Figure 1 summarizes the results for PCD density (see
Supplementary Table S1 for participant-level data). The mean ± SD PCD value across all participants at visit 1 was 187,000 ± 20,000 cones/mm
2 (15,800 ± 1790 cones/deg
2), and at visit 2 it was 189,000 ± 21,700 cones/mm
2 (16,100 ± 2020 cones/deg
2). The smallest intervisit percentage change in linear density was 0.56%, and the largest was 13.6%. Eleven out of 19 participants had a PCD linear density difference of less than 5%. Angular density percentage change in PCD ranged from 0% to 17.4%, with 12 out of 19 participants having a PCD angular density difference of less than 5% (all 11 participants who had less than 5% change in PCD linear densities also had less than 5% change in the PCD angular density). The PCD did not differ significantly between visits with regard to linear density (
t = 0.66;
df = 18;
P = 0.52, paired
t-test) or angular density (
t = 0.99;
df = 18;
P = 0.33, paired
t-test). There was strong correlation between visit 1 and visit 2 values with regard to linear density (Pearson correlation
r = 0.85; 95% confidence interval [CI], 0.65–0.94;
P < 0.0001) and angular density (Pearson correlation
r = 0.85; 95% CI, 0.64–0.94;
P <0.0001). Bland–Altman analysis of linear (
Fig. 1A) and angular (
Fig. 1B) PCD showed good agreement between visits as the 95% CI of the mean bias included zero; for linear density, the mean bias was 1740 cones/mm
2 (95% CI, −3770 to 7240 cones/mm
2), and, for angular density, the mean bias was 247 cones/deg
2 (95% CI, −274 to 769 cones/deg
2). There was no significant correlation between the intervisit interval time and absolute differences in PCD for linear density (Pearson correlation
r = −0.21; 95% CI, −0.61 to 0.27;
P = 0.38) or angular density (Pearson correlation
r = −0.12; 95% CI, −0.55 to 0.35;
P = 0.61).