Glaucoma in the UK Biobank: A Comparison of Diagnostic-Based versus Treatment-Based Definitions

To compare, in the context of a population-scale biobank, associations of established risk factors and severity biomarkers with different glaucoma disease definitions.

Observational cohort study.

Five hundred three thousand three hundred twenty-five participants in the UK Biobank.

Odds ratios (ORs) for associations of established risk factors (corneal-compensated intraocular pressure [IOPcc] and polygenic risk score [PRS]) and clinical biomarkers of disease severity (macular retinal nerve fiber layer [mRNFL] and macular ganglion cell layer [mGCL] thickness) with different glaucoma definitions.

Participants without glaucoma (n = 481 772) were compared with those with: (1) a diagnosis of glaucoma based on self-report or hospital inpatient or primary care records (diagnosis-only group; n = 16 154); or (2) a diagnosis plus evidence of treatment with glaucoma-specific medication or surgery (diagnosis and treatment group; n = 7012), either at baseline or during follow-up.

For participants in the highest 5% of IOPcc compared with the remainder, ORs were 23.6 (95% confidence interval [CI], 21.3-26.2) in the diagnosis and treatment group versus 11.4 (95% CI, 10.6-12.3; P < 1 × 10^-10 for difference) in the diagnosis-only group. Corresponding ORs were: 8.0 (95% CI, 6.9-9.2) in the diagnosis and treatment group vs 4.7 (95% CI, 4.2-5.3; P = 2.32 × 10^-7 for difference) in the diagnosis-only group for the thinnest 5% mRNFL; 7.1 (95% CI, 6.1-8.3) in the diagnosis and treatment group vs 4.2 (95% CI, 3.7-4.7; P = 2.53 × 10^-7 for difference) in diagnosis-only group for the thinnest 5% mGCL; and 8.4 (95% CI, 7.9-8.8) in the diagnosis and treatment group versus 5.6 (95% CI, 5.4-5.9; P < 1 × 10^-10 for difference) in the diagnosis-only group for the highest 5% PRS. Similar patterns were found for the other risk factors studied.

Definitions of glaucoma that combine both diagnostic and treatment evidence yielded substantially stronger ORs with IOPcc, mRNFL, and mGCL than definitions based on diagnostic evidence alone, suggesting implications for studies of glaucoma in population biobanks.

Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.