgical microscope as a ring meaning that the big bubble has been formed (b). Then a lamellar dissection with a Crescent blade of the anterior stroma is performed (c) followed by the removal of the stromal tissue above the DM plane with spatula and scissors (d), making first sure to exchange the air in the supradescemetic plane with viscoelastic to avoid inadvertent puncture of the DM . When all of the stromal tissue is successfully removed, the DME membrane exposed is characteristically smooth (e), and the donor cornea without its DM and endothelium is then sutured with the preferred suture technique (f)
23.4.1.2 Melles Manual Method
This technique is based on the air–endothelium interface [8]. First the anterior chamber is filled with air. Then, using a series of curved spatulas through a scleral pocket, the stroma is carefully dissected away from the underlying DM . The difference in refractive index between the air and the corneal tissue creates a reflex in front of the surgical spatulas, and the distance between the instrument and the reflex is used to judge the amount of remaining tissue. Viscoelastic is injected through the scleral incision into the stromal pocket. Once the desired plane is reached, the superficial stroma is removed using trephine and lamellar dissection (Fig. 23.4).
Fig. 23.4
DALK Melles Technique: first the anterior chamber is filled with air and a partial trephination of 70 % of the corneal stromais performed (a). Then, using a series of curved spatulas through a scleral pocket, the stroma is carefully dissected away from the underlying DM (b). The difference in refractive index between air and corneal tissue creates a reflex of the surgical spatulas, and the distance between the instrument and reflex is used to judge the amount of remaining underlying tissue (b, arrows). Viscoelastic is injected through the scleral incision into the stromal pocket and the dissection can be completed through the trephination edge (c). Once it is completed, the superficial stroma is removed (d), the DME membrane exposed, (e) and the donor cornea sutured (f)
Since the original descriptions, there have been many variations to the standard technique. Lamellar dissection can be made with diamond knife, nylon wire, microkeratome [37], or femtosecond laser. To help guiding the dissection plane trypan blue, ultrasound pachymetry [38] or real time optical coherence tomography [39] (OCT) has been tried. Partharsathy et al. describe the “small bubble” technique for confirming the presence of the big bubble [40].
For corneas with extreme peripheral thinning, a modified procedure has been proposed dubbed “tuck-in lamellar keratoplasty” [41, 42]. In this technique, the central anterior stromal disc is removed and a centrifugal lamellar dissection is performed using a knife to create a peripheral intrastromal pocket extending 0.5 mm beyond the limbus. The donor cornea is prepared in such a way that it has a central full thickness graft with a peripheral partial thickness flange. The edges of a large anterior lamellar graft are tucked in below to add extra thickness.
23.4.2 Outcomes
Most studies have found equivalent visual and refractive results between PK and DALK provided stromal dissection reaches the level or close to the DM [12, 43–48], although 20/20 vision seems more likely after PK [12, 48]. For instance, in a recent study from Australian patients including 73 consecutive patients with keratoconus, the mean BCVA was not significantly different for DALK (0.14 logMAR, SD 0.2) versus PK (0.05 logMAR, SD 0.11) [12, 44]. A review of published literature that included 11 comparative studies on DALK and PK found that visual and refractive outcomes are comparable if the residual bed thickness in DALK cases is between 25 and 65 μm [14].
In those studies where the visual outcomes of DALK were inferior to PK [49], the dissection plane was “predescemetic” and the incomplete stromal dissection and the not fully baring of the DM had a negative impact in the results [49]. The problem seems to be related to the depth of the undissected stromal bed rather than to its smoothness as predescemetic DALKs performed by laser ablation did not outperform those dissected manually.
The recently published Australian graft registry data compared the outcomes of PKs and DALKs performed for KC over the same period of time and found that overall, both graft survival and visual outcomes were superior for PK. In a recent study from the UK, Jones et al. compared the outcomes after PKP and DALK for keratoconus [12]. The risk of graft failure for DALK was almost twice that for PKP . Probably, in the day-to-day clinical practice, visual outcomes with DALK, although comparable with PK, may be just slightly inferior or less predictable compared with PK, given surgical inexperience, and unpredictable issues regarding residual stromal thickness and DM folds. Nonetheless, elimination of risk of endothelial rejection compensates for this difference.
Lastly, one of the important advantages of DALK is a lower rate of endothelial loss compared with PK. The reported endothelial cell loss is as high as 34.6 % after PK, whereas it was 13.9 % after DALK [50].
23.4.3 Complications
Allograft reactions are less frequent in DALK than in PK and less likely to result in graft failure if correct treatment is initiated. Subepithelial and stromal rejection after DALK has been reported in the range of 3–14.3 % whereas in PKP ranges from 13 to 31 % in the first 3 years after surgery [11]. Endothelial rejection is not an issue in DALK.
Increases in IOP following DALK has been reported to be only 1.3 % of operated eyes, compared with 42 % of eyes after PK [50]. Development of glaucoma may also be up to 40 % less [51]. It is attributed to the lower steroid requirement of DALK [52].
Urrets–Zavalia Syndrome first reported following PK in KC and causing fixed, dilated pupil with iris atrophy is a rare entity following DALK [53].
There are also a few complications that are unique to DALK and the presence of a donor–host interface. One of the major problems with DALKs is intraoperative DM perforation, which may occur in 0–50 % of the eyes [11]. Surgeon’s inexperience, corneal scarring near the DM, and advanced ectasias with corneal thickness less than 250 μm increase the risk [54, 55]. Depending on the size of the perforation, conversion to PK may be required to avoid double anterior chamber and persistent corneal edema, especially when the rupture leads to the collapse of the anterior chamber (macroperforation). Incidence of pseudoanterior chamber or double anterior chamber is in the range of 1 % [56]. It can occur because of retention of fluid secondary to breaks in the DM, or, because of incomplete removal of viscoelastic in the interface [57]. Large pseudo chambers must be managed surgically by drainage of the fluid and anterior chamber injection of air or gas [58]. The presence of DM folds caused by a mismatch between donor button and the recipient bed is usually transient and disappear over time, but interface wrinkling when central and persistent may affect quality of vision [59]. Occasionally an eye with anatomically correct DALK may require a reoperation secondary to interface haze and poor visual acuity, usually stemming from incomplete or predescemetic stromal dissection [11]. Interface keratitis is a serious complication of DALK and it is caused mainly by Candida [60] but Klebsiella pneumonia [61], and nontuberculous mycobacteria [62] have also been isolated in several cases. Conservative treatment is usually unsuccessful and most cases need a therapeutic PK [60]. Interface vascularization can occur because of inflammatory, infective, and traumatic episodes and can be treated with injection of bevacizumab [63].
23.5 Femtosecond Laser-Assisted Keratoplasty for Keratoconus
The capability of femtosecond laser energy to create different cutting patterns with a controlled level of biological interaction and minimal tissue trauma has provided a new possibility for corneal surgeons in both penetrating and nonpenetrating keratoplasty procedures. From case to case the cutting profile can be more convenient in one specific shape, what was impossible to be made before with the manual trephination techniques. The potential advantages of femtosecond laser-assisted keratoplasy are the following [64, 65]:
- 1.
More precise and regular cuts
- 2.
No risk of injury for intraocular structures
- 3.
Perfect donor–recipient size matching
- 4.
Less injury to donor endothelium
- 5.
Customization of the cutting pattern
- 6.
More donor–host tissue interaction promoting better wound healing
- 7.
Potentially less induction of surgically induced astigmatism
- 8.
Shorter visual rehabilitation time
- 9.
Stronger and probably more stable wounds with earlier suture removal
23.5.1 Femtosecond-Assisted Keratoplasty Incision Profiles
Intralase (AMO) provides the more sophisticated and complex patterns when compared to the other technologies. The combination of simpler incisions (posterior side cut, anterior side cut, and lamellar cut) can be combined in limitless number of complex edge profile graft combinations [66] with a very high level of precision in dimensions and concentration which is not possible with other techniques [67].
Top-hat-based edge profile
This type of architecture maximizes the posterior tissue to be transplanted. Not suitable for keratoconus.
Mushroom-based edge profile
The inverted version of a top-hat is known as mushroom (Fig. 23.5) and is often used for anterior surface surgery, being composed of a narrower posterior side cut and a wider anterior side cut both intersected by the ring lamellar cut. The broader anterior section maximizes the anterior stroma area to be transplanted, what makes it suitable for keratoconus.
Fig. 23.5
Mushroom-based edge profile
Zigzag-based edge profile
Zigzag profile (Fig. 23.6) is composed of a slanted anterior and posterior side cuts connected by the ring lamellar cut. It can be adapted to maximize either anterior or posterior surface, which makes it suitable to a wide variety of situations, including keratoconus.
Fig. 23.6
Zigzag-based edge profile
23.5.2 Basics on Femtosecond Graft Architecture
Regardless of the profile being used, it is important to understand the basics of femtosecond incisions and stepped graft architecture [64].
Incision intersection
Femtosecond laser-assisted keratoplasty is formed by a combination of straight incisions that need to interact with each other in order to obtain a clear and continuous graft cutting profile that is easy to dissect, handle, and be later extracted. The overlapping degree is generally set between 10 and 30 μm.
Lamellar cut depth
As a thumb rule, lamellar cut depth is set at 50 % of average pachymetry in both donor and recipient corneas.
Oversizing
Femtosecond laser is able to create identical sized grafts in terms of diameter regardless of the keratometry readings, which has led to recommend using the exact same diameter in donor and recipient corneas [66, 68]. It is then up to the surgeon to decide oversizing primarily just in terms of postoperative spherical refraction needs.
Deepest posterior point
When designing the donor graft in PKP , posterior depth must be set below the maximum pachymetric reading in the incision area in order to assure a clean and full thickness incision. On the other hand, recipient might be moved from laser room to the main surgical operating room. In these cases, patient movement makes advisable to intentionally avoid performing a full thickness incision to prevent any pressure leaking during transportation from the Lasik room to the suturing surgical theater. Thus, posterior depth is generally set 70 μm above the thinnest pachymetric reading in the incision area, which is generally enough tissue to prevent wound leakage [69].
23.5.3 Surgical Technique
Donor: an artificial anterior chamber (AAC) is required to hold the donor tissue in proper position and pressure similar to a real patient during a lasik procedure; Recipient: as previously described, it is necessary to leave a safety gap of posterior noncut tissue on the recipient cornea which prevents any aqueous humor leakage during transportation [69]; Graft manipulation: a Sinskey hook is used to dissect the different incisions being the procedure identical in both donor and recipient corneas. However, recipient cornea partial thickness cut will be completed with a diamond blade and curved scissors.
23.5.4 Femtosecond-Assisted Penetrating Keratoplasty
The initial published evidence comparing manual and femtosecond laser-assisted PK (f-PK ) showed a faster visual rehabilitation together with an improved best corrected visual acuity, and a lower refractive and topographic astigmatism in the laser group [70–73]. Nevertheless, these papers had a limited follow-up of the cases, generally shorter than a year. Little evidence still exists about the long-term outcomes of f-PK : Chamberlain et al. published their results with 2-year follow-up and using a “zig-zag” edge profile [74]. They could demonstrate a topographic astigmatism significantly lower in the f-PK group but only during the first 6 postoperative months. Afterward no significant differences were observed regarding the refractive or topographic astigmatism and visual acuity. Only a few papers have been published comparing the different cutting edge profiles [75]. Our impression is that there are not significant differences in the visual or refractive outcomes regarding the preferred edge profile, although studies comparing the “zig-zag” and “mushroom” profiles in keratoconus are still required (Fig. 23.7).
Fig. 23.7
Femtosecond laser-assisted penetrating keratoplasty with a “Zig-Zag” edge profile (a, b; courtesy of Abbott Medical Optics, USA). Postoperative clinical picture (c) and an anterior segment OCT capture (d) where it is possible to appreciate the zig-zag edge profile at the host–donor interface with a perfect coalescence of the edges
23.5.5 Femtosecond-Assisted Deep Anterior Lamellar Keratoplasty
The use of the femtosecond laser in DALK (f-DALK ) avoids manual trephination and allows more precise identification of tissue depth and insertion of the air needle by following the plane between the lamellar and posterior laser side cuts. As variability in stromal thickness in eyes with advanced keratoconus, ectasia, or dense and deep stromal scars may limit the ability of the femtosecond laser to produce a uniform lamellar plane, we use the laser only to create the side cut both in donor and recipient cornea, while leaving a minimal amount of residual corneal tissue. With this we try to control the potential risk of creating a DM perforation with the femtosecond laser.
Femtosecond laser mushroom configuration is the preferred profile for DALK (Fig. 23.8). For the side cut a full thickness mushroom configuration cut is made on the donor cornea first and then a nonpenetrating mushroom configuration on the recipient. In the recipient cornea, the depth of the anterior side cut is about 60 % of the thinnest corneal pachymetry, the depth of the posterior side cut about 80 % of the thinnest corneal pachymetry, leaving a ring lamellar cut of 1 mm (Fig. 23.9). In the donor cornea, the Descemet’s membrane (DM ) and endothelium is debrided assisted by trypan blue dye.
Fig. 23.8
Femtosecond laser-assisted DALK with a “Mushroom” edge profile (up: courtesy of Abbott Medical Optics, USA). Postoperative anterior segment OCT capture (down) where it is possible to appreciate the mushroom edge profile at the host–donor interface
Fig. 23.9
Mushroom DALK configuration: the depth of the anterior side cut is about 60 % of the thinnest corneal pachymetry, the depth of the posterior side cut about 80 % of the thinnest corneal pachymetry, leaving a ring lamellar cut of 1 mm
Femtosecond laser-assisted DALK might have an advantage in keratoconus cases because it provides a larger amount of donor–recipient tissue to interact for the purpose of corneal wound healing consistency, and it has been demonstrated recently by our group that f-DALK shows a more active wound healing leading to leucomatous wounds [76]. We established a grading for the side cut corneal healing pattern as observed by slit lamp examination (Table 23.1), and we could observe that 52 % of f-DALK cases showed a healing pattern grade 3 or 4 [76]. The reasons for this could be either due to the larger area of contact between the donor and recipient tissues and/or to femtosecond laser-related biological activation of the corneal tissues, which should be related to the level of energy used for the creation of the side cut.
Table 23.1
Analysis of femtosecond laser side cut corneal wound healing pattern
 |
Equivalent to f-PK , f-DALK accelerates the visual rehabilitation, showing a better visual result during the immediate postoperative period, but without significant differences after the sixth postoperative month [77]. In a recent study of our group, we could not demonstrate significant differences regarding the visual or refractive outcomes after 1-year follow-up. However, we demonstrated a faster visual rehabilitation in the f-DALK group versus manual DALK as well as significant differences in the wound healing pattern between groups, being more intense in the f-DALK group [76].
Summarizing, and regarding the current evidence, femtosecond laser accelerates the visual rehabilitation after PK or DALK compared with the manual technique, obtaining better refractive and visual results along the immediate postoperative period; however, these benefits are lost later in the follow-up, not being able to improve the refractive or visual outcomes in the long term. Considering the high costs of this technology, these results do not justify its use as a gold standard for keratoplasty . Nevertheless, if available, femtosecond laser offers important intraoperative (easy wound closure) and perioperative (faster rehabilitation) advantages, together with a stronger surgical wound that allows a faster suture removal (depending on the wound healing pattern) and probably less risk of dehiscence against an ocular trauma in the long term.
23.6 Keratoconus Recurrence After Corneal Transplantation
We have already discussed the good long-term results of the different options of corneal grafting for keratoconus. Nevertheless, de Toledo et al. observed a progressive increase of keratometric astigmatism in 70 % of their cases from 10 years after suture removal, following an initial phase of refractive stability during the first 7 years after PK for keratoconus (4.05 ± 2.29 D 1 year after suture removal, 3.90 ± 2.28 D at year 3, 4.03 ± 2.49 D at year 5, 4.39 ± 2.48 D at year 7, 5.48 ± 3.11 D at year 10, 6.43 ± 4.11 D at year 15; 7.28 ± 4.21 D at year 20, and 7.25 ± 4.27 D at year 25), so a late recurrence of the disease may occur with an increasing risk over time [16]. Actually, a 20 year post-PK probability of 10 % has been reported previously, with a mean time to recurrence of 17.9–21.9 years, so given the younger age at which keratoconus patients undergo corneal transplantation, these long-term findings should be explained to patients and incorporated into the preoperative counseling [29, 34, 78] (Fig. 23.10).
Fig. 23.10
Reconstruction of corneal stroma. (a) Hematoxylin–eosin staining of a rabbit cornea with an implanted graft of decellularized human corneal stroma with h-ADASC colonization: hypocellular band of ECM without vessels or any inflammatory sign (magnification ×200); (b) human cells labeled with CM-DiI around and inside the implant that express (c) human keratocan (human adult keratocyte specific marker; magnification ×400), confirming the presence of living human cells inside the corneal stroma and their differentiation into human keratocytes (arrows); (d) phase-contrast photomicrographs showing a morphologically unaltered corneal stroma (magnification ×400); (e) the graft remains totally transparent after 12 weeks of follow-up (magnification ×2) (arrows point to the slightly visible edge of the graft). Epi epithelium, Str stroma, Lam Lamina
It is well known how other corneal stromal dystrophies, like granular or lattice dystrophy, tend to recur into the donor cornea, due to either colonization of the new stroma by the abnormal host keratocytes or epithelial secretion in early stages. In keratoconus this host keratocyte invasion has not been well stabilized as the main etiology for the post graft recurrent ectasia, but is likely in relation with the early keratoconic changes observed in the histology of explanted donor buttons after regrafting [78–80]. Postgraft ectasia is often preceded by thinning of the recipient stroma at the graft–host junction, so the disease progression at the host stroma is likely to be the underlying reason for these cases of recurrent ectasia and progressive astigmatism over time [16, 78]. In such cases, a mean keratometric sphere and cylinder increase of 4D and 3D, respectively, between final suture removal and diagnosis can be observed [78].
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
