--Doctors urge caution in diagnosis based on test results
Blood tests are becoming an increasingly popular tool in the diagnosis of food allergies, but a study led by scientists at Johns Hopkins Children’s Center reports that some tests are more accurate than others and that too-heavy reliance on blood tests alone to predict allergic reactions may not be a good idea.
Unlike food challenge testing — in which people thought to have a food allergy are fed small amounts under a doctor’s supervision to directly measure an actual allergic reaction — blood tests measure IgE antibodies, immune system chemicals involved in allergic reactions.
In a study of the three most common commercial tests, the Johns Hopkins research concluded that “some tests are more accurate than others.”
In the study, researchers sent patient blood samples with already-known levels of antibodies to either soy or peanut to labs using the three most popular systems.
They discovered substantial differences in the tests’ ability to detect the antibodies, as well as to accurately measure the amount of antibodies, according to a report in the July 15 issue of Annals of Asthma, Allergy & Immunology.
“The differences we saw in sensitivity and accuracy are clearly a red flag because reliance on results may lead to a wrong diagnosis,” says the study’s lead author, Robert Wood, M.D., director of Allergy and Immunology at the Johns Hopkins Children’s Center.
In recent years, blood tests have emerged as an alternative to such diagnostic mainstays as food challenges or skin testing.
Wood says doctors and patients should approach all blood test results with caution and “only in the context of a patient’s symptoms and clinical history.” Food challenges and skin testing still have a role in the diagnosis of food allergies, he says.
Researchers studied and compared the three most widely used commercial blood tests, ImmunoCap, Turbo RAST and Immulite. Using ImmunoCap as a reference because of its well documented past performance, Immulite failed to pick up peanut antibodies in three out of 50 positive samples, while Turbo RAST failed to detect peanut antibodies in 11 out of 50 positive samples. Immulite failed to detect soy antibodies in one out of 17 positive samples, while Turbo RAST failed to detect soy antibodies in five out of 17 positive samples.
Moreover, ImmuLite tended to overestimate the amount of antibodies, while Turbo RAST found fewer antibodies than were actually present. Overestimating the number of antibodies could lead a physician to believe that a person is allergic when he or she is not, while underestimating the antibodies could mean that no allergy would be diagnosed where there is one.
Different commercial labs use different assays, and different physicians use different labs depending on which labs the patient’s insurance contracts with. In the absence of a uniform testing system, labs should be clear about which test they use, and health-care providers need to stay up to date on test reliability.
More than 11 million Americans, mainly children, are estimated to have food allergies, principally to milk, egg, peanut, and soy.
Phadia AB, Sweden, manufacturer of ImmunoCap, paid for the lab analyses.
Other investigators in the study were Nathan Segall, M.D., private practice in Atlanta, Ga.; Staffan Ahlstedt, Ph.D., Karolinska Institute, Stockholm, Sweden, also working for Phadia AB; and P. Brock Williams, Ph.D., University of Missouri-Kansas City.