It’s a busy shift in good old A pod. A nurse comes up to you with a bag of blood tubes for a vomiting patient and asks if you want to shotgun labs and give some anti-emetics. You happily oblige, ordering a liter of normal saline, Zofran, and some labs. But what to order? A CMP to look at LFT’s and electrolytes. A lipase for pancreatitis. Type and Screen with coags in case the patient needs surgery? A CBC…just because. Maybe an EKG/CXR for possible angina equivalent of vomiting and cxr to look for air under the diaphragm.
But what about an Amylase? I mean, I know it goes up in pancreatitis, but, we are already getting a lipase. Are we just ordering two tests for the same purpose? Would this be equivalent to ordering a V/Q scan at the same time we order a CTPA? Is the lipase enough? Or do we need both tests to increase our diagnostic power?
What does “the literature” say about the value of an amylase in the work-up of pancreatitis?
Let’s First Look at Amylase
Paper 1: How sensitive is amylase in the ED for pancreatitis? In a paper diagnosing etoh pancreatitis with CT or US – 32% of patients had a NORMAL amylase!!!!
Another paper from 1989 (back when I was 6) found 20% of people with pancreatitis on CT had normal amylase in the ED on their arrival. Fortunately, of those 20%, 88% had an elevated lipase. The same study analyzed peritoneal lipase and amylase, which firstly I haven’t heard of and secondly was shown here to be no more sensitive or specific than blood lipase. So no need to tap their pancreases in the ED.
Conclusion: Amylase may not have a high sensitivity.
Problems with specificity
Serum amylase is elevated often in celiac disease, HIV, lymphoma, UC, RA, and monoclonal gammopathy, renal failure, ectopic pregnancy, salpingitis, anorexia/bulimia, parotitis, bowel obstruction/infarction, tumors involving/inflitrating the salivary glands, amylase producing tumors. Numerous studies have discovered that amylase has much less sensitivity if symptoms for greater than 24 hours, alcoholic pancreatitis, or pancreatitis from hyperlipidemia. Amylase sensitivity is increased specifically in gallstone pancreatitis, where it is elevated more often and more significantly than in other forms of pancreatitis. I have been told (but found no evidence for this), that amylase levels can go up from vomiting alone.
Unfortunately lipase is elevated in HIV, DKA, bowel obstruction/infarction, renal failure. Shorter, but still not 100% specific.
Conclusion: Specificity is not great for amylase. According to many studies, lipase ranges from more specific to no significant difference in specificity. Lipase is at least as good, possibly better.
What about comparing Amylase to Lipase
A retrospective chart review found over 1000 patients with the diagnosis of pancreatitis. Receiver operator curves showed improved diagnostic accuracy of lipase over amylase using lipase >208, (nml < 190). 20% of patients did not have elevated amylase on their initial ED visit, while only 3% did not have an elevated lipase.
Numerous papers have compared the sensitivities of lipase vs amylase – 95% vs 79%, 100% vs 72%, 100% vs 55%. And “the cumulative literature supports the replacement of amylase with lipase”.
Why?
Theory has it that the amylase levels rise as early as lipase does, but drops off much faster (some papers say in <24 hours). Also, when the pancreas gets burned out from chronic inflammation, it stops producing amylase much sooner than when it stops producing lipase. One study found a 91% decrease in amylase production vs a 26% decrease in lipase production in patients with chronic pancreatitis
So lipase is better than amylase. But are both better than lipase alone?
A study (that also came out when I was 6), found lipase to have increased sensitivity over amylase, and discovered that adding the amylase to the lipase “does not enhance discrimination of acute pancreatitis”.
Another study, compared amylase vs lipase in pancreatitis diagnosed by CT showed lipase to be an in general better test. Here they looked at the curves of sens/spec that we learned year 1 of medical school. On day one, lipase had better sensitivity/specificity. On day two of hospitalization, amylase became a much worse test, for example: lipase specificity of 82% vs 68% for amylase.
Both studies concluded that adding amylase to the lipase on the initial evaluation was not helpful.
A review in 1999 concludes “the obtaining of both serum amylase and lipase levels is not warranted.
The next step- STOP ORDERING AMYLASE
This information is apparently so common knowledge that there are actually papers published on the best way to stop ER doctors from ordering amylase. One analysis showed 93% of patients with lipase ordered had amylase ordered as well. An educational intervention was done which included a lecture during conference and another during faculty meeting, which both instructed physicians not to order an amylase in the evaluation of pancreatitis. After the meetings, ordering an amylase with a lipase decreased a whopping 2 % – down to 91%.
Finally, this hospital removed the amylase check box from the order sheet, making ED physicians write in amylase if they wanted it ordered. This intervention lowered the amylase ordering rate to 25%. Removing it from the pre-set “trauma labs” labs lowered it even more – to 14%.
At their shop, the amylase test costs 58$, but with health care the way it is today, who knows how much the patients got charged?
Conclusion
We have two tests for pancreatitis. Lipase is better and adding an amylase to the lipase does not help. Amylase tests cost money. Education interventions such as lectures (or writing a blog about it), does not seem to reduce ED physicians ordering amylase. A more effective stratagem would be to make it less easy to order.
You do not order a V/Q scan and a CTPA together for PE. Why would you order an amylase and a lipase together for pancreatitis?
Resources
Spechler SJ, Dalton JW, Robbins AH, Gerzof SG, Stern JS, Johnson WC, Nabseth DC, Schimmel EM. Prevalence of normal serum amylase levels in patients with acute alcoholic pancreatitis. Dig Dis Sci. 1983 Oct;28(10):865-9.
Smith et al. Should serum pancreatic lipase replace serum amylase as a biomarker of acute pancreatitis? Anz J Surg. 2005:75(6):399-404
Yadav et al. A Critical Evaluation of Laboratory Tests in Acute Pancreatitis. The American Journal of Gastroenterology. 2002; 97:1309-18
Wener H et al. Strategic use of individual and combined enzyme indicators for acute pancreatitis analyzed by receiver operator characteristics. Clin Chem 1989; 967-71
Keim V et al. A comparison of aylase and lipase in the diagnosis of acute panceratitis in patients with abdominal pain. Pancrease 1998;16:45-9
Vissers et al. Amylase and Lipase in the Emergency Department Evaluation of Acute Pancreatitis. Journal of Emergency Medicine. 1999;17(6): 1027-37
Volz et al. Eliminating Amylase Testing from the Evaluation of Pancreatitis in the Emergency Department. Western Journal of Emergency Medicine. 2010; Volume XI (4): 344-347
andygrock
- Resident Editor In Chief of blog.clinicalmonster.com.
- Co-Founder and Co-Director of the ALiEM AIR Executive Board - Check it out here: http://www.aliem.com/aliem-approved-instructional-resources-air-series/
- Resident at Kings County Hospital
Latest posts by andygrock (see all)
- A Tox Mystery…. - May 26, 2015
- Of Course, US Only for Kidney Stones… - May 18, 2015
- Case of the Month 11: Answer - May 12, 2015
- Too Classic a Question to Be Bored Review - May 5, 2015
- Case of the Month 11: Presentation - May 1, 2015
Nice job, Grock.
I’ll take THIS one step further. With a reliable history/physical, basic ECG interpretation and bedside sonography skills, you don’t need a lipase or any of those other serum tests for that matter. If the patient denies EtOH abuse and has no associated stigmata on exam, bedside sonography shows no cholelithiasis and normal CBD, and ECG shows normal intervals (thus suggesting lack of significant electrolyte abnormalities), then you can reassess the patient after symptomatic treatment (if you’re working with me this means, NO INTRAVENOUS PPI – pure waste of money). You can then take some dark pleasure (as I sometimes do) in chucking that bag of blood tubes in a (red) trash can, and instead ask that eager nurse to straight catheterize her other 85 year-old with altered mental status.
Let me throw our ED some praise: as a Downstate MS3 I learned not to order amylase! As an intern, I don’t think I’ve ordered it once.
Nice review! Anyone know if amylase is used for any other purpose? If not, why do labs still do the test?