A retrospective study of 90 cases of small-sized (less than 3 mm on the printed CT film) acute (within 24 hours) subdural hematoma (SASDH) is presented. From March 1985 to December 1986, the SASDH were immediately operated on (operation rate: 86.0%). From January 1988 to December 1989, we attempted to treat them conservatively (operation rate: 49.1%). The patient population for this study consisted of 38 surgically-treated patients in the first period (Group I), 26 surgically-treated patients in the second period (Group IIs), and 26 conservatively-treated patients in the second period (Group IIc). We compared the clinical features, radiologic findings, and outcome of these 3 groups. The clinical features of Group I, including age, sex, Glasgow Coma Scale (GCS) score on admission, pupillary status on arrival, and interval from injury to the CT, did not differ significantly from those of Group II (P greater than 0.01). The only difference was the timing of the operation. In Group I, 20 patients (52.6%) received an operation within 4 hours, while in Group IIs, only 7 patients (26.9%) underwent surgery within 4 hours (P less than 0.05). The radiologic findings of Group I, including the thickness and volume of the hematoma, the degree of midline shift, and the frequency of skull fracture, also did not differ from those of Group II (P greater than 0.1). However, the outcome of Group II strikingly differed from that of Group I. The mortality rate was 76.3% in Group I, while it was 44.2% in Group II (P less than 0.005).(ABSTRACT TRUNCATED AT 250 WORDS)