Screening tests for detecting open angle glaucoma: a systematic review and meta-analysis

Methodological quality of the review: Medium confidence

Author: Mowatt G, Burr JM, Cook JA, Siddiqui MA, Ramsay C, Fraser C, Azuara-Blanco A, Deeks JJ; OAG Screening Project

Geographical coverage: United States of America (USA), Canada, Spain, Italy, Greece, Sweden, Australia, United Kingdom (UK), Norway, India, Netherlands, Belgium

Sector: Glaucoma

Sub-sector: Open angle glaucoma

Equity focus: None specified

Review type: Systematic review

Quantitative synthesis method: Meta-analysis

Qualitative synthesis method: Not applicable

Background: Early detection and diagnosis of glaucoma, especially open angle, will delay disease progression and blindness. There are many tests or combinations of tests for detecting glaucoma, and it is very important to identify the optimal test or combination of tests.

Objectives : To assess the comparative accuracy of potential screening tests for open angle glaucoma (OAG).

Main findings: 40 studies enrolling over 48,000 people reported nine tests were included. 20 studies were population-based and representative of a screening setting, whereas 20 studies were considered representative of patients with suspected glaucoma referred from primary care, of which eight were cohort studies and 12 were case-control studies. Seven studies used the first and best reference standard of open-angle glaucoma (OAG) confirmed on longitudinal follow-up, whereas the remainder used ophthalmologist-diagnosed OAG.

Frequency-doubling technology (FDT; C-20-1) was significantly more sensitive than ophthalmoscopy (30, 95% credible interval [CrI] 0–62) and Goldmann applanation tonometry (GAT; 45, 95% CrI 17–68), whereas threshold standard automated perimetry (SAP) and Heidelberg Retinal Tomograph (HRT II) were both more sensitive than GAT (41, 95% CrI 14–64 and 39, 95% CrI 3–64, respectively).

GAT was more specific than both FDT C-20-5 (19, 95% CrI 0-53) and threshold SAP (14, 95% CrI 1-37). Judging performance by diagnostic odds ratio, FDT, oculokinetic perimetry and HRT II are promising tests. Ophthalmoscopy, SAP, retinal photography and GAT had relatively poor performance as single tests. The interpretation of these findings should take into account heterogeneity and the limited quality of the data, and the consequent associated considerable uncertainty.

The authors concluded that there was no specific test or group of tests clearly superior for glaucoma screening. Further research is needed to evaluate the comparative accuracy of the most promising tests.

Methodology: The following databases were used for this review: MEDLINE, EMBASE, Biosis, Science Citation Index and the Cochrane Library. In addition, full-text electronic searches of the American Journal of Ophthalmology, British Journal of Ophthalmology, Investigative Ophthalmology and Visual Science, and the Journal of Glaucoma were undertaken. Searches were restricted to English language publications. To identify additional potentially relevant reports, the reference lists of included studies were also scanned. Studies assessing candidate screening tests for detecting OAG in persons older than 40 years that reported true and false positives and negatives were included.

Two reviewers extracted the data from identified studies in a single form. In the event of uncertainty, the other reviewer provided advice and validated the data extraction. Quality assessment was done by two reviewers independently using a version of QUADAS adapted for assessing reports of the accuracy of screening tests for OAG. QUADAS is a quality-assessment tool for use in systematic reviews of diagnostic studies. Disagreements were resolved by consensus or arbitration by a third reviewer.

Meta-analysis was undertaken using the hierarchical summary receiver-operating characteristic model.

Applicability/external validity: Authors did not discuss the applicability/external validity of the results

Geographic focus: This review did not focus specifically on low/middle-income countries. However, data from Spain and Greece were presented. In the context of this assessment, conclusions and outcomes will benefit low/middle-income countries.

Summary of quality assessment: Authors conducted comprehensive searches to identify relevant studies. However, language bias was not avoided, as only English articles were included. Two researchers examined the title and abstract of each paper; however, it is not clear if this was done independently or not. Overall, the method used to analyse findings is clearly reported, and limitations are appropriately acknowledged by the reviewers. Thus, medium confidence was attributed to the conclusion about the effects of this review.