Monte Minute‎ > ‎Monte Minute‎ > ‎

09/19/2011 - lipase and pancreatitis

posted Sep 28, 2011, 6:02 PM by Chief Resident   [ updated Sep 28, 2011, 7:01 PM by Purnema Madahar ]
Many a morning we talk about the significance of amylase and lipase the ED so freely provides. We don't seem to be the only ones: in the attached paper, investigators in Nottingham, UK, performed a chart review on 1598 patients who had an amylase and lipase assay over a 62 day period. The bulk was ordered in the ED (85.5%), the rest on the surgical unit (14.5%). The authors are unfortunately not very clear on the 'gold standard' for pancreatits but found a diagnosis of acute pancreatitis in 44 charts (2.9%). Of these, 28 were associated with an elevation of amylase and lipase > 3 x upper limit of normal (ULN), the remainder was diagnosed by imaging alone (CT and sono). In 41 patients with lipase > 3 x ULN, no radiologic evidence of pancreatitis was found and other GI causes were deemed causative, e.g. peptic ulcer disease.

This paper (much like any other I could find on the topic) has some obvious methodologic flaws:

1. It retrospectively includes patients who had the test of interest (lipase, amylase) already done. In general, one has to suppose that there was a reason for ordering the test, which is obscured in a retrospective analysis. Ironically, in this particular situation and very similar to our setting, this might be less of a concern: only 6.8% of the tests ordered in the ED were based on the clinical suspicion of pancreatitis!

2. No 'gold standard' was defined, although it seems that in all studied cases a diagnosis was made. However, it's not clear how diagnoses were made, so the potential for misclassification is significant.

Keeping these flaws in mind and recognizing the absence of better literature, one might conclude that:

1. a lipase level > 3 x ULN has a 64% sensitivity and 97% specicificity for acute pancreatitis
2. in the study population, this only translates into a PPV of 41%, but a NPV 99%
3. there is NO additional benefit in testing both lipase and amylase together
4. the indiscriminate use of these tests (similar to our ED), resulted in 115 patients needed be screened for 1 diagnosis of acute pancreatitis at a cost of $800 for lab tests alone!

The last 2 observations are addressed in the second attached paper, which describes an intervention at BI in Boston to reduce co-ordering of lipase and amylase and achieved a 91% reduction of co-ordering, resulting in $350,000/year in patient charges averted. Change IS possible!
The role of routine assays of serum amylase and lipase for the diagnosis of acute abdominal pain.
Paul A Sutton, David J Humes, Gemma Purcell, Janette K Smith, Frances Whiting, Tom Wright, Linda Morgan, Dileep N Lobo (2009). Annals of the Royal College of Surgeons of England 91 (5) p. 381-4

Eliminating amylase testing from the evaluation of pancreatitis in the emergency department.
Kathryn A Volz, Daniel C McGillicuddy, Gary L Horowitz, Richard E Wolfe, Nina Joyce, Leon D Sanchez (2010). The western journal of emergency medicine 11 (4) p. 344-7