Predictive biomarkers for the prognosis of phacoemulsification and posterior chamber intraocular lens implantation in Fuchs endothelial corneal dystrophy Authors: Young In Yun, Chang Ho Yoon, Joo Youn Oh, Hyuk Jin Choi, Mee Kum Kim Published: 2025 ABSTRACT BackgroundTo investigate the clinical outcomes and related risk factors following phacoemulsification and posterior chamber intraocular lens implantation (PE & PCL) in East Asian patients with Fuchs endothelial corneal dystrophy (FECD).MethodsOne hundred eighty eyes of 112 FECD patients who underwent PE & PCL at a tertiary hospital in Korea were included. Demographic and ocular biometric characteristics including best-corrected visual acuity (BCVA), central corneal thickness (CCT), and endothelial cell density (ECD) were collected. Risk factors were analyzed for (1) Descemet membrane endothelial keratoplasty (DMEK) requirement and (2) postoperative BCVA < 0.5, using univariate and multivariate regression analyses and receiver operating characteristic curves.ResultsMean LogMAR BCVA improved from 0.51 to 0.20, CCT increased by 2.6% from baseline, and ECD decreased by 14.6% (all P < 0.05). Ten eyes (5.6%) underwent subsequent DMEK within three months. Baseline CCT was a risk factor for the need for DMEK (odds ratio 1.58, P < 0.05) with a cutoff value of 620 μm, achieving 100% sensitivity and 91% specificity. Baseline BCVA, CCT, and ECD predicted limited postoperative BCVA (all P < 0.05), with cutoff values of 0.35 logMAR (91% sensitivity and 59% specificity), 584 μm (57% sensitivity and 82% specificity), and 1244 cells/mm2 (65% sensitivity and 77% specificity), respectively.ConclusionsIn our cohort of Korean FECD eyes that underwent routine PE & PCL, progression to DMEK in the early postoperative period was relatively low. Baseline CCT values provide strong predictive power for both the necessity of DMEK and the likelihood of visual improvement after surgery.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12886-025-04521-x. Introduction Fuchs endothelial corneal dystrophy (FECD), the most common bilateral corneal dystrophy worldwide, is an endothelial disorder characterized by the presence of guttae, which are drop-like excrescences on the Descemet membrane [1]. It leads to a progressive loss of the endothelium pump functions, resulting in corneal edema, decreased vision, and blindness in later stages [2]. The prevalence of FECD varies significantly across geographic regions and ethnicities, with higher frequencies observed among Caucasians and inhabitants of Iceland, while it is relatively uncommon among Singaporean-Chinese populations and rare in East Asian countries, including Korea and Japan [3–7]. These variations may be attributed to genetic differences in genes linked to FECD such as TCF4, TCF8, LOXHD1, and SLC5A11, as well as environmental factors [8–13]. In the presence of a cataract in FECD patients, the decision to perform phacoemulsification and posterior chamber intraocular lens implantation (PE & PCL) poses a challenge, as reduced vision may stem from the cataract, corneal edema related to FECD, or a combination of both [14]. Additionally, phacoemulsification has been shown to reduce endothelial cell count by 8.5% to 16% in normal eyes, and significantly accelerate the rate of endothelial cell loss in FECD patients [7, 15, 16]. Historically, penetrating keratoplasty was considered the gold standard for definitive treatment of corneal decompensation in FECD. However, advancements in surgical techniques, particularly Descemet membrane endothelial keratoplasty (DMEK), now offer quicker visual recovery with reduced risks of complications such as bleeding, infection, astigmatism, corneal denervation, and tissue rejection [17, 18]. Despite these advancements, eye banks struggle to meet the global demand for corneal donor tissue, especially in many Asian countries, leaving a substantial number of patients without access to transplants [19, 20]. A survey by Gain et al. [19]. revealed approximately 12.7 million people awaiting corneal transplants worldwide, highlighting a significant mismatch between supply and demand, with a staggering ratio of 1:70. Consequently, the decision to proceed with PE & PCL is approached with great caution, particularly in regions facing severe donor cornea shortages. Factors such as low endothelial cell count, stromal edema, guttae, and epithelial microcystic changes have been associated with an increased risk of corneal decompensation following cataract surgery [21–23]. Specifically, pachymetry measurements exceeding 640 μm and endothelial cell density below 1000/mm2 have been used to predict corneal decompensation. However, the advancement of phacoemulsification techniques from longitudinal to torsional methods, as well as improvement in phacoemulsification machines, has enhanced efficiency and minimized endothelial cell damage, potentially altering the prognosis [24, 25]. Moreover, much of the existing literature on this topic has primarily focused on Western countries, with less understanding of its implications in Asian countries where the incidence of FECD is lower and the ocular dimension is smaller. Ethnic and geographic variations in FECD prevalence suggest that biomarkers for visual prognosis and risk factors for DMEK progression require separate evaluation across different populations [26]. These genetic and environmental differences indicate potential variability in the outcomes of PE & PCL procedures among Asian FECD patients. Currently, there is insufficient evidence to inform patients and surgeons about the visual prognosis of performing PE & PCL in FECD patients, as well as the prevalence of progression to DMEK and its associated risk factors in the East Asian population. This study aims to investigate the clinical outcomes following PE & PCL in patients with FECD and to identify biomarkers that could predict the need for DMEK or limited visual improvement following PE & PCL in the East Asian patients. Methods The present study was a retrospective, longitudinal, observational study, approved by the Institutional Review Board (IRB) of Seoul National University Hospital (IRB No. 2308-131-1459). The study followed the tenets of the Declaration of Helsinki (1964), and written informed consent was waived due to the retrospective design and absence of any additional medical intervention. Study subjects Inclusion criteria comprised patients with FECD who underwent PE & PCL from January 1, 2006 to December 31, 2023 at the Department of Ophthalmology in Seoul National University Hospital, a tertiary hospital in Seoul, Korea. Medical records of 221 eyes from 143 consecutive FECD patients who underwent PE & PCL were reviewed. All patients had FECD grade 2 or higher, as determined by a modified Krachmer scale, characterized by more than 12 central or paracentral nonconfluent corneal guttae. Both eyes with predominantly nonconfluent guttae and eyes with confluent central guttae were eligible for inclusion. Eyes with previous or concurrent ocular surgeries, or those with any other coexisting ocular pathology affecting vision, such as amblyopia, anterior corneal dystrophy, glaucoma, exudative age-related macular degeneration and retinal vascular disease, were excluded. Eyes with frank, visually disabling corneal decompensation judged to require immediate endothelial keratoplasty were not scheduled for standalone PE & PCL and were therefore not included. However, eyes with only minimal stromal edema and increased CCT, without clinically severe epithelial edema or bullae, were considered eligible for PE & PCL if cataract was thought to be visually significant and the patient elected a cataract-first approach. At 1 month postoperative, the decision to proceed to DMEK was made based on the patient’s request and the presence of persistent corneal stromal edema with suboptimal vision function despite cataract surgery, after excluding any other concurrent ocular conditions that might impair the patient’s vision. In total, 10 eyes of 10 patients underwent DMEK within 3 months after PE & PCL. Surgical procedure All PE & PCL procedures were performed by four experienced anterior segment surgeons, each with more than 10 years of experience in cataract surgery. PE & PCL was performed using a standard technique of 2.75 mm corneal incision with torsional phacoemulsifier (Infinity or Centurion Cataract Surgery System (Alcon Laboratories, Fort Worth, TX, USA)) under topical anesthesia and/or monitored anesthesia care. In all cases, a monofocal aspheric single-piece intraocular lens (IOL) was placed inside the capsular bag without complication. Postoperatively, topical 0.1% fluorometholone acetate (Flarex; Novartis, Basel, Switzerland) and 1.5% levofloxacin eye drops (Cravit; Santen, Osaka, Japan) were administered four times daily for 1 month. The patients were followed up at 1 day, 1 week and 1 month after the surgery. Data collection Demographic information including age, gender, and the presence of underlying disease, including diabetes mellitus and hypertension, were gathered. Preoperative baseline variables were measured, including best-corrected visual acuity (BCVA), cataract severity, central corneal thickness (CCT), endothelial cell density (ECD), coefficient of variation (CV) and hexagonality, axial length (AL), white-to-white distance (WTW), lens thickness (LT), and anterior chamber depth (ACD). BCVA was tested using a 5-m Hahn’s visual acuity chart mounted on a light box under standard illumination conditions, after correcting for any refractive error measured by autokeratometry (KR-8900; Topcon, Tokyo, Japan). The decimal equivalent was determined for each chart reading by identifying the last line where the majority of letters were read correctly, and then converted to a logMAR value. Cataract severity was graded on a scale of 1 to 6 based on nuclear color using the Lens Opacities Classification System (LOCS) III under slit-lamp microscope. Optical biometry, including AL, WTW, LT and ACD, was determined using the IOLMaster 500 (Zeiss, Jena, Germany) before 2016 and the IOLMaster 700 (Zeiss) from 2016 onwards. The CCT was measured using the IOLMaster 700 (Zeiss); prior to its introduction, ultrasound pachymeter (Pocket II; Quantel Medical, Clermont-Ferrand, France) was used. Central ECD, CV and hexagonality were measured using a specular microscope (Noncon ROBO CA; Konan Medical, Hyogo, Japan) by a trained examiner using a standardized protocol. Because confluent guttae in FECD can cause segmentation failure and overcounting with fully automated specular microscopy, automated outputs were reviewed and, when necessary, corrected by manual border tracing in analyzable areas adjacent to confluent central guttae. Measurements were accepted only if at least 100 contiguous cells with clearly defined borders could be identified. Frames with poor reflection, marked distortion, or extensive confluent guttae were excluded and reacquired at the same visit. When adequate central images could not be obtained, a peripheral ECD reading was attempted; however, only central ECD was used in the main analyses. All baseline measurements, including slit-lamp examination and ocular biometric characteristics, were collected within 1 month prior to PE & PCL. In a subset of eyes, preoperative anterior segment tomography was also reviewed to assess for subclinical corneal edema. A total of 125 eyes had analyzable preoperative tomography with reliable pachymetry and posterior elevation data obtained from CASIA II swept-source anterior segment OCT, Orbscan scanning-slit/Placido, or Galilei dual-Scheimpflug/Placido systems. Eyes in which posterior corneal elevation could not be reliably assessed were excluded from this tomographic analysis. For each of these 125 eyes, three previously described tomographic features of subclinical edema in Fuchs endothelial corneal dystrophy were graded: (1) loss of parallel isopachs, (2) displacement of the thinnest point, and (3) focal posterior corneal surface depression. Each eye was then categorized according to the number of features present (0, 1, 2, or all 3) [27]. These tomographic data were collected to explore whether baseline subclinical edema was associated with early postoperative corneal decompensation. Postoperative BCVA was assessed at 1 week and 1 month following PE & PCL. CCT, ECD, CV and hexagonality were measured at 1 month follow-up after PE & PCL. Additionally, slit-lamp examinations were performed at both postoperative visits to detect any complications. Retinal examination, including ocular coherence tomography, was conducted at the 1-month visit in order to rule out macular pathologies such as cystoid macular edema. No significant complications, aside from corneal edema, were observed in any of the cases. Assessment and statistical analyses To assess postoperative outcomes, BCVA at 1 week and 1 month postoperatively and CCT, ECD, CV, and hexagonality at 1 month postoperatively were compared with baseline BCVA, CCT, ECD, CV, and hexagonality. The percent change for CCT and ECD from baseline was calculated using the following formula: ((Postoperative - Baseline) / Baseline) x 100%. Because some patients contributed both eyes, these longitudinal comparisons were analyzed using eye-level linear mixed-effects models with a patient-level random intercept, and P values were obtained using Kenward–Roger small-sample correction. To evaluate possible risk factors for DMEK after PE & PCL, eyes were divided into two groups based on the progression to DMEK within 3 months following PE & PCL. Patient and ocular characteristics were compared between the two groups. Because some patients contributed both eyes, between-group comparisons were performed using patient-clustered models: linear regression with cluster-robust covariance for continuous variables, logistic regression with cluster-robust covariance for binary variables, and proportional-odds regression with cluster-robust covariance for the ordinal nuclear color grade. To identify predictors of (1) early DMEK after PE & PCL and (2) limited visual outcome worse than 0.30 logMAR at 1 month, univariate and multivariate logistic regression models were fit using generalized estimating equations with a binomial family and logit link, specifying patient ID as the clustering unit. Factors with P < 0.05 in univariate models were included in the multivariable models. For predictors that remained significant in multivariable analysis, receiver operating characteristic (ROC) curves were generated, and the area under the curve (AUC), optimal cutoff (Youden index, maximizing sensitivity + specificity − 1), sensitivity, specificity, and 95% confidence intervals were estimated using patient-level cluster bootstrap resampling that retained both eyes from the same patient [28]. Data analysis was performed using the Statistical Package for Social Sciences (version 26.0; IBM Corp, Armonk, NY, United States) and R software (R version 4.4.1., available at: http://www.r-project.org; accessed June 2024). Except where stated otherwise, the data are presented as the mean ± standard deviation, and the level of statistical significance was set at P < 0.05. Results Demographic characteristics The present study included 180 eyes of 112 subjects, with an average age of 68.55 ± 8.03 years at the time of PE & PCL (Table 1). 26% of these patients were male, and 74% were female. More than 95% of the patients had a nuclear color grading of 4 or lower as determined by the LOCS III. Baseline BCVA was 0.51 ± 0.38 logMAR. Average CCT and ECD were 568.60 ± 51.96 μm and 1806.3 ± 784.4 cells/mm2, respectively. Central ECD was measurable in 123 eyes, while in 57 eyes, central ECD measurements were unfeasible due to advanced progression of FECD. Among 180 eyes, 78 eyes (43.3%) showed primarily nonconfluent guttae, whereas 102 eyes (56.7%) demonstrated confluent central guttae at the time of cataract surgery. Table 1Baseline characteristics N Mean ± SD or N (%)Age, years18068.55 ± 8.03Sex, male/female18037/143 (25.6/74.4)Underlying disease180Diabetes26 (14.4)Hypertension70 (38.9)Nuclear color grading by LOCS III180NC145 (25.0)NC239 (21.7)NC352 (28.9)NC436 (20.0)NC55 (2.8)NC63 (1.7)Baseline BCVA, logMAR1800.51 ± 0.38Baseline AL, mm18023.70 ± 1.80Baseline WTW, mm13011.59 ± 0.45Baseline LT, mm1304.64 ± 0.44Baseline ACD, mm1532.92 ± 0.45Baseline CCT, µm168568.60 ± 51.96Baseline ECD, cells/mm21621806.3 ± 784.4Baseline CV, %16139.1 ± 13.9Baseline hexagonality, %16150.5 ± 11.6LOCS, Lens Opacity Classification System; NC, nuclear color; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variationN denotes the number of eyes with analyzable baseline data for each parameter. For ECD/CV/hexagonality, central specular images were ungradable in some eyes due to confluent guttae/poor reflection Table 1Baseline characteristics N Mean ± SD or N (%)Age, years18068.55 ± 8.03Sex, male/female18037/143 (25.6/74.4)Underlying disease180Diabetes26 (14.4)Hypertension70 (38.9)Nuclear color grading by LOCS III180NC145 (25.0)NC239 (21.7)NC352 (28.9)NC436 (20.0)NC55 (2.8)NC63 (1.7)Baseline BCVA, logMAR1800.51 ± 0.38Baseline AL, mm18023.70 ± 1.80Baseline WTW, mm13011.59 ± 0.45Baseline LT, mm1304.64 ± 0.44Baseline ACD, mm1532.92 ± 0.45Baseline CCT, µm168568.60 ± 51.96Baseline ECD, cells/mm21621806.3 ± 784.4Baseline CV, %16139.1 ± 13.9Baseline hexagonality, %16150.5 ± 11.6LOCS, Lens Opacity Classification System; NC, nuclear color; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variationN denotes the number of eyes with analyzable baseline data for each parameter. For ECD/CV/hexagonality, central specular images were ungradable in some eyes due to confluent guttae/poor reflection Baseline characteristics LOCS, Lens Opacity Classification System; NC, nuclear color; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation N denotes the number of eyes with analyzable baseline data for each parameter. For ECD/CV/hexagonality, central specular images were ungradable in some eyes due to confluent guttae/poor reflection Postoperative changes Postoperative changes of BCVA, CCT, ECD, CV and hexagonality are shown in Table 2. BCVA showed marked improvement at both 1 week and 1 month post-PE & PCL, with mean logMAR values improving from 0.51 ± 0.38 to 0.20 ± 0.33 and 0.13 ± 0.22, respectively (P < 0.001 for both time points). CCT increased by an average of 2.6% from a baseline of 567.9 ± 48.05 μm to 583.3 ± 66.6 μm at 1 month postoperatively (P < 0.001). ECD demonstrated a significant decrease of 14.6%, dropping from 1949.2 ± 742.4 cells/mm² at baseline to 1663.4 ± 797.5 cells/mm² at 1 month postoperatively (P < 0.001). The CV showed no significant postoperative change, remaining relatively stable (38.5 ± 14.9% to 38.3 ± 9.3%, P = 0.89). Similarly, the change in hexagonality was not statistically significant, decreasing slightly from 51.7 ± 11.5% to 49.7 ± 12.1% (P = 0.09). Table 2Postoperative changes of best-corrected visual acuity, central corneal thickness, endothelial cell density, coefficient of variation, and hexagonality in eyes with Fuchs endothelial corneal dystrophy N BaselinePostoperative*% Change from baselineP value†1-week BCVA, logMAR1800.51 ± 0.380.20 ± 0.33-< 0.0011-month BCVA, logMAR1720.51 ± 0.380.13 ± 0.22-< 0.001CCT, µm130567.9 ± 48.05583.3 ± 66.62.6 ± 4.7< 0.001ECD, cells/mm21231949.2 ± 742.41663.4 ± 797.5-14.6 ± 21.1< 0.001CV, %12038.5 ± 14.938.3 ± 9.3-0.890Hexagonality, %12051.7 ± 11.549.7 ± 12.1-0.094Data are presented as mean ± standard deviation*Post-phacoemulsification and posterior chamber intraocular lens implantation†P values from eye-level linear mixed-effects models with patient-level random intercepts and Kenward–Roger small-sample correctionBCVA, best-corrected visual acuity; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variationN denotes the number of eyes with paired measurements at baseline and 1 month postoperative. For ECD/CV/hexagonality, central specular images were ungradable in some due to confluent guttae/poor reflection Table 2Postoperative changes of best-corrected visual acuity, central corneal thickness, endothelial cell density, coefficient of variation, and hexagonality in eyes with Fuchs endothelial corneal dystrophy N BaselinePostoperative*% Change from baselineP value†1-week BCVA, logMAR1800.51 ± 0.380.20 ± 0.33-< 0.0011-month BCVA, logMAR1720.51 ± 0.380.13 ± 0.22-< 0.001CCT, µm130567.9 ± 48.05583.3 ± 66.62.6 ± 4.7< 0.001ECD, cells/mm21231949.2 ± 742.41663.4 ± 797.5-14.6 ± 21.1< 0.001CV, %12038.5 ± 14.938.3 ± 9.3-0.890Hexagonality, %12051.7 ± 11.549.7 ± 12.1-0.094Data are presented as mean ± standard deviation*Post-phacoemulsification and posterior chamber intraocular lens implantation†P values from eye-level linear mixed-effects models with patient-level random intercepts and Kenward–Roger small-sample correctionBCVA, best-corrected visual acuity; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variationN denotes the number of eyes with paired measurements at baseline and 1 month postoperative. For ECD/CV/hexagonality, central specular images were ungradable in some due to confluent guttae/poor reflection Postoperative changes of best-corrected visual acuity, central corneal thickness, endothelial cell density, coefficient of variation, and hexagonality in eyes with Fuchs endothelial corneal dystrophy Data are presented as mean ± standard deviation *Post-phacoemulsification and posterior chamber intraocular lens implantation †P values from eye-level linear mixed-effects models with patient-level random intercepts and Kenward–Roger small-sample correction BCVA, best-corrected visual acuity; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation N denotes the number of eyes with paired measurements at baseline and 1 month postoperative. For ECD/CV/hexagonality, central specular images were ungradable in some due to confluent guttae/poor reflection Patient and ocular characteristics based on DMEK progression At 1 month postoperative, the decision to proceed with DMEK was made for 10 eyes (5.6%) of 10 patients, while the remaining patients did not undergo any additional surgery. All 10 eyes (100%) in the DMEK (+) group had confluent central guttae preoperatively, whereas among the 170 eyes that did not proceed to early DMEK, 92 eyes (54.1%) showed confluent central guttae at baseline. A comparison of baseline characteristics between these groups is presented in Table 3. Patients in the DMEK (+) group were significantly older than those in the DMEK (-) group, with 74.21 ± 7.54 years in the former and 68.22 ± 7.96 years in the latter (P = 0.014). No difference was found in the distribution of sexes. Additionally, there was no difference in the proportion of patients with diabetes or hypertension. The distribution of nuclear color grading differed markedly between the two groups (P = 0.020). In the DMEK (-) group, a notably larger proportion of patients exhibited lower grading, whereas in the DMEK (+) group, patients were evenly distributed across nuclear color grade 2 to 6. Baseline BCVA was significantly worse in the DMEK (+) group (0.88 ± 0.40 logMAR) compared to the DMEK (-) group (0.48 ± 0.37 logMAR) (P = 0.002). Although baseline AL was shorter in the DMEK (+) group (22.77 ± 1.05 mm) compared to the DMEK (-) group (23.76 ± 1.82) (P = 0.008), there were no significant differences in baseline WTW, LT, and ACD between the two groups. However, baseline CCT was significantly thicker in the DMEK (+) group (681.10 ± 49.88 μm) compared to the DMEK (-) group (561.47 ± 41.99 μm) (P < 0.001). Additionally, baseline ECD was significantly lower in the DMEK (+) group (1027.5 ± 543.2 cells/mm2) compared to the DMEK (-) group (1826.0 ± 780.7 cells/mm2) (P = 0.002). No notable differences were observed in baseline CV and hexagonality between the two groups. Table 3Comparison of baseline characteristics based on the progression to Descemet membrane endothelial keratoplasty after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophyDMEK (-)DMEK (+)P valueMean ± SD N Mean ± SD N Age, years68.22 ± 7.9617074.21 ± 7.54100.014†Sex, male/female33/1371704/6100.143‡Underlying disease17010Diabetes2420.621‡Hypertension6640.942‡Nuclear color grading (LOCS III)170100.020§NC1450NC2372NC3502NC4342NC532NC612Baseline BCVA, logMAR0.48 ± 0.371700.88 ± 0.40100.002†Baseline AL, mm23.76 ± 1.8217022.77 ± 1.05100.008†Baseline WTW, mm11.60 ± 0.4312011.45 ± 0.63100.446†Baseline LT, mm4.63 ± 0.441214.78 ± 0.2990.169†Baseline ACD, mm2.93 ± 0.441432.72 ± 0.49100.187†Baseline CCT, µm561.47 ± 41.99158681.10 ± 49.8810< 0.001†Baseline ECD, cells/mm21826.0 ± 780.71581027.5 ± 543.240.002†Baseline CV, %39.1 ± 14.115737.3 ± 5.040.466†Baseline hexagonality, %50.7 ± 11.715742.8 ± 9.440.058†LOCS, Lens Opacity Classification System; NC, nuclear color; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation; DMEK, Descemet membrane endothelial keratoplasty†P values from subject-level cluster-robust linear regression (continuous variables)‡P values from subject-level cluster-robust logistic regression (binary variables)§P value from subject-level cluster-robust proportional-odds regression for the ordinal nuclear color gradeN denotes the number of eyes with analyzable baseline data for each parameter. For ECD/CV/hexagonality, central specular images were ungradable in some eyes due to confluent guttae/poor reflection Table 3Comparison of baseline characteristics based on the progression to Descemet membrane endothelial keratoplasty after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophyDMEK (-)DMEK (+)P valueMean ± SD N Mean ± SD N Age, years68.22 ± 7.9617074.21 ± 7.54100.014†Sex, male/female33/1371704/6100.143‡Underlying disease17010Diabetes2420.621‡Hypertension6640.942‡Nuclear color grading (LOCS III)170100.020§NC1450NC2372NC3502NC4342NC532NC612Baseline BCVA, logMAR0.48 ± 0.371700.88 ± 0.40100.002†Baseline AL, mm23.76 ± 1.8217022.77 ± 1.05100.008†Baseline WTW, mm11.60 ± 0.4312011.45 ± 0.63100.446†Baseline LT, mm4.63 ± 0.441214.78 ± 0.2990.169†Baseline ACD, mm2.93 ± 0.441432.72 ± 0.49100.187†Baseline CCT, µm561.47 ± 41.99158681.10 ± 49.8810< 0.001†Baseline ECD, cells/mm21826.0 ± 780.71581027.5 ± 543.240.002†Baseline CV, %39.1 ± 14.115737.3 ± 5.040.466†Baseline hexagonality, %50.7 ± 11.715742.8 ± 9.440.058†LOCS, Lens Opacity Classification System; NC, nuclear color; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation; DMEK, Descemet membrane endothelial keratoplasty†P values from subject-level cluster-robust linear regression (continuous variables)‡P values from subject-level cluster-robust logistic regression (binary variables)§P value from subject-level cluster-robust proportional-odds regression for the ordinal nuclear color gradeN denotes the number of eyes with analyzable baseline data for each parameter. For ECD/CV/hexagonality, central specular images were ungradable in some eyes due to confluent guttae/poor reflection Comparison of baseline characteristics based on the progression to Descemet membrane endothelial keratoplasty after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophy LOCS, Lens Opacity Classification System; NC, nuclear color; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation; DMEK, Descemet membrane endothelial keratoplasty †P values from subject-level cluster-robust linear regression (continuous variables) ‡P values from subject-level cluster-robust logistic regression (binary variables) §P value from subject-level cluster-robust proportional-odds regression for the ordinal nuclear color grade N denotes the number of eyes with analyzable baseline data for each parameter. For ECD/CV/hexagonality, central specular images were ungradable in some eyes due to confluent guttae/poor reflection An exploratory tomographic analysis was performed in a subset of 125 eyes with adequate preoperative anterior segment tomography. Each eye was categorized according to the number of subclinical edema features identified on pachymetry and posterior elevation maps (loss of parallel isopachs, displacement of the thinnest point, and focal posterior corneal surface depression). Of these 125 eyes, 57 eyes demonstrated all three features, 16 eyes demonstrated two features, 32 eyes demonstrated one feature, and 20 eyes demonstrated none. Early postoperative corneal decompensation requiring DMEK occurred only in eyes with all three features (9 of 57 eyes) and was not observed in eyes with two or fewer features (0 of 68 eyes combined). A summary of these findings is provided in Supplementary Table S1. Factors associated with postoperative DMEK progression or limited visual recovery Logistic regression analyses were performed to identify risk factors associated with either the progression to DMEK or postoperative BCVA worse than 0.30 logMAR, Among the 180 eyes with FECD that underwent PE & PCL, 10 eyes (5.6%) required DMEK within 3 months. Risk factors associated with the progression to DMEK and their odd ratios (OR) with 95% confidence intervals (CI) are presented in Table 4. In the univariate analysis, older age (OR 1.11, 95% CI 1.01–1.23, P = 0.035), higher nuclear color grading (OR 2.78, 95% CI 1.42–5.45, P = 0.003), worse baseline BCVA (OR 1.20, 95% CI 1.08–1.32, P < 0.001), thicker baseline CCT (OR 1.62, 95% CI 1.26–2.09 P < 0.001), and lower baseline ECD (OR 0.85, 95% CI 0.73–0.99, P = 0.036) were identified as factors increasing the odds of progression to DMEK. In the multivariate analysis, baseline CCT remained the sole significant risk factor (OR 1.58, 95% CI 1.19–2.10, P = 0.001). Additionally, a ROC curve was generated, and the AUC value was calculated to determine the optimal CCT threshold for predicting progression to DMEK. The AUC for baseline CCT was 0.975 (P < 0.001), indicating strong predictive capability. A cutoff CCT value of 620 μm was identified, achieving 100% sensitivity and 91% specificity (Fig. 1). Figure 2 shows the distribution of eyes that required or did not require DMEK after PE & PCL, according to preoperative CCT. Table 4Factors associated with the progression to Descemet membrane endothelial keratoplasty after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophyUnivariateMultivariateOR (95% CI)P value†OR (95% CI)P value†Age, years1.11 (1.01–1.23) 0.035 1.07 (0.80–1.45)0.636Sex, male2.77 (0.71–10.72)0.141NANADiabetes1.53 (0.29–8.01)0.613NANAHypertension1.06 (0.28–3.99)0.928NANANuclear color grading2.78 (1.42–5.45) 0.003 1.77 (0.47–6.74)0.400Baseline BCVA, logMAR (0.1)1.20 (1.08–1.32) < 0.001 1.06 (0.74–1.53)0.756Baseline AL, mm0.52 (0.25–1.09)0.085NANABaseline WTW, mm0.47 (0.07–3.13)0.432NANABaseline LT, mm2.21 (0.72–6.73)0.164NANABaseline ACD, mm0.33 (0.06–1.81)0.201NANABaseline CCT, µm (10)1.62 (1.26–2.09) < 0.001 1.58 (1.19–2.10) 0.001 Baseline ECD, cells/mm2 (100)0.85 (0.73–0.99) 0.036 0.85 (0.59–1.22)0.369Baseline CV, %0.98 (0.88-1.00)0.594NANABaseline hexagonality, %0.94 (0.88-1.00)0.068NANAOR, odds ratio; CI, confidence interval; NA, not included in the multivariable model; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation†Odds ratios, 95% confidence intervals, and P values were obtained from eye-level logistic regression models fit using generalized estimating equations (binomial family with logit link, exchangeable working correlation), clustering on patient ID to account for correlation between fellow eyes. The multivariable model included age, nuclear color grading, baseline BCVA (per 0.1 logMAR), baseline CCT (per 10 μm), and baseline ECD (per 100 cells/mm²) Table 4Factors associated with the progression to Descemet membrane endothelial keratoplasty after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophyUnivariateMultivariateOR (95% CI)P value†OR (95% CI)P value†Age, years1.11 (1.01–1.23) 0.035 1.07 (0.80–1.45)0.636Sex, male2.77 (0.71–10.72)0.141NANADiabetes1.53 (0.29–8.01)0.613NANAHypertension1.06 (0.28–3.99)0.928NANANuclear color grading2.78 (1.42–5.45) 0.003 1.77 (0.47–6.74)0.400Baseline BCVA, logMAR (0.1)1.20 (1.08–1.32) < 0.001 1.06 (0.74–1.53)0.756Baseline AL, mm0.52 (0.25–1.09)0.085NANABaseline WTW, mm0.47 (0.07–3.13)0.432NANABaseline LT, mm2.21 (0.72–6.73)0.164NANABaseline ACD, mm0.33 (0.06–1.81)0.201NANABaseline CCT, µm (10)1.62 (1.26–2.09) < 0.001 1.58 (1.19–2.10) 0.001 Baseline ECD, cells/mm2 (100)0.85 (0.73–0.99) 0.036 0.85 (0.59–1.22)0.369Baseline CV, %0.98 (0.88-1.00)0.594NANABaseline hexagonality, %0.94 (0.88-1.00)0.068NANAOR, odds ratio; CI, confidence interval; NA, not included in the multivariable model; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation†Odds ratios, 95% confidence intervals, and P values were obtained from eye-level logistic regression models fit using generalized estimating equations (binomial family with logit link, exchangeable working correlation), clustering on patient ID to account for correlation between fellow eyes. The multivariable model included age, nuclear color grading, baseline BCVA (per 0.1 logMAR), baseline CCT (per 10 μm), and baseline ECD (per 100 cells/mm²) Factors associated with the progression to Descemet membrane endothelial keratoplasty after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophy OR, odds ratio; CI, confidence interval; NA, not included in the multivariable model; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation †Odds ratios, 95% confidence intervals, and P values were obtained from eye-level logistic regression models fit using generalized estimating equations (binomial family with logit link, exchangeable working correlation), clustering on patient ID to account for correlation between fellow eyes. The multivariable model included age, nuclear color grading, baseline BCVA (per 0.1 logMAR), baseline CCT (per 10 μm), and baseline ECD (per 100 cells/mm²) Next, risk factors associated with postoperative BCVA worse than 0.30 logMAR were examined using both univariate and multivariate logistic regression analyses (Table 5). Out of 180 eyes, 33 eyes exhibited postoperative BCVA worse than 0.30 logMAR. In the univariate analysis, higher nuclear color grading (OR 1.95, 95% CI 1.35–2.80, P < 0.001), worse baseline BCVA (OR 1.25, 95% CI 1.13–1.38, P < 0.001), thicker baseline CCT (OR 1.38, 95% CI 1.22–1.56, P < 0.001), lower baseline ECD (OR 0.87, 95% CI 0.81–0.94, P < 0.001), lower baseline CV (OR 0.93, 95% CI 0.88–0.99, P = 0.031), and lower baseline hexagonality (OR 0.96, 95% CI 0.94–0.98, P = 0.001) showed associations with postoperative BCVA worse than 0.30 logMAR. In the multivariate analysis, worse baseline BCVA (OR 1.43, 95% CI 1.04–1.97, P = 0.026), thicker baseline CCT (OR 1.22, 95% CI 1.04–1.43, P = 0.013), and lower baseline ECD (OR 0.83, 95% CI 0.74–0.92, P < 0.001) were identified as independent risk factors for postoperative BCVA worse than 0.30 logMAR. ROC curves were generated for each risk factor. The AUC for baseline BCVA was 0.794 (P < 0.001), with a cutoff BCVA of 0.35 logMAR achieving 91% sensitivity and 59% specificity (Fig. 3A). For baseline CCT, the AUC was 0.748 (P < 0.001), with the optimal cutoff value of 583.5 μm with 57% sensitivity and 82% specificity (Fig. 3B). Baseline ECD showed an AUC of 0.714 (P = 0.020), with a cutoff value of 1244 cells/mm2 yielding 65% sensitivity and 77% specificity (Fig. 3C). Table 5Factors associated with 1-month postoperative best-corrected visual acuity worse than 0.3 LogMAR after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophyUnivariateMultivariateOR (95% CI)P value†OR (95% CI)P value†Age, years1.05 (0.98–1.12)0.137NANASex, male0.99 (0.35–2.77)0.980NANADiabetes0.59 (0.16–2.17)0.425NANAHypertension0.67 (0.28–1.59)0.361NANANuclear color grading1.95 (1.35–2.80) < 0.001 1.13 (0.59–2.14)0.714Baseline BCVA, logMAR (0.1)1.25 (1.13–1.38) < 0.001 1.43 (1.04–1.97) 0.026 Baseline AL, mm0.99 (0.77–1.28)0.952NANABaseline WTW, mm0.96 (0.32–2.89)0.938NANABaseline LT, mm1.58 (0.71–3.51)0.265NANABaseline ACD, mm0.56 (0.22–1.48)0.245NANABaseline CCT, µm (10)1.38 (1.22–1.56) < 0.001 1.22 (1.04–1.43) 0.013 Baseline ECD, cells/mm2 (100)0.87 (0.81–0.94) < 0.001 0.83 (0.74–0.92) < 0.001 Baseline CV, %0.93 (0.88–0.99) 0.031 0.90 (0.91–1.06)0.197Baseline hexagonality, %0.96 (0.94–0.98) 0.001 0.95 (0.90–1.01)0.108OR, odds ratio; CI, confidence interval; NA, not available; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation†Odds ratios, 95% confidence intervals, and P values were obtained from eye-level logistic regression models fit using generalized estimating equations (binomial family with logit link and exchangeable working correlation), clustering on patient ID to account for correlation between fellow eyes. The multivariable model included nuclear color grading, baseline BCVA (per 0.1 logMAR), baseline CCT (per 10 μm), baseline ECD (per 100 cells/mm²), baseline CV (%), and baseline hexagonality (%) Table 5Factors associated with 1-month postoperative best-corrected visual acuity worse than 0.3 LogMAR after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophyUnivariateMultivariateOR (95% CI)P value†OR (95% CI)P value†Age, years1.05 (0.98–1.12)0.137NANASex, male0.99 (0.35–2.77)0.980NANADiabetes0.59 (0.16–2.17)0.425NANAHypertension0.67 (0.28–1.59)0.361NANANuclear color grading1.95 (1.35–2.80) < 0.001 1.13 (0.59–2.14)0.714Baseline BCVA, logMAR (0.1)1.25 (1.13–1.38) < 0.001 1.43 (1.04–1.97) 0.026 Baseline AL, mm0.99 (0.77–1.28)0.952NANABaseline WTW, mm0.96 (0.32–2.89)0.938NANABaseline LT, mm1.58 (0.71–3.51)0.265NANABaseline ACD, mm0.56 (0.22–1.48)0.245NANABaseline CCT, µm (10)1.38 (1.22–1.56) < 0.001 1.22 (1.04–1.43) 0.013 Baseline ECD, cells/mm2 (100)0.87 (0.81–0.94) < 0.001 0.83 (0.74–0.92) < 0.001 Baseline CV, %0.93 (0.88–0.99) 0.031 0.90 (0.91–1.06)0.197Baseline hexagonality, %0.96 (0.94–0.98) 0.001 0.95 (0.90–1.01)0.108OR, odds ratio; CI, confidence interval; NA, not available; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation†Odds ratios, 95% confidence intervals, and P values were obtained from eye-level logistic regression models fit using generalized estimating equations (binomial family with logit link and exchangeable working correlation), clustering on patient ID to account for correlation between fellow eyes. The multivariable model included nuclear color grading, baseline BCVA (per 0.1 logMAR), baseline CCT (per 10 μm), baseline ECD (per 100 cells/mm²), baseline CV (%), and baseline hexagonality (%) Factors associated with 1-month postoperative best-corrected visual acuity worse than 0.3 LogMAR after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophy OR, odds ratio; CI, confidence interval; NA, not available; BCVA, best-corrected visual acuity; AL, axial length; WTW, white-to-white distance; LT, lens thickness; ACD, anterior chamber depth; CCT, central corneal thickness; ECD, endothelial cell density; CV, coefficient of variation †Odds ratios, 95% confidence intervals, and P values were obtained from eye-level logistic regression models fit using generalized estimating equations (binomial family with logit link and exchangeable working correlation), clustering on patient ID to account for correlation between fellow eyes. The multivariable model included nuclear color grading, baseline BCVA (per 0.1 logMAR), baseline CCT (per 10 μm), baseline ECD (per 100 cells/mm²), baseline CV (%), and baseline hexagonality (%) Fig. 1Receiver operating characteristic curve for the prediction of Descemet membrane endothelial keratopathy requirement within 3 months after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophy based on baseline central corneal thickness. Dotted circle on the curve represents the optimal cutoff point, determined by the Youden index. AUC: area under the curve; CCT: central corneal thickness Receiver operating characteristic curve for the prediction of Descemet membrane endothelial keratopathy requirement within 3 months after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophy based on baseline central corneal thickness. Dotted circle on the curve represents the optimal cutoff point, determined by the Youden index. AUC: area under the curve; CCT: central corneal thickness Fig. 2Histogram showing the distribution of eyes that required Descemet membrane endothelial keratoplasty (red) and those that did not (blue) after phacoemulsification and posterior chamber intraocular lens implantation, according to preoperative central corneal thickness. CCT: central corneal thickness; DMEK: Descemet membrane endothelial keratoplasty Histogram showing the distribution of eyes that required Descemet membrane endothelial keratoplasty (red) and those that did not (blue) after phacoemulsification and posterior chamber intraocular lens implantation, according to preoperative central corneal thickness. CCT: central corneal thickness; DMEK: Descemet membrane endothelial keratoplasty Fig. 3Receiver operating characteristic curves for the prediction of postoperative best-corrected visual acuity (BCVA) worse than logMAR 0.30 after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophy based on (A) baseline BCVA, (B) baseline central corneal thickness, and (C) baseline endothelial cell density. Each dotted circle on the curve represents the optimal cutoff point, determined by the Youden index. AUC: area under the curve; BCVA: best-corrected visual acuity; CCT: central corneal thickness; ECD: endothelial cell density Receiver operating characteristic curves for the prediction of postoperative best-corrected visual acuity (BCVA) worse than logMAR 0.30 after phacoemulsification and posterior chamber intraocular lens implantation in eyes with Fuchs endothelial corneal dystrophy based on (A) baseline BCVA, (B) baseline central corneal thickness, and (C) baseline endothelial cell density. Each dotted circle on the curve represents the optimal cutoff point, determined by the Youden index. AUC: area under the curve; BCVA: best-corrected visual acuity; CCT: central corneal thickness; ECD: endothelial cell density Discussion In this study, we investigated the prognosis of PE & PCL in Korean FECD patients, focusing on the risk of progression to DMEK and outcomes in visual acuity. Our findings reveal significant improvements in visual acuity at 1 week post-PE & PCL, with further enhancements noted at 1 month, consistent with previous studies on FECD [21, 29–31]. Notably, only 5.6% of FECD eyes in our cohort required DMEK within three months after PE & PCL, contrasting markedly with rates reported in American (10%) and European (33.3%) populations [21, 32]. This discrepancy could be attributed to various factors, including genetic differences, variations in surgical techniques, and the differing thresholds for undergoing DMEK based on the severity of corneal decompensation and patient preference [26, 33]. In Korea, where donor cornea availability is limited and the cost of surgery is high, patients may opt for DMEK at a lower visual acuity threshold. For instance, our DMEK patients underwent surgery at an average visual acuity of 1.04 logMAR, while DMEK was recommended in patients with visual acuity equal to or worse than 0.30 logMAR in the European study [32]. At 1 month post-PE & PCL, our study observed a 14.6% decrease in ECD and a 2.6% increase in CCT, which is comparable to a Japanese study that reported a 13.5% reduction in ECD following PE & PCL [34]. These ECD and CCT change rates align with previous literature on the effects of phacoemulsification in eyes without ocular pathology, where ECD loss is typically in the range of 10–16% and CCT increase is 1.8–3.2% [15, 16, 29, 34–39]. Therefore, in our cohort of Korean FECD patients, PE & PCL appears to affect endothelial cells similarly to its effects in normal eyes, while providing significant visual improvement and a relatively low risk of progression to DMEK. We further identified baseline CCT as a significant predictor of DMEK progression post-PE & PCL, with a threshold of 620 μm demonstrating 100% sensitivity and 91% specificity. Baseline BCVA, CCT, and ECD also emerged as critical risk factors for suboptimal postoperative vision, with respective cutoff values of 0.35 logMAR (91% sensitivity, 59% specificity), 584 μm (57% sensitivity, 82% specificity), and 1244 cells/mm² (65% sensitivity, 77% specificity). In 2005, Seitzman et al. [21]. recommended a CCT value of 640 μm as a criterion for increased risk of keratoplasty, advising a combined corneal transplant and cataract surgery based on the patient’s visual needs. However, their study focused on the necessity for keratoplasty within one year after cataract surgery, while our study examined the need for an immediate DMEK following cataract surgery, with the decision to perform DMEK made at 1 month postoperatively. Additionally, their study included 136 eyes with 5 undergoing keratoplasty, whereas our study involved a larger cohort of 180 eyes, with 10 undergoing DMEK, providing a more robust dataset. Furthermore, geographical and genetic differences may influence outcomes, as certain FECD genetic variations that are more susceptible to phacoemulsification might be more prevalent in the Korean population compared to the US. A study in Taiwan, which examined cataract surgery in patients with FECD and corneal decompensation, found no significant differences in baseline BCVA and CCT between the DMEK and non-DMEK groups [40]. However, ROC curve analysis indicated that a baseline BCVA of 0.44 logMAR could predict progression to DMEK with 82% sensitivity and 75% specificity. While this baseline BCVA cutoff is comparable to our predictive cutoff value for low postoperative vision, the small sample size of 16 patients limits the statistical power and generalizability of their findings. In a recent report by Arnalich-Montiel et al., a corneal thickness of 645 μm at the pupil center, measured by Pentacam, was found to predict progression to DMEK with 56% sensitivity and 95% specificity [32]. This higher cutoff may be due to Pentacam’s tendency to measure corneal thickness higher compared to ultrasound pachymetry [41, 42]. Additionally, setting the specificity at 95% lowers sensitivity and increases the cutoff value. Scheimpflug-based and related anterior segment imaging systems have been shown to detect localized stromal and posterior corneal alterations—such as loss of parallel isopachs, displacement of the thinnest point, and focal posterior surface depression—that are considered tomographic indicators of subclinical corneal edema before clinically evident edema is seen at the slit lamp [27]. In the present study, an exploratory subgroup analysis of 125 eyes with adequate preoperative tomography showed that early postoperative corneal decompensation requiring DMEK was observed only in eyes that demonstrated all three of these tomographic edema features preoperatively, and was not observed in eyes with two or fewer features (summarized in Supplementary Table S1). Although this analysis was restricted to a subset of the cohort and the number of events was small, these findings suggest that preoperative tomography may provide additional information for estimating the risk of early postoperative corneal decompensation in FECD eyes undergoing phacoemulsification, in addition to increased baseline CCT identified in our study. Importantly, our findings should not be interpreted to mean that cataract surgery alone is sufficient for all FECD eyes with guttae. Even before overt stromal edema is visible at the slit lamp, advanced FECD with confluent central guttae can already show pronounced endothelial failure and increased posterior corneal backscatter, including a hyperreflective “fibrillar layer” on Scheimpflug tomography [43, 44]. Such changes are associated with localized thickening of the posterior cornea and correspond to areas of severe endothelial cell loss in advanced FECD. These optical changes are linked to subjective visual disability — for example, glare, reduced quality of vision, and early-morning myopic shift — despite only moderate increases in central corneal thickness and sometimes relatively preserved Snellen acuity [43, 44]. Patients with this pattern of visually significant backscatter and confluent guttae may have limited visual potential even after cataract extraction alone and are often better served by early or combined endothelial keratoplasty. Preoperative posterior corneal backscatter and poorer preoperative visual acuity have both been shown to predict slower and less complete visual recovery after DMEK [45, 46]. The limitations of our study include its focus on a single ethnic group, the Korean population. While this approach enhances the specificity of our findings regarding FECD within this demographic, it may limit the generalizability of our results to other ethnicities with different genetic predispositions to FECD. Moreover, treatment decisions in our cohort reflect local resource constraints: in Korea, donor corneas for DMEK are almost exclusively imported tissue, availability is limited, and the procedure carries substantial financial and logistical burden. As a result, patients and surgeons often pursue a sequential strategy—initial cataract surgery, followed by DMEK only if there is clinically evident corneal edema, persistent visual dissatisfaction, and patient willingness to undergo keratoplasty. This practice pattern means that certain eyes with minimal preoperative edema were still managed with cataract surgery first, which contributes to the relatively high preoperative CCT values observed in the subgroup that later required early DMEK. This “cataract-first, DMEK-if-needed” approach may not generalize to healthcare systems with readily available donor tissue and lower procedural barriers, where earlier or combined keratoplasty may be favored. Secondly, due to the advanced stage of FECD in some patients, central ECD measurements were not feasible, which is a common challenge in studies involving endothelial dystrophies. In such cases, peripheral ECD measurements were attempted; however, because FECD typically begins centrally and progresses peripherally, peripheral counts tend to be higher and are less representative of central disease severity, and thus were not included in the primary analyses. Nonetheless, our data remains valuable for predicting DMEK progression and visual outcomes in cases where ECD measurements are obtainable. Also, our study did not prospectively quantify corneal backscatter, corneal densitometry, glare disability, or contrast sensitivity, which are known to drive patient dissatisfaction in FECD and can guide the timing of DMEK. Another potential limitation is the change in pachymetric device from ultrasound pachymetry to IOLMaster 700 during the long-term study, which could introduce bias. However, due to the rarity of FECD in Korea, conducting a long-term study was inevitable, and previous studies have demonstrated significant linear correlation and acceptable repeatability between these devices [47, 48]. In addition, we did not prospectively record a standardized Fuchs grade for each eye beyond documenting that all included eyes met at least grade ≥ 2 by a modified Krachmer definition. Furthermore, intraoperative factors such as surgeon experience, phacoemulsification settings, and choice of instruments varied depending on the surgeons in the study. However, the basic setting was generally standardized, and effect of these intraoperative factors on the clinical outcome appears to be minimal and within a tolerable range. Future research addressing these factors could provide further insights into optimizing outcomes in FECD patients undergoing PE & PCL. Cataract surgery in FECD patients poses unique challenges for anterior segment surgeons, particularly in regions with limited access to donor corneas. Risk stratification plays a crucial role in facilitating meaningful discussions with patients about personalized risks and managing expectations regarding potential unfavorable outcomes. It also aids in optimizing the allocation of donor corneas; if unavailable, it allows for the postponement of cataract surgery until a transplant become accessible. Our study underscores the effectiveness and relative safety of performing torsional PE & PCL in improving vision, with an average DMEK progression rate of 5.6%. However, it is important to note that eyes with a preoperative CCT value exceeding 620 μm are at a high risk of progression to DMEK. Additionally, patients with preoperative vision worse than 0.35 logMAR BCVA, CCT greater than 584 μm, or ECD less than 1244 cells/mm² should be informed about the potential limitations in visual improvement following PE & PCL. These findings enhance patient counseling and decision-making in managing FECD eyes with cataracts in resource-constrained settings. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1