Penetrating ocular injuries in children: visual outcome and prognostic factors

Editor, Although open-globe injuries are one of the major reasons for visual loss in children, there are little published data analysing the prognostic factors in such injuries according to internationally standardized terminology (Kuhn et al. 2002). This review was undertaken to identify the potential prognostic factors that relate to final visual outcome after penetrating ocular injury in children. The study included 74 patients younger than 14 years admitted to the Edouard Herriot Hospital, Lyon, France, between January 1998 and December 2002. Patients were followed for a median of 22 months (range 18–30 months). Annual incidence was 1.5 per 100 000 inhabitants. Males were affected in 76% of cases. The average age at admission was 7 years. The highest proportion of injuries (61%) occurred at home and sharp objects were the most frequent causative agents (66%). The overall outcome showed 45% of patients attained visual acuity (VA) of 20/40 or better. For statistical analysis, the final clinical outcome was defined as poor if final VA was less than 20/100. The most common preoperative factors were studied using univariate analysis (Table 1). Two variables were predictive of poor visual outcome: initial lens damage (p < 0.001), and intravitreal haemorrhage (p < 0.001). Our findings concerning annual incidence, the high incidence of accidents occurring in the home and sex ratio are consistent with those reported in the literature (MacEwen et al. 1999; Thompson et al. 2002). The final visual outcome recorded in this study also agrees with the findings of previous reports (Alfaro et al. 1994; Jandeck et al. 2000). However, comparison of the results of various studies remains difficult because of the probable diversity of injuries. In our analysis, a poor preoperative VA had no significant effect on the final visual results. This may be due to the lack of patient compliance with initial VA evaluation (Rudd et al. 1994). As previously reported (Baxter et al. 1994; Rostomian et al. 1998), we found that a delay of up to 24 hours or more between injury and surgery did not influence poor final VA. However, our study did not allow precise evaluation of this factor because the injury-to-surgery interval was always up to 72 hours. In our series, the small number of cases of scleral involvement (eight cases) and retinal prolapse (one case) did not allow statistical evaluation of these two potential prognostic factors. Children who presented with damage to the lens had significantly worse outcomes compared with others. Previous investigators (Alfaro et al. 1994; Baxter et al. 1994) also found lens damage to be a reliable indicator of poor final VA. The investigators suggested that young children with traumatic cataract have a less favourable outcome because of the problems associated with correcting aphakia and subsequent amblyopia. In our report, the poor visual outcome observed in the 12 patients with initial lens damage may be explained by the following reasons: retinal detachment after lensectomy surgery (three cases); problems with occlusion therapy in very young children (three cases); insufficient compliance by children with contact lens wear (two cases); contact lens intolerance (three cases), and corneal abscess associated with contact lens wear (one case). Initial vitreous haemorrhage is also a predictor of poor visual prognosis. The poor visual outcome observed in 13 patients with initial vitreous haemorrhage may be due to the following reasons: traction retinal detachment (one case); epiretinal membrane formation (two cases); contusive maculopathy (one case), and occlusion amblyopia despite vitrectomy surgery in very young children (nine cases). Although intravitreous haemorrhage has been recognized as an important prognostic factor in the adult population, this fact has been documented in only one previous study of paediatric open-globe injuries (Farr et al. 2001).