Some patients are prone to persisting intraabdominal infection regardless of initial eradication of the source of infection. Our aim was to characterize patients who had to undergo relaparotomy for persisting abdominal sepsis using simple clinical parameters and to define those patients who are susceptible to benefit of aggressive surgical treatment by early and repeated reoperations to control multiple organ dysfunction syndrome (MODS) caused by ongoing intraabdominal infection. Persisting abdominal sepsis was the cause of death in all of our patients who had to undergo relaparotomy. Controlling persisting abdominal sepsis should achieve a reduction in the tremendously high mortality rate. Performing a case-control study, we retrospectively reviewed 523 consecutive patients with secondary peritonitis treated from 1986 to 1996 and focused our attention on 105 patients, in whom standard surgical treatment of secondary peritonitis failed and who had to undergo relaparotomy for persisting abdominal sepsis (study group). Overall, there was no significant difference in the postoperative mortality rate between "planned relaparotomy" and "relaparotomy on demand" (54.5% versus 50. 6%). Equally clear risk estimations were given preoperatively by both the Acute Physiology and Chronic Health Evaluation (APACHE) II and the Goris scores. There was a significant difference between patients of the control group and patients of the study group with regard to preoperative APACHE II score, Goris score, age >70 years, albumin <30 g/L, extent of peritonitis, and outcome (p = 0.0001). Reexploration performed more than 48 hr after the initial operation resulted in a significantly higher mortality rate (76.5% versus 28%; p = 0.0001). However, the time of reoperation had no significant impact on survival in patients with an APACHE II score of > or = 26, because physiologic derangement is such that only a few patients could benefit from reoperation. The lowest mortality rate (9%) was achieved in patients who underwent reoperation on demand within 48 hr. We conclude that patients >70 years of age with secondary peritonitis extending over the entire abdomen and a greater degree of physiologic compromise (serum albumin levels <30 g/L, preoperative APACHE II scores >20, and existing organ failure measured by the Goris score) are at high risk for developing persistent intraabdominal infection. Our data show that timely relaparotomy provides the only surgical option that significantly improves outcome. However, aggressive surgical treatment has reached its limit in patients whose source of infection could not be controlled at the initial operation. To improve overall survival the decision to perform a relaparotomy on demand after an initially successful eradication of the source of infection must be made within 48 hr, at least before MODS emerges.