Tese de Doutoramento Medicina 2024 Faculdade de Ciências Médicas, Universidade NOVA de Lisboa

Introduction: The goal of this thesis was to report the ten-year experience in Descemet’s membrane endothelial keratoplasty (DMEK) performed by a single, experienced corneal surgeon in a centre of reference for corneal transplantation in Spain. In this thesis, the long-term outcomes and safety in a large case series of primary DMEK surgeries were characterized and critically analysed to determine factors associated with postoperative visual outcomes, as well as risk factors for graft failure. In addition, subgroups of eyes considered complex cases for DMEK were analysed, including eyes with corneal decompensation related to phakic intraocular lens (PIOL) and eyes with failed Descemet stripping automated endothelial keratoplasty (DSAEK) grafts and other ocular comorbidities. Specific postoperative complications after DMEK and their management were also analysed, including immune rejection (IR), postoperative visually significant cystoid macular oedema (vs- CMO), post-DMEK intraocular lens opacification (IOLop) requiring explantation, residual ametropia and refractive enhancement with laser vision correction (LVC), and laser in situ keratomileusis (LASIK) interface fluid syndrome (IFS). An analysis of the potential risk of IR episodes following severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccination based on the available literature was also performed. Finally, the feasibility and outcomes of repeat DMEK surgery in eyes with failed DMEK graft were analysed and compared with a control group of successful primary DMEK eyes. Methods: This Thesis included several different studies conducted at the Cornea and Refractive Surgery Unit of the Instituto de Microcirugía de Barcelona (IMO Barcelona Grupo Miranza) between December 2019 and December 2022. Study #1 was a retrospective case-control study that aimed to determine the incidence and risk factors for primary graft failure (PGF, defined as absence of corneal clearing within the first 3 postoperative months, despite a correctly positioned and oriented DMEK graft) using multivariate logistic regression analysis, and secondary graft failure (SGF, defined as corneal decompensation after a period of functioning DMEK graft) using multivariate Cox survival regression analysis. In primary DMEK eyes without graft failure and with a minimum of 12 months of postoperative follow-up (F-U), mixed effect model for repeated measures (MMRM) was performed to identify determinants of postoperative best-corrected visual acuity (BCVA) after DMEK, and to analyse the proportion of eyes reaching higher levels of postoperative BCVA (≤ 0.10 logMAR, or ≥ 20/25 Snellen). Study #2 was a retrospective case series of eyes that underwent DMEK surgery for PIOL-related corneal decompensation, where visual outcomes, postoperative complications, corneal endothelial cell (CEC) loss and graft survival were analysed. Study #3 was a retrospective comparative study where the outcomes and safety of repeat DMEK surgery eyes were analysed and compared with a control group of successful, pseudophakic primary DMEK eyes to analyse differences in visual outcomes, postoperative complications, and graft survival. Study #4 was a retrospective case series of complex eyes with previously failed DSAEK grafts, where visual outcomes, postoperative complications, CEC loss and graft survival were analysed. An addendum was added to include one case that occurred after the study period, a case of LASIK IFS occurring secondary to a failed DSAEK graft which was managed with DMEK surgery. Study #5 was a retrospective study where the incidence of vs-CMO was estimated. The outcomes after management of vs-CMO were analysed, including BCVA and central retinal thickness (CRT) on macular SD-OCT. Study #6 was a cross-sectional study of eyes that underwent IOL explantation due to post-DMEK IOLop causing loss of at least 1 line of BCVA and that had a minimum of 6-month postoperative F-U after IOL explantation. Outcomes measured at the study visit were BCVA, complications, CEC loss, and DMEK graft survival. Study #7 was a case report where LVC was performed to correct residual post-DMEK ametropia. Study #8 was a systematic review of the literature aiming to characterize the potential risk of corneal graft IR episodes following SARS-CoV-2 vaccination. Results: Study #1 included 234 primary DMEK eyes. Intraoperative complications were observed in 11.5% of cases; rebubbling rate was 19.6%; PGF occurred in 8.97% of eyes; 7.3% of eyes had IR episodes; and 18.8% of eyes had SGF. The final multivariate logistic regression model for PGF included rebubbling (HR = 13.260, p = 0.001), preoperative BCVA ≥ 0.56 logMAR (HR = 7.228, p = 0.005), comorbid ocular surface or corneal disease (HR = 8.799, p = 0.008), recipient diabetes mellitus (HR = 6.047, p = 0.010), DMEK surgery performed after mid- 2017 (HR = 0.178, p = 0.045), preoperative IOP ≥ 19.5 mmHg (HR = 3.608, p = 0.045), and intraoperative complications (HR = 3.593, p = 0.068 – weak association). The final multivariate Cox regression model for SGF included IR episodes (HR = 3.403, p = 0.001), history of glaucoma (HR = 3.227, p = 0.003), more severe corneal endothelial (CE) failure (HR = 4.073, p = 0.004), and previous corneal graft (HR = 3.867, p = 0.006) A statistically significant global decrease in BCVA logMAR units was observed at all postoperative time points (p < 0.001) and on multivariate mixed model analysis younger patient age (p = 0.094), baseline preoperative BCVA < 0.375 logMAR (p = 0.003), postoperative time (months) (p < 0.001) and Fuchs’ endothelial corneal dystrophy [(FECD), p = 0.017] were significantly associated with an increased possibility of reaching postoperative BCVA ≤ 0.10 logMAR; absence of retinal comorbidities was associated with an increased probability of better postoperative logMAR BCVA (Coef = -0.1007; p < 0.001). In Study #2, sixteen eyes of 14 patients were analysed. Mean (SD) time to PIOL explantation was 9.3 (5.0) years. In eyes without PGF (n = 15), mean postoperative F-U after DMEK was 37.3 (29.6) months. In these eyes, median (P25-P75) BCVA at last F-U improved from 0.80 (0.60- 1.08) logMAR to 0.15 (0.00-0.35) logMAR (p = 0.002). Nine eyes had more than 2-year postoperative F-U, where median final BCVA was 0.10 (0.00-0.15) logMAR. In eyes with more than 2 years of postoperative F-U, and after excluding one eye that had LEGF at 4-year F-U, mean total CEC loss was 61.7 (11.7) %. Two eyes (12.5%) required rebubbling, one of which ended in PGF (6.2%); of the initially functioning grafts (n = 15), one eye (6.7%) evolved to late endothelial graft failure (LEGF) at 4-year F-U. One case underwent photorefractive keratectomy due to residual ametropia (Study #7), significantly improving unaided distance visual acuity and with good refractive and keratometric stability, without accelerated CEC loss. In Study #3, fourteen repeat DMEK eyes of 13 patients were retrospectively analysed. There was a statistically significant improvement in mean postoperative BCVA compared with vision before repeat DMEK [1.30 (0.65) logMAR before repeat DMEK versus 0.09 (0.26) after repeat DMEK; p = 0.003). There were no cases of PGF after repeat DMEK, and three regrafts (21.4%) had SGF (IR in two eyes; LEGF in one eye). No statistically differences were observed between the interest group and the control group in terms of postoperative BCVA during the first postoperative year (p = 0.647), proportions of eyes reaching BCVA ≤ 0.10 logMAR at 1-year F-U (p = 0.793), rebubbling (p = 0.669) or IR (14.3% of eyes in repeat DMEK group versus 5.6% in the control group; p = 0.580). In Study #4, four cases of complex eyes with failed DSAEK grafts were analysed. Before DMEK, BCVA ranged from 1.30 to 1.85 logMAR. In all cases, DMEK was surgically feasible. In the early postoperative period, one eye (25%) required rebubbling and ended in PGF. In the remaining three eyes, no postoperative complications were observed and at last F-U observation (range: 18-78 months) BCVA ranged from 0.12-0.40 logMAR with clear corneas and CEC loss ranged from 39 to 61%. The “addendum” case report was a complex eye with post-LASIK IFS which underwent successful DMEK surgery, improving BCVA from 1.60 to 0.15 logMAR at 6-month F-U. In Study #5, we analysed 14 eyes of 13 patients (estimated rate of vs-CMO = 4.8%). One eye had preexisting CMO which aggravated after DMEK; in the remaining cases, median time to detection of vs-CMO was 4 (3-10) weeks, with all cases detected within the first postoperative 6 months. In eyes with resolved CMO after treatment (n = 12 eyes), first-line treatment was effective in 83.3% of eyes, and two eyes resolved after intravitreal corticosteroid. Median BCVA improved after treatment from 0.60 (0.40-0.80) logMAR to 0.30 (0.15-0.40) logMAR (p = 0.002), and median CRT improved from 582.5 (400-655) μm to 278 (258-294) μm (p = 0.005). During the recruitment period for Study #6, four eyes (4 patients) were analysed. Two eyes had hydrophilic acrylic IOLs, and two eyes had “hybrid” IOLs (hydrophobic and hydrophilic monomers). Time between DMEK surgery and recognition of IOLop ranged between 2 and 44 weeks, and IOL exchange was performed between 11 and 24 months after DMEK. At the study visit, median postoperative F-U after IOL exchange was 45 (35.5-86.8) months. All corneas were clear with functioning grafts. Median BCVA improved from 0.25 (0.19- 0.41) logMAR to 0 (0.00-0.10) logMAR (p = 0.041) and CEC loss ranged between 57.7 and exchange was performed between 11 and 24 months after DMEK. At the study visit, median postoperative F-U after IOL exchange was 45 (35.5-86.8) months. All corneas were clear with functioning grafts. Median BCVA improved from 0.25 (0.19-0.41) logMAR to 0 (0.00-0.10) logMAR (p = 0.041) and CEC loss ranged between 57.7 and 85.3%. Study #8 analysed 16 published articles (n = 36 eyes, of which 6 were DMEK eyes) and found: 1) IR after SARS-CoV- 2 vaccination was rare; 2) well-known risk factors for IR were present in several cases, including previous corneal grafts (38.9%), history of IR (11.1%), no topical immunosuppressants (20.7%); 3) complete anatomical resolution occurred in 77.8% of eyes after management, but that 4 eyes (11.1%) had partial resolution and 5 eyes (13.9%) ended in graft failure; 4) 83.3% of the DMEK eyes had complete anatomical and functional resolution after management, and no cases evolved to graft failure. Conclusions: The findings of the studies included in this Doctoral Thesis can be summarized as follows: DMEK results in rapid visual recovery which is continuous over time. Younger patient age, higher preoperative BCVA, FECD, absence of retinal comorbidities and absence of postoperative complications are predictive of better postoperative BCVA. The risk of PGF in eyes undergoing primary DMEK is increased in cases with more severe CE failure or worse preoperative BCVA, cases with corneal comorbid conditions, intraoperative complications and in cases requiring rebubbling. Eyes with more severe CE failure, previous corneal grafts, history of glaucoma and IR episodes after DMEK had a significantly higher risk of SGF after DMEK. DMEK demonstrated to be feasible, effective and safe in complex eyes with CE failure, including eyes that undergo DMEK due to PIOL-related corneal decompensation, eyes with complex anatomy or low visual potential and eyes with previous EK grafts. In this regard, post-LASIK IFS is a novel indication for DMEK surgery, but on the reverse side of the coin, a failed or detached EK graft can be a cause of IFS in eyes with previous LASIK. Postoperative vs-CMO and IOLop with significant impact in vision are uncommon complications and their correct management is associated with good visual and graft outcomes. However, IOL explantation can cause increased CEC loss and pose a risk for LEGF. Surface ablation LVC is a potentially safe and effective option in eyes with residual ametropia after DMEK, but caution is recommended before considering keratorefractive surgery in these cases. DMEK allograft IR episodes after SARSCoV- 2 immunisation consist only of sporadic reports, and the prognosis with early recognition and prompt treatment was good in most eyes