BACKGROUND Treatment to decrease the rebleeding rate in traumatic hyphema remains controversial. Although antifibrinolytics recently have been shown to reduce secondary hemorrhage rates, their routine use has not been widely applied because of adverse side effects and the relatively low frequency of severe hyphema complications. Alternatively, their use may be restricted to patients at high risk, but prognostic factors for rebleeding have not been clearly identified. METHODS From a prospective ocular trauma survey, 371 patients with traumatic hyphema were identified, and Fisher's exact test was applied to test for significant differences between patients who did and did not rebleed for various characteristics. Significant factors contributing to rebleeding were fitted into a multiple logistic regression model, and odds ratios (OR) and 95% confidence intervals (95% CI) were calculated. RESULTS Secondary hemorrhage occurred in 8% of patients and was significantly more frequent in those with visual acuities of 20/200 or less (OR = 3.1; 95% CI = 1.3,7.5), initial hyphema more than one third of the anterior chamber (OR = 2.8; 95% CI = 0.9,8.0), delayed medical attention more than 1 day after injury (OR = 2.9; 95% CI = 1.0,8.4), and elevated intraocular pressure at time of first examination (OR = 2.9; 95% CI = 1.1,7.9). The secondary hemorrhage rate rose from 5% without any of these specified factors to 15% with at least one factor present. No statistical associations were found for age, injury-related iris abnormalities, or aspirin usage. CONCLUSIONS Using multivariate logistic regression in populations with low rates of secondary hemorrhage, a predictive model may be used to categorize patients who have higher rebleeding rates, for whom possible benefits may outweigh the risks of prophylactic treatment, and those with lower rebleeding rates, who may not necessarily benefit from treatment.