Triple-contrast CT scanning (3-CT) is a diagnostic modality that has been introduced recently for the work-up of patients with penetrating injuries to the back or flank. Triple-contrast CT consists of giving oral, intravenous (IV), and rectal contrast medium. Our hypothesis was that this test is an accurate predictor of the absence of a retroperitoneal injury requiring surgical repair. We prospectively enrolled 88 clinically stable patients who sustained penetrating wounds to the back or flank. Seventy-eight received a diagnostic peritoneal lavage (DPL) before 3-CT. The scans were classified according to the risk of a retroperitoneal injury requiring repair. Patients who did not go to the operating room (OR) were observed for 48 hours. Of 88 patients entered, nine had high-risk scans, five of these patients underwent exploratory laparotomy, two of whom had significant injuries. Seventy-nine patients had non-high-risk scans. Seventy-seven were observed without complication, and two were explored for positive DPL, with no significant lesion found. The negative predictive value of a low- or moderate-risk 3-CT scan is 100% +/- 11%.