Acute kidney injury (AKI) is a common and serious problem in critically ill patients, and is known to be an independent risk factor for mortality. Renal replacement therapy (RRT) is the mainstay of supportive treatment of patients with severe acute kidney injury. The goal of RRT is to achieve adequate correction of uremia, electrolyte abnormalities, and volume overload while ensuring good hemodynamic tolerance. The advantages of continuous renal replacement therapy (CRRT) are increased time-averaged dialysis dose, less hemodynamic instability, and possibly, removal of high molecular weight solutes, such as inflammatory cytokines. Solute removal can occur by several different mechanisms in CRRT. For relatively small solutes, the importance of diffusion and convection is emphasized, for solutes of larger molecular weight, the importance of convection and adsorption is emphasized. The ability of a specific CRRT to remove a certain solute is determined by membrane characteristics. But actual measurements of middle molecule clearance in large clinical trials have not been available in most trials.

During CRRT, blood is conducted through an extracorporeal circuit, circuit clotting is a major problem in daily practice of CRRT, increasing blood loss, workload, and costs. Early clotting is related to bioincompatibility, critical illness, vascular access, CRRT circuit, and modality. Therefore, one major intervention to influence circuit survival is anticoagulation. However, systemic anticoagulation, usually with heparin, can produce hemorrhagic complications in patients at high risk of bleeding. To minimize the risk of bleeding, a number of alternative regimens has been proposed, however, each of those methods has its own limitations and complication. Nafamostat mesilate, a serine proteinase inhibitor, while inhibiting various clotting factors in filter circuit, is characterized by short half life resulting in little systemic anticoagulation effect. A recently developed CRRT AN69ST membrane® (Gambro Inc) is coated with a polyethylene imine (PEI, cationic biopolymer) on the membrane surface. Once adsorbed onto the membrane, heparin keeps its anticoagulant properties. Therefore CRRT has been managed without systemic administration of heparin.

The investigators will conduct a multicenter prospective randomized controlled open-label trial which compares the difference in circuit survival between between nafamostat infusion and heparinized saline priming as anticoagulation for CRRT. The primary end-point of this study is circuit survival, the time of 1st membrane exchange. The secondary end-point is clearance of small molecule (urea) and middle molecule (β2 microglobulin) at 0, 1, 6, 24h, ACT(activated coagulation time) measurements after 1hr of the CRRT, Hemorrhagic complication. This is a noninferiority trial. The aim is to demonstrate that nafamostat infusion is not inferior to the heparinized saline priming. For this purpose, at least 80 subjects (a total of 160) would be required for each group if type I error rate is 5% and type II error is 20% given 20% of drop-out rate during the study period. Block randomization will be used by means of a dedicated website.

There are still conflicting data on the effective exchange time of circuit membrane. Our study may help to improve prognosis in patients with severe AKI.

Inclusion Criteria:

1. Injury stage of RIFLE criteria or more: > 2-fold increase in the serum creatinine or urine output < 0.5 mL/kg/hr for 12 hours
2. Patients with any dialysis treatment before admission to the ICU or patients with end-stage renal failure and receiving dialysis
3. Informed consent has been obtained.

Exclusion Criteria:

1. patient age < 20 years or > 85 years
2. life expectancy less than 3 months (ex. terminal stage of malignancy)
3. Child-Pugh class C liver cirrhosis
4. pregnancy or lactation
5. history of anticoagulation prior to the randomization
6. bleeding tendency (platelet count < 50,000/ul, INR > 2.5, PTT > 65, or fibrinogen < 1.00 g/L)
7. history of hemorrhagic disease (ex. GI bleeding, cerebral hemorrhage, pulmonary hemorrhage)