CT prognostic factors in acute subdural haematomas: the value of the 'worst' CT scan

The relationship between radiological findings and outcome in patients with acute posttraumatic subdural haematomas (SDH) has been based on CT obtaine d upon hospital admission. This study was undertaken to investigate the eff ects on prognosis of SDH patients of lesions not present on admission, but detected by subsequent CT. We have also studied those findings present on a dmission CT that could predict worsening of the associated lesions. From 1 May 1989 to 30 April 1996, we admitted 206 patients harbouring acute SDH of thickness 5 mm or more. The admission GCS score ranged from 3 to 15. Each patient underwent CT on admission (always within 3 h from injury). Follow-u p CT was performed within 12-24 h after injury and in the following days ta n average of 4.3 examinations for each patient). These examinations were re viewed by a neuroradiologist and the 'worst' CT was determined. We defined the 'worst' examination as that showing the largest haematoma thickness/mid line shift and/or with the most extensive degree of parenchymal damage. Cli nical factors related to prognosis in this series are age, hypoxia/hypotens ion, GCS motor score and pupillary abnormalities. Time from injury to treat ment was found relevant only in patients with isolated SDH. CT findings on admission that correlated with outcome were haematoma thickness, midline sh ift and status of the basal cisterns. Prognosis was also worsened by the pr esence of associated lesions; SAH alone or associated with brain contusions . The last of these was the single most powerful predictor of worse outcome s (Odds ratio 0.37, p<0.004). Whereas the first CT showed parenchymal assoc iated damage in 56 patients, the 'worst CT' showed such damage in 105 patie nts. Presence of SAH on admission was found significant (p<0.02) in predict ing evolving parenchymal damage. Haematoma thickness, midline shift, status of the basal cisterns and presence of SAH are related to outcome when iden tified on the initial (early) CT examination. However, early (within 3 h fr om injury) CT under-estimates the ultimate size of parenchymal contusions. Patients with SAH on early CT are these at highest risk for associated evol ving conclusions. The use of sequential CT should be included in the routin e management of head-injured patients.