The evolution of selective laparotomy in children sustaining blunt abdominal trauma has been highly controversial. This report describes our experience and policy change during this transitional period. Emergency laparotomies performed in the pediatric age group (less than 14 yr) between 1980 and 1984, based on peritoneal lavage, were reviewed. Of 16 such patients, six (37.5%) had injuries identified at laparotomy necessitating operation, (four greater than Grade III spleen, one hepatic vein, one small bowel). The remaining ten patients (67.5%) had injuries which probably could have been managed nonoperatively (eight less than or equal to Grade II spleen, two less than or equal to Grade II liver). We additionally reviewed 46 peritoneal lavages done in children during 1984, and noted a 100% sensitivity but 86% specificity when considering essential laparotomies. Based on these data, we established a selective management protocol and initiated a prospective study in January 1985. The protocol consisted of: 1) routine peritoneal lavage (DPL) in children at high risk for abdominal injury, 2) immediate laparotomy for DPL positive for blood in conjunction with hemodynamic instability, 3) selective laparotomy for DPL positive for blood in a stable child, additionally evaluated by abdominal CT scan (major mechanism) or liver/spleen scan (minor mechanism), and 4) mandatory laparotomy for DPL effluent positive by criteria other than blood. This policy reduced unnecessary laparotomy, otherwise warranted by DPL, to 18% (2/11); both patients had Grade II splenic injuries. Five children sustaining low-energy trauma were managed nonoperatively following peritoneal aspiration of gross blood with L-S scan confirming minor solid visceral injury.(ABSTRACT TRUNCATED AT 250 WORDS)