Association of reticular pseudodrusen burden and dark adaptation in non-advanced age-related macular degeneration: a cross-sectional and longitudinal study.

PURPOSE: Dark adaptation (DA) has emerged as a potential functional outcome for assessing age-related macular degeneration (AMD). While reticular pseudodrusen (RPD) have been linked to poorer DA, most studies to date have focused on qualitative RPD presence/absence or comprised small cross-sectional cohorts. This work aimed to assess how automated quantification of RPD area and volume relates to rod-mediated DA and its longitudinal decline in non-advanced AMD. DESIGN: Prospective cross-sectional and longitudinal study. PARTICIPANTS: Patients with non-advanced AMD and controls, all aged above 50, from the Massachusetts Eye and Ear AMD Biomarkers Study. METHODS: All subjects underwent baseline multimodal retinal imaging and DA testing with a 20-minute extended protocol (AdaptDx, MacuLogix). Rod intercept time (RIT) and area under the dark adaptation curve (AUDAC) were calculated. A deep learning algorithm was employed to quantify RPD en face area (mm2) and volume (mm3), which were square- and cube-root transformed respectively for analysis. RPD extent within the DA testing locus was also measured. Linear mixed-effect models evaluated the relationship of RPD burden with DA metrics, accounting for age, sex, smoking, classic drusen volume and inter-eye correlation. MAIN OUTCOME MEASURES: Association between RPD burden and dark adaptation (RIT and AUDAC) on a cross-sectional and longitudinal basis RESULTS: We included 545 eyes (52 early AMD, 285 intermediate AMD, 208 controls) from 310 patients (mean 69.6±7.4 years). 128 (23.5%) had RPD. Greater RPD en-face area and volume were associated with prolonged RIT (β 1.59, p=0.002 and β 9.29, p=9.78 x 10-5 respectively) and higher AUDAC (β 0.032, p=2.15 x 10-4 and β 0.166, p=5.88x10-5 respectively). Overall macular RPD burden, rather than local density at the DA testing locus, was independently associated with worse AUDAC (β 0.005, p=0.001). In longitudinal analyses (n=204 eyes, mean 2.8±1.8 years), RPD presence (β 0.011, p=0.0067), en face area (β 0.012, p=4.48 x 10-4), and volume (β 0.054, p=2.11 x 10-4) were all significantly associated with a faster rate of AUDAC decline over time, and were predictive of AUDAC at any subsequent visit (β 0.048, p=6.72 x 10-5 and β 0.198, p=2.02 x 10-4). CONCLUSIONS: Quantitative RPD burden is independently associated with impaired rod-mediated DA and with subsequent functional decline, supporting its value as a prognostic structural biomarker for risk stratification and progression monitoring in non-advanced AMD.