Methodological quality of the review: Low confidence
Author: Hollands H, Johnson D, Hollands S, Simel DL, Jinapriya D, Sharma S
Geographical coverage: Not reported
Sector: Diagnosis and early intervention for glaucoma
Sub-sector: Primary open angle glaucoma
Equity focus: None specified
Review type: Other review
Quantitative synthesis method: Meta-analysis
Qualitative synthesis method: Not applicable
Background: Early identification of individuals at risk of glaucoma could potentially lead to earlier detection and prevent associated vision loss, as early treatment is effective in diminishing glaucoma progression.
Objectives: To quantify the diagnostic accuracy of examination findings and relevant risk factors in identifying individuals with primary open angle glaucoma (POAG), the most common form of glaucoma in North America.
Main findings: In this systematic review, 57 articles from 41 separate studies were identified. 34 studies consisted of random samples of the population (level 1 and level 2 quality) and were included in the primary analysis. Studies of screening examinations performed by general physicians in a routine setting were not identified.
Authors found that the prevalence of glaucoma in the highest-quality studies was 2.6% (95% confidence interval (CI), 2.1%-3.1%). Results showed that high myopia (OR, 5.7; 95% CI, 3.1-11) and family history (OR, 3.3; 95% CI, 2.0-5.6) were found to have the strongest association with glaucoma. Black race (OR, 2.9; 95% CI, 1.4-5.9) and increasing age (>80years; OR, 2.9; 95% CI, 1.9-4.3) were also associated with an increased risk. As cup-to-disc ratio (CDR) increased, the likelihood for POAG increased with a likelihood ratio (LR) of 14 (95% CI, 5.3-39) for CDR of 0.7 or greater. Increasing CDR asymmetry was also associated with an increased likelihood for POAG (LR, 7.3; 95% CI, 3.3-16).
No single threshold for CDR or asymmetry ruled out glaucoma. The presence of a disc haemorrhage (LR, 12; 95% CI, 2.9-48) was highly suggestive of glaucoma, but the absence of a haemorrhage was non-diagnostic (LR, 0.94; 95% CI, 0.83-0.98). At the commonly used cut-off for high IOP (≥22), the LR was 13 (95% CI, 8.2-17), while lower IOP made glaucoma less likely (LR, 0.65; 95% CI, 0.55-0.76). There were no studies of screening examinations performed by generalist physicians in a routine setting.
The authors conclude that the presence of increased CDR, CDR asymmetry, disc haemorrhage and elevated IOP – as well as family history, black race and advanced age – are associated with increased risk for POAG, but their absence does not effectively rule out POAG. The most accurate way to detect glaucoma is examination by an ophthalmologist.
Methodology: OVID MEDLINE was searched for the English-language literature from January 1950 to January 2013, for which MeSH headings ‘glaucoma’ [diagnosis, epidemiology] or ‘glaucoma, open-angle’ [diagnosis, epidemiology] were used. Case reports, comments, letters, editorials and review articles were excluded.
For each retrieved study, a quality score was assigned consistent with the approach used for the Rational Clinical Examination series. Only studies of methodological quality 1 and 2 consisting of random samples of the population were used for the main analysis. The primary outcome measures were prevalence and odds ratios for demographic features and risk factors, and likelihood ratios for examination findings. Point estimates, range, univariate random-effects measures and meta-analysis were used to summarise findings.
Authors note that studies which counted eyes as the individual observation (resulting in two observations per study participant) were reviewed but not included in the meta-analysis as these studies would have inappropriately narrowed the CIs of the outcome measures. Review authors also assessed statistical heterogeneity of included data for individual findings that were evaluated in at least three studies. This was assessed using Q statistic, P value and I2 parameter that describe the between-study variance.
Applicability/external validity: No methods were used either to assess the applicability/external validity of the results or to discuss how generalizable the results are.
Geographic focus: The review doesn’t focus specifically on low/middle-income countries but the results could be applicable to these settings, as identifying risk factors and early diagnosis will prevent glaucoma progression anywhere.
Summary of quality assessment: Low confidence in the conclusions about the effects was attributed to the review as important limitations were identified. Review authors did not report searching the literature in a comprehensive manner to ensure that all studies were identified and included in the review. Authors did not avoid language bias as only database was searched, and authors did not report reviewing reference lists and contacting authors/experts as part of the search strategy. In addition, authors did not report the methods used to select studies for inclusion in the review, or whether data extraction was conducted by two reviewers.