Object: This retrospective comparative cohort study was aimed at discovering the risk factors associated with subdural hygroma (SDG) following decompressive craniectomy (DC) to relieve intracranial hypertension in severe head injury.
Methods: Sixty-eight of 104 patients who had undergone DC during a 48-month period and survived > 30 days were eligible for this study. To assess the dynamics of subdural fluid collections, the authors compared CT scanning data from and the characteristics of 39 patients who had SDGs with the data in 29 patients who did not have hygromas. Variables significant in the appearance, evolution, and resolution of this complication were analyzed in a 36-week longitudinal study.
Results: The earliest imaging evidence of SDG was seen during the 1st week after DC. The SDG volume peaked between Weeks 3 and 4 post-DC and was gradually resolved by the 17th week. Among the mechanisms of injury, motor vehicle accidents were most often linked to the development of an SDG after DC (p < 0.0007), and falls were least often associated (p < 0.005). Moreover, patients with diffuse brain injury were more prone to this complication (p < 0.0299) than those with an evacuated mass (p < 0.0001). There were no statistically significant differences between patients with and without hygromas in terms of age, sex, Glasgow Coma Scale score, intraventricular and subarachnoid hemorrhage, levels of intracranial pressure and cerebral perfusion pressure, timing of decompression, and the need for CSF diversion. More than 90% of the SDGs were ipsilateral to the side of the craniectomy, and 3 (8%) of 39 SDGs showed evidence of internal bleeding at approximately 8 weeks postinjury. Surgical evacuation was needed in 4 patients with SDGs.
Conclusions: High dynamic accidents and patients with diffuse injury were more prone to SDGs. Close to 8% of SDGs converted themselves into subdural hematomas at approximately 2 months postinjury. Although SDGs developed in 39 (approximately 60%) of 68 post-DC patients, surgical evacuation was needed in only 4.