Association between capsulectomy shunt revision and Ahmed glaucoma valve implant exposure——Shanghai Ahmed glaucoma valve implant exposure (AGVIE) study Authors: Feng Gao, Manman Jin, Xiaobo Yu, Junyi Chen, Yi Dai, Yuhong Chen, Jiajian Wang, Xueli Chen, Li Wang, Yuan Fang, Kaidi Wang, Tingting Liu, Ye Lin Kwan, Shaohong Qian, Xinghuai Sun, Xiangmei Kong Published: 2026 ABSTRACT AbstractBackgroundThe Ahmed Glaucoma Valve (AGV) is a drainage device commonly used in the treatment of refractory glaucoma. Implant exposure is a major cause of surgical failure. This study aimed to identify factors associated with AGV exposure.MethodsThis retrospective cohort study included all inpatients who underwent primary AGV implantation at the Eye & ENT Hospital of Fudan University between January 2015 and December 2024. Baseline demographic and surgical data were collected. Factors associated with AGV exposure were evaluated using univariate and multivariate logistic regression.ResultsIn this comprehensive cohort study, involving a total of 2,905 eyes, with participants having a mean age of 49.2 ± 18.0 years and 64% being male, we observed that AGV exposure occurred in 75 eyes, accounting for an incidence rate of 2.6% (95% CI, 2.0–3.2%). Capsulectomy shunt revision (CSR) was associated with AGV exposure in both univariable analysis (OR 3.72, 95% CI 1.73–7.98; P < 0.001) and multivariable analysis (OR 3.95, 95% CI 1.79–8.69; P < 0.001).ConclusionsThe crude incidence of AGV exposure was 2.6%. CSR was statistically associated with AGV exposure, although this finding should be interpreted cautiously given the retrospective design and methodological limitations. Careful surgical planning and close postoperative monitoring may therefore be warranted.Clinical trial registrationNot applicable.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12886-026-05060-9.Keywords: Ahmed glaucoma valve, Exposure, Incidence, Capsulectomy shunt revision Background Glaucoma, the leading cause of irreversible blindness worldwide, is projected to affect approximately 111.8 million individuals by 2040 [1]. When intraocular pressure (IOP) cannot be adequately controlled with medications or laser treatment, surgical intervention becomes necessary. Following the landmark Tube Versus Trabeculectomy study [2], glaucoma drainage devices (GDDs) have been increasingly utilized as first-line surgical options [3]. Although GDDs eliminate the bleb-related complications associated with trabeculectomy [4], they introduce specific device-related challenges. Among these, conjunctival erosion with tube or plate exposure is the most vision-threatening, as it predisposes patients to chronic inflammation, infection, and even endophthalmitis [5–7]. Therefore, a thorough understanding of the risk factors contributing to implant exposure is crucial to prevent this potentially devastating complication. Reported exposure rates for GDDs range from 1.1% to 7.7%, with these variations being influenced by factors such as device design, surgical technique, and patient-specific variables [8–12]. Previous studies have identified multiple factors associated with GDD exposure, including preoperative factors (e.g., sex [13, 14], age [13, 15], previous ocular surgeries [16, 17], neovascular glaucoma [18, 19], and diabetes [20]); intraoperative variables (e.g., concomitant ocular surgeries [16, 21], patch grafts [22], and implant location [23]); as well as postoperative influences (e.g., the use of topical steroids [24] and lubricants [16]). In clinical practice, postoperative management following GDD implantation may include not only medications and digital massage but also additional interventions, such as revision procedures to restore aqueous outflow [25]. However, the potential impact of postoperative interventional procedures has not been well established. In China, the Ahmed glaucoma valve (AGV) is the only device available for clinical use, yet data on the incidence and risk factors for Ahmed valve exposure in the Chinese population remain limited. Therefore, this study aimed to comprehensively evaluate factors associated with AGV exposure, including baseline characteristics, intraoperative variables, and postoperative interventions. A better understanding of these factors may help identify higher-risk eyes and inform strategies to reduce this potentially serious complication. Methods Study design and study population We conducted this study at the Eye, Ear, Nose, and Throat Hospital (Eye & ENT Hospital) of Fudan University, Shanghai. All study protocols adhered to the tenets of the Declaration of Helsinki and received approval from the Ethics Committee of the Eye & ENT Hospital, Fudan University (approval ID 2025079). As this was a retrospective study, the requirement for informed consent was waived by the Ethics Committee in accordance with national regulations; therefore, individual informed consent from participants was not required. This study initially considered all patients who underwent AGV implantation at the Eye & ENT Hospital between January 2015 and December 2024 as potentially eligible. Only patients with a minimum postoperative follow-up of 6 months were included. All surgeries were performed by experienced glaucoma specialists with more than 10 years of surgical experience in glaucoma management. The exclusion criteria were as follows: (1) primary AGV implantation performed at another institution; (2) a history of two or more AGV implantations prior to the occurrence of implant exposure; and (3) non-Han ethnicity. AGV exposure was defined as conjunctival erosion with subsequent exposure of any part of the device (the tube, the plate, or both). We screened a total of 4,471 AGV FP7 procedures. After applying the exclusion criteria, we excluded 444 eyes due to repeat implantation in the same eye, and 1,122 eyes due to primary surgery performed elsewhere or non-Han ethnicity. Consequently, the final study cohort comprised 2,905 eyes of 2,663 patients (Fig. 1). Study design flowcharts Data collection Patient medical records were reviewed to collect data on baseline clinical characteristics, intraoperative factors, and postoperative variables. Baseline data included demographics, systemic comorbidities (such as diabetes and hypertension), laterality, glaucoma subtype, and previous ocular surgery. Intraoperative variables encompassed combined surgery, operative technique, implantation quadrant, and antimetabolite use. The postoperative variable of interest was capsulectomy shunt revision (CSR), defined as a surgical intervention performed after the initial AGV implantation and before the occurrence of AGV exposure, involving surgical excision of the fibrotic capsule overlying the valve plate to restore aqueous outflow. The procedure was performed under standard surgical conditions. The conjunctiva overlying the tube was incised and dissected posteriorly. Tenon’s capsule was incised and meticulously dissected to expose the dense fibrotic capsule encasing the plate. The fibrous capsule was then opened, and free aqueous flow was observed intraoperatively, confirming adequate device patency and function. The periplate fibrotic capsule was excised as extensively as feasible to reduce aqueous outflow resistance, while carefully preserving adjacent normal tissues. The use of antimetabolites was at the discretion of the surgeon. Finally, Tenon’s capsule and the conjunctiva were anatomically closed in layers. Statistical analysis Continuous variables are reported as mean ± standard deviation and were compared using t-tests; categorical variables are reported as number (percentage) and were compared using χ² tests or Fisher’s exact tests, as appropriate. Crude incidence and 95% confidence intervals (CIs) were estimated using the Clopper–Pearson method. Univariable logistic regression was first used to evaluate candidate baseline, intraoperative, and postoperative variables. Variables included in the multivariable analysis were selected based on previous literature and clinical relevance. Considering the limited number of exposure events, a restricted set of clinically important variables was entered into the multivariable model, including age at implantation (< 30, 30–49, and ≥ 50 years), sex, diabetes, previous ocular surgery, combined surgery at AGV implantation, implantation site, antimetabolite use, and CSR. Hypertension, laterality, glaucoma subtype, and operative technique were not included in the primary model because of concerns regarding model stability and overfitting; these variables were subsequently evaluated in sensitivity analyses by adding them individually to the primary model. Multicollinearity was assessed using variance inflation factors (VIFs), with VIF > 5 considered suggestive of relevant multicollinearity. Additional sensitivity analyses included refitting the model without CSR, leave-one-covariate-out analyses, and modeling age as a continuous variable. All analyses were performed using R version 4.6.0 (R Foundation). A P-value of less than 0.05 (two-sided) was considered to indicate statistical significance. Results A total of 2,905 eyes from 2,663 patients who underwent AGV implantation between January 2015 and December 2024 were consecutively enrolled in this study. The mean age at surgery was 49.2 ± 18.0 years (range, 0.16–90 years), and 64.0% of patients were male. A total of 75 eyes developed implant exposure following AGV implantation, corresponding to a crude incidence of 2.6% over the study period. Among these cases, the mean follow-up duration was 31.3 ± 23.1 months, and the mean time to exposure was 16.4 ± 18.1 months (median, 10.2 months; range, 0.1–90.9 months). Exposure sites were distributed as follows: tube exposure in 41 cases, plate exposure in 21 cases, and exposure of unspecified location in 13 cases due to incomplete documentation. Subgroup analysis demonstrated variation in incidence across age groups, with proportions of 3.3% in patients < 30 years, 1.8% in those aged 30–49 years, 3.0% in those aged 50–59 years, 2.9% in those aged 60–69 years, and 2.4% in those aged 70–90 years (Supplementary Table 1). Eyes with non-superotemporal implantation sites showed a higher proportion of exposure compared to superotemporal placement (3.9% [95% CI, 1.6–7.8%] vs. 2.5% [95% CI, 1.9–3.2%]). Similarly, eyes that underwent CSR demonstrated a higher proportion of exposure than eyes without CSR (8.3% [95% CI, 3.7–15.8%] vs. 2.4% [95% CI, 1.9–3.0%]). However, these comparisons were unadjusted and should be interpreted with caution, as differences in follow-up duration and potential confounding factors were not accounted for. The mean age of patients who experienced AGV exposure was 48.4 ± 20.0 years, and 60.0% were male. Among these patients, 24% had diabetes mellitus, and 31% had hypertension. A majority (65%) had a history of prior ocular surgery. Neovascular glaucoma was the most common underlying diagnosis in both the exposure and non-exposure groups (24% vs. 33%, P = 0.59), followed by primary open-angle glaucoma (17% vs. 14%, P = 0.59). A small proportion of patients (9%) underwent combined surgery at the time of AGV implantation. The most commonly used surgical technique was the self-scleral flap technique (64% vs. 62%, P = 0.94), followed by the short scleral tunnel technique (20% vs. 22%, P = 0.94). The majority of AGV devices were implanted in the superotemporal quadrant (91% vs. 94%, P = 0.26). The use of antimetabolites was similar between groups, with mitomycin C applied in 71% of eyes in the exposure group and 73% in the non-exposure group, while 5-fluorouracil used in 15% of eyes in both groups (P = 0.82). Between-group comparisons revealed no significant differences except for CSR, which was significantly more frequent in the exposure group (11% vs. 3%, P < 0.001) (Table 1). Table 1Participant characteristics, intraoperative factors and postoperative factors associated with AGV exposure in the retrospective cohort studyVariableEyes With AGV Exposure (n = 75)Eyes Without AGV Exposure (n = 2830)P ValueAge at time of implantation, y, mean ± SD48.4 ± 20.048.7 ± 18.40.91Sex, n (%)0.51 Male45 (60%)1804 (64%) Female30 (40%)1026 (36%)Systemic disease, n (%) Diabetes18 (24%)779 (28%)0.50 Hypertension23 (31%)854 (30%)0.93Eye laterality, n (%)0.16 Right33 (44%)1476 (52%) Left42 (56%)1354 (48%)Type of glaucoma, n (%)0.59 POAG13 (17%)407 (14%) PACG3 (4%)209 (7%) PCG9 (12%)306 (11%) NVG18 (24%)945 (33%) Post-vitreoretinal surgery glaucoma7 (9%)232 (8%) Post-traumatic glaucoma6 (8%)204 (7%) Uveitic glaucoma8 (11%)217 (8%) Others11 (15%)310 (11%)Previous ocular surgeries49 (65%)1637 (58%)0.20Combined surgery, n (%)7 (9%)232 (8%)0.72Operative technique0.94 Self-sclera flap48 (64%)1744 (62%) Long scleral tunnel5 (7%)233 (8%) Short scleral tunnel15 (20%)609 (22%) Human donor sclera flap7 (9%)244 (9%)Site of implant0.26 Superotemporal68 (91%)2656 (94%) Non-superotemporal7 (9%)174 (6%)Use of antimetabolites0.82 MMC53 (71%)2060 (73%) 5-FU11 (15%)424 (15%) None11 (15%)346 (12%)Capsulectomy shunt revision, n (%)8 (11%)88 (3%)< 0.001*AGV, Ahmed glaucoma valve; POAG, Primary open-angle glaucoma; PACG, Primary angle-closure glaucoma; PCG, Primary congenital glaucoma; NVG, Neovascular glaucoma; MMC, mitomycin C; 5-FU, 5 Fluorouracil*Significant results Participant characteristics, intraoperative factors and postoperative factors associated with AGV exposure in the retrospective cohort study AGV, Ahmed glaucoma valve; POAG, Primary open-angle glaucoma; PACG, Primary angle-closure glaucoma; PCG, Primary congenital glaucoma; NVG, Neovascular glaucoma; MMC, mitomycin C; 5-FU, 5 Fluorouracil *Significant results Logistic regression analysis was performed to evaluate factors associated with AGV exposure. CSR was significantly associated with AGV exposure in univariate analysis (OR 3.72, 95% CI 1.73–7.98; P < 0.001) and remained independently associated with exposure in multivariate analysis (OR 3.95, 95% CI 1.79–8.69; P < 0.001). Other variables, including age, sex, diabetes, previous ocular surgery, combined surgery at the time of AGV implantation, implantation site, and use of antimetabolites, were not significantly associated with AGV exposure in this analysis (Table 2). Table 2Univariate and multivariate logistic analysis of risk factors for AGV exposure in the retrospective cohort studyVariablesUnivariate AnalysisMultivariate AnalysisOR (95%CI) P OR (95%CI) P Age at time of implantation <301.01.0 >=30, < 500.52 (0.25, 1.07)0.0740.55 (0.26, 1.14)0.107 >=500.84 (0.46, 1.52)0.5570.99 (0.52, 1.85)0.967Sex (Female vs. male)1.17 (0.73, 1.87)0.5061.09 (0.68, 1.76)0.714Diabetes (yes vs.no)0.83 (0.49, 1.42)0.5000.86 (0.49, 1.50)0.585Hypertension (yes vs.no)1.02 (0.62, 1.68)0.927--Left eye1.39 (0.87, 2.20)0.165--Type of glaucoma POAG1.0-- PACG0.45 (0.13, 1.59)0.216-- PCG0.92 (0.39, 2.18)0.851-- NVG0.60 (0.29, 1.23)0.161-- Post-vitreoretinal surgery glaucoma0.94 (0.37, 2.40)0.905-- Post-traumatic glaucoma0.92 (0.34, 2.46)0.869-- Uveitic glaucoma1.15 (0.47, 2.83)0.754-- Others1.11 (0.49, 2.51)0.801--Previous ocular surgeries (yes vs.no)1.37 (0.85, 2.22)0.1961.26 (0.76, 2.08)0.368Combined surgery (yes vs.no)1.15 (0.52, 2.54)0.7241.17 (0.52, 2.61)0.710Operative technique Self-sclera flap1.0-- Long scleral tunnel0.78 (0.31, 1.98)0.600-- Short scleral tunnel0.89 (0.50, 1.61)0.711-- Human donor sclera flap1.04 (0.47, 2.33)0.920--Site of implant (No-Sup.temp. vs. Sup.temp.)1.57 (0.71, 3.47)0.2641.53 (0.67, 3.46)0.310Use of antimetabolites MMC1.01.0 5-FU1.01 (0.52, 1.95)0.9800.92 (0.47, 1.80)0.809 None1.24 (0.64, 2.39)0.5291.10 (0.57, 2.16)0.771Capsulectomy shunt revision (yes vs.no)3.72 (1.73, 7.98)< 0.001*3.95 (1.79, 8.69)< 0.001*AGV, Ahmed glaucoma valve; POAG, Primary open-angle glaucoma; PACG, Primary angle-closure glaucoma; PCG, Primary congenital glaucoma; NVG, Neovascular glaucoma; MMC, mitomycin C; 5-FU, 5 Fluorouracil. Variables included in the multivariable analysis were selected based on previous literature and clinical relevance. Hypertension, laterality, glaucoma subtype, and operative technique were excluded from the primary multivariable model due to limited exposure events and concerns about model stability; these variables were assessed individually in sensitivity analyses*Significant results Univariate and multivariate logistic analysis of risk factors for AGV exposure in the retrospective cohort study AGV, Ahmed glaucoma valve; POAG, Primary open-angle glaucoma; PACG, Primary angle-closure glaucoma; PCG, Primary congenital glaucoma; NVG, Neovascular glaucoma; MMC, mitomycin C; 5-FU, 5 Fluorouracil. Variables included in the multivariable analysis were selected based on previous literature and clinical relevance. Hypertension, laterality, glaucoma subtype, and operative technique were excluded from the primary multivariable model due to limited exposure events and concerns about model stability; these variables were assessed individually in sensitivity analyses *Significant results CSR had been performed in 8 of the 75 eyes with AGV exposure (11%) and in 88 of the 2,830 eyes without exposure (3%). In the AGV exposure group, the mean interval from AGV implantation to CSR was 23.5 months (median, 11.5 months; range, 1.7–76.7 months), while the mean interval from CSR to AGV exposure was 12.1 months (median, 7.6 months; range, 0.9–28.4 months). No relevant multicollinearity was detected in the primary multivariable model, with all VIFs below 5 and a maximum VIF of 2.22 (Supplementary Table 2). In a sensitivity model excluding CSR, the direction and statistical interpretation of the other covariates were not materially changed, suggesting internal stability of the covariate estimates (Supplementary Table 3). The estimated association between CSR and AGV exposure also remained stable after removal of individual covariate groups from the primary model or addition of excluded variables individually, including hypertension, laterality, glaucoma subtype, and operative technique (CSR OR range, 3.65–4.03). Similar results were observed when age was modeled as a continuous variable rather than as a categorical variable (OR 3.76, 95% CI 1.72–8.24; P < 0.001) (Supplementary Table 4). Discussion In this large retrospective cohort study of 2,905 eyes from a high-volume tertiary glaucoma center conducted over a 10-year period, the crude incidence of AGV exposure was 2.6%. CSR was statistically associated with AGV exposure after adjustment for multiple perioperative factors. This finding highlights a clinically actionable postoperative factor that has been underexplored in previous literature. However, interpretation of this association requires caution, as AGV exposure is inherently time-dependent and standardized longitudinal follow-up was incomplete across the cohort. Accordingly, the observed OR of 3.95 should be interpreted as a non–time-adjusted measure of association rather than a definitive estimate of effect magnitude. With respect to incidence, the observed exposure rate of 2.6% in our cohort is consistent with previous reports. Stewart et al. [12] reported an exposure rate of 2.7% with a mean follow-up of approximately 26 months, while Budenz et al. [ 26] observed a rate of 2.9% over five years. The similarity of these estimates across different populations and surgical settings supports the external validity of our findings. The observed association between CSR and implant exposure is both biologically and surgically plausible. Repeated conjunctival manipulation during revision procedures may lead to cumulative microvascular damage, resulting in localized ischemia and impaired wound healing. In addition, the use of adjunctive antimetabolites, such as mitomycin C, may further suppress fibroblast activity and contribute to progressive conjunctival thinning [27–29]. From a surgical perspective, disruption of the established fibrous capsule and repeated tissue dissection may destabilize the implant–tissue interface, thereby increasing susceptibility to mechanical erosion over time [10, 28, 30, 31]. Interestingly, tube exposure was more frequently observed than plate exposure in our cohort. One possible explanation is the surgical approach used during CSR, in which the conjunctival incision is typically created over the tube and extended posteriorly, potentially leading to greater conjunctival disruption in this region. However, given the retrospective design and incomplete data regarding exposure location, this observation should be considered hypothesis-generating rather than mechanistic evidence and warrants further investigation in prospective studies. From a clinical perspective, the findings highlight the need to weigh the risks and benefits of postoperative revision procedures carefully. CSR may be necessary for managing refractory ocular hypertension or device dysfunction. In non-urgent cases, management often follows a stepwise approach, including digital massage, bleb needling with 5-fluorouracil, IOP-lowering medications, and other less invasive strategies before surgical revision is considered. When surgical revision is unavoidable, efforts should be made to minimize the risk of subsequent implant exposure. Practical considerations include careful selection of the revision indication, gentle tissue handling, reinforcement with scleral or fascial patch grafts when appropriate, cautious use of antimetabolites, and closer postoperative surveillance for early conjunctival thinning or erosion. These measures are based on surgical reasoning and the present observational findings and require further prospective validation. Several additional limitations should be acknowledged. Residual confounding remains possible in this retrospective study. CSR may function as a surrogate marker for more severe or refractory glaucoma, compromised conjunctival integrity, or a greater burden of postoperative interventions. Surgical heterogeneity may also have influenced outcomes, as factors such as tissue handling, antimetabolite application, and the extent of conjunctival dissection during CSR were not fully captured. In addition, the relatively small number of CSR cases may have affected the precision of the estimated association. The extreme age range of the cohort may have also influenced estimates from age-stratified analyses, particularly due to the pediatric subgroup. Furthermore, exclusion of non-Han patients may limit the generalizability of the findings. Prospective studies with standardized follow-up intervals, complete time-to-event data, detailed postoperative treatment records, and more granular documentation of conjunctival status and exposure location are needed to confirm and refine these findings. Conclusions In conclusion, in this large retrospective cohort, the crude incidence of AGV exposure was 2.6%. CSR was significantly associated with AGV exposure and may indicate a higher-risk subgroup of patients. However, because complete standardized follow-up was unavailable and the CSR estimate was based on sparse events, the OR should be interpreted as a non-time-adjusted association rather than a causal effect. Careful surgical planning and vigilant postoperative monitoring remain important when postoperative revision procedures are considered. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Acknowledgements We thank Dr. Yue Zhou from Nantong First People’s Hospital, Jiangsu Province, for assistance with verification of the statistical analyses. Abbreviations Glaucoma drainage device Ahmed glaucoma valve Intraocular pressure Capsulectomy shunt revision Primary open-angle glaucoma Primary angle-closure glaucoma Primary congenital glaucoma Neovascular glaucoma Mitomycin C 5 Fluorouracil Author contributions Conception: Xiangmei Kong; Methodology: Feng Gao; Data collection: Manman Jin, Xiaobo Yu, JunYi Chen, Yi Dai, Yuhong Chen, Jiajian Wang, Xueli Chen, Li Wang, Yuan Fang, Kaidi Wang, Tingting Liu, Shaohong Qian; Statistical analysis: Feng Gao, Manman Jin; Manuscript drafting: Feng Gao, Ye Lin Kwan; Manuscript reviewing and editing: Xiangmei Kong; Funding acquisition: Xiangmei Kong, Xinghuai Sun, Yuhong Chen, Feng Gao; Supervision: Xiangmei Kong, Xinghuai Sun. F.G. and M-M.J. contributed equally to this work. All authors read and approved the final manuscript. Funding The project was supported by the National Natural Science Foundation of China (Grant No. 82070957, No.81800820, No.82371056, No. 82030027, No.81790641); National Key Research and Development Program of China (2020YFA0112700); Clinical Research Plan of SHDC (SHDC2020CR6029). The sponsor or funding organization had no role in the design or conduct of this research. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate All study protocols adhered to the tenets of the Declaration of Helsinki and received approval from the ethics committee of the Eye & ENT Hospital, Fudan University (approval ID 2025079). As this was a retrospective study, informed consent from the patients was not required. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Feng Gao and Manman Jin contributed equally to this work. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.