Intravenous fluid therapy in acute pancreatitis: a critical review of the randomized trials.

INTRODUCTION: Fluid management is a cornerstone of treatment in acute pancreatitis (AP).

METHODS: Identification of existing randomized prospective trials of patients with AP, in which intravenous fluid management was a significant parameter in the experimental design, was undertaken using the PubMed and ENDOBASE databases.

RESULTS: Included patients in the seven studies identified were on the whole very unwell with deaths occurring in six trials. Fluid regimens in AP included crystalloid alone, combinations of crystalloid and colloid and in two studies, plasma. In most studies, patients with premorbid major organ failure and advanced age were excluded. Study entry within 6 h of presentation occurred in three trials. Initial fluid administration rates varied from 1 to 2 mL/kg/h to 15 mL/kg/h. Rapid fluid rates were associated with increased morbidity and mortality except in one study in which a high fluid regimen (20 mL/kg bolus within 4 h of presentation followed by 3 mL/kg) led to a better clinical outcome than a more conservative regimen. Use of Ringer's lactate led to improved surrogate outcome markers compared to that of normal saline in one study. Administration of colloid as part of the fluid management was associated with lower morbidity and lower fluid requirements.

CONCLUSION: Rapid infusion of crystalloid very early in the course of AP may be beneficial but rapid infusion of fluid later in the course of the illness may be deleterious. Colloid administration and the use of Ringer's lactate solution in preference to normal saline may improve outcome.