College of American Pathologists

  In coagulation, reflexive testing algorithms do the trick


CAP Today




September 2010
Feature Story

Karen Lusky

Lab utilization management (see “Pulling back the reins on superfluous testing,”) has a clear role to play in helping clinicians with coagulation testing.

As an example of what can happen without order entry support for clinicians, Massachusetts General Hospital’s Kent Lewandrowski, MD, presented a copy of a coagulation testing requisition during his presentation at the AACC annual meeting. Most of the boxes on the requisition had been checked, indicating the clinician was clueless about what to order and was thus “trying to catch a fish,” he said. The requisition included $3,000 to $4,000 worth of testing in what he described as a shotgun approach.

Using an algorithmic approach, clinicians can order tests in sequence or let the lab cut to the chase in figuring out what’s causing the patient’s coagulation problem.

As an example, Massachusetts General offers a prolonged PTT reflexive testing algorithm that allows the clinician to check one box on the requisition. Then the lab does a series of tests, starting with one to determine if heparin is the culprit. If it is not, the next step is a two-hour incubated PTT mixing study (combining a 1:1 ratio of the patient’s specimen and normal plasma). If the PTT normalizes, the lab measures factors VIII, IX, XI, and XII. If the PTT continues to be lengthened, the lab would test for an inhibitor, which would most commonly be a lupus anticoagulant. If the PTT normalized initially but became prolonged again, the lab would test for a factor VIII inhibitor. However, the laboratory approach is not always that simple: Sometimes the mixing study corrects despite a lupus anticoagulant, and factor inhibitors can cause false-positive lupus anticoagulant tests.

“Without this prolonged PTT algorithm,” says Massachusetts General’s Elizabeth M. Van Cott, MD, “clinicians tend to order just pieces of the puzzle without getting the complete answer. Or they erroneously order tests that are unrelated to the PTT, such as factor VII.”

As another possibility, a clinician might just order factor VIII and a lupus anticoagulant test, which would not provide a complete answer to why the patient’s PTT is prolonged, says Dr. Van Cott, director of the coagulation laboratory and medical director of the core laboratory. That would be particularly true, she says, if both of those tests were normal.

Using the algorithm, however, the lab can provide the clinician an answer “without performing unnecessary tests, and from a single blood draw,” avoiding delays.

The possible causes of a prolonged PTT, which Dr. Van Cott can rattle off at the drop of a hat, include “a deficiency of factor VIII, IX, XI, XII, prekallikrein, or high-molecular-weight kininogen, a lupus anticoagulant, heparin, direct thrombin inhibitor therapy, such as argatroban, hirudin, or bivalirudin, or a specific factor inhibitor, i.e., an antibody against factor VIII, IX, XI, or XII.”

Reflexive testing algorithms are also useful for coagulation tests where clinicians can choose either an activity or antigen assay. That includes testing for protein C, protein S, and antithrombin to look for hypercoagulability. Without an algorithm, clinicians or labs who send out for such tests tend to choose antigen testing instead of starting with activity testing, Dr. Van Cott observes.

At Massachusetts General, when a clinician orders protein C, protein S, or antithrombin, the laboratory performs an activity test, reserving the antigen test for situations in which the activity is low, and the patient has no evidence of an acquired etiology affecting coagulation (liver dysfunction, warfarin, or vitamin K deficiency, for example).

If it appears the patient has an acquired etiology, and if more than one of the three proteins (protein C, protein S, or antithrombin) is low, the lab does not routinely perform antigen testing, and instead the pathologist interpretation suggests the clinician repeat the activity test when the patient has recovered from the condition causing the problem, if possible.

An activity test can detect both qualitative (type II) and quantitative (type I) deficiencies, whereas an antigen assay can detect only a quantitative (type I) deficiency. (Qualitative means the protein is present but nonfunctional; quantitative means the protein is reduced in quantity, Dr. Van Cott says.)

Generally, determining whether someone has a type I or type II deficiency for protein C, protein S, or antithrombin isn’t all that clinically important, but there are a couple of caveats to consider, Dr. Van Cott says. For one, when the decrease is quantitative, obtaining a low antigen result helps confirm that the low activity result is valid.

“In type II antithrombin deficiencies, a normal antigen result might alert the clinician to consider a ‘heparin-binding’ variant, which is typically asymptomatic in the heterozygote. Other type II antithrombin deficiencies are symptomatic in the heterozygote.”

By contrast, when testing someone for von Willebrand disease, a hereditary bleeding disorder, it is important from a treatment perspective to know if the patient has a type 1 or type 2 deficiency, Dr. Van Cott says. And to determine the type of deficiency, the clinician has to order both activity and antigen testing.

“Type 2 deficiencies are often more symptomatic, and treatment approaches can vary considerably. For example, type 2 deficiencies are more likely to fail to respond to DDAVP [desmopressin] therapy,” Dr. Van Cott says. “And some experts consider DDAVP contraindicated in a particular type 2 deficiency called type 2B deficiency.”

At Massachusetts General, the clinician can order a von Willebrand panel to get a complete answer about whether a patient has von Willebrand disease and, if so, the type. The panel includes the von Willebrand activity test (the ristocetin cofactor) and antigen tests, as well as the following, says Dr. Van Cott:

  • Factor VIII because factor VIII becomes low when von Willebrand factor is low.
  • Blood type because von Willebrand levels vary with blood type.
  • Fibrinogen because von Willebrand factor, factor VIII, and fibrinogen become elevated during an acute phase reaction, potentially masking a diagnosis.

In other words, it gets complicated. The preceding testing algorithms also include a pathologist’s interpretation, which a study at Massachusetts General shows decreased the number of incorrect test orders significantly (Laposata M, et al. Arch Pathol Lab Med. 2004;128:1424–1427).

Karen Lusky is a writer in Brentwood, Tenn.

Related Links Related Links