Prognostic factors in open eye injury managed with vitrectomy: retrospective study.

Aim: To evaluate surgery results and establish prognostic factors that predicted final functional (good or poor vision) and anatomic (final retinal detachment) outcome in open eye injury involving the posterior segment managed with pars plana vitrectomy.
Methods: Medical records of 52 consecutive patients with open eye injury involving the posterior segment were retrospectively reviewed. Specific variables of a system for classifying mechanical injuries of the eye were analyzed: the type of injury (defined by the mechanism of injury), grade of injury (defined by initial visual acuity), zone of injury (defined by the location of the wound), and relative afferent pupillary defect. Additional variables, such as wound length, retinal detachment, endophthalmitis, and timing of vitrectomy, were also included in the analysis. Final visual outcome and retinal attachment rate were recorded. Data were analyzed with chi-square test, univariate analysis for predictors, and multivariate logistic regression analysis.
Results: After a mean follow up of 20.0+/-10.5 months, 50% of eyes achieved visual acuity 0.5 or better, 27% of eyes achieved visual acuity 0.1 or worse, and 10% of eyes had final retinal detachment. According to univariate analysis results, the following parameters were predictors of good vision (visual acuity > or =0.5 in comparison with visual acuity <0.5): grade of injury (p=0.008), zone of injury (p=0.01), afferent pupillary response (p<0.001), wound length (p=0.002), and initial retinal detachment (p=0.009). The predictors of poor vision (visual acuity < or =0.1 in comparison with visual acuity >0.1) were zone of injury (p<0.001), relative afferent pupillary defect (p<0.001), wound length (p=0.002), and initial retinal detachment (p<0.001). Relative afferent pupillary defect (p=0.003) and initial retinal detachment (p<0.001) were predictors of final retinal detachment with proliferative vitreoretinopathy. However, multivariate logistic regression analyses revealed that relative afferent pupillary defect was the only significant factor for poor visual acuity (odds ratio, 10.3; 95% confidence interval, 1.1-92; p=0.04). On the other hand, none of the variables was a significant independent predictor for either good visual acuity or final retinal detachment.
Conclusion: Half of the eyes with a good final visual outcome in our study were successfully managed with pars plana vitrectomy for open eye injury after trauma. The classification system may become useful prognostic tool for visual outcome in posterior segment ocular injuries managed with vitrectomy. Relative afferent pupillary defect as a functional test is a good predictor for visual outcome.