Across all subjects, photopsias were indicated during 47% of test sessions. In all, 5 (14%) subjects never noted photopsias, while 5 (14%) subjects noted photopsias at every test session. A total of 24 (67%) of the subjects experienced photopsias frequently (>20% of test sessions) during the PC-based vision tests. The most common type of photopsia was (a group/series of) quick flashes of light, noted during 17% of test sessions, followed by phosphenes, described as slow, localized dots or shapes, during 11% of sessions. White-out glare was noted at 8% of sessions, fluorescence or background glow was experienced at 6% of sessions, and static noise (like on a TV without reception) was seen at 5% of sessions.
compares the characteristics of subjects who noted photopsias frequently (>20%) during test sessions to those who experienced them less frequently. The vast majority of males (89%) had an increased frequency of photopsias, which was statistically significantly higher (P=0.005) than that of females (44%). We explored whether there were any statistically significant differences among males vs females for lab-based ETDRS VA (P=0.96), Pelli–Robson CS (P=0.69), GVF (P=0.31), mean PC-based VA (P=0.84), mean PC-based CS (P=0.34), mean PC-based VF (P=0.26), or the duration of vision loss (P=0.59) or night vision loss (P=0.40). We did not find evidence of any statistically significant differences in the level of vision according to gender, indicating that males and females were comparable according to their severity of RP. However, the males in this study on average reported statistically significantly greater levels of depressive symptoms on the BDI (7.4 points higher than females; 95% CI: 1.8, 13.1; P=0.01). There was also a trend toward a significant difference according to gender for the mean PSS score (males 2.7 points higher than females; 95% CI: −0.5, 5.9; P=0.10) and mean positive mood score (males 4.7 points lower than females; 95% CI: −9.8, 0.5; P=0.07).
RP subjects' characteristics: photopsias frequency across test sessions
There were no statistically significant differences in the frequency of photopsias in relation to subjects' age, vision, and sleepiness measures (ESS and SSS). Subjects with photopsias that occurred more frequently had a mean PSS score that was 4.3 points higher on average than those with less frequent photopsias (noted <20% of the time) (95% CI: −1.7, −6.9; P=0.002). Individuals with photopsias that occurred more frequently had a mean positive mood score that was 7.7 points lower on average than those with less frequent photopsias (95% CI: 2.8, 12.5; P=0.003). Participants with photopsias that occurred more frequently had a mean negative mood score that was 2.7 points greater on average than those with less frequent photopsias (95% CI: −0.2, −5.3; P=0.04). The depressive symptoms score was 5.8 points greater on average for subjects with more frequent photopsias as compared with those with less frequent photopsias (95% CI: 0.7, 10.7; P=0.03).
shows the crude and adjusted odds of frequent photopsias (>20% of test sessions) for each mean psychosocial variable. After adjusting for age, gender, VA, and GVF, for every 1-point increase in the mean perceived stress score, there was a statistically significant 57% increase in the odds of noting photopsias frequently (95% CI: 1.04, 2.4; P=0.03). For every 1-point increase in the mean positive mood score, there was a statistically significant 38% decrease in the odds of noting photopsias frequently (95% CI: 0.39, 0.98; P=0.04). There was a non-statistically significant trend for negative mood, such that for every 1-point increase in the mean negative mood score, there was a 44% increase in the odds of noting photopsias frequently (95% CI: 0.95, 2.2; P=0.09). The odds of noting photopsias frequently did not increase statistically significantly according to the sleepiness measures or depressive symptoms.
Crude and adjusted odds of increased frequency of photopsias (during >20% of sessions) for each psychosocial variable
provides a graphical depiction of the data, indicating a lower median and interquartile range for the mean positive mood score in subjects with photopsias during >20% of sessions. also shows a higher median and interquartile range for the mean perceived stress score for those with photopsias noted during >20% of sessions.
Figure 1 Box plots of the frequency of photopsias in relation to (a) positive mood and (b) perceived stress. The bottom and top of the box are the 25th and 75th percentiles (the lower and upper quartiles, respectively), and the band near the middle of the box (more ...)
displays the adjusted odds of experiencing photopsias at a test session in relation to 1 and 3 point changes in the psychosocial variables, as well as light exposure. There were no significant or qualitative differences in the crude vs adjusted odds in and therefore only the adjusted odds are reported. After adjusting for age and gender, for every 3-point increase in perceived stress and positive mood score, there was a statistically significant 16% increase and 17% decrease in the odds of experiencing photopsias, respectively. The within-subject mean SD for the perceived stress and positive mood scores across test sessions was 4.5 points (range 1.3–10.8) and 5.1 points (range 1.6–8.7), respectively.
Table 3 Adjusteda odds of RP subjects (n=36) experiencing photopsias at a test session
To further illustrate the scale of the psychosocial questionnaires, we determined relative changes from the mean for 1- and 3-point score changes. Across subjects, a 1-point score change was roughly equivalent to an 8% change from the mean for the perceived stress and positive mood scores, while 3-point score changes were roughly equivalent to 18 and 25% changes from the mean for the perceived stress and positive mood scores, respectively. Subjects had a change of at least 3 points from their mean score across 45% of sessions for the PSS and 49% of sessions for positive mood score. depicts the unadjusted increasing probabilities of photopsias with increasing perceived stress score and the decreasing probabilities of photopsias with increasing positive mood score.
Probability of experiencing photopsias in relation to (a) positive mood and (b) perceived stress.
The results in are based on random intercept models. Although theoretically it may seem appropriate to allow each subject to have their own random slope in addition to a random intercept, as effects may vary across individuals, exploratory data analyses supported using a random intercept only model for the final analyses. The coefficients from the random slope and intercept models were not significantly or qualitatively different than the random intercept only models.
Negative mood, sleepiness, and light exposure (dim or bright) in the past hour were not statistically significantly associated with the occurrence of photopsias at a test session. Subjects had a change of at least 3 points from their mean score across 38% of sessions for negative mood score, 21% of sessions for the ESS, and 2% of sessions for the SSS. Across all subjects, 27 and 17% of the PC tests were taken after dim and bright light exposure in the past hour, respectively.
We were able to determine whether within-session variability occurred during VA and CS as these tests measured two thresholds simultaneously and independently through Bayesian estimation within a single session. We examined whether the presence of photopsias affects test performance within session. During CS testing, at times when there was no within-session variability (ie, the two thresholds were the same; during 31% of all sessions), subjects were less likely to report photopsias (12% of all sessions) (χ2 test: P=0.01). At times when there was no within-session test variability, on average the mean CS was better by 0.023 logCS (95% CI: −0.04, −0.003; P=0.02). However, at times when the subjects reported photopsias during the vision tests, we did not find a statistically significant reduction in CS on average. We also did not find any statistically significant effect on VF, VA, or within-session variability of VA when photopsias were noted.