Long-term results of Boston keratoprosthesis surgery in Polish patients

The attempts to use an artificial cornea to treat corneal blindness date back to the eighteenth century, but surgeries with acceptable rates of success were introduced in the last century. The Boston type I keratoprosthesis (BKPro) is the most commonly implanted keratoprosthesis worldwide. BKPro was first used in 1965 by Claes H. Dohlman [1]. In 1992, the U.S. FDA approved the use of the BKPro for corneal surgery, confirming the safety of its use. After more than 10 years, the device received a CE mark in 2014, allowing its commercial use in Europe.

The majority of patients with corneal blindness could be treated effectively with corneal transplantation. However, there is a group of patients with reduced graft survival rates [2, 3, 4]. Before the introduction of keratoprosthesis surgery, this group of patients was fated to lifetime blindness. BKPro has become a viable option for patients for whom corneal transplantation is high risk. At present, the main indications for BKPro surgery include: corneal graft failure, and congenital or acquired corneal limbal stem cell deficiency, including aniridia and ocular burns or vascularized corneal leucoma [5, 6, 7, 8, 9]. There are significant complications that can occur in eyes after keratoprosthesis surgery, including prosthetic failure, retroprosthetic membrane (RPM), glaucoma, keratolysis, vitreoretinal complications, endophthalmitis, cystoid macular edema, uveitis, sterile vitritis, and hypotony/phthisis [5, 6, 7, 8, 9, 10, 11]. A relatively high percentage of postoperative complications may influence the long-term success rate of this surgery.

Patients after BKPro surgery must follow a strict postoperative regimen and should be monitored carefully in case of an occurrence of listed complications. Modern anterior eye segment imaging techniques, such as optical coherence tomography (OCT), could be useful in monitoring the anterior chamber parameters in such patients [12]. OCT provides in vivo anterior eye segment imaging with the axial resolution from 18 μm with time domain OCT (TD OCT) to 5,0 μm with spectral domain OCT (SD OCT). OCT is proven to provide reliable anterior eye segment morphology and morphometry results [13, 14, 15].

The paper presents long-term (10 year) outcomes of BKPro surgery performed in six Polish patients, since to this date very few studies showed visual outcomes or retention rates in patients over the period of 5 to 10 years; a majority of these outcomes were based on estimated, not real, survival rates.

This was a retrospective study conducted at Chair and Clinical Department of Ophthalmology, Faculty of Medical Sciences in Zabrze, Medical University of Silesia in Katowice, Poland. The study was conducted in accordance with the ethical standards stated in the 1964 Declaration of Helsinki and was approved by the Ethics Committee of the Silesian Medical Chamber (resolution 14/2010/07/06/2010), since in 2010 the BKPro surgery was considered as experimental treatment in European countries.

1.3

Postoperative care

Patients were examined at day 1, 1 week, and 2 weeks within the first postoperative month, every month during one year after surgery, followed by every three months during the second year and every 6 months during the follow-up of 10 years. Last recorded visits were performed in November of 2020, when all patients met criteria for over 10 years of the follow-up period. Postoperative medication regimen obligatory for all patients included topical 0.5% moxifloxacin, 0.5% vancomycin and 0.1 % dexamethasone four times daily within the first month, 2 times daily within the second postoperative month. After two months, 0.5% vancomycin and 0.5% moxifloxacin were prescribed once daily. Postoperative care also included the use of a disposable soft contact lens (PureVision 2 - balafilcon A; BC=8,6 mm; Bausch and Lomb) over the keratoprosthesis in all patients. Visual acuity, slit-lamp examination with the detailed ocular surface assessment, AS OCT, fundoscopy and intraocular digital pressure measurement were performed at each visit.

Fig. 1

Main outcome measures in the postoperative period were stability of the keratoprosthesis, best corrected visual acuity (BCVA), and occurrence of complications.

Retention of keratoprosthesis was achieved in all patients during the follow-up period of 10 years. The pre- and postoperative data of patients were summarized in Table 1.

2.6

Anterior segment OCT

Preoperatively AS OCT examination revealed ACD ranged from 1.8 to 4.5 mm (median = 2,98 mm), ACA 0° and 180o ranged from 0° to 41o and from 0o to 37o respectively. Preoperative OCT data showed some degree of anterior synechial angle closure in 3 patients (PIII, PIV, PVI). Postoperative data showed the formation of anterior synechiae with ACA closure in 5 patients (all except PI). At last recorded visit ACA 0° and 180o ranged from 0o to 39o and from 0° to 41o respectively. In patients after AGV implantation, OCT allowed the tube visualization in the anterior chamber. In patient PII, in the early postoperative period, we observed a BKPro protrusion with the front plate elevated over the level of the anterior corneal surface. The protrusion was not noted on the slit lamp exam. It resolved 6 months after the surgery. In patient PIII, part of the retroprosthetic membrane could be visualized behind the implant. Figure 2 contains TD OCT representative scans showing complications which occurred in the presented case series.

Fig. 2

In the last ten years, there have been an increasing number of publications regarding intraoperative and postoperative complications of BKPro surgery, indicating that BKPro is a viable option allowing a long-term restoration of vision in patients with high risk for corneal transplantation. However, the majority of papers were focused on the short-term outcomes, with only few studies investigating long-term outcomes, and most of these were based on estimated, not real, survival rates [6, 7, 19, 20, 21, 22, 23, 24, 25]. The largest previously published longitudinal cohorts have demonstrated a range for the probability of device retention between 84% and 89% at 2 years, 67% at 4 years, and 67% at 7 years, depending on the study group size, length of the follow-up period, concomitant diseases, and compliance of patients [19, 20, 21, 25]. In the recent single-center, retrospective cohort study of 68 patients, the 5 and 10 years estimated probability of BKPro retention was 89.2%. The authors also reported a 72.6% probability of the initial device being retained [23]. In the follow-up period of 10 years, we observed a 100% BKPro retention rate. In the multicenter study on BKPro retention, three factors were found to be independent predictors of failure: autoimmune diseases, exposure keratopathy requiring a concurrent tarsorrhaphy and an increased number of prior failed penetrating keratoplasties. The most common cause of keratoprosthesis failure was infection and tissue necrosis. Cicatrizing conditions such as Stevens-Johnson syndrome and ocular cicatricial pemphigoid portend a worse prognosis [21]. The main indication for the surgery in our case series was failed graft surgery (3 patients), followed by ocular burn and vascularized leucoma after herpetic keratitis, which is consistent with all other published series [5, 6, 7, 8, 9]. We did not diagnose autoimmune diseases or exposure keratopathy in any of our patients, which could be a reason for the 100% retention rate.

As indicated previously by other authors, the BKPro surgery is usually complex with a high incidence of intraocular complications. The procedures most commonly performed in conjunction with the BKPro implantation reported in multicenter case series were: anterior vitrectomy, tube shunt implantation, cataract extraction, irydoplasty, pupilloplasty and intraocular lens removal [5, 20]. In our case series, only one out of six patients had BKPro implantation without additional procedures. The additional procedures were: cataract extraction (4 patients) and glaucoma valve implantation (2 patients). No significant intraoperative complications occurred, although it was the first BKPro surgery performed in Poland. The most probable reason is that the surgeries were performed in the corneal, clinical department and the Polish experienced corneal surgeon (EW) was assisted and guided by an experienced BKPro surgeon during the first five procedures.

Postoperative visual acuity outcomes were comparable to other published data [5, 6, 7, 8, 9, 19, 20, 21, 23, 24, 25]. In the latest study of the long-term BKPro outcomes published in 2020, the probability of maintaining or improving vision was 75.0% at 5 years and 66.7% at 10 years [23]. In our study, final postoperative vision was improved compared to the preoperative vision in 3 patients (PI–PIII), was unchanged in one patient (PIV), and decreased in two patients (PV, PVI). The poor visual outcomes were achieved in three patients with preoperative ocular surface injury after chemical and thermal burns. The diagnosis of the end-stage glaucoma was responsible for the irreversible decrease in visual acuity in our case series. The above observation was not consistent with the previously published data on the BKPro in ocular burns [26]. The authors of this research, based on a study group of 10 patients with a mean follow-up of 2 years, reported postoperative BCVA better than 20/200 in 90% of the patients and better than 20/60 in 60% of the patients. Six patients were suspicious of the glaucoma preoperatively, five of them were confirmed to have glaucoma after the surgery, and three had progression of underlying glaucoma. In the table with the list of BKPro complications, the authors mentioned glaucoma occurrence only in one patient. In the contrary, in our case series, the diagnosis of glaucoma was established in three patients preoperatively and in all patients postoperatively. Two patients had AGV implantation at the time of BKPro surgery (PIII, PIV), one patient prior to the procedure (PVI). The disease caused a progression to NLP vision in two patients (PV, PVI) and was responsible for no improvement in vision in one patient (PIV). Other authors also reported glaucoma to be the most common pre-existing comorbidity with an incidence ranging from 40% to 73% [5, 6, 7, 8, 11, 19, 20, 21, 24, 25, 26, 27]. The damage of the angle during the BKPro surgery was reported to be responsible for intraocular pressure elevation, anterior iris synechiae formation, and further glaucoma progression [11, 27]. Based on OCT scans, we observed anterior iris synechiae formation with ACA closure in 5 patients (all except PI). The progressive angle closure, shallowing of the anterior chamber, and synechiae formation, not visible on slit lamp exam, was also reported previously based on serial OCT analysis [28].

Another complication, very common after BKPro surgery, is the retroprosthetic membrane formation, which occurred in one patient (PIII) from our study group. It was treated successfully with YAG laser membranotomy. As indicated by the results of a multicenter study focused on the pathogenesis of this complication, the strongest risk factor for the retroprosthetic membrane development was infectious keratitis, while chemical injuries and adjuvant intracameral steroids were proved to have a protective role against the formation [29]. Our patient (PIII), who was diagnosed with the retroprosthetic membrane was after two failed corneal grafts, initially diagnosed with Axenfeld-Rieger syndrome with corneal leucoma, iris atrophy, glaucoma and nystagmus. Two patients with previous infectious keratitis did not develop a retroprosthetic membrane in the follow-up period. The retroprosthetic membrane formation was also the most common complication occurring in over 60% of patients in the recently published report on the long-term BKPro results [23].

Another important issue in the postoperative care is the role of the noninvasive imaging techniques, such as optical coherence tomography, in the detection and monitoring of anterior eye segment complications. During the qualification and follow-up, we used OCT analysis, which allowed visualization of the anterior iris synechiae, AGV tube, protrusion of the BKPro and the retroprosthetic membrane, not detectable during the slit lamp examination. Other authors also underline the role of OCT in anterior eye segment monitoring and detection of complications not visible on clinical examination [1, 28, 30, 31, 32, 33].

Ocular surface homeostasis pre- and postoperatively is a major factor responsible for the BKPro retention and visual outcomes. Severe ocular surface disease, such as keratoconjunctivitis sicca, cicatrizing conjunctivitis from chemical or thermal trauma, autoimmune or inflammatory conditions such as mucous membrane pemphigoid, Stevens-Johnsons syndrome, or atopic disease are significantly more common in eyes that require a device explantation, compared to those that retained the device (p<0,001) [19]. The lack of the tear film may result in sterile keratolysis, and ultimately endophthalmitis, which leads to the device removal and could cause the irreversible loss of vision. That is why in the qualification process for the BKPro type I surgery, it is crucial to assess and exclude severe DED. In patients with severe ocular surface disease who have poor prognosis for other types of corneal surgery, the BKPro type II or other types of keratoprostheses, such as the modified osteo-odonto Kpro or the osteo-Kpro should be considered. The BKPro type II is a modified version of the BKPro type I and has an additional anterior nub that is designed for implantation through surgically closed eyelids [34, 35].

DED is a multifactorial disease of the ocular surface characterized by a loss of homeostasis of the tear film, and accompanied by ocular symptoms in which tear film instability and hyperosmolarity, ocular surface inflammation and damage, and neurosensory abnormalities play etiological roles. DED is a disease of many etiological factors, multiple forms, and differing levels of severity. The established group of risk factors associated with it, which may influence the occurrence or intensification of DED, has been reported in the literature [36]. Those include, among many others, the factors common for patients qualified for BKPro surgery: environmental factors, ocular surgical interventions in the past, contact lens wear, ocular burns, neurotrophic corneal abnormalities and ocular uptake of certain groups of eye drops: antiglaucomatous, antibiotics, anti-inflammatory–containing preservatives. The main pathological form of DED is an evaporative DED developing in consequence to MGD at different severity levels. All our patients met the criteria for the MGD diagnosis preoperatively: mild in patients PI and PII, moderate in patients PIII, PV and PVI and severe in patient III. We also observed a significant progression of DED and MGD in the follow-up period. In our opinion this diagnosis of MGD worsening over time provides a rationale for implementing DED treatment in all BKPro patients. The main purpose of the treatment is to restore an ocular surface homeostasis. This can be achieved by supporting the stability of the tear film layers, normalizing their osmolarity, and establishing continuous biophysical protection for the ocular surface. All our patients were advised to use ocular lubricants without preservatives at least every three hours in the daytime and were instructed to use eyelid margin hygiene. Additionally in cases of severe MGD the patients were prescribed repeated courses of oral doxycycline treatment (50 mg for three months) during the follow-up. Tetracycline and its analogues (doxycycline) were proved in multiple studies to decrease the activity of collagenase, phospholipase A2, and several MMPs; to decrease the production of inflammatory mediators such as IL-1 and TNF-a in a wide range of tissues, including the corneal epithelium; and also to inhibit lipase production, which may contribute to improvement in clinical parameters in patients diagnosed with MGD [22]. Other studies also underline the progression of complication occurrence over time. Kanu et al. observed that all complications had increasing incidence beyond 5 years; in particular, corneal melt, surgical glaucoma interventions, and endophthalmitis tended to have late presentations [23]. This increase in prevalence of complications over time highlights the importance of long-term follow-up. Compliance of patients is another important factor influencing complication occurrence, underlined by many authors. Patients after BKPro surgery should be monitored and treated very carefully.

The conducted study is not without certain limitations. The main limitation of this study is the small sample size, with only six patients in the study group. However, since corneal blindness in patients with high risk for corneal transplantation is very rare, most reported studies also were based on small case series. The main advantage of our study is a long observation time, exceeding 10 years. To date only few studies reported a follow-up period above 5 years, with only one study published recently by Kanu et al. with a follow-up period of 10 years for part of their study group. Moreover, in 2010, BKPro surgery was an experimental treatment strategy and those six surgeries were first performed in Poland in order to provide a treatment for patients with irreversible blindness. At this time, due to many efforts of our team led by Professor Edward Wylegala, BKpro surgery became refundable in Poland in December 2019. We continue to provide this treatment option in the larger group of Polish patients.

In summary, our study of a 10-year results of BKPro surgeries in Poland provided a long-term observation of retention, visual outcomes, and possible complications related to BKPro type 1. We underlined the role of anterior optical coherence tomography in monitoring and detection of complications not visible on the slit lamp examination. We also provided a rationale for implementing DED treatment in all BKPro patients. Finally, our clinical results support the use of the BKPro in managing corneal blindness in patients with high risk for corneal transplantation.