From a series of 112 cases (64 men and 48 women, aged 37 - 91 years) of chronic subdural haematoma (CSDH) in a 2-year period from January 1998 to December 1999, we have prospectively studied a group of 30 patients, who were managed non-operatively: 26 patients were treated with dexamethasone (Group 1) and four patients expectantly (Group 4). Nineteen patients (73%) from Group 1 were confused or had focal neurological deficits on admission. The mean maximum thickness of the CSDH was 12 mm. Only one of these cases (4%) required surgical drainage 6 weeks after steroid therapy. One patient died of an unrelated stroke (mortality = 4%). Two patients (8%) were left severely disabled. No significant complication from steroid therapy was documented. Out of the 85 surgically treated patients, 69 patients underwent surgical drainage in addition to steroid therapy (Group 2). Thirteen patients were treated with burr-hole drainage only (Group 3). The mean maximum thickness of the CSDH for these two groups were both 16 mm. Comparing with group 1, the redrainage rate of Group 2 [4% (3/69, p = 1)] and that of Group 3 [15% (2/13, p = 0.253)] were not significantly different. 50% of patients in Group 4 (2/4, p = 0.039) required delayed surgical drainage. The mortality rates of Groups 2, 3 and 4 were 3% (2/69, p = 1), 15% (2/13, p = 0.253) and 50% (2/4, p = 0.039), respectively. Our results suggest that steroid treatment in a selected group of patients is a good option, particularly in patients with co-morbidity.