Safety and efficacy of subretinal visual implants in humans: methodological aspects.

Background: Replacing the function of visual pathway neurons by electronic implants is a novel approach presently explored by various groups in basic research and clinical trials. The novelty raises unexplored methodological aspects of clinical trial design that may require adaptation and validation.

Methods: We present procedures of efficacy and safety testing for subretinal visual implants in humans, as developed during our pilot trial 2005 to 2009 and multi-centre clinical trial since 2010.

Results: Planning such a trial requires appropriate inclusion and exclusion criteria. For subretinal electronic visual implants, patients with photoreceptor degeneration are the target patient group, whereas presence of additional diseases affecting clear optic media or the visual pathway must be excluded. Because sham surgery is not possible, a masked study design with implant power ON versus OFF is necessary. Prior to the efficacy testing by psychophysical tests, the implant's technical characteristics have to be controlled via electroretinography (ERG). Moreover the testing methods require adaptation to the particular technology. We recommend standardised tasks first to determine the light perception thresholds, light localisation and movement detection, followed by grating acuity and vision acuity test via Landolt C rings. A laboratory setup for assessing essential activities of daily living is presented. Subjective visual experiences with the implant in a natural environment, as well as questionnaires and psychological counselling are further important aspects.

Conclusions: A clinical trial protocol for artificial vision in humans, which leads a patient from blindness to the state of very low vision is a challenge and cannot be defined completely prior to the study. Available tests of visual function may not be sufficiently suited for efficacy testing of artificial vision devices. A protocol based on experience with subretinal visual implants in 22 patients is presented that has been found adequate to monitor safety and efficacy.