Purpose
This study aims to systematically review and perform a single-arm meta-analysis to comprehensively evaluate the effect of removal of intraluminal stent from the Paul glaucoma implant (PGI) on intraocular pressure (IOP) changes and to summarize its long-term efficacy and safety.
Clinical Relevance
Glaucoma remains one of the leading global causes of irreversible vision loss, necessitating effective surgical management for refractory cases. Glaucoma drainage devices are critical for these complex forms, and intraluminal stent removal from the PGI serves as a key strategy to titrate aqueous outflow and optimize long-term IOP control.
Methods
A systematic review was conducted by searching EMBASE, Medline, and CENTRAL. The meta-analysis ultimately included 5 observational studies. The Risk of Bias in Nonrandomized Studies of Interventions (ROBINS-I) tool was used to assess the quality of the included studies. Continuous outcomes (eg, postremoval stent IOP reduction) were analyzed using the mean difference (MD) with 95% confidence intervals (CI); dichotomous outcomes (eg, hypotony incidence) were evaluated using pooled proportion with 95% CI. All meta-analyses employed a random-effects model.
Results
Our meta-analysis included 5 studies, involving a total of 283 eyes. The meta-analysis revealed that removal of the stent leads to a significant and immediate IOP reduction. The IOP reduction (immediate postremoval stent IOP minus preremoval stent IOP) was significant, with a pooled MD of −9.09 mm Hg IOP (MD = 9.09; 95% CI = [−11.77, −6.41]). In the preremoval stent high-IOP (>21 mm Hg) subgroup, the IOP reduction is −11.82 (95% CI = [−10.14, −13.51]). This reduction was significantly greater than the preremoval stent low-IOP (<21 mm Hg) subgroup, whose reduction is −6.66 mm Hg (95% CI = [−5.55, −7.77]). Regarding safety, the pooled proportion of clinically hypotony was low and highly consistent ( I ² = 0.00%) at only 0.03 (95% CI = [0.00, 0.15]).
Conclusions
Intraluminal stent removal from the PGI is an effective and safe IOP-lowering intervention. Its IOP reduction effect is significantly more pronounced in patients with preremoval stent high IOP, providing crucial clinical guidance for managing persistently high IOP after PGI implantation.
INTRODUCTION
G laucoma remains one of the leading global causes of irreversible vision loss. , The first-line treatment for glaucoma primarily relies on medical or laser therapy. , Surgical intervention is rarely used as a primary treatment option. , However, once the medicine or laser is failed, surgery can serve as an alternative treatment. , Among surgical options, glaucoma drainage devices (GDDs) have become increasingly important for managing complex forms of glaucoma by lowering intraocular pressure (IOP). ,
The Paul glaucoma implant (PGI) represents a recently developed GDD for IOP control. A key feature of the PGI is the ability to insert a 6-0 or 7-0 polypropylene stent within the tube lumen. This intraluminal stent serves as a protective mechanism against early postoperative hypotony. However, once the stent is removed, this protective mechanism is lifted. This increases the likelihood of hypotony and related complications, including choroidal detachment and suprachoroidal hemorrhage. ,
While intraluminal stent removal is viewed as a key strategy for titrating aqueous outflow in eyes with suboptimal IOP, the reported clinical outcomes regarding its efficacy and safety are heterogeneous across published studies. ,,,, For instance, the immediate drop magnitude of IOP is highly variable. Mendel et al and Chan et al reported IOP decreases approximately 6 mm Hg after removal of stent in 24 hours. In contrast, Weber et al, Richardson et al, and Labay-Tejado et al reported mean IOP decreases of over 10 mm Hg. The safety profile is also diverse. Specific cohorts report no incidents of hypotony or the need for stent reinsertion following the procedure. ,, However, other large series cite clinically significant hypotony as the most common complication, sometimes requiring surgical intervention or reinsertion. ,
The effect of PGI stent removal in IOP reduction and safety remains unclear, due to the small number of participants and the conflicting conclusions. Therefore, this study aims to conduct a single-arm meta-analysis to synthesize available data on intraluminal stent removal from the PGI. This study aims to systematically evaluate the magnitude of IOP reduction at various follow-up periods, cf the differential effect in high- and low-IOP subgroups, and accurately quantify the safety outcomes, thereby providing robust, consolidated evidence to guide glaucoma surgeons in optimizing PGI postremoval stent care.
MATERIALS AND METHODS
study design and search strategy
This study complied with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodological and reporting requirements. This study conducted comprehensive electronic database searches in EMBASE, Medline, and the Cochrane Central Register of Controlled Trials (CENTRAL) databases from their inception up until September 2025. The search aimed to identify studies reporting outcomes following intraluminal stent removal after PGI surgery. The search terms and MeSH terms included combinations of: “Paul Glaucoma Implant,” “PGI,” “stent removal,” “ripcord,” “intraocular pressure,” and “IOP.” A detailed search strategy is provided in the Supplemental Table 1. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist is presented in Supplemental Table 2.
study selection
Database searches were imported into EndNote (version 21), where duplicate citations were removed. The remaining studies were screened by two reviewers (C.H.L. and H.L.T.) independently through evaluation of titles and abstracts. Potentially eligible articles were subsequently retrieved for full-text review and assessed according to predefined inclusion criteria: (1) observational studies or randomized controlled trials, (2) clinical studies reporting IOP outcomes following removal of an intraluminal stent or ripcord from the PGI, and (3) availability of mean IOP values with corresponding SDs at a minimum of one preremoval and one postremoval follow-up time point. Studies were excluded if they were editorials, case reports, reviews, commentaries, conference abstracts lacking full-text data, or did not provide quantitative IOP outcomes required for the meta-analysis.
data extraction and quality assessment
Two reviewers (C.H.L. and H.L.T.) extracted data independently. Extracted variables included study characteristics (author, published year, study design, sample size, intervention type, follow-up duration) and outcome measures. Disagreements were resolved by consensus with the senior author (W.T.Y.). The risk of bias was evaluated using the ROBINS-I tool for nonrandomized studies.
statistical analysis
Meta-analysis was conducted using the “meta” and “metafor” packages in R software (version 4.5.1). A random-effects model was applied to account for between-study variability. Since all included studies are single-arm studies, this study analyzed the change in IOP relative to the preremoval baseline (continuous outcome) and the incidence of complications (dichotomous outcomes). For continuous outcome measures, including the magnitude of IOP reduction (eg, postremoval stent IOP minus preremoval stent IOP), mean differences (MD) with 95% confidence intervals (CI) were calculated. As for dichotomous outcomes (eg, hypotony incidence), pooled proportions with 95% CI were used.
Heterogeneity was quantified using the I ² statistic. Values of 0% to 40% were considered minimal, 30% to 60% moderate, 50% to 90% substantial, and >75% considerable heterogeneity, according to the Cochrane Handbook. In this study, the high IOP group was defined as >21 mm Hg. Subgroup analyses were prespecified and performed based on the preremoval IOP status: high-IOP (IOP > 21 mm Hg) and low-IOP (IOP ≤ 21 mm Hg) subgroups, to assess the differential IOP reduction effect.
RESULTS
literature search
A total of 595 citations were retrieved from EMBASE, Medline, and CENTRAL databases ( Figure 1 ). Deduplication yielded 544 records for screening, and 515 were excluded based on title and abstract review. After evaluating 20 articles in full, 5 satisfied the inclusion criteria and were retained in the final dataset (Chan et al ; Mendel et al ; Richardson et al ; Weber et al ; and Labay-Tejado et al ) in the meta-analysis. A total of 283 eyes were included in the meta-analysis. The search results are presented in the PRISMA flow diagram ( Figure 1 ).
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. The PRISMA flow diagram summarizes the process of literature identification, screening, and study selection for randomized controlled trials and observational studies. It outlines the sequential steps used to determine study’s eligibility in the systematic review and meta-analysis. Ultimately, five studies met the inclusion criteria.
characteristics of included studies
Table outlines the primary baseline characteristics of the five studies that met the inclusion criteria. A total of 285 eyes from five eligible studies were evaluated in the pooled analysis. Collectively, these studies reported on patients who underwent PGI stent removal, with sample sizes ranging from 6 to 145 eyes and follow-up durations spanning from the immediate postremoval period to 36 months. The mean preremoval IOP values varied from 18.8 to 25.6 mm Hg, whereas the mean postremoval IOP values ranged from 6.1 to 15.6 mm Hg.
TABLE
Characteristics of Included Studies.
| Study | Y | Method | Country | Population | Number of Eyes (Preremoval) | Stent Removal Time | Mean Preremoval Stent IOP (mm Hg) | Mean Postremoval Stent IOP (mm Hg) |
|---|---|---|---|---|---|---|---|---|
| Labay-Tejado et al | 2024 | Observational, retrospective | Spain | 67.8 ± 17.6 y, with refractory glaucoma | 6 | 6.9 ± 5.6 mo (range 0.75-23 mo) | 25.6 ± 5.2 | 15.6 ± 3.6 |
| Richardson et al | 2024 | Observational, retrospective | United Kingdom | 45.8 ± 19.8 y, with uveitic glaucoma | 21 | 3 mo | 24.7 ± 7.0 | 14.3 ± 2.7 |
| Chan et al | 2025 | Observational, retrospective | Singapore | >18 y with any type of glaucoma | 43 | Not specified | 18.8 ± 3.6 | 12.4 ± 4.3 |
| Mendel et al | 2025 | Observational, retrospective | United Kingdom | >18 y with any type of glaucoma | 145 | 181.9 ± 131.4 d (range 19-631) | 19.5 ± 6.7 | 12.64 ± 6.1 |
| Weber et al | 2025 | Observational, retrospective | Germany | 63.2 (14-90 y) years, with any type of glaucoma | 70 | 4.07 mo (0-26 mo) | 23.93 (11-40) | 11.22 (2-23) |
The table includes the baseline characteristics of the five included studies, spanning diverse populations in Europe and Asia. Across all studies, postremoval stent IOP values were consistently lower than preremoval stent levels.
risk of bias in included studies
Risk of bias across the included observational studies was examined using the ROBINS-I instrument, as depicted in Figure 2 . Overall, five studies were classified to have a moderate overall risk of bias. ,,,,
Risk of Bias in Nonrandomized Studies of Interventions (ROBINS-I) tool assessment for included observational studies. This figure presents the risk of bias assessment for the observational studies included in the systematic review, evaluated using the Risk of Bias in Nonrandomized Studies of Interventions (ROBINS-I) tool.
Bias due to missing data was rated as low to moderate, reflecting inconsistent follow-up durations and unequal data completeness across participants, particularly at later time points. Finally, each study demonstrated a moderate risk for outcome reporting bias. The lack of prespecified statistical analysis plans contributed to uncertainty regarding selective reporting. As acknowledged in the Risk of Bias in Nonrandomized Studies of Interventions (ROBINS-I) guidance, retrospective designs rarely achieve a “low risk” rating in the domain of selective reporting, since predefined statistical analysis plans are typically unavailable and analytical choices are made after data collection.
efficacy outcomes-immediate iop reduction and subgroup analysis
The meta-analysis of continuous outcomes consistently demonstrated a significant reduction in IOP following intraluminal stent removal. The pooled immediate IOP reduction (immediate postremoval stent IOP minus preremoval stent IOP) across all included studies was −9.09 mm Hg (95% CI: [−11.77, −6.41]) ( Figure 3 ). The most critical finding was the immediate IOP reduction (immediate postremoval stent IOP minus preremoval stent IOP), which was highly dependent on the patient’s preremoval stent IOP status. The IOP reduction amount was significantly different between the preremoval stent IOP subgroups. For the pre-op high-IOP (>21 mm Hg) subgroup, the pooled MD in IOP reduction was −11.82 mm Hg (95% CI: [−13.51, −10.14]) ( Figure 3 ). This reduction showed minimal heterogeneity ( I ² = 0.00%), suggesting significant homogeneity in the magnitude of IOP lowering within this high-IOP cohort. For the preremoval stent low-IOP (≤21 mm Hg) subgroup, the pooled MD was −6.66 mm Hg (95% CI: [−7.77, −5.55]) ( Figure 3 ). The reduction was numerically smaller than that in the high-IOP group and exhibited high homogeneity ( I ² = 0.00%).
