Highlight
– Moderate-certainty evidence: bilateral disease predicts faster visual-field (VF) progression in open-angle glaucoma.
– Low-certainty evidence: presence of optic disc haemorrhage and female sex are associated with higher hazard of progression; active glaucoma treatment is associated with reduced hazard.
– Most other candidate prognostic factors (baseline IOP, central corneal thickness, systemic vascular disease, migraine, Raynaud’s) showed inconsistent or very low-certainty evidence.
Background and clinical context
Glaucoma is a chronic, progressive optic neuropathy and a leading cause of irreversible blindness worldwide. Primary open-angle glaucoma (POAG) is the commonest form; pseudoexfoliative glaucoma (PXFG) is a frequent secondary open-angle glaucoma that often progresses more rapidly. Accurate identification of prognostic factors for structural and functional deterioration enables clinicians to tailor monitoring and treatment intensity, prioritize high-risk patients for earlier intervention, and design prognostic models for shared decision-making.
Study design and methods (brief)
The Cochrane review by Piyasena et al. (2025) systematically identified prognostic-factor studies of adults (≥18 years) with POAG, normal-tension glaucoma (NTG), and pseudoexfoliative glaucoma, excluding eyes with prior glaucoma surgery. Included designs were cohorts, case–control studies and randomized trials that provided prognostic-factor estimates with at least 2 years of follow-up and sample sizes ≥200. Searches of CENTRAL, MEDLINE, Embase and trial registries (to 15 August 2024) yielded 22 eligible studies (123 reports) with 6,082 participants. Progression endpoints were primarily visual-field (VF) deterioration (16 studies) and combined functional plus structural outcomes (6 studies, e.g., retinal nerve fibre layer changes by OCT).
Key findings
The review evaluated many candidate prognostic factors and pooled estimates where data were sufficiently homogeneous. Risk-of-bias appraisal (QUIPS) judged 19 of 22 included studies at high overall risk of bias. Certainty of evidence was rated using GRADE and ranged from very low to moderate.
Consistent prognostic factors
– Bilateral disease (presence of glaucoma in both eyes). Pooled adjusted hazard ratio (HR) for VF progression 1.77 (95% CI 1.35–2.32; 771 participants; 2 studies). GRADE: moderate certainty. Interpretation: patients with bilateral involvement at baseline had ~75% higher hazard of VF progression than those with unilateral disease, suggesting baseline disease burden or systemic predisposition as drivers of risk.
– Optic disc haemorrhage. Pooled adjusted HR 2.03 (95% CI 1.55–2.67; 1,068 participants; 3 studies); pooled unadjusted HR 1.51 (95% CI 1.12–2.02; 961 participants, 3 studies). GRADE: low certainty. Interpretation: disc haemorrhages—small splinter haemorrhages on the optic disc margin—were associated with about a doubled hazard of progression in adjusted analyses. Clinically, their appearance remains an important red flag requiring closer surveillance or treatment review.
– Treatment for glaucoma (pharmacologic). Pooled adjusted HR 0.44 (95% CI 0.31–0.61; 961 participants; 3 studies); unadjusted HR 0.56 (95% CI 0.44–0.72; 771 participants; 2 studies). GRADE: low certainty. Interpretation: treated eyes had substantially lower hazard of VF progression. This likely reflects a genuine protective effect of lowering intraocular pressure (IOP), but the authors appropriately note potential confounding by indication and treatment heterogeneity across studies.
Factors with mixed or uncertain evidence
– Baseline intraocular pressure (IOP). Pooled adjusted HR per unit change 1.08 (95% CI 1.03–1.13; 913 participants; 3 studies; low certainty) suggested a small effect, but pooled adjusted ORs did not show an association (adjusted OR 0.96, 95% CI 0.84–1.10; 458 participants; 2 studies) and many individual studies found no IOP effect. Differences in IOP metrics (mean IOP, peak IOP, diurnal fluctuation), measurement protocols and the influence of treatment likely contributed to inconsistency.
– Age. Pooled adjusted HR per 1-year increase 1.01 (95% CI 0.97–1.05; 865 participants; 4 studies) with very low certainty; individual studies conflicted. Age remains biologically plausible as a risk modifier, but evidence for incremental hazard per year in established open-angle glaucoma is weak.
– Sex. Combined adjusted analyses of two studies suggested higher hazard in females (HR 1.64, 95% CI 1.15–2.34; 961 participants; low certainty). However, other estimates were inconsistent, and the biological basis is uncertain. Sex differences may reflect sampling, phenotype differences (e.g., NTG predominance), hormonal factors, or care-seeking behaviour.
– Central corneal thickness (CCT), systemic hypertension, cardiovascular disease, migraine, and Raynaud’s phenomenon: pooled estimates were generally non-significant with very low to low certainty. For example, adjusted HR for CCT 1.13 (95% CI 0.85–1.51; 425 participants; 2 studies) and for systemic hypertension adjusted HR 1.33 (95% CI 0.68–2.60; 731 participants; 3 studies). These results do not rule out small or subgroup-specific effects but indicate inconsistent evidence in the existing literature.
Scope and magnitude of evidence
Most studies used VF-based progression; only six included structural OCT outcomes. Median follow-up varied but all studies required at least two years. Overall, despite a sizeable pooled sample, the evidence base is limited by heterogeneity in definitions of progression, variable adjustment sets in multivariable analyses, and high risk of bias in study conduct and analysis.
Expert commentary and interpretation
The review provides a pragmatic synthesis clinicians can use to prioritise risk features during follow-up. Two signals stand out: optic disc haemorrhage and bilateral disease. Disc haemorrhages have long been recognised clinically as markers of active or unstable disease; this synthesis quantifies the associated increased hazard and strengthens the recommendation to intensify surveillance or treatment review when they appear. Bilateral disease likely captures both higher baseline burden and systemic susceptibility.
The finding that treatment is associated with reduced hazard is consistent with decades of trial evidence that IOP lowering slows progression. However, cohort data are susceptible to confounding by indication (patients with more severe disease may be more intensively treated) and to variability in adherence, treatment intensity, and surgical interventions excluded by the review criteria.
Baseline IOP’s inconsistent relationship with progression in these prognostic studies may surprise clinicians used to trial data demonstrating benefit of IOP lowering. Explanations include: (1) heterogeneity of IOP measurement and reporting (mean vs peak vs fluctuation); (2) the modifying influence of treatment initiated during follow-up; (3) differences between predictors of incident glaucoma versus progression in established disease; and (4) variable inclusion of NTG and PXFG subgroups, which have different IOP-progression relationships.
Other systemic vascular risk markers (hypertension, cardiovascular disease, migraine, Raynaud’s) remain plausible mechanistically—particularly for NTG—but current observational data do not provide consistent, high-certainty prognostic estimates. Central corneal thickness remains an important factor in risk stratification for ocular hypertension (conversion to glaucoma) but its role as an independent progression predictor in established glaucoma is less clear based on this review.
Clinical implications and practical recommendations
– High-risk signals: treat new optic disc haemorrhage as evidence of increased progression risk. Consider earlier re-evaluation of target IOP, shorter VF/OCT intervals (e.g., 3–4 months until stability confirmed), and treatment intensification where appropriate.
– Bilateral disease: consider baseline bilateral involvement a marker for closer follow-up and potentially lower target IOP, particularly when other risk features are present.
– Use of treatment: this synthesis reinforces the protective role of treatment; clinicians should continue to pursue evidence-based IOP lowering, monitor adherence, and document response.
– IOP assessment: individual IOP metrics (peak, fluctuation) and context (treated vs untreated) matter—do not rely solely on single baseline IOP measurement to predict future course.
– Structural monitoring: because several studies relied on VF alone, incorporating OCT-based structural monitoring may detect progression earlier for some patients; clinicians should combine structural and functional data where possible.
Limitations of the evidence and review
– High risk of bias across most included studies (selection, prognostic factor measurement, confounding and analysis reporting).
– Heterogeneity in definitions of progression and outcome measures (VF algorithms, event- or trend-based analyses, OCT parameters).
– Potential confounding by indication in treatment–outcome associations and variable adjustment for baseline disease severity.
– Limited numbers of studies for many prognostic factors and few studies using contemporary OCT metrics.
– Generalisability: study populations varied in demographics, glaucoma subtypes (POAG, NTG, PXFG), and care settings—findings may not apply uniformly across regions or clinical contexts.
Research and practice priorities
– Standardise progression definitions across cohorts (harmonised VF event/trend criteria and OCT structural endpoints).
– Prospectively designed prognostic-factor studies with pre-specified variable selection, adequate sample sizes, robust confounder adjustment (including time-updated treatment), and external validation.
– Investigate the prognostic value of modern structural metrics (OCT ganglion cell/retinal nerve fibre layer) and composite structural-functional models.
– Explore mechanistic links between disc haemorrhage and progression to identify potential therapeutic targets beyond IOP lowering.
– Develop and validate multivariable risk calculators incorporating the most robust prognostic factors to inform individualized monitoring and treatment decisions.
Conclusion
The Cochrane synthesis identifies bilateral disease as a moderate-certainty prognostic indicator for VF progression in open-angle glaucoma, with lower-certainty evidence implicating optic disc haemorrhage and female sex as risk markers, while active pharmacologic treatment is associated with reduced hazard. For most other candidate factors, current evidence is inconsistent or of low certainty. Clinicians should consider these findings when stratifying follow-up intensity and therapeutic thresholds, but also recognise the limitations imposed by study heterogeneity and bias. Well-designed prospective prognostic studies and harmonised outcome measures are needed to improve risk prediction and guide personalized glaucoma care.
Funding and trial registration
Funding for the Cochrane review: NIH (NEI: UG1EY020522), USA; Health Research Board (HRB), Ireland; HSC PHA (CBES-2018-001), Ireland. Protocol registration: doi.org/10.1002/14651858.CD015436.
References
1. Piyasena MP, Daka Q, Qureshi R, Roberti G, Michelessi M, Liu SH, Li T, Takwoingi Y, Azuara-Blanco A, Virgili G; supported by the Cochrane Eyes and Vision Group. Prognostic factors associated with progression of open-angle glaucoma in adults. Cochrane Database Syst Rev. 2025 Dec 10;12(12):CD015436. doi: 10.1002/14651858.CD015436.pub2. PMID: 41370133; PMCID: PMC12694756.
2. World Health Organization. World Report on Vision. Geneva: WHO; 2019. (Provides global data on vision impairment and blindness burden.)
3. National Eye Institute (NEI). Glaucoma — What Is Glaucoma? https://www.nei.nih.gov/ (Accessed 2024). (Authoritative overview of glaucoma epidemiology and clinical management.)
