Retinal biomarkers for the detection of cognitive frailty in older adults: a systematic review

Background. Population ageing in Europe is increasing the burden of cognitive impairment, yet early detection remains constrained by time-intensive neuropsychological assessment. The retina, embryologically part of the central nervous system, can be imaged non-invasively and may carry detectable signals of incipient cognitive change. Objective. To synthesise the evidence on quantitative retinal biomarkers (optical coherence tomography [OCT], OCT-angiography [OCT-A], fundus photography) for the detection of cognitive frailty in adults aged 60 and older. Methods. PubMed, Embase, Web of Science, Scopus and Cochrane Library were searched from January 2013 (IANA/IAGG cognitive frailty consensus) to the cut-off date. Two reviewers independently screened records, extracted data and appraised risk of bias with Newcastle-Ottawa, AXIS or QUADAS-2 according to design. Reporting follows PRISMA 2020. Overall certainty was rated with GRADE. The protocol was prospectively registered with PROSPERO (provisional ID 1379317). Results. Fifteen pivotal studies clustered into three groups: large population cohorts, memory-clinic case-control studies and OCT-A or fundus-vascular work. Peripapillary retinal nerve fibre layer (RNFL) and macular ganglion cell-inner plexiform layer (GCL-IPL) thinning were consistently associated with cognitive impairment (HR up to 1.44 per SD decrease in RNFL). OCT-A vascular signals were heterogeneous. Deep-learning fundus models reached internal-validation AUCs above 0.9 but dropped externally. GRADE certainty: very low to moderate. Conclusions. Retinal biomarkers, particularly structural OCT measurements, show reproducible associations with cognitive impairment and warrant prospective validation focused on cognitive frailty in institutionalised geriatric populations.