This study was undertaken to confirm the safety and efficacy of diagnostic peritoneal lavage (DPL) for trauma patients. A prospectively maintained database of all DPLs performed in the past 75 months was analyzed. A red blood cell count of 100,000/mm(3) was considered positive for injury in blunt trauma; 10,000/mm(3) was considered positive for peritoneal penetration in penetrating trauma. Information relative to type of injury, DPL result, laparotomy result and complications, was analysed to determine if DPL was more or less suited to any specific indication or type of patient. Over a 75 month period, 2501 DPLs were performed at our urban level I trauma center. The overall sensitivity, specificity and accuracy for the above thresholds were 95, 99 and 98%. The majority (2409, 96%) were performed using percutaneous or "closed" seldinger technique. Ninety-two (4%) were performed using open technique because of pelvic fractures, previous scars and pregnancy. Open DPL was less sensitive than closed DPL in patients who sustained blunt trauma (90 vs 95%) but slightly more sensitive in determining penetration (100 vs 96%). Overall, there were 21 complications (0.8%). There was no difference in complication rate between open and closed DPL. In conclusion, DPL remains a highly accurate, sensitive and specific test with an extremely low complication rate. It can be performed either open or closed with comparable results. We recommend its use in the evaluation of both blunt and penetrating trauma.