College of American Pathologists
Printable Version

  Allergy testing: from skin to tube to chip


CAP Today




December 2009
Feature Story

William Check, PhD

In the personal communications arena, there is a great range of options based on varying degrees of electronic sophistication: landlines connected to conventional handsets; basic mobile phones; BlackBerries that provide Web ­access; the iPhone; and voice-over-Internet protocols. Users can choose the optimal mode for their own needs and budgets.

In the field of allergy testing, too, the current range of technologies covers a wide spectrum. Office-based physicians often use prick or puncture skin testing, a technique that dates from 1880. In laboratories, in vitro serologic tests use the EIA method to measure total IgE and specific IgE antibodies that react with extracts from allergenic substances, a method based on the discovery in the 1960s of IgE and its role in allergic reactivity. And for the past several years researchers have been developing and validating a component-based method to help diagnose allergies: detection of IgE antibodies to recombinant or highly purified individual protein molecules, rather than the complex mixture of substances contained in an extract. Molecular components are typically immobilized on a micro­array chip, called Immuno Solid Phase Allergen Chip, or ISAC.

In the past 40 years, progress in measuring total and allergen-specific IgE by in vitro testing has been “remarkable,” says Robert G. Hamilton, PhD, DABMLI, professor of medicine and pathology at Johns Hopkins University School of Medicine and director, Johns Hopkins Dermatology, Allergy and Clinical Immunology Reference Laboratory. IgE was defined in 1967 as a unique immunoglobulin class that was “the gatekeeper of the allergic response,” Dr. Hamilton says. The first assay for total and specific allergen-related IgE became available in 1968. Since then these assays have been improved by increased use of automated analyzers, enhanced quantitation of IgE antibody measurement for more than 200 allergen specificities, and analytically more sensitive assays. “Technology has brought measurement of IgE antibodies to a high level of quality,” Dr. Hamilton says. “You can see that in College of American Pathologists’ proficiency Surveys.”

Most recently, there has been investigation of component allergens on a chip, a technology Dr. Hamilton calls “the new kid on the block.” Most of this work is being done in Europe. One pioneer investigator in this field is Adriano Mari, MD, chairman of the European Academy of Allergy and Clinical Immunology allergy diagnosis interest group and head of the Center for Clinical and Experimental Allergology at IDI-IRCCS, Rome. “We are moving from extract-based to component-based analysis of allergy,” says Dr. Mari, who has worked with molecular components of allergens for 15 years. “We need to remove as much of the variation from allergy testing as we can.” Because each allergen extract contains many individual reactive molecules, comprehensive testing with components requires many more reagents per allergen than testing with whole extracts. “We can’t do that with standard methods,” Dr. Mari says. “So we need to go to multiplex platforms. We need to transition from extracts to molecules and from singleplex to multiplex testing.”

Pathologist Henry A. Homburger, MD, who has observed and participated in many important changes in allergy testing during his career, provides a historical perspective on these advances. “I’ve watched this field for over 30 years,” says Dr. Homburger, professor emeritus in the Department of Laboratory Medicine and Pathology at Mayo College of Medicine and former director of the immunology and antibody laboratories at Mayo Clinic. “I spent my career directing the laboratory that implemented in vitro allergy testing.” To Dr. Homburger, “It is clear that allergists in the U.S., by virtue of their training, oftentimes prefer to use skin testing for diagnosis. For many years there was a question of how well in vitro tests compared to skin testing. For the last 10 years or so it has been quite clear that in vitro test reagents give results comparable to skin testing.

“What has happened,” Dr. Homburger continues, “is that there have been changes in health care delivery such that primary care physicians are often asked to perform the initial evaluation of patients with suspected allergy. And they have begun increasingly to use in vitro tests.”

Dr. Homburger emphasizes the technological advances in IgE testing, saying that in vitro diagnostic methods have gotten better in the past several years. “With currently available commercial methods we have the ability to quantify specific IgE antibody. Also, there has been a lot of progress in coming up with better preparations of allergens such that results are more consistent from lot to lot and batch to batch.” As the latest in this progression, Dr. Homburger cites the use of recombinant DNA technology to prepare quantities of individual allergen proteins that make up complex allergens, which are then arrayed on a chip, such as ISAC. He is familiar with this technology as medical director of the Phadia Immunology Reference Laboratory, launched earlier this year by Phadia, a co-developer of ISAC. “This advance should help us greatly in years to come to distinguish between sensitization and cross-reactivity and to identify antibodies associated with persistent allergy and severe allergic disease,” he says. “It is anybody’s guess how rapidly that will occur in the U.S. It appears that it will occur more rapidly in Europe.”

Drs. Hamilton and Mari spoke this past July at an AACC symposium on “The Role of the Laboratory in the Diagnosis and Management of Human Allergic Disease.” In their talks they evaluated the pros and cons of various kinds of allergy testing and how the laboratory can help the clinician diagnose this complex condition. “Clinical history drives the diagnosis of allergy,” Dr. Hamilton said. “The clinician attempts to relate symptoms to exposure.” While diagnostic test results are helpful, he said, they cannot replace a carefully collected clinical history. Detecting allergen-specific IgE antibody, whether by skin test or serology, “serves strictly as confirmation for a clinical history-driven diagnosis.” By identifying specific IgE antibodies, laboratory serology can also help in planning treatment. And, in the case of food allergy, the concentration of specific IgE antibody has predictive value.

A history of a patient presenting with apparent allergic symptoms often elicits evidence for atopy—a genetic predisposition to develop sensitization to allergens, usually early in life, and generally characterized by one of several clinical conditions, such as atopic dermatitis, asthma, or allergic rhinitis. However, “Individuals who are not atopic can develop an allergic response, too,” Dr. Hamilton cautions. He himself illustrates this phenomenon: He has no atopic conditions, and yet, within a few years of becoming a beekeeper in 1980, he became sensitized to honeybee venom from repeated stings.

Dr. Hamilton notes that a laboratory that offers allergy testing may perform other tests in addition to total and specific IgE, such as mast cell tryptase, IgG antibody assessment for Hymenoptera venom, and aeroallergen quantitation for mite, cat, and dog. Tryptase may be ordered for insect sting-induced anaphylaxis: Total tryptase (α + β) reflects mast cell number, while β tryptase indicates mast cell degranulation, a hallmark of anaphylaxis. For this assay, serum must be collected between 30 minutes and four hours after the onset of a suspected reaction.

As for IgG, Dr. Hamilton and his colleagues showed some years ago that 3 µg/mL of venom-specific IgG or greater identifies a person with a reduced risk of reaction to stings. “IgG antibody is necessary but not sufficient for protection,” he cautions. Measuring specific IgG antibody also helps monitor and guide immunotherapy: Successful aeroallergen immunotherapy is almost always accompanied by elevated allergen-specific IgG.

Current clinical systems used in the U.S. for serologic IgE assays include Immulite 2000/2500 (Siemens Healthcare Diagnostics), ImmunoCAP 250/1000 (Phadia), and Hy-Tec (Hycor-Agilent). Phadia is working with VBC Genomics on the Immuno Solid Phase Allergen Chip. “All [of these platforms] have strengths and weaknesses, and all have an important role to play,” Dr. Hamilton said.

All three systems are based on the same biochemistry but use different solid supports. More important, all three measure different populations of IgE antibodies, according to a 2007 publication (Wood RA, et al. Ann Allergy Asthma Immunol. 2007;99:34–41). In that study, the three systems showed “extensive variability” in quantitation, Dr. Hamilton says. However, he adds, “Since 2007 we’ve come a long way.” As evidence, he cites results of two recent College of American Pathologists Surveys (SE-C-08 and SE-A-09) in which about 160 licensed diagnostic allergy laboratories in North America took part. For total IgE, agreement among the assays was “excellent” by most criteria, including precision, Dr. Hamilton reported in his AACC talk. Analytical sensitivity was comparable at about 2 IU/mL. For allergen-specific IgE, agreement among the assays was also excellent for precision, reproducibility, and linearity. However, there was marked variability in quantitation among methods. “For IgE antibody testing, there is still a lot of discrepancy among these three methods in terms of what they measure,” Dr. Hamilton told CAP TODAY. (Data have been accepted for publication in Archives of Pathology & Laboratory Medicine.) This marked variability “may be ameliorated with component allergen use,” he suggests.

Dr. Hamilton endorses the recommendation of the 2007 study authors: “It is important that all laboratories clarify which system they are using. Just because two systems present their results in the same units does not mean the results are necessarily correct or interchangeable.”

According to Robert W. Reinhardt, MD, U.S. medical affairs director at Phadia and associate professor of medicine at Michigan State University College of Medicine, “Phadia ImmunoCAP technology is widely utilized in research and is far and away the market leader in clinical laboratories in the U.S.” Dr. Reinhardt’s assertion is supported by the distribution of participants in the CAP Survey: About three-quarters used an ImmunoCAP system. Dr. Reinhardt adds, “ImmunoCAP looks very good when compared in head-to-head studies with all other allergy testing systems, even the acknowledged gold standard, challenge testing.” He cites the 2007 Wood study and other publications (for example: Wang J, et al. J Allergy Clin Immunol. 2008;121:1219–1224; Wood RA, et al. J Allergy Clin Immunol. 1999;103:773–779).

In the College’s Survey eight participants used Hycor’s medium-through­put instrument, the 288. “We are encouraging more of our customers to participate in the quantitative IgE part of the Survey so those numbers more accurately represent our presence in the U.S. market,” says Skip Thune, general manager for the Hycor division of Agilent. Thune estimates Hycor’s share at about 14 percent. Because Phadia systems dominate this market and because there is no recognized independent gold standard for specific IgE quantitation, Thune says, “People think of Phadia as the gold standard. This raises issues for clinicians. We have been working hard to make sure that our results correlate well with their system.”

At the AACC meeting, Hycor-Agilent launched its Ultra-Sensitive EIA System for allergen-specific IgE, along with a new high-throughput instrument, the Hy-Tec 1536. The 1536 runs more than 1,500 assays on each cycle on up to 238 patient serums.

Regarding the Siemens systems, Debra Hovanec-Burns, PhD, vice president of allergy diagnostics, Siemens Healthcare Diagnostics, points out that Immulite platforms are “unique” in also being able to cascade from panel test results by reflex testing to specific allergens or, alternatively, from specific allergens to individual allergenic components. In addition, allergy and non-allergy EIAs, such as thyroid, tumor markers, reproductive hormones, and infectious disease markers can be simultaneously performed on the Immulite systems.

Dr. Hamilton’s laboratory is the only one of its kind in the U.S.—a facility focused on translational research related to allergy testing. It also acts as a reference laboratory. “It is more common to have allergy testing done in a general immunology laboratory,” Dr. Hamilton says. “Once a laboratory director learns the principles, it is not usually a problem. On the other side of the coin,” he continues, “there are very few general immunology labs where they have knowledge of allergen specificities and extent of cross-reactivity. There is minimal interpretation that a lab can do for allergen testing. A positive is a positive, but interpretation of a positive falls into the realm of the allergist or clinician.”

As an example he cites birch tree pollen, which contains a number of very allergenic molecules. One of them, Bet v 1, has structural similarity to molecules in many other substances, such as Cor a 1, the major allergen in hazelnut. “This can be very important to a pediatric allergist,” Dr. Hamilton says. A positive reaction to Cor a 1 could be interpreted as sensitivity to hazelnut but may actually be due to cross-reaction to Bet v 1.

For two major types of allergy—food and respiratory—serologic assays and skin testing are both acceptable methods. “It is a remarkable step forward that we have been able to get serum IgE testing accepted as equivalent to puncture skin testing for respiratory allergy,” Dr. Hamilton says.

For food allergy, serum testing has two technical advantages over skin testing: Food extracts are complicated and sometimes unstable and, Dr. Hamilton says, “children don’t like to be prodded and poked, so one blood draw is often better accepted than multiple skin tests.”

Another, more important strength of serology for food allergy is its clinical utility. “The perception of having food allergy is much more common than having an IgE-mediated clinical event provoked by food,” Dr. Hamilton says. “About 20 percent of people have some sort of intolerance or problem with food, but only three to six percent have IgE-mediated food allergy.” Evidence shows that increasing concentrations of specific IgE antibody predict which children have symptomatic allergy to egg, milk, peanut, and fish (Sampson HA. J Allergy Clin Immunol. 2001;107:891–896) as well as the probability of wheeze and reduced lung function (Simpson A, et al. J Allergy Clin Immunol. 2005;116:744–749; Marinho S, et al. Allergy. 2007;62:1379–1386). “The ability of IgE to predict which individuals have a high probability of a positive challenge test has been a tremendous step forward,” Dr. Hamilton says. “It can sometimes eliminate the need to do a provocation test.” He notes that this type of data has been published only for ImmunoCAP. As Julie Wang, MD, and her colleagues wrote, “The predictive values for clinical reactivity associated with food-specific IgE levels determined by ImmunoCAP should not be applied to results from other assays” (J Allergy Clin Immunol. 2008;121:1219–1224).

Ironically, articles in major U.S. newspapers published this year cast doubt on this exact value of serologic testing. One writer claimed that “Many doctors, especially primary care doctors who aren’t allergy specialists, are far more likely to do blood tests, which are much less accurate and more difficult to interpret” and that this trend has led to misdiagnoses of food allergies. Dr. Hamilton counters: “Basically the writer got it wrong. Serologic data are often not the problem. It can be misinterpretation of data by allergists. Knowledgeable allergists know that IgE antibody can be present in the absence of clinical manifestations of allergic disease. Just because IgE antibody is positive doesn’t mean the patient will have a problem with that food.”

In the case in question, based on positive IgE tests an allergist diagnosed a two-year-old boy with allergies to nuts, eggs, wheat, beef, peas, rice, and corn and told the mother to eliminate them all from the child’s diet. “The child was basically subsisting on potatoes, fruit, and hypoallergenic formula,” Dr. Hamilton says. The distraught mother consulted a second allergist, who had experience with food allergy and restored all the foods except eggs and nuts. “The IgE antibody test was blamed for being unreliable when it was actually incorrect interpretation by an allergist,” Dr. Hamilton says. “We can make the IgE antibody measurement as accurate as possible, but we can’t interpret the result. That would be practicing medicine.” It is the responsibility of the allergist to interpret the test result. Unfortunately, Dr. Hamilton says, “Some allergists don’t have enough expertise in food allergy to interpret test results correctly, so if they see any positive results they will have the patient eliminate that food. An educated allergist knows to base a decision on serology plus history.”

In contrast, specific IgG and IgG4 antibody testing is generally not useful in diagnosing food allergy, Dr. Hamilton says. “IgG antibody test results do not correlate with the results of double-blind placebo-controlled food challenges. We have companies trying to get an FDA claim for measuring IgG antibody for diagnostic purposes, but the FDA fortunately won’t approve them.”

While the value of serologic testing by EIA using whole allergen extracts has been established, the molecular technique, which Rome’s Dr. Mari calls component-resolved diagnosis (CRD), typically done on a microarray chip, is still in its developmental stages. Dr. Mari points out that the immune system recognizes only a small part of any molecule, called an epitope, which is usually no more than five or six amino acids long. “We need to look carefully at what interacts with the immunoglobulins we are interested in,” he says. “As allergists and immunologists we are interested in IgE. We are just in the beginning of this recognition story.” As a first practical step, Dr. Mari and many others are isolating individual molecules that react with allergen-specific IgE antibodies and making them in pure form, often by recombinant DNA technology.

One advantage of this molecular approach is the potential to reduce variability. “When we use material from living sources, the quality of the antigenic extracts can vary, which affects the skin test,” Dr. Mari says. For instance, molecular analysis showed that the amount of the main apple allergen, Mal d 3, varies widely among 88 cultivars (Sancho AI, et al. Int Arch Allergy Immunol. 2008;146:19–26). Also, cat allergen extracts vary depending on whether they are obtained from a male or female. When Dr. Mari reviewed this field recently, he found more than 50 articles published since 2006 on differences in extracts relating to the source.

Peanut is a widespread and important allergen. Dr. Mari joined with a group of German investigators to determine whether a potent allergenic molecule called a lipid transfer protein that is found in many foods and pollens is present in peanut. It was not found with conventional extraction techniques, but it was if they did an acidic extraction (Krause S, et al. J Allergy Clin Immunol. 2009;124:771–778). Moreover, when the peanut lipid transfer protein, named Ara h 9, was included as a single component in CRD analysis, it turned out to be important for detecting sensitization in peanut-allergic individuals in Southern Europe. “Extracts are variable and depend on the method you use for extraction,” Dr. Mari reiterates. “What is important is that we recognize all the allergenic molecules in the source.”

A technical advantage of the Immuno Solid Phase Allergen Chip is that it is a micro-method and requires only small amounts of material. “One milligram of molecules can make 10,000 chips,” Dr. Mari says. However, using molecular components demands constant vigilance. With antigens derived from natural sources, purity must always be monitored. A recombinant molecule, on the other hand, is not always the same as a natural molecule. “We had a bad experience with a ragweed antigen provided by a commercial source,” Dr. Mari says. It gave no positive test results. Eventually, the company acknowledged that the antigen was not good. “So we must double-check all antigens put onto the chip,” Dr. Mari stresses. Also, not all component antigens can be made by recombinant methods. An example is the common worldwide allergen cypress/ juniper. “No one has been successful in producing it as a recombinant,” Dr. Mari says.

Evaluation of an early version of the allergen chip showed good correlation with skin testing (Hiller R, et al. FASEB J. 2002;16:152–158). The current version contains 103 components (native and recombinant) from 47 allergens. Dr. Reinhardt says Phadia hopes to take the microarray to the Food and Drug Administration sometime next year. Phadia is also developing a highly resolved version of ImmunoCAP that incorporates component-resolved diagnosis in a different way. Standard extracts are complemented by component proteins of many allergens. The company expects to go to the FDA with this technology also in 2010.

A number of interesting results have been obtained with the allergen chip. Dr. Mari and his coworkers did a survey of allergy frequencies among almost 17,000 Italians. Most results were predictable—allergies to cypress (Cup a 1), mite (Der f 2), and grass pollen (Phl p 1) were most common. However, there were a couple of surprises: More than 10 percent of people showed reactivity to a peach allergen (Pru p 3, a lipid transfer protein) and about five percent to birch (Bet v 1). The latter was especially surprising, Dr. Mari says, since “there is no birch in Rome.” Probably this is an example of cross-reactivity.

In a clinical application of the allergen chip, it identified milk allergy in a patient with urticaria. This was unusual because the patient was a 43-year-old man and milk allergy is overwhelmingly found in children. The chip was vindicated when the urticaria disappeared after the man eliminated milk from his diet.

With specific components it is possible to define the basis of cross-reactions. For instance, Dr. Mari says, Japanese individuals who were allergic to Japanese cedar pollen expected to obtain relief when they visited Rome, where the plant doesn’t grow. They were unpleasantly surprised to discover they were having allergic symptoms. The chip revealed cross-reactivity to cypress. And a Viennese group constructed a microarray chip with recombinant wheat seed and grass pollen allergens that allowed them to discriminate baker’s asthma from wheat-induced food allergy and grass pollen allergy (Constantin C, et al. Allergy. 2009;64:1030–1037).

Dr. Mari, who runs a specialty allergology clinic, has been using the allergen chip in a clinical context for four or five years. By June 2010 he expects to see a commercial chip with 126 molecules. “This technology has theoretically no limits,” he says. He considers it a “democratic” test because it is capable of recognizing a broad diversity of allergies and doesn’t discriminate against “minorities”—allergies that are present in only a small percentage of the population.

Dr. Hamilton says the allergen chip has been in Europe for quite a few years. “In the U.S. right now there is only one lab besides ours that has it, and that’s a lab run by Phadia, a co-developer of the chip.” A “definite strength” of the chip, according to Dr. Hamilton, is that it uses a matrix of recombinant allergens and so can look at cross-reactivity among foods in a unique way. “Its strength is also its weakness,” he adds. For peanut, ISAC has four component antigens: Ara h 1, 2, 3, and 8, which together cover most of the major peanut allergens. On the other hand, he says, “Take ragweed. There is only one ragweed component antigen on ISAC. So you could miss many ragweed-allergic persons.” In sum, ISAC is highly specific but you need to put many components on the chip to cover all allergic people. This is being done, but it takes time to identify all components and to make them by recombinant methods and put them on the chip. “In my opinion,” Dr. Hamilton says, “ISAC’s strength lies in the area of food allergy, since it can dissect cross-reactivity among foods in a way that would almost not be possible with extract-based tests.” He adds one crucial consideration: “The verdict is not yet in on cost.”

Because the chip uses reagents that have developmental status in the U.S., Phadia’s reference laboratory uses it primarily to provide data for clinical investigations, says Dr. Homburger. “Most of our work is with clinical investigators who wish to use these reagents in experimental protocols,” he explains. Throughout the rest of the world, in contrast, component reagents on the allergen chip and on hybrid ImmunoCAP platforms are widely used. This is particularly the case in Europe. “At the present time the U.S. is lagging in implementation of this technology,” Dr. Homburger says. “I see literally dozens of publications from around the world pointing strongly to the value of this type of testing. It tends to be used in initial investigation of patients to detect polysensitization or cross-reactivity.” When the hybrid ImmunoCAP platform is used, he says, “it provides a way to drill down and get quantitative results for IgE antibodies to some of the more important individual components.”

Ultimately, use of component-resolved diagnosis and the allergen chip will be driven by empirical data. “If it helps clinicians manage patients and make appropriate referrals to allergy specialists, it will flourish,” Dr. Homburger says. “But the technology itself is very compelling.”

“When you consider the finding of positive test results to food allergens, the frequencies of positive IgE antibody tests exceed the frequencies of clinical allergy. For example, in the case of peanut, the frequency of positive test results approaches 10 percent, yet fewer than 1.5 percent of people have signs of allergy on peanut ingestion. What accounts for this discrepancy?” One explanation, he says, is that certain proteins in peanut, such as storage proteins, are potent allergens and elicit antibodies that are more strongly associated with allergic reactions than other proteins. “Panallergens, in contrast to storage proteins, are also found in foods, are often responsible for cross-sensitization, and do not typically cause severe allergic reactions to foods including peanut,” Dr. Homburger explains. Panallergens are proteins found widely in pollens and foods that have considerable homology. Sensitization to a panallergen elicits antibodies that cross-react in serologic assays. “Component-resolved diagnostics should help greatly to identify those individuals who have significant clinical allergy as opposed to those who have been sensitized by exposure to a cross-reacting component and are at lesser risk of having an allergic food reaction,” he says.

Like Dr. Hamilton, he adds, “A lot will depend on whether this technology is cost-effective.”

The eventual fate of component-resolved diagnosis and the allergen chip is an issue for the future. When one focuses on the present, other more fundamental issues arise. “Virtually every effective modality of allergy treatment starts with accurate identification of the allergens to which the patient is sensitive,” Hycor’s Thune says. “That’s the opportunity we are trying to address with expanded utilization of lab-based IgE testing. Most physicians are not trained in skin testing.” Faced with a patient who has respiratory symptoms that may be due to allergy, primary care practitioners may believe their only options are to refer the patient to an allergist or to do presumptive treatment. “Putting people on antihistamines may be unnecessary,” Thune says. “Testing for the responsible allergen or allergens could lead to a simple avoidance message or a determination that allergy may not be cause of the symptoms.” Skin testing is not necessary in most of these patients. “Clinical studies have indicated that serologic testing for a panel of about 12 of the most common allergens can generate reliable data to exclude allergy from the diagnosis,” Thune says.

Dr. Reinhardt sees a similar challenge. In the debate over skin testing versus laboratory testing, he identifies a third option—no testing. “A group of nonallergist clinicians in the U.S. manage patients with empirical treatment strategies,” he says. “These patients never get any kind of testing.” In Dr. Reinhardt’s experience, many front-line health care providers—family physicians, pediatricians, pulmonologists, nurse practitioners—aren’t familiar with serologic testing. “For the last decade our company has been involved in attempts to educate these providers,” he says. Evidence for the value of serologic testing exists. In the Inner-City Asthma Study, for example, measures were put in place to identify trigger allergens for each child and to reduce exposure to these triggers. “The environmental intervention group had impressive improvements in outcome measures relative to the control group,” Dr. Reinhardt says. Environmental intervention provided 34 more symptom-free days in the first year of followup (Morgan WJ, et al. N Engl J Med. 2004;351:1068–1080). “Testing leading to appropriate intervention gave them back 10 percent of their year,” Dr. Reinhardt says.

While allergic rhinitis and atopic dermatitis are important, “asthma is more important in the larger health care scheme,” Dr. Reinhardt says. “Asthma is where this kind of testing might make the biggest impact.” Yet even here it is not being used, he says: “Clinicians often know about the stepwise medication protocol, but they typically don’t know they are supposed to get allergy testing.”

William Check is a medical writer in Wilmette, Ill.