Although we have reported the beneficial effect of continuous regional arterial infusion (CRAI) of protease inhibitor and antibiotic on acute necrotizing pancreatitis (ANP), the optimal timing of the initiation of CRAI therapy has not been clarified. The present study was conducted to evaluate whether the difference of the timing of CRAI therapy may affect the clinical course and outcome in ANP. 73 patients with ANP were stratified into three groups according to the interval between the onset and initiation of CRAI therapy as follows: group I (32 patients in whom CRAI therapy was initiated within 48 h after the onset); group II (22 patients in whom CRAI therapy was initiated between 48 and 72 h after the onset), and group III (19 patients in whom CRAI was initiated more than 72 h after the onset). The mortality rate was 3.2% in group I, 9.1% in group II, and 26.3% in group III. The mortality rate was significantly low in group I compared with that in group III. The frequency of respiratory failure in group I was also significantly low compared with that in group III. CRP and APACHE II score were reduced rapidly in both groups I and II after the initiation of CRAI therapy. These results suggested that the optimal timing of CRAI therapy in ANP should be considered to be within 72 h after the onset.