An ideal laboratory test in the evaluation of a patient with acute pancreatitis (AP) should, in addition to accurately establishing the diagnosis of AP, provide early assessment of its severity and identify the etiology. None of the tests available today meet all these criteria, and presently there is no biochemical test that can be considered the "gold standard" for the diagnosis and assessment of severity of AP. In the diagnosis of AP, serum amylase and lipase remain important tests. Advantages of amylase estimation are its technical simplicity, easy availability, and high sensitivity. However, its greatest disadvantage is its low specificity. A normal amylase would usually exclude the diagnosis of AP, with the exception of AP secondary to hyperlipidemia, acute exacerbation of chronic pancreatitis, and when the estimation of amylase is delayed in the course of the disease. The major advantage of lipase is an increased sensitivity in acute alcoholic pancreatitis and in patients who initially present to the emergency room days after the onset of the disease, as lipase remains elevated longer than amylase. Although once considered to be specific for AP, nonspecific elevations of lipase have been reported in almost as many disorders as amylase, thus decreasing its specificity. Simultaneous estimation of amylase and lipase does not improve the accuracy. Other enzymes for the diagnosis of AP--pancreatic isoamylase, immunoreactive trypsin, and elastase--are more cumbersome and expensive and have no clear role in the diagnosis of AP. No enzyme assay has a predictive role in determining the severity or etiology of AP. Once the diagnosis of AP is established, daily measurements of enzymes have no value in assessing the clinical progress of the patient or ultimate prognosis and should be discouraged. A host of new serological and urinary markers have been investigated in the last few years. Their main use is in predicting the severity of AP. At present, serum C-reactive protein at 48 h is the best available laboratory marker of severity. Urinary trypsinogen activation peptides within 12-24 h of onset of AP are able to predict the severity but are not widely available. Serum interleukins 6 and 8 seem promising but remain experimental.