We reviewed medical records and films of all 196 trauma patients who underwent computed tomography (CT) between June 1982 and October 1986 to see whether CT achieved the level of accuracy attributed to it, whether diagnostic peritoneal lavage (DPL) performed in conjunction with CT was a useful diagnostic test for blunt abdominal trauma, and whether laparotomy was mandatory when pelvic fluid collections were seen by CT after blunt trauma. A total of 36 patients underwent DPL, 29 before and seven after CT. There were seven false-negative CTs that were clinically significant. Diagnostic peritoneal lavage was positive in three patients who had false-negative CTs. Although overall accuracy was excellent, CT was not reliable in detecting bowel injury. Diagnostic peritoneal lavage was helpful in detecting injuries missed by CT. Most stable patients with moderate or large intraperitoneal fluid collections on CT accompanying solid viscus injury were treated successfully without laparotomy.