RÉSUMÉ
ObjectiveTo study the early diagnosis and prevention of fungal infection in severe acute pancreatitis(SAP). Method 1.SAP patients from July 1998 to June 2002 were prospectively randomized into 3 groups: garlicin prevention group, fluconazole (low dosage) prevention group and control group, the incidence of fungal infection in SAP was compared between the groups. For fungal infection patients, the fungal clearance and mortality rate were observed. 2.Clinical data of SAP patients with fungal infection and with simple bacterial infection was compared by multivariate logistic regression, and clinical characters and risk factors of fungal infection were evaluated. Results 1.There were lower incidences of fungal infection in garlicin group (16% vs. 30%,P