ABSTRACT
OBJECTIVES: Our objective was to investigate the use of serum lipase levels >10,000 U/L as a tool for predicting the etiology of acute pancreatitis (AP) and to further address the relationship between lipase elevation and disease severity. METHODS: We compared patients with AP and serum lipase >10,000 U/L (HL) with patients with AP and lower serum lipase levels (855-10,000 U/L). The etiology and severity of AP were recorded. Differences between groups were calculated. RESULTS: Of the 114 patients in the HL group, the common etiologies of AP were biliary (68%), iatrogenic trauma (14%), and idiopathic (10%). Only one patient had alcoholic AP. Conversely, the common etiologies of AP in the 146-patient comparison group (lipase 855-10,000 U/L) were broader: biliary (34%), idiopathic (23%), alcohol (14%), and iatrogenic trauma (10%). Biliary AP was twice as common in the HL group (P < 0.0001) whereas alcoholic AP was significantly less common (P < 0.0001). The positive predictive value (PPV) for biliary AP of lipase >10,000 U/L was 80% whereas the negative predictive (NPV) for alcoholic AP was 99%. No difference between groups was observed in the severity markers including ICU admission, length of hospital stay, complications, or mortality. CONCLUSIONS: In AP a serum lipase of >10,000 U/L at presentation is a useful marker and portends a biliary etiology while virtually excluding alcoholic AP. Therefore, if ultrasonography is negative for stones in this population, these data suggest workup with MRCP or EUS is warranted to evaluate for microlithiasis or sludge given the high likelihood of occult stone disease in these individuals.
