Problems in Diagnosing Allergies

[Please see the revised, expanded, and more readable version of this post here.]

Pediatric Grand Rounds is up, and is worth the read. The Granola Rules did a great job putting together the PGR, and even included a yummy-looking granola recipe. I'm pleased to have one of my posts included.

NHS Blog Doctor's contribution to PGR is about cow's milk allergy and formula marketing. While NHS Blog Doctor makes some good points, the post brought out some things to which I've been giving considerable thought.

One is illustrated by a statement from The British Dietic Association that NHS BlogDoc linked to. It says, "The only reliable way to test for a true allergy is an IgE blood test, a skin prick test or a patch test."

That's a common misconception among doctors and even allergists. The scientific community largely agrees that none of those tests are really all that accurate. Skin prick tests are generally considered to be the most reliable, but even those have only somewhere between a 50% and a 90% accuracy.

Variables such as the tools used, how the test is performed, medications taken by the patient, how recently the person has been exposed to the allergen, recent anaphylactic events, and even the age and quality of allergen extracts used can affect the reliability of testing.

Several studies have indicated that commercial allergen extracts are less effective and reliable in allergy skin-prick testing than a prick-to-prick method (poking the actual food and then the person's arm), especially with fruit.

Incidentally, some experts point out that "false positives" are not really false positives in that they do reliably indicate the presence of IgE antibodies to that substance. In some cases a person may have antibodies with no discernible clinical reaction; in others this could be an indication that the person may develop a more serious allergy with continued exposure to that item.

Test results can provide useful information, but clinical symptoms trump testing.

#############

An elimination diet (taking out a food to see if symptoms subside, and then reintroducing it to see if they return) can be used with some reliability to diagnose adverse food reactions, and is often used in conjunction with other testing to confirm results.

The gold standard for diagnosing allergies is a double-blind oral challenge, giving the person something (when neither the doctor nor the patient knows whether it's the food or a placebo) and seeing how they react. If the person has an onset of symptoms consistent with allergies after consuming the food, they can be considered allergic no matter what the test says.

Severe anaphylaxis that is clearly the result of consuming one particular food usually requires no other testing. A skin prick or oral challenge in that case can be too dangerous. Blood testing can still be used to help confirm the diagnosis if desired.

Unfortunately, too many doctors and allergists believe the hype about allergy testing. Far too many interact with their patients based on statistics rather than on that person's individual history, and assume that if something is rare then it must not be present in their patient.

When confronted with a patient who has an anaphylactic reaction or consistent allergic symptoms when consuming a particular food, but has a negative result to the allergy-testing, too many doctors will tell the patient that they don't truly have an allergy and don't really need to avoid the food.

Unfortunately, this kind of advice could land a patient in the emergency room or the morgue.

I know of someone who "forgets to breathe" when she is exposed to popcorn fumes, along with other symptoms of anaphylaxis, had huge welts on the skin prick test, and yet still has trouble getting doctors to take her allergies seriously.

She says, "Think about this. I have the highest weal scores possible, confirmed reaction to at least one corn product (popcorn) through an ER visit, and yet, I cannot get the majority of the medical community to take me seriously. Is it any wonder others have issues?"

This allergy sufferer took part in a study about the prevalence of corn allergy. It was the Tulane study that's often quoted as the only definitive measure of the prevalence of corn allergy. My understanding is that the study was funded by corn companies.

All of the participants had positive skin prick tests and a clinical history of adverse reactions to corn. Of 22 people that fit the research criteria and agreed to participate, 3 were excluded from the study because of negative skin test results just prior to the test. Of the 19 remaining, 2 did not complete the challenge (perhaps because of the severity of their reactions?).

The researchers defined "allergic reaction" so tightly that they defined many clear typical allergic reactions right out of the study.

"If no objective reactions were observed at the end of the DBPCFC, open challenges were done with corn chips followed by popcorn. Subjects without reactions were considered negative."

Only about 6 or 7 of the 16 common symptoms of moderate to severe allergic reactions listed on Medline would have qualified under their standards. After all, itching, dizziness, difficulty swallowing, chest tightness, nausea and abdominal pain are not empirically observable or quantifiable.

Airway restriction was counted only if it resulted in more than a 15% decrease in peak airflow. Vomiting, diarrhea and congestion were, according to information from a participant, not considered a reaction at all.

The study summary says [definitions in brackets added by me]: "Five patients had positive challenges, two during the blinded portion, three with the open challenge. Reactions included: anaphylaxis, urticaria [hives], conjunctivitis [redness of the eyes] and rhinorrhea [runny nose], perioral urticaria [rash on the face or mouth area] and edema [swelling], and wheezing (>15% decrease in peak flow)."

Several of the ingredients in the study's placebo were also foods that often contain hidden corn (sugar often contains corn syrup, apple sauce and apple juice almost always contain corn-derived citric acid and/or corn-derived sweeteners, etc.), so I wonder if the placebo was truly a good control.

The person I know who took part in the study says that she started vomiting within moments of the first administration of the test product, and the researchers considered that a non-reaction and continued the test. She spent the rest of the night huddled into a ball, crying from abdominal pain and having severe diarrhea and gastrointestinal distress. She got immediate mucus and congestion in her airways and sinuses as well. Because of her history of anaphylaxis when exposed to popcorn fumes, she did not take the risk of participating in the open popcorn challenge.

The participant was categorized in the "no reaction" portion of the study results.

Vomiting and diarrhea are known, widely-accepted symptoms of anaphylactic allergic reaction, and yet these researchers and many doctors (including my own allergist) discount them as allergic symptoms.

This participant said, "As a matter of fact, the nurse said I was definitely highly 'atopic' but was considered not reactive by their standard. Um... atopy is allergy."

If this is the kind of research we have, how accurate do you suppose the figures of allergy prevalence based on such studies (1-2% of adults, 6-8% of children) really are?

Unfortunately the attitude from doctors that allergic reactions aren't real because they're uncommon, don't fit into a certain tightly-defined set of symptoms (usually limited to athsma, hives and airway restriction), or are unverified by a particular test, is all too common.

I've experienced this attitude personally.

My allergist believes that because it's rare for a child to have more than 4 food allergies, some of Baby E's positive skin tests confirmed by clinical observations "must be false positives", because she can't possibly really be allergic to that many foods.

Our allergist insists that many of my daughter's and my own adverse reactions to specific foods must be "coincidence" and refuses to consider even the possibility of an intolerance, much less an allergy. Those reactions include swelling and itching of the mouth and throat, all-over body itching, eczema, rashes (but not the magical giant hives), severe immediate vomiting and/or diarrhea, and other reactions happening consistently after the food is consumed.

This board-certified pediatric allergist believes that because Baby E's positive skin test welts were "not that big" that she obviously "isn't very allergic" to any of the foods she tested positive to. He says that despite the clinical evidence of a severe, immediate reaction (sometimes bordering on anaphylaxis) to the smallest trace of any corn or soy derivative. He has encouraged me to go ahead and give my child things with corn and soy derivatives in them, and was not willing to try to find an allergy medicine that doesn't contain corn or soy for her.

He must not be aware of the research and clinical evidence indicating that the size of the wheal is not directly correlated to the severity of the reaction, and that allergic reactions can take place even with a completely negative skin test. A person can have a very small wheal with a severe anaphylactic reaction to that food.

Also, it is common knowledge that "the wheal size induced by both positive control solutions and allergen-induced prick tests tend to be smaller in infants than adults." Young children are also known to have a higher number of false negatives than adults.

I was told that I wasn't allergic to any foods because none of my welts reached 6/12. I had a positive result of 3/9 to the control, which meant that none of my 5/7 or 4/10 welts "counted" as positive results--despite a clinical history of adverse reactions to those foods which meant I hadn't consumed some of them for a year or more.

Interestingly enough, the control and/or the carrier used in allergy testing often contains glycerine--a substance often derived from common allergens such as coconut, corn, soy or any number of oils derived from vegetable, animal or petroleum sources.

Several studies have found that a glycerine-containing control has a different rate of positive reactions than saline alone. There are documented cases of allergy to saline solution, glycerine, and the preservatives (especially thimerosol) in saline solution.

Allergies to "glycerine" are noted repeatedly both in anecdotes gleaned in a quick web search, and in scientific literature. Glycerin can cause allergic reactions all on its own.

Strangely enough, I could find little if any mention of trying to differentiate between sources of glycerine when an allergic reaction is confirmed. Saying someone is "allergic to glycerine" without looking at the food from which the glycerine is derived is like saying someone is "allergic to oil" because they had a reaction to peanut oil.

My positive reaction to the control was clearly not the result of dermographism, because other pricks registered 0. I don't get a raised welt from being poked with a bare needle.

There is no protocol in place for distinguishing between dermographism, a false positive for other reasons, and an actual allergic reaction to the negative control in allergy skin testing.

It would be so simple and inexpensive to rule out dermographism in such cases just by adding a prick with no substance or with purified water to the test. I don't know why that's not standard procedure.

Most sources agree that any positive reaction to the negative control makes interpreting test results difficult if not impossible. Yet a common standard is "3 mm larger than the negative control" for a positive result. Many allergists don't take into account that the control should be at or very close to 0 for results to be reliable, despite the fact that dermographism is considered by some to be a contraindication to skin testing.

Apparently there is some precedence in the scientific community for using the smallest wheal as the control if the saline or glycerine control has a larger wheal than one or more of the allergens being tested. The study linked above noted: "If the wheal reaction to the glycerin control was greater than the wheal diameter of the allergen reaction, the wheal size for that allergen was set to zero."

In my case, that would have changed the results of my tests, giving me a number of positive reactions instead of all "negative" results as my allergist claimed.

Even with no reaction to the negative control, the statistics I've seen for reliability of skin prick testing range from about 50% to 90% accuracy, or sometimes as low as 15% in accuracy and specificity.

An article published by a Mayo Clinic Doctor in American Family Physician says, "In a study(7) where the gold standard for allergy was a double-blind food challenge to the suspected allergen (e.g., egg, milk, peanut, soy, wheat, or fish), the sensitivity of percutaneous tests was 76 to 98 percent, with specificity ranging from 29 to 57 percent, depending on the food extract used for testing."

Most sources agree that the other methods of testing are even less accurate.

Yet what are the remaining 10% to 50% or more of patients to do when their doctor believes that such testing is the only accurate way to diagnose allergy?

They find themselves in situations like my recent experience: I had a clear anaphylactic reaction to a particular food after multiple occasions of less severe but increasingly obvious allergic reactions to that food. There were no other possible causes for the reaction at that time.

Within seconds of ingesting the food I had severe itching and tingling of the mouth and throat, all-over body itching, tightness of the throat, difficulty swallowing, hoarseness and coughing, among other symptoms also including dizziness, rapid heartbeat, swelling eyelids and immediate intestinal distress.

A liquid antihistamine helped significantly within about 10 minutes of the dose, suggesting that the reaction was histamine-related.

Perhaps because months earlier I had tested "negative" to that food (the welt was only 4/7), my allergist didn't seem to believe I could really be allergic to that food. He wouldn't even see me about the reaction, although he could call in a prescription for an epi-pen "if I really wanted him to, even though he didn't think it was necessary."

He added that if I wanted to, we could do a skin-prick test at some point in the unspecified future "to see if I was really allergic to it."

I suspect that if the skin test came back negative, he would tell me to go ahead and eat it.

On another occasion, this same allergist told me that my daughter's facial swelling and rash after receiving a gas anesthetic was not any kind of allergic reaction "because if it was an anaphylactic reaction, she would have most likely had trouble breathing."

Like many others, he seems to have little frame of reference for an allergic reaction that falls somewhere in the spectrum between nothing and respiratory arrest--at least if systemic hives or athsma are not involved.

A different allergist refused to include rice in the panel of grains she was testing my daughter for, even though Baby E was having symptoms after consuming rice.

Her reasoning? "Nobody is ever allergic to rice," she said. "In all my 30 years of testing I've never had a patient who tested positive to rice, so I just took it off the panel. I don't even test for it."

Of course, people can be and are allergic to rice . . . the prevalence varies in different studies from less than 1% in groups of adults with generalized allergies to 69% in Finnish children allergic to wheat. The prevalence is about 10% or more in areas where rice is frequently consumed, just as Japan.

Rice is most commonly cross-reactive with other grasses and grains. Rice allergy is much more common in patients who are also allergic to wheat, corn, soy or oats. Studies have indicated that 50% of patients with corn allergy, 65% with soy allergy, and 35% with oat allergy were also allergic to rice.

Baby E is allergic to corn, soy and oats, and we suspected allergies to wheat, rye and barley as well. The allergist ran tests for just about every grain under the sun except for rice. Baby E was definitely a case where testing for hypersensitivity to rice (an allergen not included in routine panels) would have been wise. Leaving it off her panel made no sense.

Like too many doctors, this allergist looked at overall statistics about prevalence of rice allergy in the general population instead of considering my child's specific risk factors and symptoms. Wouldn't it be far better to treat the patient as an individual instead of a statistic?

Issues with allergy testing and doctors' attitudes toward atopic patients are bad enough when the patient is experiencing only bothersome symptoms. But the problems with diagnosis become most dangerous in the midst of a life-threatening anaphylactic reaction.

Too many doctors and emergency personnel have an appallingly poor understanding of allergic reactions.

One woman recently had a severe anaphylactic reaction involving loss of consciousness, drop in blood pressure, vomiting, and systemic changes in skin color, among other symptoms. When she woke up she could not get up off the floor to get help, and could have died if her husband had not come home a few minutes later.

The EMTs and then the emergency room doctors kept insisting that what she was experiencing could not be an allergic reaction, despite the patient's allergy alert bracelet and her protestations that she was certain she was experiencing an allergic reaction. Even though she told them the location of her epi-pen, they did not administer it.

When she got to the hospital the ER personnel kept telling her that passing out was not a symptom of anaphylaxis.

The patient did not have hives, swelling or airway restriction (her unconsciousness and weakness was from the drop in blood pressure), so she did not fit the doctors' preconceptions about what a severe allergic reaction looks like.

It took more than two hours before she was finally given epinephrine and antihistamines, or treated in any way for an allergic reaction.

Unfortunately, this type of scenario is all too common. According to studies such as this one, anaphylaxis is grossly underdiagnosed. Even when anaphylaxis is correctly diagnosed, epinephrine and other lifesaving measures go unused in far too many cases.

The medical community and the general public need to be better educated about allergies. We need better systems for diagnosing allergies and for dealing with them. Until then, patients will continue to have difficulty getting allergies properly diagnosed and treated.

[This article is cross-posted at News for Corn Avoiders.]