Purpose To analyse the functional and anatomical outcomes of different types of keratoprostheses in eyes with retained silicone oil following vitreoretinal surgery.
Methods Retrospective chart review of patients operated with any type of permanent keratoprosthesis (Kpro) in silicone oil-filled eyes between March 2003 and June 2017 were analysed.
Results 40 silicone oil-filled eyes underwent keratoprostheses, of which 22 were type 1 and 18 were type 2 Kpros (Lucia variant—nine, modified osteo odonto kerato prosthesis (MOOKP)—four, Boston type 2—three and osteoKpro—two) with a mean follow-up of 61.54 , 42.77, 45.25 , 25 and 37 months, respectively. Anatomic retention of the primary Kpro was noted in 33 eyes (82.5%). A best-corrected visual acuity of better than 20/200 and 20/400 was achieved in 26 (65%)+32 (80%) eyes. Retroprosthetic membrane (RPM) was the most common complication noted in 17 eyes (42.5%). Perioptic graft melt was noted in 4 of 22 eyes of the type 1 Kpro (2 (10.5%) without associated ocular surface disorder (OSD)) and in 1 eye each of Boston and Lucia type 2 Kpro. Laminar resorption occurred in one eye each of the MOOKP and OKP groups. Endophthalmitis and glaucoma did not occur in any eye.
Conclusion Appropriately chosen keratoprosthesis is a viable option for visual rehabilitation in eyes post vitreoretinal surgery with retained silicone oil-induced keratopathy not amenable to conventional penetrating keratoplasty. Kpro melt among type 1 Kpro did not occur in 89.5% eyes without associated OSD (19 of 22 eyes), despite the lack of aqueous humour and presence of RPM (4 eyes), two factors considered to play a significant role in the causation of sterile melts. Of interest to note was the absence of infection in any of these eyes. The possible protective role of oil from endophthalmitis is interesting, though yet to be ascertained.
- retinal detachment
- ocular hypotony
- silicone oil
- boston keratoprosthesis
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The visual prognosis in eyes post vitreoretinal procedures with the need for long-term silicone oil retention is guarded and is further compromised by the silicone oil-induced keratopathy. Keratoprosthesis (Kpro) procedure can be performed along with vitreoretinal surgeries that require simultaneous silicone oil injection to reattach the retina and/or to address the associated chronic hypotony secondary to ocular comorbidities.1 2 Kpro can also be performed in eyes with retained oil wherein the oil cannot be removed to avoid a re-retinal detachment or counter the hypotony, an indication for the type 1 Kpro that had been described by the authors earlier.3 Retained silicone oil is known to be a cause for keratopathy as well as for failure of corneal grafts.4 This prevents good long-term visual recovery in these eyes despite the retinal health allowing for better visual outcome. Earlier few reports are for the simultaneous implantation of Boston type 1 Kpro with silicone oil injection for the associated vitreoretinal condition.1–3However, silicone oil-filled eyes that are not suitable for the type 1 Kpro would require one of the type 2 Kpros, either the modified osteo odonto kerato prosthesis (MOOKP) or the Boston type 2 and its variants or the osteokpro for visual rehabilitation.
In addition to postoperative issues that are common to the type of Kpro implanted in any eye except for differences in any specific complication that might be noted, oil-filled eyes have additional preoperative and intraoperative considerations to be kept in mind. The power of the Kpro to be implanted is indeterminate. The best vision achieved after silicone oil injection prior to initiation of keratopathy is of utmost importance in determining the possible visual potential in the eye (termed expected visual outcome) and therefore the decision to implant a Kpro. Knowledge about the viscosity of the oil in the eye helps determine the need to replace or refill during the Kpro surgery. Intraoperatively, the need for a vitreoretinal colleague and the placement of the oil infusion cannula are mandatory necessary additions over a routine Kpro procedure. In eyes with chronic hypotony due to a reduced aqueous production, emulsification of the silicone oil is not an issue. This possibility is further reduced with a higher viscosity of the oil. Any additional vitreoretinal procedures are based on intraoperative findings and one needs to be prepared for the same.
Herein, we present our outcomes with different types of Kpros in silicone oil-filled eyes, analysed separately for each type of Kpro, over a long-term follow-up as well as the observation of certain variations in outcomes from the other indication groups with possible explanations for the same.
Approval of the Institutional Review Board was obtained for this study which adhered to the tenets of the Declaration of Helsinki.
The records of 40 eyes of 40 patients who underwent any type of permanent Kpro implantation with silicone oil injection/replacement between March 2003 and June 2017 were analysed. All patients were bilaterally blind with no functional vision in the other eye. All eyes had undergone previous vitreoretinal surgeries (38 eyes underwent these surgeries in our institute with all previous records available) with inability to remove the silicone oil due to associated hypotony or risk of re-RD.
Each Kpro was performed as has been described earlier along with the vitreoretinal surgeon at the appropriate stage of surgery to replace or top up with 5000 cs silicone oil.3 5–7
Preoperative and Postoperative evaluation
Based on our records or previous records from patients, the best achieved visual acuity following vitreoretinal surgery before onset of corneal/graft changes was noted among patients who underwent the type 1 Kpro (22 eyes). Assessment of visual potential was not possible for patients with severe OSDss who underwent vitreoretinal procedures with plan for subsequent type 2 Kpro due to the need for a total tarsorrhaphy or conjunctival hooding after the vitreoretinal procedure (using a temporary Kpro) with a penetrating keratoplasty.
For the type 1 Kpro candidates, a detailed ophthalmic evaluation was carried out which included present best-corrected visual acuity, slit lamp examination, intraocular pressure measurement and fundus examination with indirect ophthalmoscope through hazy corneal medium (for eye to be operated) corroborated by B-scan ultrasound examination of the posterior segment (for both eyes). Finger tension was documented in eyes with non-recordable intraocular pressure.
The eye was examined for the presence of adequate tears by performing Schirmer’s I test and noting the amount of wetting. Patency of nasolacrimal duct was confirmed to rule out any adjacent focus of infection. The completeness of the blink was confirmed as was the absence of any cause for exposure.
For the type 2 Kpro candidates, MOOKP was the first choice of procedure in eyes with a relatively good prognosis. For those not eligible for the same, the other kpros were decided on based on the health of the oral mucosa, fitness for general anaesthesia and condition of the eyelids to decide between the osteoKpro and the Boston type 2 Kpro and based on the condition of the eyelids and affordability of the patient to decide between the Boston type 2 Kpro and the Lucia type 2. This decision-making was not influenced by the presence of oil in the eye; rather it was based on the general guidelines followed to choose a particular type of Kpro in these eyes. However, the Lucia type 2 Kpro was the preferred choice in patients with a guarded visual prognosis based on intraoperative fundus findings by virtue of its low cost and the technique being not as demanding as the MOOKP.
In addition to a detailed ophthalmic evaluation including confirmation of perception of light with accurate projection of rays, normal intraocular pressure noted by means of digital tonometry and a B-scan for posterior segment evaluation, an oromaxillofacial surgeon evaluated the oral health along with a spiral CT scan of the canines in case of the MOOKP.
Details regarding patients’ acceptance for regular follow-ups, compliance with postoperative medications, realistic expectations and adherence to strict hygienic postoperative care were ensured prior to planning any Kpro surgery.
Postoperative care and follow-up for each of the Kpro was as described earlier.
The primary outcome measures included anatomical and functional success. Anatomical success was defined as retention of the first Kpro implanted at last follow-up. Functional outcome was recorded in terms of the best-corrected visual acuity achieved and maintained with respect to the expected visual outcome. Postoperative complications and the necessary surgical interventions for the same were secondary outcome measures.
All statistical analyses were performed using SPSS V.14. Success rates were reported as percentages with 95% CIs or Kaplan-Meier survival probability/rates.
The type 1 Kpro was performed in 22 oil-filled eyes (15 Boston type 1 and 7 Lucia variant of the type 1 Kpro). Of the remaining 18 oil-filled eyes that underwent the type 2 Kpros, the Lucia type 2 Kpro, MOOKP, Boston type 2 and the osteoKpro were performed in nine, four, three and two eyes, respectively. The demographics including the aetiology are tabulated in tables 1 and 2 for the type 1 and 2 Kpros, respectively.
The mean follow-up was calculated for each type of Kpro. It was noted to be 61.54 months (for the type 1 Kpros) and 42.77, 45.25, 25 and 37 months, respectively, for each of the type 2 Kpros.
Of the 22 eyes with the type 1 Kpro, 3 eyes had an associated ocular surface disease (chemical injury 2, Stevens-Johnson syndrome (SJS) 1). The remaining 18 eyes had a relatively normal ocular surface with no clinically evident limbal stem cell deficiency. Prior penetrating keratoplasty had been performed in 14 out of the 22 eyes. In the remaining eight eyes, type 1 Kpro was performed as the primary procedure to address silicone oil-induced keratopathy. There was no difference in the outcome noted in these eyes.
In total, 17 of the 18 eyes that received the type 2 Kpro were in patients with OSDs (chemical injury 12, SJS -5).
Anatomical retention of the primary Kpro was noted in 33/40 eyes (82.5%) (type 1 Kpro—18/22; Boston type 2—3/3, Lucia type 2— 8/9, OKP—1/2, MOOKP—3/4).
A functional visual outcome of >20/200 was achieved in 26 eyes (65%) of which 16 eyes (40%) maintained it till last follow-up (figure 1A). Among the type 1 Kpros, 13 out of the 22 eyes (59.1%) achieved a BCVA of >20/200 of which 7 eyes (31.8%) maintained the same till last follow-up.Of the six eyes that experienced a drop in vision, two had a recurrent RD, the eye with SJS had a corneal melt with perforation and RD that was settled, one eye with chemical injury had a Kpro extrusion, one eye developed recurrent retroprosthetic membrane (RPM) requiring frequent membranectomies and one eye had a drop in vision associated with disc pallor.
For the type 1 Kpro, the expected visual outcome was documented to be less than 20/200 in 15 eyes (68%). A visual acuity of better than the expected visual outcome was achieved in 6 eyes (27.27%) with a BCVA of better than 20/200 achieved in 13 eyes (59.1%).
In the type 2 Kpro group, 13 out of the 18 eyes (72.2%) achieved a BCVA of >20/200. This was maintained in 10 eyes (76.9%) at the time of last follow-up. Of the three eyes that had a decrease in vision, the causes included RPM in two eyes (Boston type 2 Kpro—1,Lucia type 2—1) and extrusion of the lamina in one eye with the MOOKP.
A functional visual outcome of >20/400 was achieved in 32/40 eyes (80%) and maintained in 23/40 eyes (57.5%). Among the type 1 Kpro, 18/22 (81.8%) eyes achieved better than 20/400 and maintained in 11/22 eyes(50%). Among the type 2 Kpros, BCVA>20/400 was achieved in 14/18 eyes (77.7%) and maintained in 12/18 eyes (66.7%) (figure 2).
A hyperopic refractive outcome was noted in 35 out of 40 eyes (90%).
The details of the functional and anatomical outcomes for type 1 and type 2 Kpros are given in tables 1 and 2, respectively.
Perioptic melt (defined as carrier graft melt in the type 1 and Boston and Lucia type 2 Kpro and laminar resorption in the MOOKP and OKP) was noted in four eyes (18.1%) among the type 1 Kpro patients and one each in all four type 2 Kpros. Of the four eyes in type 1 Kpro, two eyes belonged to chemical injury (figure 1C,D) and SJS aetiology and two eyes developed a graft melt secondary to repeated bandage contact lens displacement. Of these four eyes, a reKpro was done in three eyes with improvement in vision in two eyes to the earlier achieved vision and one patient refused further intervention.
Among the type 2 Kpro melts/resorption, the Boston type 2 Kpro patient underwent a lamellar patch graft to address the melt, the Lucia type 2 Kpro patient had a reKpro done twice with restoration of vision to the earlier achieved vision, the OKP patient had the Kpro replaced with recovery of vision and the MOOKP patient with Kpro extrusion refused any further intervention.
A reKpro was required in seven out of the eight eyes of which five underwent a secondary Kpro (type 1 Boston Kpro—2, Lucia type 1—1, Lucia type 2—1 and OKP—1), with maintenance of BCVA achieved with the primary Kpro, and two refused further intervention.
Retroprosthetic membrane was noted in 17 eyes (42.5%). Of these eight eyes (36.4%) belonged to the type 1 Kpro (7/15 Boston type 1 eyes and one out of seven Lucia type 1 eyes) (figure 1B). Among the type 2 Kpros, two eyes with the Boston type 2 (figure 3A) and OKP group each, five eyes in the Lucia type 2 Kpro (figure 3B) group and none in the MOOKP group had RPM. RPM was not noted to have any correlation with either the primary aetiology (injury, uveitis, rhegmatogenous RD in myopic eyes) or with the duration between the last performed vitreoretinal procedure with silicone oil injection and the Kpro procedure. However, among the type 1 Kpros, five of the eight eyes that had RPM developed it within 3–4 months of the Kpro procedure requiring surgical membranectomy through the pars plana by the vitreoretinal surgeon in all eyes within 6 months.
Re-retinal detachment occurred in three eyes with the type 1 Kpro requiring surgical intervention with partial recovery of vision. This was secondary to the primary aetiology in one eye and following a sterile melt with leak in two eyes.
Glaucoma was not seen to be associated with any of these eyes that were hypotonous to begin with.
Endophthalmitis or graft infiltration was not seen in any eye across all types of Kpros.
The options for visual rehabilitation for corneal decompensation in eyes with retained silicone oil either due to chronic hypotony or a risk of retinal re-detachment are not too many. The long-term presence of oil in the eye inevitably leads to corneal decompensation and causes similar damage with time to the transplanted cornea too. In the presence of a functional normal contralateral eye, no intervention is pursued in the affected eye due to the known poor outcomes. In bilaterally blind patients, single eyed or otherwise, Kpro forms a valuable alternative to penetrating keratoplasty in the eye with retained silicone oil as described earlier.1–3 The advantages of the same have been discussed earlier, and the authors have presented their short-term outcomes with the type 1 Kpro in oil-filled eyes.4
The choice of Kpro depends on the primary aetiology as well as the ocular surface status. A type 1 Kpro is the procedure of choice in normal eyes with retained oil without associated OSDs, with adequate tears and a normal blink. The type 2 Kpros are indicated in severe OSDs such as chemical injuries and SJS. Among the type 2 Kpros, various factors determine the choice of the Kpro which are adopted as applicable to any eye undergoing a Kpro procedure and to elaborate the same is beyond the purview of this paper. The indications for type 1 and type 2 Kpros are different and their outcomes cannot be compared. The aim of this paper was not to compare the outcomes of the different types of Kpros; rather it is to highlight the encouraging outcomes in these oil-filled eyes with a variety of available options.
In oil-filled eyes, certain specific issues exist that differ from the other eyes that undergo Kpro, especially with regards to those that receive the type 1 Kpro in an otherwise normal eye. These are eyes that have had a retinal detachment and/or intraocular inflammatory processes leading to chronic hypotony. Hence these are eyes that would be expected to have a subnormal visual recovery. It is therefore of importance to estimate the possible extent of visual recovery in these eyes preoperatively in order to project the target visual outcome. This is termed as the expected visual outcome and should be determined from the best visual acuity achieved following the vitreoretinal procedure prior to onset of corneal decompensation. This might not always be accurate as was noted in our series with 27.27% eyes achieving a vision better than the expected outcome due to clarity of the Kpro and reduced astigmatism compared with the cornea or graft in an oil-filled eye.
With a posterior segment pathology requiring silicone oil to be retained long term, the expected visual outcome was less than 20/200 in most eyes. Though Kpro outcomes are usually expressed in terms of visual acuity better than 20/200, a BCVA of 20/400 allows a patient to be self-ambulatory and hence the visual outcome in our series was also analysed with this yardstick which is the limit for legal blindness. The functional success for a BCVA of better than 20/400 was maintained in 50% of eyes compared with 31.8% with a BCVA of better than 20/200 for type 1 Kpros over a mean follow-up period of >5 years.
Simultaneous Kpro can be undertaken in eyes with posterior segment pathology when, the primary aetiology warrants a Kpro for the corneal condition such as aniridia,2 the patient understands that the visual prognosis depends on the final retinal status and is willing to follow the rigmarole involved in Kpro postoperative care irrespective of visual outcome. It is also important that the delay involved in the time taken to procure the Kpro in terms of cost and distance should not have an adverse effect on the outcome of the posterior segment intervention. All or some of these were prohibitive/non-conducive factors in our series and all eyes in our series received the Kpro subsequent to the vitreoretinal procedure. The visual outcome with respect to the expected visual outcome describes the functional outcome in these eyes better than the absolute best-corrected visual acuity which might appear much less than reported outcomes following type 1 Kpros.
Estimating the expected visual outcome in eyes with OSDs like chemical injuries or SJS with concurrent retinal detachment that requires the silicone oil to be retained in the eye is usually not possible due to the need to prevent a corneal melt by means of a conjunctival hood or tarsorrhaphy. This does not allow the postoperative vision to be recorded. In certain instances, the eye can be left open without a conjunctival hooding or tarsorrhaphy for the first 4–5 days with the sole aim of estimating visual recovery after which a tarsorrhaphy is performed. In the presence of graft haze or oedema or intraoperative haem trickling into the vitreous cavity, an accurate estimation of vision might still be difficult within the first postoperative week.
A hyperopia of >+4D was noted in 35 eyes (type 1 Kpro—18 eyes) with the remaining eyes having a myopic or low hyperopic error. Unlike intraocular lens power calculation for cataract surgery in oil-filled eyes, where a corrective factor is applied for eyes that are to have simultaneous silicone oil removal or in eyes where an intraocular lens is to be placed with an intact PC without altering the retained oil, the preoperative calculation of the dioptric power of the Kpro in an oil-filled eye may not be relevant with respect to the final refractive outcome. In oil-filled eyes, the amount of oil in the eye will change intraoperatively, thereby leading to an over or under fill compared with the preoperative status which will alter the dynamics and thereby the dioptric power calculation in these eyes. The oil-Kpro meniscus too has a role to play in determining the final refractive outcome, which in turn will depend on the amount of oil filled at the end of surgery as well as the curvature of the backsurface of the optic which is plano in the type 1 Kpro (so also in the Lucia and Boston type 2 Kpros) and convex in the optical cylinder used for the MOOKP and the OKP. However, based on the largely hyperopic refraction noted in most eyes in our series, it might be prudent to choose a Kpro in the range of axial length of 19–21 for hypotonic oil-filled eyes that require a Kpro. All the Lucia type 2 Kpros in our series were of a standard axial length of 23 D with final refractive outcome of more than +4D and the Boston type 1 Kpros implanted were in the range of 21–23 D.
RPM was noted to be the most common complication following any type of Kpro in oil-filled eyes (36.3% following type 1 Kpro).8 9 However, these rates are similar to those seen after a type 1 Kpro for any other indication (>50%) and hence may not represent a specific role of the oil in the formation of the RPM. Though the numbers are too small to compare, none of the MOOKP eyes developed RPM.
Two issues were noted in these eyes which were of particular interest. One being the relatively rare occurrence of sterile melts in these eyes despite the relative lack of aqueous humour in these eyes largely replaced by silicone oil, and the other being the non-occurrence of endophthalmitis in all eyes, over a considerable period of follow-up.
Sterile graft melts have been reported in 10%–25% of eyes with type 1 Kpro.10 11 Apart from dessication related to lack of BCL and OSDs, one of the most important causes cited for melts is the presence of RPM that prevents the access of aqueous humour to nourish the graft. Among the 19 out of the 22 eyes with the type 1 Kpro in our series with no associated ocular surface disease, 5 had RPM and 2 had a sterile melt (of which 1 eye had RPM). This indicated that 89.4% of eyes not associated with OSDs prone to sterile melts had no graft melt over a mean follow-up of 5 years, of which four had RPM. These were eyes that were marginally hypotonous despite the oil in the eye that had not undergone emulsification indicative of very minimal to nil aqueous production. This raises questions regarding the nutritive role attributed to the aqueous humour as well as its lack in causing sterile melts and probably needs to be studied further to understand a causative role.10 11 Sequestration of inflammatory mediators by the oil preventing it from causing a sterile melt should have held truer for RPM which was seen to occur at the same rate as for other eyes. Moreover, it has been the lack of nourishment more so that has been held responsible for the sterile melts rather than presence of inflammatory mediators.
Of the four eyes that required a re-Kpro due to an oil leak (two type 1 Kpro/two type 2 Kpros), the recovery of vision was as much as following the first Kpro. This probably could be due to the slow leak that occurs in an oil-filled eye compared with aqueous leak which leads to a sudden hypotony and its consequences on the retina.
Another very interesting aspect noted in our series was the absence of endophthalmitis in silicone oil-filled eyes across all types of Kpros over the duration of follow-up. Sight-threatening complication like endophthalmitis has been reported to occur in upto 15% of eyes with the type 1 Kpro.11 The potential antimicrobial properties of silicone oil have been described. The highly hydrophobic nature of the oil makes it impervious to cells and bacteria. Its high interfacial tension limits space for free movement of infectious agents. It maintains cells and infective agents in close contact with the ciliary body and retinal vessels, which might improve the efficacy of human defence mechanisms, with higher concentration of biochemical mediators, antibodies and inflammatory cells within the limited aqueous phase of the vitreous cavity.12 13 The possible protective role of oil from endophthalmitis is interesting, though yet to be ascertained.
Prior reports1 2 other than the authors′3 are primarily of simultaneous implantation of silicone oil and type 1 Kpro in patients with coexisting corneal/OSDs and retinal pathology requiring oil injection. Some of these indications would warrant silicone oil removal later. Our series is representative of the long-term outcome in eyes with pre-Kpro presence of oil, with silicone oil-induced keratopathy being the indication for all the type 1 Kpros. This series also highlights the use of various type 2 Kpros in oil-filled eyes not reported earlier. These outcomes are definitely encouraging, do raise questions regarding our understanding of perioptic tissue melt related to lack of aqueous nourishment and highlight the possible role of silicone oil in curbing infections. It also focuses on the varied choices of Kpro for these eyes based on the severity of the ocular surface disease.
The authors thank Professor James Chodosh, MEEI, Boston, for continued support and guidance with type 2 Kpros (Boston and Lucia variants).
Contributors GI, BS and SA designed the study, provided the material, analysed and interpreted the data, wrote, proofed and revised the article. RP analysed and interpreted the data ER, PR and SS provided the material and proofed the article VN provided the statistical analysis.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Ethics approval Approval of the Institutional Review Board was obtained for this study which adhered to the tenets of the Declaration of Helsinki.
Provenance and peer review Not commissioned; externally peer reviewed.