Doc Talk

Beware unproven allergy tests

Some tests are unvalidated and don't get to real cause of problem

The puzzled look on the parents' faces told me I had a lot of explaining to do. What was the problem they faced, though?

The father of the happy seven-year-old girl took out a sheaf of stapled papers, replete with colourful bar charts and a long list of foods to avoid.

I nodded. I had seen this test enough times to recognise it. Such tests can easily be bought online.

The parents said they decided to send her for allergy testing because of her intermittent tummy pain. They got the results, tried avoiding everything on the list and noticed that there was no difference in the tummy pain anyway.

After we discussed the cause of the pain - something unrelated to allergy - I asked them if I could toss the test results in the bin.

A few years ago, I had alluded to the fact that getting a careful history and examination is the most accurate tool a doctor could employ.

But, sometimes, this tool is not enough. In an allergy practice, there are two types of tests that can help - skin and blood tests for specific Immunoglobulin E (IgE), an antibody. Both tests are proven and accurate. People with an allergy will produce too much of this substance.

I would caution against the use of other tests to check for an allergy and of doctors untrained in allergy giving allergy diagnoses based on any test. There are two good reasons for this.

The first is that many tests, especially allergy tests, do not give straightforward "yes or no" results.


Doctors who specialise in allergy spend years training, reading and following up with patients to become proficient in interpreting tests.

It is like a jigsaw puzzle. These test results are just one piece of the jigsaw. They are interpreted together with other pieces of the jigsaw puzzle - clues in the history and physical examination - to arrive at the answer. Even then, we are not 100 per cent accurate all the time, myself included.

The second reason is that anyone can claim anything on the Internet. Many tests claim to be able to detect allergies. Some examples are Alcat (antigen leukocyte antibody test), bioresonance and IgG (Immunoglobulin G).

There are two good reasons for this (not using other tests to check for an allergy). The first is that many tests, especially allergy tests, do not give straightforward "yes or no" results. The second reason is that anyone can claim anything on the Internet.

Despite the snazzy marketing on the websites, these tests are unproven and unvalidated. This means it has not been shown conclusively that the tests actually look for an allergy.

It is akin to flipping a coin to decide if you have an allergy. The coin has a 50 per cent chance of landing on either side, and whether it is head or tail has nothing to do with the question of allergy.

The most common unvalidated test that patients give me results for is the IgG test. This is a blood test that detects the production of IgG antibodies in response to various types of food and ingredients.

What often comes back is a positive result to a long list of foods. This list is often composed of things the patient has been eating often or interacting with his whole life.

This is no coincidence. IgG is produced by our body upon exposure to various substances.

It is best known in protection against infections. Naturally, the IgG test would provide a list of the things the patient usually eats or is exposed to, and other random substances he has encountered.

In contrast with IgE, which is produced when the body develops an allergy to a substance, IgG is produced when the body is exposed to a substance, regardless of whether you are allergic to it.

However, when the patient is worried about allergy, the point of doing the test is to look for the allergy, not a list of what he has been exposed to in his lifetime.

When the patient is told to avoid everything on that list, it makes his life restrictive and difficult.

From a logical, mathematical or diagnostic point of view, it is impossible to have 10 to 30 culprits causing his health problem. It is more likely just one or two culprits.

Many patients find it impossible to avoid so many substances and foods. So, they start eating some of the foods again and realise there is no difference in their illness, whether they avoid the foods totally or ate half the things on the list.

Others doubted the results (correctly) as the "culprits" were far too many, while noticing that the items on the list happened to be things they ate most of the time, often without any problem at all.

In avoiding the whole list, the patient may feel better temporarily - perhaps one item on the list was indeed causing a reaction.

Another possible explanation is the placebo effect, where the belief that something will help, even if it actually doesn't, makes the patient feel better anyway.

On rare occasions, asking for a patient's history and skin tests did reveal one single culprit hiding among the list of 10 to 30 foods. In such situations, it would still have been better to do the validated, proven allergy test. This way, the correct culprit could be picked out from the outset and patients need not avoid an endless list of foods.

At best, unproven tests create unnecessary restrictions. At worst, they can be dangerous.

One of my patients with a peanut allergy went to see someone who did an unproven test on him. The child was told he was not allergic to peanuts. But he later had a severe allergic reaction and had to go to a hospital emergency department for treatment.

I am glad he is still alive, as a severe allergic reaction can kill.

  • Dr Soh Jian Yi is a consultant in the division of paediatric allergy, immunology and rheumatology at the National University Hospital.
A version of this article appeared in the print edition of The Straits Times on June 13, 2017, with the headline 'Beware unproven allergy tests'. Print Edition | Subscribe