Food allergy testing: Do those big panels work?
© 2014 Roy Benaroch, MD
Theresa asked about eczema (see last post), and also asked about food allergy panels. Does all of that testing help?
Well, usually, no.
It turns out that allergy testing (whether done by skin scratches or blood tests) doesn’t even test for allergy. We sort of smudge through that distinction, because it’s complicated. But it’s time for someone to spill the beans about this kind of testing, and here at The Pediatric Insider, we’re all about telling you guys “the real deal”—the inside info that docs typically keep to ourselves. Here’s something you may not want to hear: “allergy” testing is, well, just not very good.
Let’s start basic. An “allergy” is an adverse reaction (a bad thing, some kind of symptom) triggered by an exposure to something, and caused by a reaction of the immune system. It requires symptoms of some kind, and requires the symptoms to be caused by an immune reaction of some sort.
A broader term, “adverse reaction,” includes allergies but also non-immunologic reactions to things. For instance, many people get bloaty and gassey after ingesting milk. That’s typically caused by lactose intolerance, an inability to digest milk sugar that has nothing to do with any allergy. Still, lactose intolerance and gluten sensitivity and many other reactions are sometimes lumped in with allergies. They’re not allergic, and no “allergy testing” of any kind could possibly identify non-allergic reactions.
Also, allergy – an allergic reaction – requires some kind of symptom, and the symptom ought to be reproducible after exposures. Allergy “testing” often reveals “positives” to foods or other things that upon exposure doesn’t actually trigger a reaction. If eating a food you’ve tested positive for doesn’t cause a reaction, you are not allergic to the food. Period. Allergy requires symptoms.
And that’s the biggest issue with allergy testing—because these tests don’t test for allergy. They test for “sensitization”. They show that your immune system has the capacity to react to the substance, but that doesn’t necessarily mean that you will react to the substance.
I can’t tell you how many times I’ve heard from families who tested positive for food XX, though they’ve never had a reaction to XX—and in fact, they used to eat XX all the time. If your child eats peanuts routinely and doesn’t have a reaction, it doesn’t matter what the allergy testing shows. He is not allergic. The test shouldn’t have been done in the first place.
There are other problems with allergy testing. Though individual tests can fairly accurately say who is sensitized, each test does have a chance of a false result. If the rate of an incorrect test is, let’s say, 5%, that might sound pretty good. But what if you do a panel of 40 foods, each of which has a 5% chance of a false result? There’s a 88%* chance you’ll get at least one false positive. If the panel is large, you’re going to get false results. The larger the panel, the more wrong answers you’ll get. What then?
Another problem: there are a lot of ways to do allergy testing, and they’re not all the same, and I most families have no idea what they’re getting into. The current, most reliable blood testing for sensitivity is called a Cap-RAST or ImmunoCap, and it tests for specific IgE molecules. Older tests are far less accurate, and some still in wide use are as worthless as flipping a coin. If you are going to do allergy testing, you should at least do the best one. Skin testing, too, has different reagents and methods, and it can be difficult to know if what’s being done is the most-reliable testing.
About skin tests: they’ll be unreliable if the patient has been taking antihistamines, and they’ll be less reliable if the patient has a skin condition like eczema, which causes high overall IgE levels and a sort of overall hypersensitivity that leads to many false positives on both skin and blood testing. Testing children with eczema is fraught with peril, which is one reason many dermatologists aren’t keen on addressing the possible allergy-eczema connection.
Still, allergy testing can sometimes help, if you keep these points in mind:
- Consult with a board-certified, genuine allergist. Many companies market allergy kits and tests and things for general practitioners and ENTs and who knows who else. They’re a big money maker, but we really don’t know what we’re doing with them or how well they work. Please stay away from alt-med practitioners who claim to be able to diagnose allergies by holding your hands or waving vials about or using some kind of elecromagnetic hyperscience quantumconfusionating walletemptier.
- Test only for foods or other allergens that might be causing a reaction. Foods that are eaten routinely without problems are NOT allergies (and testing will lead only to confusion); foods that always do cause a reactions ARE allergies, and don’t have to be tested. Only test the grey-zone, “maybe” foods, or you’re asking for trouble.
- Test a limited number of foods. Almost all food allergies are to a small number of candidates: milk, egg, wheat, soy, shellfish, fish-fish, peanuts, and treenuts. Tests for mustard and plantain and okra and lamb are unlikely to yield useful results, and might upset the lambs.
- Think of allergy testing as a starting point, for clues to what might be a real allergy (positive tests), and clues for what are probably not causes of allergy (negative tests.) Under most circumstances, unless there’s been a life-threatening reaction, you have to confirm ‘allergy testing’ with deliberate exposures. And if there has been a life-threatening reaction, you probably already know what the trigger was—so why do the test?
* It has been a long time since I did statistics. What I did was figured the 5% chance of a false result on one test means that 19/20 times the test will be correct. So if you’re doing that 40 times I figure the chance of 40 correct tests in a row is 19/20 raised to the 40th power = ~ .12, and the chance of that not happening is 1-.12 = 88%. Am I even close? At The Pediatric Insider we welcome comments that point out boneheaded mistakes. Be gentle.