Outcomes of Deep Anterior Lamellar Keratoplasty Using the Big-Bubble Technique in Various Corneal Diseases


To report clinical outcomes of deep anterior lamellar keratoplasty (DALK) using the big-bubble technique in various original diagnoses.


Prospective interventional study.


Single hospital study of 115 unselected consecutive patients (131 eyes) with various diagnoses undergoing DALK using the big-bubble technique. The main outcome measures were intraoperative and postoperative complications, postoperative best spectacle-corrected visual acuity (BSCVA), and endothelial cell density (ECD).


Descemet membrane was exposed successfully in 25 eyes (80.6%) with advanced keratoconus, 11 (73.3%) with chemical or thermal burns, 20 (71.4%) with corneal dystrophy, 21 (70%) with a herpes simplex (HSK) keratitis scar, 4 (36.4%) with moderate keratoconus, and 5 (31.3%) with a bacterial keratitis scar ( P < .05). Loosening of the sutures occurred in 24 eyes (23.8%) between postoperative4 and 7 months. Epithelial rejection was observed in 2 eyes and stromal rejection occurred in 5 eyes. BSCVA was improved in HSK scarring and corneal dystrophy vs keratoconus and corneal burns at 1 year ( P < .05), but not after a mean follow-up of 21.4 months ( P < .05). Patients who experienced stromal rejection had lower ECD than patients with no rejection at 18 and 24 months ( P < .05). At all follow-up times after 6 months, patients with multiple air injection attempts had lower ECD than patients with 1 injection attempt ( P < .05).


Different Descemet membrane exposure rates were observed in different diagnoses using the big-bubble technique. Both severe stromal rejection and additional manipulation may have a deleterious effect on the corneal endothelium.

Deep anterior lamellar keratoplasty (DALK) has been proposed as an alternative to penetrating keratoplasty (PK) for the treatment of various corneal diseases that do not affect the corneal endothelium. The main advantage of DALK is preservation of endothelium, reducing the risk of immunologic reactions and graft failure; however, attempts to isolate Descemet membrane can be complicated by perforations, which can lead to PK or collection of aqueous fluid in the double anterior chamber recipient–donor interface, leading to stromal opacity.

Anwar and Teichmann proposed the big-bubble technique, which allows the surgeon to gain safe and direct access, with the advantages of shortening surgical time, reducing risk of perforation, and exposing a smooth, even surface of excellent optical quality.

The big-bubble technique represents the most important recent development in DALK. Previous studies have indicated DALK for various corneal stromal diseases with unaffected endothelium; however, no large-scale study has been conducted to compare the success rate and visual outcomes of DALK in various indications where the corneal endothelium is affected. The aim of this study was to compare the therapeutic outcomes after DALK in patients with different diagnosis. To our knowledge, this is the first study to investigate the success rate of the big-bubble technique in the treatment of various ocular pathologies.


Between September 3, 2003 and August 20, 2008, a total of 115 consecutive patients (131 eyes) with various diagnoses were included in a prospective clinical study that aimed to evaluate clinical outcomes, especially the success rate of Descemet membrane exposure using the big-bubble technique.

Patient Data

The 115 patients with various diagnoses assigned to undergo DALK had a mean ± standard deviation (SD) age of 33.6 ± 10.7 years at surgery and mean ± SD follow-up of 24.8 ± 7.8 months. The indication for grafting was to improve vision in all patients. Table 1 summarizes the characteristics and operative data of patients in the study.


Demographic Data of Patients Undergoing Deep Anterior Lamellar Keratoplasty Using the Big-Bubble Technique

Advanced Keratoconus Moderate Keratoconus HSK Scar Corneal Dystrophy Bacterial Scar Burn
Number of eyes (n) 31 11 30 28 16 15
Male:female 19:12 5:6 17:13 18:10 10:6 10:5
Age (y) (mean ± SD) 24.4 ± 10.8 22.5 ± 7.2 39.4 ± 10.2 47.3 ± 11.6 36.5 ± 9.7 37.4 ± 11.2
Follow-up (m) (mean ± SD) 24.2 ± 5.6 25.3 ± 7.4 22.4 ± 7.4 22.6 ± 6.5 25.6 ± 7.7 23.7 ± 5.8
Preoperative BSCVA 0.186 ± 0.104 0.248 ± 0.126 0.08 ± 0.022 0.1 ± 0.036 0.04 ± 0.062 0.04 ± 0.058

BSCVA = best spectacle-corrected visual acuity; HSK = herpes simplex keratitis; SD = standard deviation; y = years.

Primary pathologies consisted of 31 eyes with advanced keratoconus, 11 with moderate keratoconus, 30 with stromal scar attributable to herpes simplex keratitis (HSK), 16 with stromal scar attributable to bacterial keratitis, 28 with corneal dystrophy (including 6 macular, 10 granular, and 12 lattice), and 15 with corneal clouding attributable to chemical or thermal burn. In the patients with keratoconus, we limited the indication of DALK to moderate to advanced cases intolerant to contact lenses and with poor spectacle-corrected visual acuity. Presence of apical scarring not involving the Descemet membrane was the indication for DALK. In other groups, we limited the indication of DALK to corneas in which the opacity did not involve the Descemet membrane under slit-lamp examination.

Preoperative and postoperative eye examinations, including the Snellen best spectacle-corrected visual acuity (BSCVA), slit-lamp examination, tonometry, fundus examination, and specular microscopy, were carried out by 1 of 2 investigators (H.T., W.Y.). Visual function was retested 1, 3, 6, 12, 18, and 24 months after surgery; specular microscopy was repeated at 6, 12, 18, and 24 months follow-up.

Corneal endothelium was photographed before and after surgery with a noncontact specular microscope (Topcon SP2000p; Topcon Corp, Tokyo, Japan). Endothelial cell density (ECD) was calculated by marking 40 cells within a predefined standard region of interest. Intraoperative and postoperative complications and secondary interventions were recorded.

All patients were operated on by a single surgeon who was experienced in full-thickness and lamellar graft surgery (H.T.) at the same hospital (Zhongshan Ophthalmic Center, Guangzhou, China). The big-bubble technique described by Anwar and Teichmann was used. All procedures were performed under retrobulbar anesthesia.

In brief, a partial-thickness trephination of variable diameter, between 7.5 and 8.0 mm, was performed with the Barron trephine (Katena, Denville, New Jersey, USA) set to cut 400 μm into the corneal stroma. A superficial keratectomy was performed with a crescent knife to reduce the corneal stroma by half.

A 30-gauge needle bent 60 degrees 5 mm from the tip with the bevel facing down was inserted on a 5-mL syringe filled with air. The needle was introduced into the deep stroma starting at the bottom of the trephination groove and advanced approximately 3 to 4 mm toward the center of the cornea following the corneal curvature. Air was injected progressively into the stroma, with the aim of forming a large air bubble between the Descemet membrane and posterior stroma. A peripheral paracentesis was performed, allowing some aqueous to escape and lower intraocular pressure. Viscoelastic material was injected into the space between posterior stroma and Descemet membrane to maintain the space. Blunt-tipped scissors were used to divide posterior stroma into 4 sections, which were then removed by cutting each quadrant at the edge of the trephination, exposing the Descemet membrane. In cases where a big bubble was not achieved after 3 to 6 attempts, a layer-by-layer manual stromal dissection was performed with a crescent knife to expose the Descemet membrane.

Fresh donor sclerocorneal buttons were used for the preparation of donor cornea. The cornea was punched out from the endothelial side with the Barron donor punch (Katena, Denville, New Jersey, USA) with a blade of a diameter 0.2 to 0.3 mm larger than the recipient. Descemet membrane with endothelium was gently stripped off with a dry sponge. The button was sutured in place using 1 continuous 20-bite 10-0 nylon suture or 16 interrupted sutures. At the end of surgery, intraoperative suture adjustment was performed with the Malone hand-held keratoscope. All sutures were removed at 18 months; earlier removal was performed in cases where the sutures loosened and/or when increased vascularization of the host cornea occurred.

When the cornea was reepithelialized, the recipient eye was treated with topical dexamethasone 0.1% 4 times a day for a month. Artificial tears were instilled 4 times a day for 6 months. Topical steroid was tapered over a period of 12 weeks.

All excised diseased host tissues were fixed in 10% neutral buffered formalin (NBF) for 24 hours at 4 C, then automatically dehydrated, embedded in paraffin (Citadel 2000; Shandon, Cheshire, England), and stored at room temperature. Sections (5 μm) were cut using a Leica microtome (American Optical Company, New York, New York, USA) and stained with hematoxylin-eosin. The corneal buttons excised from corneal dystrophy were prepared for histochemical examinations including Masson’s trichrome and Congo red staining.

Statistical Analysis

Statistical analysis was performed with SAS for Windows Version 8.1 (SAS Institute Inc, Cary, North Carolina, USA). The Kruskal-Wallis test was used to compare multiple samples with a nonparametric multiple comparison test post-test. Data on loosening of sutures and vascularization of the host cornea were expressed as proportions and analyzed with Fisher exact test. The Mann-Whitney U test was used to compare ECD outcome variables between patients experiencing stromal rejection and patients with no rejection episode, or patients who had multiple attempts at air injection and patients who had only 1 attempt; P < .05 was considered statistically significant.


True Descemet Membrane Exposure Frequency

Overall, the DALK surgical procedures were carried out on 101 of 131 eyes (77.1%), including 86 eyes (65.6%) using the big-bubble technique. In the remaining 45 eyes where air injection did not result in big-bubble formation, layer-by-layer manual stromal dissection was required to expose Descemet membrane. Using this technique, Descemet membrane exposure was successfully achieved in 15 eyes (15/45, 33.3%). In the remaining 30 cases (22.9%) conversion to PK during surgery was required as a result of Descemet membrane tearing or perforation of deeper stroma.

Using big-bubble technique, Descemet membrane was successfully exposed in 25 eyes (25/31, 80.6%) with advanced keratoconus, 11 (11/15, 73.3%) with chemical or thermal burns, 20 (20/28, 71.4%) with corneal dystrophy, 21 (21/30, 70%) with HSK scar, 4 (4/11, 36.4%) with moderate keratoconus, and 5 (5/16, 31.3%) with bacterial keratitis scars. Advanced keratoconus had the highest Descemet membrane exposure success, whereas bacterial keratitis and moderate keratoconus corneal scars had the lowest incidence of Descemet membrane exposure ( P < .05).


Operative complications are summarized in Table 2 . The most common complications in this study, all of which of all occurred during layer-by-layer manual stromal dissection, were Descemet membrane tearing in 6 of 45 patients (13.3%) and perforation of the deeper stroma in 24 of 45 eyes (53.3%). Based on our previous experience, we converted to PK to avoid postoperative double anterior chamber, Descemet membrane detachment, or papillary block attributable to injection of air into the anterior chamber in all cases where Descemet membrane tearing and deeper stroma perforation occurred.


Postoperative Complications After Deep Anterior Lamellar Keratoplasty Using the Big-Bubble Technique in Various Corneal Diseases

Complications Advanced Keratoconus (n = 31) Moderate Keratoconus (n = 11) HSK Scar (n = 30) Corneal Dystrophy (n = 28) Bacterial Scar (n = 16) Burn (n = 15) Total (n = 131)
Interface opacity 1 1 2 1 1 6
DM folds 4 2 2 1 9
Sutures loosening 7 3 6 3 5 24
Vascularization 5 4 3 3 15
Graft rejection 1 2 1 1 2 7
Delayed reepithelialization 1 2 1 2 6
Recurrent epithelial erosion 3 3

DM = Descemet membrane; HSK = herpes simplex keratitis.

Clear graft-host interface was observed in the early postoperative period in 86 of 101 eyes (85.1%) in which DALK was conducted. Although true exposure of Descemet membrane was achieved during DALK, mild graft-host interface opacity was recorded in 6 of 101 eyes (5.9%). Interface opacities cleared during the first 3 months of follow-up in 4 eyes and remained unchanged in 2 eyes at the last visit. Poor final visual outcome associated with persistent interface opacities occurred in 1 eye with macular corneal dystrophy attributable to opacities at the center of the recipient bed. The other persistent interface opacity was observed in a chemical burn, without involvement of the visual axis. Descemet membrane folds were observed in 9 of 101 eyes (8.9%), mostly (6 eyes, 66.7%) in keratoconus with severe preoperative ectasia, and all Descemet membrane folds disappeared within 3 to 6 months after surgery.

In the 101 eyes in which DALK was successfully conducted, loosening of the sutures was recorded in 24 eyes (23.8%) at 4 to 7 months, including 10 with keratoconus, 6 with HSK scar, 5 with chemical or thermal burns, and 3 with corneal dystrophy. In our series, continuous sutures were used most often in keratoconus and corneal dystrophy patients, while the interrupted sutures most often in HSK, bacterial scar, and corneal burn patients. Thus, suture loosening after DALK may be much more common in interrupted than in continuous suture technique. Vascularization of the host cornea occurred in 15 eyes (15/101, 14.9%), including 9 with keratoconus, 3 with HSK scar, and 3 with chemical or thermal burns. Compared with tight sutures, loosening of sutures increased vascularization of the host cornea (9/24, 37.5% vs 6/77, 7.8%, P < .05).

Delayed re-epithelialization (beyond 2 weeks after surgery) occurred in 2 eyes with HSK scar, 2 with chemical or thermal burns, 1 with corneal dystrophy, and 1 with keratoconus. Recurrent epithelial erosion at between 4 and 7 months was observed in 3 eyes with chemical or thermal burns. Both delayed re-epithelialization and recurrent epithelial erosion were successfully managed with a bandage of soft contact lens.

Epithelial rejection showing epithelial rejected line was observed in 2 eyes at 6 and 8 months after surgery, respectively. Stromal rejection associated with decreased vision occurred in 5 eyes between 6 and 15 months, characterized by stromal haze and edema. Except for 1 graft performed on a host with chemical burns, which failed at 7 months as a result of irreversible stromal rejection, the others responded well to intensive topical corticosteroids, and grafts cleared after 2 to 3 weeks of treatment.

Visual Outcome

Two cases were excluded for visual outcome analysis, including 1 with persistent interface opacities involving the visual axis and 1 with failed graft as a result of irreversible stromal rejection. Table 3 shows the eyeglasses corrected changes in BSCVA in various pathologic entities after DALK. Postoperative BSCVA stabilized at 6 months in advanced and moderate keratoconus, and at 3 months in the remaining pathologic entities. No statistical difference was observed in BSCVA among the various diagnoses at a mean follow-up of 21.4 months (range 18-24), at least 6 months after all sutures were removed ( P > .05). At 1 year, when the sutures were still in place, BSCVA was better in HSK scar and corneal dystrophy recipients vs advanced and moderate keratoconus and corneal burn patients ( P < .05).

Corneal Endothelial Cell Density

Images of the corneal endothelium were obtained preoperatively in all patients with moderate keratoconus, and in only 23 with corneal dystrophy and 22 with advanced keratoconus attributable to severe central corneal conical protrusion. No images of the corneal endothelium were available preoperatively in patients with HSK scar, corneal burns, or bacterial keratitis scar because of heavy corneal clouding. Five patients who experienced stromal rejection showed a relatively lower ECD than patients with no rejection episode at 18 and 24 months follow-up, including the patient who developed graft failure because of irreversible stromal rejection ( P < .05, Figure 1 ) . After 6 months, the 18 patients who had more attempts (range 3 to 6 times) at air injection into the deeper stroma showed significantly lower ECD than patients in which the first attempt was successful ( P < .05, Figure 2 ) . Excluding the 5 patients who experienced stromal rejection and 18 patients in whom more than 1 attempt of air injection occurred, the remaining cases showed a small drop in the physiologic rate within 6 months after surgery. ECD stabilized after 6 months and no progressive endothelial cell loss was noted after 6 months ( Tables 4 and 5 ).

Jan 12, 2017 | Posted by in OPHTHALMOLOGY | Comments Off on Outcomes of Deep Anterior Lamellar Keratoplasty Using the Big-Bubble Technique in Various Corneal Diseases

Full access? Get Clinical Tree

Get Clinical Tree app for offline access