Objectives: To determine the safety, efficacy, and evidence for halting disease progression for retinal prosthesis systems (RPSs) and the outcomes that are and could be assessed in future studies of these devices.
Data sources: We searched Medline, EMBASE, Cumulative Index to Nursing and Allied Health (CINAHL), the Cochrane Library, PubMed (unprocessed records only), and gray literature sources, including conference proceedings from specialty societies, for studies of RPS devices published from January 1, 2000, through April 25, 2016.
Review methods: We performed redundant title and abstract screening with one reviewer's selection required for full-text article retrieval. Dual independent review was performed on all full-text articles, with disagreements resolved by consensus. Data extraction was performed by a single reviewer and was fully verified by a second reviewer. Extracted data included study design, psychometric properties assessment methods based on the COSMIN checklist, patient blinding to experimental condition, outcome assessor blinding to experimental condition, experimental condition randomly presented, number of outcome assessors, country/site, number of patients enrolled, patient inclusion criteria, patient exclusion criteria, RPS treatment details, prior treatment, concurrent treatment, study duration, diagnosis, age at diagnosis, age at implantation, eye implanted, time from implantation to study participation, sex, race, visual acuity at time of implantation, outcomes, and outcome definitions. We assessed risk of bias of individual studies for the outcomes of interest, and graded the overall strength of evidence using Evidence-based Practice Centers guidance.
Results: Eleven studies of RPS effectiveness were included. Although some patients clearly improve on tests of visual function, visual acuity, visual field, color vision, laboratory-based function, and day-to-day function from an RPS, the evidence was insufficient to estimate the proportion of patients who would benefit. Intraoperative adverse events were typically mild but some serious adverse events were reported, including intraocular pressure increase, hypotony, and presumed endophthalmitis. Three studies pointed to the possibility that RPSs may provide neuroprotection. Of the 74 outcomes reported in the 11 included studies, only 4 (Early Treatment of Diabetic Retinopathy Study visual acuity test [ETDRS], Grating Acuity Test [GAT], Chow Color Test [CCT], and Functional Low-Vision Observer Rated Assessment [FLORA]) had evidence of validity and/or reliability. Measures with evidence of validity and reliability that could be used in future RPS studies include full-field flash test, Grating Contrast Sensitivity (GCS), FAST instrument (Functional Assessment of Self-Reliance on Tasks), Very Low Vision Instrumental Activities of Daily Living (IADL-VLV), Modified National Eye Institute Visual Function Questionnaire 25-item (NEI-VFQ-25) plus supplement, and the Modified Impact of Vision Impairment (IVI).
Conclusions: Some patients clearly benefit from RPSs. Future studies of retinal prosthesis should make an effort to report valid and reliable measures of day-to-day function and quality of life.