In Vitro Allergy Testing: An Overview 1997

In vitro assays for detecting allergy are appealing as an alternative to in vivo tests for several reasons. They can be used to detect allergy with no risk of reaction by the patient, and are particularly useful when the allergens being tested are exquisitely sensitizing, toxic, or irritating. In vitro tests are also useful when skin disease or use of drugs makes skin testing difficult or impossible, or for testing patients fearful of skin tests, or for rapidly screening large populations. Finally, in vitro tests are both objective and quantitative.

In vitro allergy tests are the most recent development in the very old field of allergy. Modern allergy began in 1835, when Wyman reported that pollen was responsible for the symptoms of hay fever, and showed that only certain of his family members were affected by pollen inhalation, while other people were unaffected. (1, 2) In the 1870s, Blackley reproduced the symptoms of allergic rhinitis in himself by inhalation of pollen, and in 1873, demonstrated immediate wheal and flare skin scratch tests. (3) In the early 1900s, Pasteur, Von Behring, and Kitasato demonstrated that passive transfer of serum that contained antibody against toxin could be clinically effective. (4) Thinking that the symptoms of hay fever were caused by a pollen toxin, Noon and Freeman in 1911 reported excellent results following immunization of grass-sensitive patients with boiled grass pollen extracts. (5,6,7) This idea of allergy immunotherapy was introduced to the United States by Cooke in 1915. (8,9,10) Then in 1921, Prausnitz and Kustner made the discovery that serum could transfer the allergic wheal and flare reaction from one individual to another. (11) Subsequently, studies during the 1930s by Cooke demonstrated that patients undergoing treatment with allergy injections developed antibodies which were capable of preventing the passive transfer reaction. (12) Eventually, this induced blocking antibody was determined to be immunoglobulin G (IgG). (13,14) But it was not until 1966 that Prausnitz and Kustner¹s skin-sensitizing antibody was shown by the Drs. Ishizaka and by Bennich and Johansson to be immunoglobulin E (IgE) . (15) With the development of the radioallergosorbent test (RAST) by Wide, Bennich, and Johansson in 1967, research into the pathophysiology of the allergic reaction began to rapidly accelerate. (16,17) In the 1970s, sensitive laboratory methods were developed to measure both allergen-specific IgE antibodies and several of the mediators released following the interaction between allergens and mast cell. (18) This was to become the dawn of a new era in the use of the laboratory as an aid in both the diagnosis and treatment of IgE-mediated diseases. (18,19)

Soon after the exact nature of IgE had been determined, and the RAST radioimmunoassay had been developed, Wide et al suggested that such a test might be helpful in the diagnosis of allergic disorders. (17) They reported that 28 of 29 patients who responded to an allergen challenge had detectable levels of specific IgE antibody in the RAST test, while, 21 of 22 nonresponding patients did not. The test results were correct in 49/51 patients, so that the RAST had a clinical sensitivity, specificity, and overall efficiency of about 96%. These early results were very exciting.

The basic principle of RAST testing is the formation of a molecular sandwich: allergen specific IgE in serum can be measured by sequential reaction of an insoluble support, a known allergen, the patient's serum containing an unknown amount of allergen-specific IgE, and anti-(human IgE), labeled in such a way that the whole sandwich can be detected and quantitated (figure 0). All in vitro allergy assays, regardless of the support medium or detection methodology, utilize this same assay technique (20).

The sandwich technique depends on the immobilization of one of the two test reagents (usually the allergen) on a solid support. Soluble allergens must be bound to supports to create stable, insoluble immunosorbents that acquire the antigenicity of the allergen. Many types of supports can be used, but the first RAST tests used small cellulose filter paper disks, to which specific allergens were bound either by strong noncovalent attraction, or by chemically activated covalent bonds. These immunosorbents with attached allergens, when incubated with a test serum, will react with specific IgE antibody to form antigen-specific immune complexes. The incubation requires sufficient time for equilibration, and conditions that favor only specific binding of IgE antibodies to their specific allergens. Because the antigen is firmly attached to the support, the entire complex also stays attached to the support. The unreacted antibodies in the serum are then washed away, leaving only specific IgE bound to it's specific allergen, which is, in turn, attached to the support. Anti-(human IgE) IgG raised in another species (such as rabbit), and radiolabeled with iodine-125, will react with antigenic determinants on the Fc portion of the patient¹s IgE antibody that is bound to the allergen-coated immunosorbent. This forms a three-layered molecular sandwich: allergen+patient¹s allergen-specific IgE+anti-IgE, and the entire sandwich is bound to the support, so that it can be easily washed, transferred, and the bound radioactivity measured. Although RAST uses a radioactive tracer, other in vitro tests employ colorimetric enzymes, fluorochromes, or other detection devices to quantitate the bound IgE.

The amount of bound label then gives a relative measurement of the amount of specific IgE that was in the patient's serum, which can be converted to the actual concentration by comparison with known IgE standards which are assayed in parallel with the patient samples. In the original research lab RAST, measured radioactivity between two and five times the amount that was due to nonspecific binding by nonatopic negative control sera were scored as positive, and above five times the control level was considered to be strongly positive (Fig. 1, left). 17 Scores below twice that obtained with the nonatopic control were considered to be negative. In practice, the exact serum concentration of allergen specific IgE is rarely reported. Instead, results are expressed in terms of degrees of positivity, or, classes. The Phadebas RAST was reported in four classes, and the Modified RAST is reported in five classes (21).

Several investigators reported a high correlation between RAST scores and the allergic history, Prausnitz-Kustner transfers, the response to allergen challenge, the results of skin endpoint titration, (22-25) the response to drug treatment,(26) and the tolerance to initial immunotherapy doses. (27,28) Aas and Lundkvist, using purified allergen from codfish, found 100% sensitivity and specificity when comparing the results of RAST and a history of either tolerance or intolerance to the ingestion of cod in 112 allergic children. (29) Wide et al report similar sensitivity in 128 patients tested against birch tree, timothy grass, cat, dog, horse, and house dust. (30) RAST was also used in the diagnosis of some chemical sensitivities. (31,32) By 1975, Gleich and Yunginger, writing about the role of the RAST in the practice of allergy, stated ". . . the introduction of RAST has been a milestone in the transition of allergy from a practice based on subjective judgments related to history and skin tests to one based on definitive biochemical information derived from the clinical chemistry laboratory." (33) And in 1979, Van Arsdele wrote that allergy testing,if properly done, should correlate well with the results of RAST, and stated that "the latter is the benchmark for testing the reliability of different skin testing methods and standardizing test allergens." (34) Unfortunately, during the period of 1979 to 1981, RAST testing became controversial.

Controversy arose as the result of the scoring system associated with the first commercially available in vitro test, the Phadebas RAST (Fig.1, second from left). For the first time, physicians were provided with all the reagents necessary to perform this test in an office setting. Included with the test materials were four IgE reference standards: standard A taken from a pooled serum from patients highly sensitive to birch pollen, standard B (a fivefold dilution of standard A), standard C (a fivefold dilution of standard B), and standard D (a twofold dilution of standard C). Each of these standards was assigned an arbitrary number of Phadebas RAST units (PRUs), ranging from 50 for reference A to 1 for reference D. In the Phadebas RAST scoring system, test sera with bound radioactivity less than that obtained with standard D (tested against a birch disk) were considered to have a nondetectable level of allergen-specific IgE. However, Nalebuff observed that reference D actually had about 10 times greater radioactivity than the negative controls. (35) Consequently, sera with bound radioactivity significantly above those of negative controls, yet below those of reference D, were erroneously assumed to have nondetectable levels of specific-IgE antibody, but, in fact, were wrongly scored as negative. This occurred in up to 50% of tested patients. Nalebuff then showed that a 10-fold dilution of reference standard D with human cord serum to 0.1 PRU still gave a binding value greater than three times that obtained with negative controls and suggested that this be adopted as a new cut-off for the detection of specific-IgE.

Deuschl and Johansson described 18 allergic rhinitis patients who were positive by provocation, history, and skin test, yet had negative Phadebas RAST results. Like Nalebuff, they also found that a fivefold dilution of standard D still gave binding significantly greater than that of their negative control. By diluting reference D fivefold and lowering the cut-off point to 0.2 PRU, they improved the sensitivity of the test. With this change, sixty-six percent of the provocation positive, Phadebas RAST negative tests were then recorded as positive. (36) In 1980, as a result of these observations and in response to criticisms leveled at the test for its low sensitivity, an updated Phadebas RAST was introduced in which the reference standards were diluted 1 : 3. As a result, the new reference D cut-off point was set at 0.35 PRU. Unfortunately, this new threshold did not solve the sensitivity problem (Fig. 1, third from left).

Nalebuff then demonstrated that specific IgE was still detectable below the new Phadebas RAST cut-off point (Table 1). (37)

These observations led to the development of the modified RAST (MRT). Several changes were made which increased the test sensitivity. First, the initial incubation period was increased from 3 to 18 hours. Second, the volume of the test serum was increased from 50 to 100 mL, which prevented disks from drying out during the incubation. Third, after the second incubation with labeled antihuman IgE and prior to counting the bound radioactivity,the coated allergen disks were removed from their original polystyrene tubes and, before counting, were placed into fresh ones. This was done to ensure that only radioactivity immunologically bound to the disk was measured, rather than radioactivity adhering to the inner surface of the tube. (37)

Test reproducibility between runs, or between laboratories was also improved by the use of a time or count control. This standardized the MRT for variations in test radioactivity due to isotopic decay. The time used to count each disk equals that required by a known 25 unit/mL IgE sample to reach 25,000 counts when tested against an anti-IgE disk run in parallel to the sera under study. When this was done, the average nonspecific binding in the original MRT assay averaged 500 counts or 2% of that of the 25 unit IgE control, and the lower limit of detectable allergen-specific IgE was 750 counts or 1.5 times the negative control. Controls were either human cord serum, serum from nonatopic patients, or serum from atopic patients tested against an inappropriate allergen.

With further improvements in the isotopic MRT, nonspecific binding was reduced to 250 counts or 1% of the time control, and 500 counts, or 2% of the time control, is the 95% confidence level for the detection of allergen-specific IgE. Scores between 500 and 750 counts are considered to be equivocal, since the low level of detected IgE may not be clinically relevant. With an enzymatic label, the nonspecific binding is somewhat higher at 375 counts. If enzymatic markers are used, a cut-off point at 750 counts, or 3% of the time control, is recommended. Scores above 750 counts are divided into distinct classes, each representing approximately a fivefold increase in the amount of specific IgE antibody present in the sample (Fig. 1, right). MRT class 1 (750 to 1600 counts) is a low level positive score. Class 1 levels are always associated with a positive skin test reaction, and 70% of affected patients will respond to a nasal or conjunctival challenge. In contrast, MRT class 2 scores and above (1600 to 40,000 counts) represent positive scores with increasing levels of detectable specific IgE antibody and degrees of clinical sensitivity, and 95% of patients with these scores respond positively to an appropriate allergen challenge test.

The modifications that produced the MRT scoring system were initially met with skepticism. (38) However, peer review of these adjustments confirmed the shortcomings in sensitivity of the commercial Phadebas RAST. (39,40) Ultimately, the American Academy of Allergy's Committee on In Vitro Tests compared the Phadebas and modified RAST. (41) Each participant was supplied with positive and negative control sera, as well as the Phadebas reference reagents, and ran the RAST in their own labs using the modified RAST protocol. The committee's data verified that the cut-off of reference D (0.35 PRU) in the Phadebas RAST is set at about 10 times that obtained with nonatopic human serum (Table 2).

The positive control sera were obtained from known patients highly allergic to either ragweed, rye grass, Alternaria, cat epithelium, milk, or house dust mite. Each of these six sera were supplied in three dilutions labeled X, Y, and Z (Table 3).

Whereas the modified RAST was able to identify the presence of allergen-specific IgE antibody in ALL 306 tests performed on these three dilutions, only 60% of test scores were above the reference D cut-off point in the Phadebas RAST, and ALL 102 tests performed on the "Z" dilutions were scored as negative by Phadebas RAST (Fig. 2).

Negative controls were also compared, including horse serum diluent (HSD), 5% human serum albumin (5% HSA), nonallergic serum (NAS), and nonallergic sera spiked with 200 units of IgE myeloma protein (IgE). The modified RAST test scores were negative in 95% (455 of 476) of negative control sera tests (Fig. 3).

If the results of two investigators are excluded (Fig. 4), the remaining 15 obtained a modified RAST specificity of 99%; that is, 415 of 420 negative control tests were scored as MRT class 0. As expected, the Phadebas RAST was 100% specific. Therefore, the changes embodied in the performance and scoring of the modified RAST increased test sensitivity 40% higher than that obtained with the Phadebas RAST with only a I% loss in test specificity.

Twelve Investigators found the modified RAST 100% specific; 15 investigators scored 415/420 tests negative for a 99% specificity; and 16 investigators scored 438/448 tests negative for a 98% specificity.

A subsequent study concluded that the MRT had a level of sensitivity approximating that of nasal provocation and skin tests. (42) As a result of these reports, the modified RAST rapidly gained popularity: today, approximately 85% of RASTs performed in the United States employ this technique.

Many alternatives to the use of radioactive isotopes have subsequently been developed (Table 4).

Since the 1980s, there has been a tendency to replace the radioactive labels used in the RAST with enzymatic markers. These enzyme labels are stable for longer periods of time and avoid the hazards of radioactive materials and the inconveniences of storage, disposal, and recordkeeping. All of the manufacturers of RAST tests now provide an enzyme alternative. These newer technologies may give results that differ from MRT . (43,44,45) Results are affected by a number of factors, such as the potency of the allergens used, the type of support employed, varying absorption by the allergens to the plastic surface, and the specificity of the anti-IgE used.

All in vitro systems are potentially capable of measuring the presence of specific IgG antibody, provided that appropriate IgG standards and anti-(human IgG) are used. This can allow the course of immunotherapy to be monitored, as allergen-specfic IgG blocking antibody is produced. Although not routinely performed, blocking antibody measurements can be employed when patients should be showing improvement on immunotherapy, but are not.

Food specific IgG levels can also be measured, but are currently practical only as a means of confirming that someone has been strictly following a prolonged exclusion diet. All individuals are exposed to large amounts of immunogenic food proteins and respond by producing IgG antibodies. (46) The normal range of IgG levels and IgG levels of allergic patients overlap, so that food specific IgG levels have no diagnostic meaning other than to indicate significant exposure to that food. Food specific IgG results can be used as a rough guide to the quantity of certain foods eaten by an individual over the past few months, but are very expensive in comparison with a simple food ingestion diary.

The Modified RAST scale incorporates the five-fold dilution system of standard titration skin testing, so that each of the five Modified RAST classes corresponds to five-fold changes in the serum IgE concentration. An example of a typical Modified RAST scale, using standard Iodine-125 radioactively labeled anti-IgE is shown in Table 5 and Fig. 5. Similar scales can be designed for any kind of label and in vitro assay system, using the five-fold dilution principle to make the classes correspond to titration skin test results.

In vitro tests are less time consuming for the patient than in vivo skin tests - a large advantage in a busy world. In vitro tests are useful in apprehensive children, in patients with dermatologic problems where skin testing is difficult, and in those patients unable to stop medication which can inhibit whealing. Other distinct advantages are that there is no patient risk, especially when testing persons who have already experienced anaphylaxis, there is a low incidence of false positive results, and less office personnel are required when utilizing in vitro testing. The major disadvantage of in vitro testing is cost. Although costs have come down, in vitro diagnosis is still more expensive per test than skin testing. However, the costs of a diagnostic workup were not found to be different between those physicians using in vitro or in vivo methods in the evaluation of allergic patients. (47) Moreover, the initially incurred diagnostic expenses are only a small portion of the overall cost of managing a chronic illness such as allergy. Finally, use of a small number of initial screening tests can reduce the expense of in vitro testing by eliminating nonatopic patients from full testing. Consequently, in vitro testing remains a very viable option.

There are situations in which data obtained by in vitro testing may be informative and yet not contribute to improved patient care. For example, there is no reason for using these tests when the diagnosis is apparent from the clinical history (eg, symptoms on exposure to cat or ingestion of a particular food) and when pharmacotherapy or avoidance is found to be effective in alleviating symptoms. Further-more, patients for whom properly performed skin tests, with controls, have given negative results are not likely to have measurable levels of allergen-specific IgE antibody.

On the other hand, in vitro tests are valuable in evaluating individual sensitivity when immunotherapy is required for control of symptoms, or, for evaluating transfer patients on immunotherapy who are moving into the area. In vitro tests are also ideal for evaluating stinging insect sensitivity, or in confirming or identifying anaphylactic food reactions. Certain drug and chemical sensitivities are also best evaluated by RAST, since skin testing with these compounds might be dangerous or cause skin slough.

Both in vivo and in vitro tests are equally acceptable for inhalant allergy testing. For anaphylactic food allergy testing, in vitro testing is far safer, and is strongly preferred. For delayed, cyclic type food allergy, skin testing using the IDPFT technique identifies more positives, but in vitro screening, especially in children, also identifies a significant number of positives. The two testing modalities are actually complementary, rather than exclusive, since they both have advantages and disadvantages. A general comparison is shown in Table 6 (21,48).

Particular in vitro assays from various companies may or may not be typical in all categories, and need careful evaluation, including comparison runs with skin tests, prior to selection of a system for use either as a reference laboratory or in your own office lab. In the evaluation of any in vitro system, there is no substitute for clinical correlation and direct comparison with carefully performed skin tests.

Acknowledgment: to Drs Fadal & Nalebuff, without whom this paper, and modern in vitro methods, would not have been possible. (49, 50)

1. Samter M (ed): Excerpts from Classics in Allergy. Columbus, OH, Ross Laboratories, 1969

2. Zeiss CR: Immunotherapy in allergic disease. In Lockey RF (ed): Allergy and Clinical Immunology. Garden City, NY, Medical Examination Publishing Co., 1979, p 977

3. Blackley CH: Experimental Researches on the Cause and Nature of Catarrhus Reactions (Hay Fever or Hay Asthma). London, Bailliere, Tindall and Cox, 1873

4. Adkinson NF: In Vitro Tests for IgE In The Practice Of Allergy. In: Grant A, (ed): Insights in Allergy 1. St Louis, CV Mosby, 1986

5. Freeman J, Noon L: Further observations on the treatment of hayfever by hypodermic inoculations of pollen vaccine. Lancet 2:814, 1911

6. Noon L: Prophylactic inoculation against hay fever. Lancet 1:1572-1573, 1911

7. Portier P, Richet C: De I'action anaphylactique de certains venins. Compt Rend Soc Biol 54:170, 1902

8. Cooke RA: Active immunization in hay fever. Laryngoscope 25:108, 1915

9. Cooke RA: The treatment of hay fever by active immunization. Laryngoscope 25:108, 1915

10. Cooke RA, VanDer Veer A Jr: Human sensitization. j Immunol 1:201-237, 1916

11. Prausnitz C, Kustner J: Studien Uber Die Ueberempfindlichkeit. Cent Backteriol 86:160, 1921

12. Cooke RA, Bamard JH, Hebald S, et al: Serologic evidence of immunity with coexisting sensitization in a type of human allergy (hay fever). J Exp Med 62:733, 1935

13. Creticos PS, Norman PS: Immunotherapy in inhalant allergy. In Lessof MH, Lee TH, Kemeny DM (eds): Allergy, An International Textbook. Great Britain, The Bath Press, 1987, p 617

14. Lichtenstein LM, et al: Antibody response following immunotherapy in ragweed hay fever:Allpyral vs. whole ragweed extract. J Allergy 41:49, 1968

15. Bennich HH, Ishizaka K, Johansson SGO, et al: Immunoglobulin E: A new class of human immunoglobulin. Bull WHO 38:151, 1968

16. Johansson SGO, Bennich H: Studies on a new class of human immunoglobulins: Immunological properties. In Killander j (ed): Gamma Globulins: Structure and Control of Biosynthesis, Nobel Symposium 3. Stockholm, Almqvist & Wiksell, 1967, p 193

17. Wide L, Bennich H, Johansson SGO: Diagnosis of allergy by an in vitro test for allergen antibodies. Lancet 2:1105, 1967

18. Gleich Gj: Introduction: Part 11. In Evans R III (ed): Advances in Diagnosis of Allergy: RAST. Miami, Symposia Specialists, 1975, p 111

19.Gleich Gj, Yunginger JW: The radioallergosorbent test: Its present place and likely future in the practice of allergy. Adv Asthma Allergy 2:1, 1975

20. Terr AI. In Vitro Tests for Immediate Hypersensitivity. Ann Rev Med 1988;39:135-145

21. Nalebuff DJ. PRIST, RAST, and Beyond. Otolaryngol Clin N Am 1985; 18:725-744

22.Berg TLO, Johansson SGO: Allergy diagnosis with the radioallergosorbent test. A comparison with the results of skin and provocation tests in an unselected group of children with asthma and hayfever. J Allergy Clin Immunol 54:209, 1974

23.Eriksson NE: Diagnosis of reaginic allergy with house dust, animal dander and allergens in adult patients. 111. Case histories and combinations of case histories, skin tests, and RAST compared with provocation tests. Int Arch Allergy Appl Immunol 53:441, 1977

24.Evans R, et al: An immunologic evaluation of ragweed sensitive patients by newer techniques. J Allergy Clin lmmunol 49:285, 1972

25.Johansson SGO, Bennich HH: The clinical impact of the discovery of IgE. Ann Allergy 48:325, 1982

26.Welsh PW, et al: Preseasonal IgE Ragweed antibody level as predictor of response to therapy of ragweed hay fever with intranasal cromolyn sodium solution. J Allergy Chn Immunol 60:104, 1977

27. Nalebuff DJ: Letter to the editor. Clin Allergy 8:99-100, 1978

28. Nalebuff DJ: An enthusiastic view of the use of RAST in clinical allergy. Immunol Allergy Pract 3:18, 1981

29.Aas K, Lundkvist U: The radioallergosorbent test with a purified allergen from cod fish. Clin Allergy 3:255, 1973

30. Wide L: Clinical significance of measurement of reaginic (IgE) antibody by RAST. Clin Allergy 3(suppl):583, 1973

31. Dolovich J, Bell B: Allergy to product(s) of ethylene oxide gas. J Allergy Clin Immunol 62:30, 1978

32. Karol MH, Alarie Y: Antigens which detect IgE antibodies in workers sensitive to toluene diisocyanate. Clin Allery 10:101, 1980

33. Gleich Gj, Yunginger JW: The radioallergosorbent test: Its present place and likely future in the practice of allergy. Adv Asthma Allergy 2:1, 1975

34. Van Arsdel PP Jr: Allergy. JAMA 241(13):1375-1376, 1979

35. Nalebuff DJ: Letter to the editor. Clin Allergy 8:99-100, 1978

36. Deuschl H, Johansson SGO: Specific IgE antibodies in nasal secretion from patients with allergic rhinitis and with a negative or weakly positive RAST on the serum. Chn Allergy 7:195-202, 1977

37. Nalebuff DJ: An enthusiastic view of the use of RAST in clinical allergy. Immunol Allergy Pract 3:18, 1981

38. Adkinson NF Jr: The radioallergosorbent test: Uses and abuses. J Allergy Clin Immunol 65:1, 1980

39. Hoffman DR: Comparison of methods of performing the radioallergosorbent test: Phadebas, Fadal Nalebuff and Hoffman procedures. Ann Allergy 45:343, 1980

40. Sanirach BA, Parker JL, Jones LT, et al: Diagnostic and therapeutic applications of a modified radioallergosorbent test and comparison with the conventional radioallergosorbent test. J Allergy Clin Immunol 67:97, 1981

41. American Academy of Allergy Committee on In-vitro Tests: 1981 study. (This unpublished data is available from Dr. D.J. Nalebuff on request.)

42. De Filippi 1, Yman L: Clinical accuracy of updated version of the Phadebas RAST test. Ann Allergy 46:249, 1981

43. Emanual 1: Comparison of in-vitro allergy diagnostic methods. Immunol Allergy Pract 7:10-483, 1985

44. Pecoud A, Peitrequin R, Fasel J, et al: Comparison of two assays for the determination of specific IgE in serum of atopic and nonatopic subjects: The Allergenetics FAST and the Phadezym RAST. Allergy 41(4):243, 1986

45. Perelmutter L, Emanuel 1: Assessment of in vitro IgE testing to diagnose allergic disease. Ann Allergy 55:762, 1985

46. AMA Council on Scientific Affairs: In Vitro Testing for allergy; Report 11 of the allergy panel. JAMA 258:1639, 1987

47. Project Hope, Center for Health Affairs: The cost implications and cost effectiveness of allergy in vitro diagnostic testing. (executive summary), October 1988.

48. Ownby DR. Allergy Testing: In Vivo Versus In Vitro. Ped Allergic Dis 1988; 35:995-1009

49. Nalebuff DJ. In Vitro Testing Methodologies: Evolution and Current Status. Otol Clin N A 1992; 25: 27- 42

50. Fadal RG. Experience with Rast-Based Immunotherapy. Otol Clin N A 1992; 25: 43-60