Purpose
To develop a simple algorithm permitting rapid identification and treatment of vision-threatening retrobulbar hemorrhage (RBH).
Methods
This is a retrospective review examining all patients seen by our consult service with radiographic evidence of RBH between January 2010 and April 2012.
Results
Forty-two patients were identified. Of these, 52.4% were observed, 25.9% were treated with drops and/or osmotic diuretics, and 16.7% received emergent canthotomy/cantholysis. Among those observed, mean IOP was 21.14 ± 9.40 mmHg. Among those requiring pharmacologic or surgical intervention, mean pressures were 33.31 ± 6.76 and 51.00 ± 13.24 mmHg, respectively. There was no significant relationship between need for intervention or degree of IOP elevation and time to consultation. Comparing treatment groups, there were no differences in GCS, trauma score, pain score, loss of consciousness, need for intubation, degree of maxillofacial trauma, or bodily trauma. The presence of lid edema, ecchymosis, chemosis, subconjunctival hemorrhage and diplopia also failed to correlate with need for intervention. The eyelids were “tight,” or difficult to open, however, in no patients who were observed, in 12.5% managed medically, and in 100% of those receiving urgent canthotomy/cantholysis (P < 0.001). All patients with “tight” lids who were managed pharmacologically also had large preseptal hematomas on CT. Unilateral proptosis was observed in 18.2%, 61.5% and 100% of patients, respectively (P < 0.001). An APD was appreciated in 9.1% of patients who were observed, 0% of those managed pharmacologically, and 42.5% who received canthotomy/cantholysis (P < 0.02). In the observation group, all patients with an APD had prior ocular injuries to explain this finding.
Conclusions
Delays often occur in recognizing and treating significant RBH. Many standard clinical clues are not useful in identifying cases needing intervention. In our study, however, 100% of the cases requiring urgent canthotomy/cantholysis were predicted using a 3-factor model requiring no specialized training or equipment. The factors are: presence of (1) unilateral or asymmetric proptosis, (2) eyelids that are “tight” or difficult to open with finger pressure, and (3) an APD if confounding factors such as preseptal hematoma are present. Adoption of this algorithm may help prevent vision loss due to delays in treatment of RBH.