Of the 49 participants in this study, 21 (43%) were controls and 28 (57%) were glaucoma patients. There was no statistical difference between controls and glaucoma patients for any of the demographic characteristics examined (
Table 1). However, VF MD in the better-eye and binocular contrast sensitivity were significantly worse in glaucoma patients than in controls (
P value ≤ 0.001 for both). Of the reading metrics examined, sustained reading speed was significantly slower among glaucoma patients as compared with controls (
P = 0.044). Reading measures of MNREAD and IReST did not differ between glaucoma and control groups. Scores on cognitive tests, MMSE and BTA, did not differ by glaucoma status in this study population (
P > 0.146, for all).
Figures 1A through
1C show the unadjusted relationship between reading speeds from MNRead, IRest, and sustained silent tests with BTA score by glaucoma status. For all three tests, these graphs indicate that there is a linear relationship between reading speeds and BTA score, and suggest that the slope of this relationship differs modestly between glaucoma and controls patients.
Results from the primary regression analyses are shown in
Table 2 for the full study population. For all three reading tests, there was no significant difference in reading speeds in models comparing glaucoma participants with controls (
P = 0.11, 0.20, 0.054 for the MNRead, IRest, and sustained tests, respectively). However, for all three tests, worse VF loss and contrast sensitivity are associated with slower reading speeds (
P = 0.03, 0.03, and 0.001 for the MNRead, IRest, and sustained tests, respectively). Worse VA in the better eye was significantly associated with slower reading speeds for the IRest test alone (
P = 0.009). For the IRest test, but not the MNRead and sustained silent reading tests, worse MMSE score was significantly associated with slower readings speeds. For the sustained silent reading tests, older age and being African-American were associated with slower speeds (
P = 0.03 and 0.01, respectively).
Stratified regression analyses were performed to examine potential effect modification of the relationship between visual metrics and reading speed outcomes by BTA score (<7 or ≥7). These models were adjusted for the same covariates as in
Table 2, and results are shown in
Figures 2A through
2D and
Table 3. For maximum reading speed on MNRead testing, a decline in log contrast sensitivity (logCS) of 0.1 was associated with a 9 wpm (95% confidence interval [CI]: −16, −2) decline in maximum reading speed among those with BTA scores below the median value. However, for those with BTA scores above the median, this effect was smaller and not statistically significant (
β = −2; 95% CI: −6, +2 wpm). Similarly, for the IRest reading test, a 0.1 decline in logCS was associated with a 12 wpm (95% CI: −20, −4) reading speed decline among those with low BTA scores, but this effect was smaller and not statistically significant (
β = −4; 95% CI: −10, +2 wpm) among those with high BTA scores. Among those with lower BTA scores, sustained silent reading speeds were 39% slower among glaucoma patients, 21% slower among those with worse VF loss, and 19% slower among those with worse CS. However, among those with high BTA scores no such associations were observed with any of the above visual measures (
P > 0.18 for all). Changes in reading speed can also be calculated for the sustained reading test. Worse VA was associated with a steeper decline in sustained silent reading speed over the 30-minute reading period among those with low BTA scores (−0.47 wpm per 0.1 worsening of logMAR VA; 95% CI: −0.89, −0.05), but not among those with high scores (0.03 wpm per 0.1 worsening of logMAR VA; 95% CI: −1.04, +1.09).