R. C. Smith MD, FRACS; J. Southwell-Keely MB BS; D. Chesher PhD, FRCPA.
SHOULD SERUM PANCREATIC LIPASE REPLACE SERUM AMYLASE AS A BIOMARKER OF ACUTE PANCREATITIS?
Article first published online: 8 JUN 2005
ANZ Journal of Surgery
Volume 75, Issue 6, pages 399–404, June 2005
How to Cite
Smith, R. C., Southwell-Keely, J. and Chesher, D. (2005), SHOULD SERUM PANCREATIC LIPASE REPLACE SERUM AMYLASE AS A BIOMARKER OF ACUTE PANCREATITIS?. ANZ Journal of Surgery, 75: 399–404. doi: 10.1111/j.1445-2197.2005.03391.x
- Issue published online: 8 JUN 2005
- Article first published online: 8 JUN 2005
- Accepted for publication 7 December 2004.
- diagnostic accuracy;
- receiver operator characteristic curve;
Background: Serum pancreatic lipase may improve the diagnosis of pancreatitis compared to serum amylase. Both enzymes have been measured simultaneously at our hospital allowing for a comparison of their diagnostic accuracy.
Methods: Seventeen thousand five hundred and thirty-one measurements of either serum amylase and or serum pancreatic lipase were made on 10 931 patients treated at a metropolitan teaching hospital between January 2001 and May 2003. Of these, 8937 were initially treated in the Emergency Department. These results were collected in a database, which was linked by the patients’ medical record number to the radiology and medical records. Patients with either an elevated lipase value or a discharge diagnosis of acute pancreatitis had their radiological diagnosis reviewed along with their biochemistry and histology record. The diagnosis of acute pancreatitis was made if there was radiological evidence of peripancreatic inflammation.
Results: One thousand eight hundred and twenty-five patients had either elevated serum amylase and or serum pancreatic lipase. The medical records coded for pancreatitis in a further 55 whose enzymes were not elevated. Three hundred and twenty of these had radiological evidence of acute pancreatitis. Receiver operator characteristic analysis of the initial sample from patients received in the Emergency Department showed improved diagnostic accuracy for serum pancreatic lipase (area under the curve (AUC) 0.948) compared with serum amylase (AUC, 0.906, P < 0.05). A clinically useful cut-off point would be at the diagnostic threshold; 208 U/L (normal <190 U/L) for serum pancreatic lipase and 114 U/L (normal 27–100 U/L) for serum amylase where the sensitivity was 90.3 cf., 76.8% and the specificity was 93 cf., 92.6%. 18.8% of the acute pancreatitis patients did not have elevated serum amylase while only 2.9% did not have elevated serum pancreatic lipase on the first emergency department measurement.
Conclusion: It is concluded that serum pancreatic lipase is a more accurate biomarker of acute pancreatitis than serum amylase.