The mean age for our study patients was 43-years-old (range, 18 to 89 years-old), which falls within the range of means reported by other Australian studies (30.4 to 44.8 years-old).(1, 2, 8) Westmead Hospital is an adult hospital and only accepts patients 14 years 9 months and older, so the demographic will be skewed towards older ages. The male preponderance (84.6%) in our cohort is slightly more than reported in other Australian studies (77.4–83.0%) (1, 2, 8) and much higher than reported in Japan (66.1%).(9) In our study, 21.3% OGI were work-related, similar to other Australian studies (18.6%-38.8%)(1, 2, 8) but less than that reported in a Japanese study (45.8%).(9) These variations could be due to the differences in culture and lifestyle. Besides that, the definition of work-related injury is loose. For instance, injuries sustained whilst working at home or not covered by workers’ compensation may not have been counted.
Six cases (5.8%) of substance related OGI were reported in our study (4 alcohol and 1 drug-related assaults with 1 alcohol related fall). Thirty percent of the OGI caused by assaults were substance related. This is half of what has been reported in another Australian study, where 76.2% of assault were alcohol related.(1) In that study, the majority of the alcohol related assaults occurred in Aboriginals and Torres Strait Islanders, a demographic that is not frequently seen at Westmead Hospital.(1)
Multivariate analysis identified three main predictors of final VA: presenting VA, type of injury and presence of a retinal detachment. Although other factors such as zones of wound location, hyphema, lens expulsion and vitreous hemorrhage were also associated with final VA outcomes, they were all correlated with each other. For example, a ruptured globe is likely to have hyphema, vitreous hemorrhage and lens expulsion as well.
The Ocular Trauma Classification Group based its classification on four variables: initial VA, mechanism of injury, zone of wound location and presence of a RAPD as accurate predictors of final VA.(6) This is broadly in agreement with our study except we didn’t analyse RAPD due to incomplete data (2 positive RAPD, 17 negative RAPD and 85 undocumented). However, we found another parameter, retinal detachment to be a strong predictive factor. It is possible that in the absence of a documented RAPD, retinal detachment may be used as a replacement for outcome prediction since a RAPD is often present when there is a large retinal detachment.
Initial VA is well established as one of the most important parameters determining final visual outcomes.(2, 6, 8, 9) Good vision reflects mild ocular damage, whereas poor vision reflects more extensive destruction which could result in injuries such as retinal detachment and vitreous hemorrhage.(6) In line with previous studies, our results show that globe rupture carries a poor visual prognosis.(2, 6, 8) Globe rupture which resulted from blunt injury causes more diffuse damage compared to sharp injuries such as penetrating eye injuries and IOFBs which cause local damage.(6) In agreement with previous studies(6, 9), Zone III wounds were associated with a poorer visual prognosis in our study. These wounds involve posterior structures such as the retina and optic nerve which heal poorly and lack regenerating ability.(6)
Retinal detachment, hyphema, lens expulsion and vitreous hemorrhage have been reported to be associated with poor visual outcomes.(2, 6, 8, 9) This usually results from severe ocular trauma in association with other ocular tissue damage. In agreement with some(2) but not all(8) studies, uveal prolapse was not shown to be a significant predictor of final visual outcome.
Eyelid laceration has previously been reported to be associated with poor outcomes.(10) Lid laceration and adnexal injuries have been associated with blunt trauma and severe ocular injury with increased likelihood of posterior globe injuries.(10) However, we did not find this association in our study and had more sharp injuries such as penetrating eye injury (n = 8) and IOFB (n = 1) than globe ruptures (n = 7) associated with lid laceration.
The numbers of vitrectomised eyes are not large enough for meaningful analysis of correlation between timing of surgery and visual acuity. However, we note that the indication for early vitrectomy (within 2 weeks post-trauma) was highest for IOFB (80%), followed by lens dislocation (50%), endophthalmitis (50%), retinal detachment (33.3%), vitreous hemorrhage (0%) and epiretinal membrane (0%). This likely reflects the urgency required to remove an IOFB, delayed presentations of endophthalmitis, retinal detachment and epiretinal membrane, delayed diagnosis of retinal detachment in a traumatised eye with media opacity and decisions to observe vitreous hemorrhage which may spontaneously resolve.
Our study has limitations including its retrospective nature, a short follow-up period due to patients being discharged to private rooms, the presence of tamponade agents (12 eyes, 11.5%) at final follow-up making visual acuity readings difficult to interpret, poor documentation of RAPD, use of pin-hole instead of refracted BCVA in some patients and poor documentation of the time of injury, meaning that the time of wound closure could only be calculated to days and not hours. Despite this, it is a reasonably large cohort with significant findings of those parameters which were accurately documented.