Ocular Biomarkers for Alzheimer Disease Dementia: An Umbrella Review of Systematic Reviews and Meta-analyses

Eliana Costanzo,Imre Lengyel,Mariacristina Parravano,Ilaria Biagini,Michele Veldsman,AmanPreet Badhwar,Matthew Betts,Antonio Cherubini,David J Llewellyn,Ilianna Lourida,Tom MacGillivray,Timothy Rittman,Stefano Tamburin,Xin You Tai,Gianni Virgili

DOI: https://doi.org/10.1001/jamaophthalmol.2022.4845

2023-01-01

Abstract:Importance: Several ocular biomarkers have been proposed for the early detection of Alzheimer disease (AD) and mild cognitive impairment (MCI), particularly fundus photography, optical coherence tomography (OCT), and OCT angiography (OCTA). Objective: To perform an umbrella review of systematic reviews to assess the diagnostic accuracy of ocular biomarkers for early diagnosis of Alzheimer disease. Data sources: MEDLINE, Embase, and PsycINFO were searched from January 2000 to November 2021. The references of included reviews were also searched. Study selection: Systematic reviews investigating the diagnostic accuracy of ocular biomarkers to detect AD and MCI, in secondary care or memory clinics, against established clinical criteria or clinical judgment. Data extraction and synthesis: The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline checklist was followed and the Risk Of Bias in Systematic reviews tool was used to assess review quality. Main outcomes and measures: The prespecified outcome was the accuracy of ocular biomarkers for diagnosing AD and MCI. The area under the curve (AUC) was derived from standardized mean difference. Results: From the 591 titles, 14 systematic reviews were included (median [range] number of studies in each review, 14 [5-126]). Only 4 reviews were at low risk of bias on all Risk of Bias in Systematic Reviews domains. The imaging-derived parameters with the most evidence for detecting AD compared with healthy controls were OCT peripapillary retinal nerve fiber layer thickness (38 studies including 1883 patients with AD and 2510 controls; AUC = 0.70; 95% CI, 0.53-0.79); OCTA foveal avascular zone (5 studies including 177 patients with AD and 371 controls; AUC = 0.73; 95% CI, 0.50-0.89); and saccadic eye movements prosaccade latency (30 studies including 651 patients with AD/MCI and 771 controls; AUC = 0.64; 95% CI, 0.58-0.69). Antisaccade error was investigated in fewer studies (12 studies including 424 patients with AD/MCI and 382 controls) and yielded the best accuracy (AUC = 0.79; 95% CI, 0.70-0.88). Conclusions and relevance: This umbrella review has highlighted limitations in design and reporting of the existing research on ocular biomarkers for diagnosing AD. Parameters with the best evidence showed poor to moderate diagnostic accuracy in cross-sectional studies. Future longitudinal studies should investigate whether changes in OCT and OCTA measurements over time can yield accurate predictions of AD onset.