Eligible patients will receive either AP or matching placebo in a double blind, randomized design and following a 2:1 ratio. All medication will be given in addition to standard care for sepsis patients. Patients will be followed for 28 days after the start of study medication administration. A blinded safety review of the study results will take place after the inclusion of 12 patients in the study.

Both bolus and continuous infusions were found to effectively reduce the LPS-induced pro-inflammatory cytokine TNFα in piglets. When AP was administered as a bolus (160 U/kg), the reduction of TNFα after administration of 10 μg/kg LPS in piglets was found to be in the order of 32%. When 100 U/kg of CIAP was administered as an infusion over 50 minutes, reaching steady state levels of 400 U/L, the reduction in TNFα after LPS challenge was substantially (70%), as compared to TNFα levels after LPS only. A series of these studies demonstrated that dosages of 100-120 U/kg administered over 50 minutes could effectively reduce LPS toxicity at steady state levels around 400 U/L, about a 10 fold of the normal alkaline phosphatase plasma levels (40 U/L in piglets). Furthermore, it is estimated that the above-mentioned dosages can effectively detoxify 10 μg LPS/kg in piglets, which represents an LPS-equivalent derived from about 1010 colony forming units (cfu) of E.coli. These amounts of circulating bacteria are not easily established in a sepsis patient. It is estimated (personal communication, Prof. S. van Deventer, Academic Medical Centre, Amsterdam) that the bacterial load in these patients is less than 107 bacteria total, which is equivalent to approx. 2x105 cfu/kg.

In summary, in piglets 160 U/kg AP was able to detoxify an amount of LPS equivalent to 1010 cfu E. Coli/kg, which is approximately 50,000 times the expected amount of bacteria in sepsis patients. We therefore expect that the dosage used in human (200 U/kg) administered over 24 hours is able to detoxify the amount of LPS present in sepsis patients. This dosage will result in 5-times normal plasma level of alkaline phosphatase which was well tolerated during the previous clinical trials. Since the effect of antibiotic treatment is expected to occur within hours of administration we decided to establish this steady state level of AP as fast as possible, which explains the initial bolus-like administration by short infusion, followed by a prolonged steady state infusion.

Inclusion Criteria:

• Patients between 18 and 80 years (inclusive);
• proven or suspected infection;
• meeting 2 of 4 of the systemic inflammatory response syndrome (SIRS) criteria;
• septic shock or one or more acute organ failures in the preceding 12 hours;
• written informed consent obtained.

Exclusion Criteria:

• Pregnant or lactating women;
• known HIV seropositive patients;
• patients receiving immunosuppressive therapy or chronically using high doses of glucocorticosteroids (defined as > 1 mg/kg/day) equivalent to prednisone 1 mg/kg/day;
• patients expected to have rapidly fatal disease within 24 h;
• known confirmed gram-positive sepsis;
• known confirmed fungal sepsis;
• chronic renal failure requiring haemodialysis or peritoneal dialysis;
• acute pancreatitis with no established source of infection;
• patients not expected to survive for 28 days due to other medical conditions such as end-stage neoplasm or other diseases;
• participation in another investigational study within 90 days prior to start of the study which might interfere with this study;
• previous administration of AP;
• known allergy for cow milk.