Visit-handling methods significantly affect reported glaucoma surgery success rates
Key Takeaways
- Visit-handling methods significantly affect reported glaucoma surgery success rates, with the highest rates seen when the lowest IOP is used and the lowest rates when peak or all IOP values are considered.
- Intermediate success rates are observed when using mean, median, or closest IOP values, highlighting variability based on analytic approach.
- IOP-based success definitions show limited correlation with visual field outcomes, as substantial overlap in progression rates persists across all methods.
Different strategies for handling multiple intraocular pressure (IOP) measurements within follow-up time windows can substantially alter reported success rates after glaucoma surgery, according to a study.
Researchers applied standard high-IOP failure criteria to 934 trabeculectomy-treated eyes and 1760 deep sclerectomy-treated eyes, with approximately 41–45 months of follow-up. Multiple visit-handling strategies (mean, median, highest, lowest, closest, or all IOP values) were compared using Kaplan-Meier success estimates.
At a 21 mm Hg threshold, the highest 5-year success rates were observed when the lowest recorded IOP was used (54.8% for trabeculectomy and 74.5% for deep sclerectomy). Lower success rates were seen with approaches using peak IOP (39.3% and 60.4%, respectively) or incorporating all visits (38.8% and 61.0%). Intermediate results were reported when using mean, median, or the closest IOP values.
In a subset of patients undergoing trabeculectomy with sufficient visual field data, eyes classified as failures showed faster progression of mean deviation compared with those labeled as successes. However, there was considerable overlap in progression rates between groups regardless of the visit-handling method used.
Reference
Rabiolo A, Khaliliyeh D, Jin SW, et al. Impact of visit schedule on estimated success rates in glaucoma surgical studies. Br J Ophthalmol. 2026;bjo-2025-329283. doi: 10.1136/bjo-2025-329283. Epub ahead of print. PMID: 42055787.
