Select subject demographics and ocular and orbital measurements are reported in
Table 1. In some subjects, CT scan slices did not extend to the inferior orbit, and it was not possible to collect all the measurements. Missing data were accounted for in the statistical analysis.
A multivariate correlation analysis was used to examine relationships between subject demographics (age and subject height) and the ocular and orbital measurements collected (OW, OH, LP, SP, LD, SD, orbital rim perimeters, and height-normalized OW and OH). Pearson product-moment correlation coefficients and
P values were computed for all combinations, comparing demographics to measurements, as well as measurements to each other. Results are presented in
Table 2 for correlations that were significant (
P < 0.05) or mildly significant (
P < 0.10). Noteworthy findings included significant positive correlations with age for the raw and height-normalized OW measurements, as well as a mildly significant association between age and the SD measurement. Several measurements (OH, height-normalized OW and OH, SD, and the inferior and total rim perimeters) correlated significantly with subject height, suggesting that subject height is a key factor in explaining variation in orbit anthropometry across individuals. In addition, mild correlations with subject height were identified for the OW and LD measurements. Many of the ocular and orbital measurements correlated significantly with each other. Measurements characterizing orbital aperture (OW, OH, and orbital rim perimeters) were found to correlate significantly with eye protrusion (LP and SP measurements) and location of the eye within the orbit (LD and SD measurements). Eye protrusion measurements (LP, SP) were not only significantly correlated with orbital aperture measurements, but also with the SD measurement describing eye location within the orbit. No correlations with
P < 0.10 were identified for the brow protrusion angle.
One-way analysis of variance (ANOVA) was used to assess the effect of the sex of the individual on each of the ocular and orbital measurements. ANOVA
F test statistics and
P-values are reported in
Table 3.
P < 0.05 indicates a statistically significant difference in sample means between genders. As
Table 3 shows, the lateral distance measurement and inferior orbital rim perimeter vary significantly between males and females. Orbital width and height measurements normalized by subject height vary significantly between the sexes. When these measurements (OW, OH) are not normalized, no significant differences are detected between the sexes, suggesting that normalizing by height unmasks the effect on orbital anthropometry of the sex of the person.