A survey of Asian Eye Institutions on perioperative antibiotic prophylaxis in cataract surgery.

PURPOSE: To determine current institutional practice patterns for the use of perioperative antibiotics and other measures to prevent infection after cataract surgery in Asia. METHODS: An online survey-based study of leading eye institutions in China, Hong Kong, India, Indonesia, Japan, Malaysia, Pakistan, Philippines, Singapore, South Korea, Taiwan, Thailand and Vietnam was conducted. The survey was administered to 26 representative key opinion leaders from prominent tertiary eye institutions that are also national academic teaching institutions in Asia. Survey responses were collated and anonymized during analysis. RESULTS: All surveyed institutions used povidone iodine for the preoperative antiseptic preparation of the eye, with notable variations in the concentration of povidone iodine used for conjunctival sac instillation. Preoperative topical antibiotics were prescribed by 61.5% and 69.2% of institutions in low-risk and high-risk cases, respectively. Regarding the use of intra-operative antibiotics, 60.0% and 66.7% of institutions administered intracameral antibiotics in low-risk and high-risk patients, respectively. Postoperative topical antibiotics use patterns were generally very similar in low-risk and high-risk patients. Over half of the institutions (52.2% and 68.0% in low-risk and high-risk patients, respectively) also indicated prolonged postoperative use of topical antibiotics (> 2 weeks). Not all surveyed institutions had established policies/protocols for perioperative antibiotic use in cataract surgery, endophthalmitis surveillance, and/or a monitoring program for emerging antimicrobial resistance. CONCLUSION: There are variations in antimicrobial prophylaxis approaches to preoperative, intra-operative and postoperative regimens in cataract surgery in Asia. More evidence-based research is needed to support the development of detailed guidelines for perioperative antibiotic prophylaxis to reduce postoperative infections.

Establishing a Virtual Corneal Clinic: A Real-Time Teleophthalmology Approach.

PURPOSE: Keratoplasty patients require regular and timely follow-ups. During this COVID-19 pandemic which restricted global travel, we developed a novel real-time, hybrid teleophthalmology approach to comanage international keratoplasty patients between Singapore and Indonesia. METHODS: A retrospective consecutive observational study of 72 corneal patients (63 were postkeratoplasty) who attended a virtual corneal clinic (VCC) between June 2020 and April 2021 at JEC Eye Hospitals (JEC) in Jakarta, Indonesia. ZOOM Meeting software (Zoom Video Communication Inc, San Jose, CA) was used to simultaneously connect the Singapore corneal specialist at Eye & Cornea Surgeons (ECS), Singapore, using a real-time approach. Clinical examinations included full panels of video-linked corneal, glaucoma, and retinal imaging and investigations performed before real-time video-linked slit-lamp examination, with immediate clinical decision making between corneal specialists and patients. RESULTS: VCC enabled effective real-time clinical evaluation and collaborative clinical decisions, with full patient interaction, with the aim of maintenance of graft clarity, visual function, and management of comorbidities-a) topical and systemic medications were adjusted in 79.2% of patients; b) further referrals to glaucoma, retinal, and oculoplastic subspecialists were made in 16.6% of cases; c) additional adjunctive surgical procedures were performed at JEC in 6.9% cases; and d) government permission was obtained for 4 patients (5.6%) to fly to Singapore for urgent corneal surgery. CONCLUSIONS: The virtual corneal clinic is a novel real-time hybrid teleophthalmology approach which is effective in the comanagement of international keratoplasty patients and represents the advances in ophthalmic telemedicine.

Retinal photograph-based deep learning algorithms for myopia and a blockchain platform to facilitate artificial intelligence medical research: a retrospective multicohort study.

BACKGROUND: By 2050, almost 5 billion people globally are projected to have myopia, of whom 20% are likely to have high myopia with clinically significant risk of sight-threatening complications such as myopic macular degeneration. These are diagnoses that typically require specialist assessment or measurement with multiple unconnected pieces of equipment. Artificial intelligence (AI) approaches might be effective for risk stratification and to identify individuals at highest risk of visual loss. However, unresolved challenges for AI medical studies remain, including paucity of transparency, auditability, and traceability. METHODS: In this retrospective multicohort study, we developed and tested retinal photograph-based deep learning algorithms for detection of myopic macular degeneration and high myopia, using a total of 226 686 retinal images. First we trained and internally validated the algorithms on datasets from Singapore, and then externally tested them on datasets from China, Taiwan, India, Russia, and the UK. We also compared the performance of the deep learning algorithms against six human experts in the grading of a randomly selected dataset of 400 images from the external datasets. As proof of concept, we used a blockchain-based AI platform to demonstrate the real-world application of secure data transfer, model transfer, and model testing across three sites in Singapore and China. FINDINGS: The deep learning algorithms showed robust diagnostic performance with areas under the receiver operating characteristic curves [AUC] of 0·969 (95% CI 0·959-0·977) or higher for myopic macular degeneration and 0·913 (0·906-0·920) or higher for high myopia across the external testing datasets with available data. In the randomly selected dataset, the deep learning algorithms outperformed all six expert graders in detection of each condition (AUC of 0·978 [0·957-0·994] for myopic macular degeneration and 0·973 [0·941-0·995] for high myopia). We also successfully used blockchain technology for data transfer, model transfer, and model testing between sites and across two countries. INTERPRETATION: Deep learning algorithms can be effective tools for risk stratification and screening of myopic macular degeneration and high myopia among the large global population with myopia. The blockchain platform developed here could potentially serve as a trusted platform for performance testing of future AI models in medicine. FUNDING: None.

Descemet Membrane Endothelial Keratoplasty With a Pull-Through Insertion Device: Surgical Technique, Endothelial Cell Loss, and Early Clinical Results.

PURPOSE: To describe a surgical technique for Descemet membrane endothelial keratoplasty (DMEK) using a pull-through, endothelium-in insertion device, the DMEK EndoGlide. We evaluated the endothelial cell loss (ECL) associated with the EndoGlide-DMEK (E-DMEK) technique in both ex vivo and prospective clinical studies. METHODS: The ex vivo study involved calcein acetoxymethyl staining and preparation of DMEK grafts, which were trifolded endothelium-in, loaded into the EndoGlide, pulled through, and unfolded in imaging dishes. Inverted fluorescent microscopy was performed, and ECL was quantified using trainable segmentation software. The prospective clinical series describes the outcomes of consecutive surgeries using the E-DMEK technique. Grafts were pulled through the EndoGlide with forceps and unfolded in the anterior chamber endothelium-down. Our main outcome measure was ECL in both studies. RESULTS: In the ex vivo study with 9 human donor corneas, mean ECL was 15.2% ± 5.4% (n = 9). In our clinical series of 69 eyes, leading indications for surgery were pseudophakic/aphakic bullous keratopathy (47.8%), previous failed grafts (23.2%), and Fuchs endothelial dystrophy (18.8%). Rebubbling and primary graft failure rates related to E-DMEK were 11.6% and 1.5%, respectively. Among eyes with at least 6 months of follow-up, mean preoperative endothelial cell density was 2772 (range 2457-3448) cells/mm, and postoperative endothelial cell density was 1830 (range 541-2545) cells/mm. Mean ECL was 33.6% (range 7.5-80.4; n = 32) at the 7.1 (range 6-11) months follow-up. CONCLUSIONS: The ex vivo and pilot clinical studies suggest that E-DMEK shows acceptable rates of ECL, with safe and promising early clinical outcomes.

Outcomes of Repeat Anterior Lamellar Keratoplasty.

PURPOSE: To describe the outcomes and complications of repeat anterior lamellar keratoplasty (ALK) after a failed primary ALK. METHODS: This was a retrospective case series. Twenty-three eyes of 22 patients who underwent repeat ALK for optical indications from January 1, 1991, to December 31, 2017, were included. Clinical data were recorded from the Singapore Corneal Transplant Study database. Outcome measures were graft survival, best-corrected visual acuity (BCVA), and complications. RESULTS: Of the 780 cases of ALK performed for optical indications during the study period, 23 cases (2.9%) underwent repeat ALK. Primary ALK comprised of 16 deep ALK and 7 automated lamellar therapeutic keratoplasty cases. After repeat ALK, 21 eyes had further dissection down to the descemetic or predescemetic plane, whereas 2 eyes had graft exchange for failed automated lamellar therapeutic keratoplasty. The graft survival of repeat ALK was 93.2%, 86.1%, and 78.3% at 1, 3, and 5 years, respectively. There were 4 cases of graft failure with a mean time to failure of 2.0 ± 1.7 years. After the repeat ALK procedure, 63.6% and 66.7% of eyes achieved a BCVA of LogMAR +0.3 (20/40) or better at 1 and 2 years, respectively. Five eyes (17.4%) had an intraoperative microperforation. CONCLUSIONS: In the case of a failed ALK graft for a range of optical indications, a repeat ALK procedure offers a viable and safe surgical option with good long-term graft survival and visual outcomes, while avoiding the various disadvantages of performing penetrating keratoplasty, especially the risk of allograft endothelial rejection.

Lamellar Dissection Technique for Descemet Membrane Endothelial Keratoplasty Graft Preparation.

PURPOSE: To describe a novel lamellar dissection technique for Descemet membrane endothelial keratoplasty (DMEK) graft preparation, and to evaluate the rate of endothelial cell loss (ECL) and graft preparation failure associated with this technique. METHODS: We conducted an ex vivo laboratory-based study comparing ECL between the lamellar dissection and peeling techniques. Eight pairs of human donor corneas underwent calcein acetoxymethyl staining-all right eyes underwent the peeling technique and all left eyes underwent the lamellar dissection technique. ECL was quantified by image analysis with trainable segmentation software and compared between groups. We also conducted a retrospective analysis of 161 consecutive DMEK graft preparations by a single surgeon using the lamellar dissection technique from 2010 to 2018. Data on donor characteristics and graft preparation failures were obtained. RESULTS: Baseline donor characteristics were comparable in both arms of the laboratory-based study. Mean (SD) ECL with the lamellar dissection and peeling techniques was 13.8% (4.2%) and 11.2% (6.1%), respectively. There was no significant difference between the two (P = 0.327). In the clinical series, there were 2 graft preparation failures in 161 cases (1.2%). Among cases performed on diabetic donor tissue, the rate of graft preparation failure was 4.7%. CONCLUSIONS: The lamellar dissection technique has a similar rate of ECL compared with the peeling technique for DMEK graft preparation. This technique also has a low rate of graft preparation failure and may be a useful technique for diabetic donor tissue.

Cytomegalovirus Corneal Endotheliitis After Descemet Membrane Endothelial Keratoplasty.

PURPOSE: To describe 4 cases of cytomegalovirus (CMV) corneal endotheliitis occurring after Descemet membrane endothelial keratoplasty (DMEK). METHODS: This is a retrospective, interventional case series. Case records of 4 patients (one eye each) diagnosed with CMV corneal endotheliitis after DMEK were reviewed retrospectively. Presenting clinical features, treatment, and outcomes were examined. RESULTS: Patients' age ranged from 68 to 77 years. Three patients underwent DMEK for failed corneal grafts and 1 for pseudophakic bullous keratopathy. Time from DMEK to presentation ranged from 5 to 15 weeks. Presenting features included corneal edema, pigmented keratic precipitates, mild anterior chamber inflammation, and raised intraocular pressure. Two cases were initially misdiagnosed as graft rejection and treated with corticosteroids. Both worsened, and delayed diagnoses of CMV corneal endotheliitis were made. The two other cases were diagnosed correctly at initial presentation. All cases were confirmed by anterior chamber paracentesis and polymerase chain reaction testing for CMV. All cases were treated initially with topical ganciclovir gel and oral valganciclovir. Three cases showed clinical resolution. Final corrected visual acuity ranged from 20/25 to 20/40. One case failed to respond to topical ganciclovir gel, oral valganciclovir, and intravenous ganciclovir and foscarnet. This patient declined further systemic treatment and was maintained on topical ganciclovir only but subsequently showed spontaneous resolution 3 months later. CONCLUSION: CMV corneal endotheliitis may occur after DMEK and can closely mimic graft rejection. Polymerase chain reaction testing of aqueous humor is advised for suspect cases. A high index of suspicion is important for prompt diagnosis and initiation of appropriate antiviral treatment.

The Asia Cornea Society Infectious Keratitis Study: A Prospective Multicenter Study of Infectious Keratitis in Asia.

PURPOSE: To survey the demographics, risk factors, microbiology, and outcomes for infectious keratitis in Asia. DESIGN: Prospective, nonrandomized clinical study. METHODS: Thirteen study centers and 30 sub-centers recruited consecutive subjects over 12-18 months, and performed standardized data collection. A microbiological protocol standardized the processing and reporting of all isolates. Treatment of the infectious keratitis was decided by the managing ophthalmologist. Subjects were observed for up to 6 months. Main outcome measures were final visual acuity and the need for surgery during infection. RESULTS: A total of 6626 eyes of 6563 subjects were studied. The majority of subjects were male (n = 3992). Trauma (n = 2279, 34.7%) and contact lens wear (n = 704, 10.7%) were the commonest risk factors. Overall, bacterial keratitis was diagnosed in 2521 eyes (38.0%) and fungal keratitis in 2166 eyes (32.7%). Of the 2831 microorganisms isolated, the most common were Fusarium species (n = 518, 18.3%), Pseudomonas aeruginosa (n = 302, 10.7%), and Aspergillus flavus (n = 236, 8.3%). Cornea transplantation was performed in 628 eyes to manage ongoing infection, but 289 grafts (46%) had failed by the end of the study. Moderate visual impairment (Snellen vision less than 20/60) was documented in 3478 eyes (53.6%). CONCLUSION: Demographic and risk factors for infection vary by country, but infections occur predominantly in male subjects and are frequently related to trauma. Overall, a similar percentage of bacterial and fungal infections were diagnosed in this study. Visual recovery after infectious keratitis is guarded, and corneal transplantation for active infection is associated with a high failure rate.

Canine and Premolar Root Dimensions in Chinese. A Reference for Osteoodontokeratoprosthesis Surgery.

PURPOSE: Osteoodontokeratoprosthesis (OOKP) surgery is used to restore vision in end-stage corneal disorders, where an autogenous tooth supporting an optical cylinder is implanted through the cornea under a buccal mucosal graft. The ideal tooth for OOKP is a healthy single-rooted permanent tooth with sufficient buccolingual/palatal root diameter to accommodate an optical cylinder. The aim of this study was to determine the buccolingual/palatal diameters of canine and premolar roots in Chinese, for selection of teeth for OOKP surgery. DESIGN: This was an anatomical study on root dimensions of extracted intact teeth. METHODS: Extracted canine and premolar teeth (excluding maxillary first premolars) were collected and the buccolingual/palatal and mesiodistal diameters of the root at the cervical line and at 2-mm intervals below the cervical line were measured with Vernier calipers. Other measurements included total tooth length, crown buccolingual/palatal diameter, and root length. Mean and minimum buccolingual/palatal root diameters were compiled for each 2-mm interval. RESULTS: A total of 415 extracted teeth (198 male, 217 female) were collected and measured. Recorded dimensions of keratoprostheses in 55 previous OOKP surgeries were used to establish acceptable lamina dimensions to ascertain root size adequacy. Premolars in Chinese female patients were undersized in a small minority. Minimal dimensions of teeth were insufficient if at 6 mm root level, the buccolingual/palatal width was less than 5 mm, or the mesiodistal width was less than 3 mm. This was noted in female mandibular first premolars (5.6%), maxillary second premolars (4.5%), and mandibular second premolars (1.5%). CONCLUSIONS: Canines have adequate dimensions for OOKP surgery. However, premolars in Chinese females may be undersized in a small minority.

Prophylactic Vancomycin Drops Reduce the Severity of Early Bacterial Keratitis in Keratoprosthesis.

BACKGROUND: Artificial cornea transplantation, keratoprosthesis, improves vision for patients at high risk of failure with human cadaveric cornea. However, post-operative infection can cause visual loss and implant extrusion in 3.2-17% of eyes. Long-term vancomycin drops are recommended following keratoprosthesis to prevent bacterial keratitis. Evidence, though, in support of this practice is poor. We investigated whether prophylactic vancomycin drops prevented bacterial keratitis in an animal keratoprosthesis model. METHODOLOGY: Twenty-three rabbits were assigned either to a prophylactic group (n = 13) that received vancomycin 1.4% drops 5 times/day from keratoprosthesis implantation to sacrifice, or a non-prophylactic group (n = 10) that received no drops. All rabbits had Staphylococcus aureus inoculation into the cornea at 7-12 days post-implantation and were sacrificed at predetermined time-points. Prophylactic and non-prophylactic groups were compared with slit-lamp photography (SLP), anterior segment optical coherence tomography (AS-OCT), and histology, immunohistochemistry and bacterial quantification of excised corneas. Corneal vancomycin pharmacokinetics were studied in 8 additional rabbits. RESULTS: On day 1 post-inoculation, the median SLP score and mean±SEM AS-OCT corneal thickness (CT) were greater in the non-prophylactic than the prophylactic group (11 vs. 1, p = 0.049 and 486.9±61.2 vs. 327.4±37.1 µm, p = 0.029 respectively). On days 2 and 4, SLP scores and CT were not significantly different. Immunohistochemistry showed a greater CD11b+ve/non-CD11b+ve cell ratio in the non-prophylactic group (1.45 vs. 0.71) on day 2. Bacterial counts were not significantly different between the two groups. Corneal vancomycin concentration (2.835±0.383 µg/ml) exceeded minimum inhibitory concentration (MIC) for Staphylococcus aureus only after 16 days of vancomycin drops. Two of 3 rabbits still developed infection despite bacterial inoculation after 16 days of prophylactic drops. CONCLUSIONS: Prophylactic vancomycin drops provided short-term benefit, but did not prevent infection. Achieving MIC in the cornea was not sufficient to prevent Staphylococcus aureus keratitis. Patients should continue to be counselled regarding the risk of infection following keratoprosthesis.

The outcomes of primary pediatric keratoplasty in Singapore.

PURPOSE: To evaluate the long-term corneal graft survival and risk factors for graft failure in pediatric eyes. DESIGN: Retrospective, interventional consecutive case series. METHODS: Unilateral eyes of 105 patients aged 16 years and below were included from the Singapore Corneal Transplant Study between April 4, 1991 and April 4, 2011. Corneal graft survival was calculated using Kaplan-Meier survival analysis, and survival distributions were compared using log-rank test. RESULTS: Mean recipient age was 8.38 ± 5.63 years (range 0.18-15.92 years). Mean follow-up time was 34.16 ± 39.10 months. Main diagnoses were corneal scar (22.9%), limbal dermoid (21.9%), anterior segment dysgenesis (15.2%), and keratoconus (14.3%). Forty-four eyes (41.9%) underwent penetrating keratoplasty (PK), 37 (35.2%) underwent anterior lamellar keratoplasty (ALK), 22 (21.0%) underwent lamellar corneal patch graft, and 2 (1.9%) underwent Descemet stripping automated endothelial keratoplasty (DSAEK). Kaplan-Meier survival rates for PK were 92.8% at 1 year, 88.9% at 2-4 years, and 80.9% at 5-16 years; survival rates for ALK were 88.0% at 1 year and 84.3% at 2-7 years; survival rates for corneal patch graft were 100% at 1-3 years and 90% at 4-10 years; these were not statistically significant (P = .362). Deep corneal vascularization (P = .012), preexisting active inflammation (P = .023), preexisting glaucoma drainage device (P = .023), and preexisting ocular surface disease (P = .037) were associated with reduced graft survival in a univariate analysis. CONCLUSIONS: We report good long-term graft survival following pediatric keratoplasty for various indications. Lamellar keratoplasty, when indicated, should be the procedure of choice in high-risk keratoplasties.

Descemet stripping automated endothelial keratoplasty with a graft insertion device: surgical technique and early clinical results.

PURPOSE: To present the surgical technique and the early clinical results of the EndoGlide, a graft insertion device for use during Descemet stripping automated endothelial keratoplasty (DSAEK). DESIGN: Prospective interventional case series. PATIENTS: Graft insertion with the EndoGlide was performed in 25 eyes of Asian patients with endothelial dysfunction suitable for DSAEK. Demographic and clinical details, best-corrected visual acuity, postoperative complications, and endothelial cell density (ECD) loss were documented. SURGICAL TECHNIQUE: The prepared graft is transferred onto a preparation base and pulled into the glide capsule with forceps. Drawing the graft into the capsule curls it into a "double-coil" configuration. The assembled EndoGlide is inserted into the eye, and the graft is pulled into the anterior chamber with forceps, where it uncoils in the correct orientation with minimal manipulation. RESULTS: Average age was 70 years (range 28-91) and the commonest diagnosis was pseudophakic bullous keratopathy (12 eyes). Preoperative donor ECD was 2957 ± 242 cells/mm(2), median graft diameter was 8.75 mm, and all surgeries were completed successfully by 2 surgeons. There was a short learning curve in loading of the graft into the EndoGlide and in uncoiling the graft in the recipient eye. No patient had graft dislocation or primary iatrogenic graft failure. At 6 months, the mean ECD was 2586 ± 338 cells/mm(2) and mean cell loss was 13.1% in 20 eyes. At 12 months, the mean ECD was 2575 ± 289 cells/mm(2) and mean cell loss was 15.6% in 10 eyes. CONCLUSIONS: Graft insertion with the EndoGlide is safe in DSAEK, with a short learning curve and a low endothelial cell loss at 6 and 12 months.