Investigation of the clinical usefulness of the intracutaneous
provocative-neutralization food test (IPFT)

William P. King, MD, Wallace A. Rubin, MD, Richard G Fadal, MD, Walter A Ward, MD, Richard J. Trevino, MD William B Pierce, MD, J Alan Stewart, MD, and John H. Boyles, Jr. MD, Corpus Christi, Texas, New Orleans, Louisiana, Teaneck, New Jersey, Winston-Salem, North Carolina, Oakland, California, Batavia, New York, Lake Jackson, Texas and Dayton, Ohio.

This study investigated the clinical usefulness of the intracutaneous provocative-neutralization food test (IPFT). Thirty seven patients were tested for five identical food allergies by eight physicians in different geographical locations. Throughout the study, comparison was made between the IPFT when interpreted by skin response (IPFT SK) and when interpreted by symptom provocation (IPFT PR). Double-blind IPFT results were compared with those of previously accomplished oral challenge food tests (OCFT). IPFT reliability was determined by a double-blind comparison of the initial IPFT, with two subsequent IPFTs performed 7 days apart. Correlation of the IPFT SK and IPFT PR with the OCFT provided validity coefficients of 0.78 and 0.61 respectively, both significant beyond the 0.01 level of confidence. Reliability of the IPFT SK and IPFT PR was shown to be 0.68 and 0.40, respectively. The IPFT SK was significant beyond the 0.01 level of confidence and the IPFT PR was significant beyond the 0.05 level of confidence. (OTOLARYNGOL HEAD NECK SURG 1988:99-263)

From the Department of Otolaryngology and Communication Sciences, Baylor College of Medicine (Dr. King), the Department of Otorhinolaryngology and Biocommunication, Louisiana State University School of Medicine (Dr. Rubin), the Department of Immunology, Holy Name Hospital (Dr. Fadal), Oak Knowl Naval Hospital (Dr. Trevino), Genesee Memorial Hospital (Dr. Pierce), and Children’s Medical Center and St. Elizabeth Medical Center and Wright State University (Dr. Boyles), Drs. King (Corpus Christi), Ward (Winston-Salem), Pierce (Batavia), and Stewart (Lake Jackson) are in private practice.

Sponsored by the American Academy of Otolaryngic Allergy.

Presented at the Annual Meeting of the American Academy of Otolaryngic Allergy, Chicago, Ill. Sept. 18 1987
Submitted for publication Jan. 26 1988: revision received April 18, 1988: accepted April 25, 1988
Reprint requests: William P. King MD, 1415 Third Street, Suite 507 Corpus Christi, TX 78410.

Food hypersensitivity has ben divided clinically into two broad categories – immediate and delayed. This classification has both clinical and immunologic implications. The immediate reaction is a Type I reagenic, IgE mediated reaction with a direct anaphylactic cause-and-effect characteristic. The immediate reaction tends to be persistent, irrespective of frequency or dose of allergen exposure. This immediate IgE dependent food reaction is frequently referred to as a fixed food allergy and accounts for about 5% of all immunologically mediated adverse reactions to foods.1,,,

Whereas a Type I reaction may be delayed, the more common delayed reaction involves, alone or in combination, Types II, III, or IV non-IgE immunologically mediated food hypersensitivity. It involves other immunoglobulins such as IgG and IgM and is often associated with complement activation.1,2, This variety of delay reaction is also referred to as cyclic food hypersensitivity and accounts for the majority of immunologically mediated adverse reactions to foods.1,3-5

Unlike fixed food allergy, in which the reactions commonly involve the respiratory, dermal, or gastro-intestinal systems, the cyclic food reactions are not only multi-mechanistic, but present manifestations that involve multiple systems and organs.1,3-5,,,, Occurrence of symptoms is more dependent on frequency and volume of food allergen exposure. The most common cyclic allergenic foods are those that have been eaten in one form or another almost daily, such as milk, wheat, corn, yeast, soy and egg.1,3,4 Cyclic food reactivity is a dynamic process and is associated with hyperacute reactivity, tolerance, and improvement or cyclic allergy symptomatology by ingestion of some of the allergenic food. The latter phenomenon is called masking.3,4 Dependence on this masking action for continued well being often results in food craving or even addition. This masking phenomenon has contributed to the confusion surrounding the delayed food sensitivity response. The multi-system symptomatology is thought to be caused by a breakdown of this masking phenomenon.3,4,11

Lee originally presented the IPFT in 1961 in an attempt to develop a simple and more efficient clinical tool for cyclic food hypersensitivity diagnosis than the 5-day OCFT, originally described by Rinkel et al.3,5 and redefined by Dickey. The latter test requires that the patient fast from all forms of the food to be tested for 4 days, then consume a large test meal of that food on the fifth day when the cyclic sensitivity becomes hyperacute, allowing maximum diagnostic advantage. Immediate post-challenge symptoms are frequently provoked by this method, though delayed symptoms may occur. Continued specific food abstinence results in reversal of sensitivity and increasingly greater tolerance.3 Lee12 sought to accomplish a similar challenge using intracutaneous food injection rather than oral feeding. As after oral challenge, occasional delayed symptom provocation occurred. He also observed that delivery of a slightly weaker dose than that which produced a positive skin response often resulted in the disappearance of provoked symptoms. He labeled this phenomenon "neutralization." and the specific volume and dilution that accomplished this, the "neutralizing dose." Thereafter, the IPFT was revised by several clinicians, resulting in numerous publicized alterations of technique.5,,, Neutralization was consistently observed to occur only on application of a dilution that produced a negative skin response, and most often on the strongest dilution that resulted in a negative skin response.

Provocation-neutralization is used for the diagnosis of delayed food hypersensitivity, and never for confirmation of an immediate food allergy obvious by history. The mechanisms of action are not fully understood, but are presently being investigated. The recent studies of Boris et al indicate that one mechanism of action is a non-immunologic attenuation of shock organ reactivity. This attenuation appears to be initiated by a rise in prostaglandin as direct response to the earlier administration of a neutralizing dose. The studies of Sobotka et al and Mendoza and Kiminobu document antigen specific desensitization of basophils for high-dose histamine release by low-dose pre-incubation. This possible mechanism of neutralization is supported by Scadding and Brostoff in their double-blind study of low-dose house dust mite treatment.

Materials and Methods

Study Objectives. The study objective was to assess the validity of the IPFT; that is whether it tests what it is supposed to test; and to assess the reliability of the IPFT; that is whether the same results are consistently obtained upon retest. Validity was to be determined by comparing the double-blind results of IPFT testing with those of OCFT results, accepting OCFT results as a standard for determining allergenicity. Reliability was to be evaluated by comparing the results of a double-blind IPFT with two other identical double-blind IPFTs on the same patients performed at 7-day intervals after the initial IPFT. The protocol for the study was approved by the Research Committee of the American Academy of Otolaryngic Allergy.

Materials. Five commonly allergenic foods were chosen for testing: wheat, corn, beef, white potato, and milk. Their extract forms are clear, and present no sight recognition problems. The commercial extract and placebo concentrates were supplied by Meridian BioMedical, Inc., Denver, Colorado. The placebo was 50% glycerin. The OCFT meals consisted of: milk, 1 pint of plain cow’s milk; corn, ½ cup of pure grits (cracked corn) prepared in distilled water; wheat, ½ cup of pure cream of wheat prepared in distilled water; white potato, 1 large or 2 small baked potatoes, plain and beef, 4 to 6 ounces of broiled beef, plain. All the cooked foods were prepared in a radar unit in glass containers.

Participants. Eight physicians experienced with IPFT technique participated in the study. Each physician was assigned the task of completing the study on five food sensitive patients. Three of the physicians contributed only four patient results. The study, therefore, consists of the data accumulated from 185 food test comparisons completed on 37 patients.

Table 1. Patients’ age and sex



No. of







Patient-participant requirements were as follows: (1) age, from 5 to 50 years (Table 1); (2) not pregnant or nursing; (3) not severely reactive, because provocation testing was begun with a maximal provoking dose; (4) not permitted to use the following medications for 2 weeks before and during the study: antihistamines, cromolyn sodium, nasal sprays and aerosols, oral decongestants, tranquilizers, and antidepressants; (5) not permitted to use the following medications for 4 weeks before and during the study: hydroxyzine, steroids, theophylines, beta blockers, beta agonists, and nonsteroidal anti-inflammatory medications: (6) have never experienced previous provocation-neutralization testing or any form of food allergy treatment; (7) read and sign a standardized release form before participation; and (8) respond positively to at least one of the five foods orally challenged.

Three individuals were used in each office – a coder, a tester and a collector – to maintain a strict double-blind for all tests except the open OCFT. Once a week, the coder randomly changed the order of the allergens and a placebo. The order of testing was also randomized. Fresh dilutions were made by the coder for each patient tested. Only the coder knew the codes until completion of the office study. The tester performed the IPFTs and recorded the results. The coder was not informed of any test results. The collector received the sole copy of all coded results. These records were not shared until completion of the office study.

Study procedure. The patients were instructed to continue eating their usual diet throughout the study period. The 5-day OCFTs were then completed, after which the five test foods were double-blind IPFT tested three times at 7-day intervals. If symptoms were provoked on the first IPFT, no neutralization was attempted. If a positive whealing response resulted without symptom provocation, the testing for this allergen was stopped without identifying the strongest dilution that produced a negative wheal. These limitations were made to assure that the second IPFT procedures would not be influenced by the previous delivery of a neutralizing dose. When symptoms were provoked on the second or third IPFT, appropriate neutralization was accomplished and recorded. When positive whealing occurred without symptom provocation, the strongest dilution that produced a negative wheal was identified as the final end point. Thereby, any effect on the third IPFT responses of providing a neutralizing or endpoint dose 7 days earlier could be observed.

Statistical analysis. For purposes of statistical evaluation, all the response data for all the tests were coded. All negative responses were assigned an arbitrary value of 0, whereas all positive responses were given a value of I. Data were then compared for all five foods for the OCFT and the three IPFTs, SK and PR. Pearson Product Moment correlations (r) were computed for each allergen. The correlation of OCFT with the initial IPFT, SK and PR represents the test for validity of the IPFT. The reliability of the IPFT was determined by the intercorrelation of the initial IPFT, SK and PR with both the second and third IPFT. There was no communication of results between physicians’ offices until the final study was completed and all results were turned into the Research Committee of the AAOA.

Study standardized IPFT. The patients were instructed to consume the five foods to be tested at lunch or dinner the day before. Such priming has been advised to elicit clearer skin responses.14-16 Five-fold dilutions of a l:10 food extract concentrate were used. The concentrates contained 50% glycerin which necessitated the use of five-fold dilutions of 50% glycerin for a negative control. Testing was begun by placing a maximal provoking 0.05 cc intracutaneous dose of the #1 dilution of the related food extract on the upper lateral arm. The usual initial screening with weaker dilutions was deleted. The reason for this modification was prevention of interference with symptom provocation by earlier administration of a neutralizing dose. If no provocation of symptoms and no positive whealing 2mm or more larger than the #1 dilution of the glycerin negative control occurred within 20 minutes, the test was recorded as negative. If the whealing response was positive without symptom provocation within 20 minutes, an application of identical sized wheals using consecutively weaker dilutions were applied every 10 minutes until the strongest dilution that produced a negative wheal was identified. This final end point dose was recorded. If the whealing response was positive and symptoms were provoked within the 20 minutes, the tester then proceeded to neutralize the provoked symptoms. This was accomplished by applying identically sized wheals via consecutively weaker dilutions at approximately 10 minutes intervals. Most often the neutralizing dose is the strongest dilution that produces a negative wheal. Occasionally, multiple applications of this same dilution (or one or two dilutions weaker) is required.15,16 A dose that produces a positive skin response cannot be a neutralizing dose.

Table 2.

Glossary of table and figure abbreviations
Oral Challenge Food Test
Intracutaneous provocative food test
Intracutaneous provocative food test interpreted by skin response
Intracutaneous provocative food test interpreted by symptom provocation
Positive vs negative skin response on intracutaneous provocative food testing.
Provocation of any variety of symptom
Final Dose
The neutralizing or endpoint final dose determined by intracutaneous provocative food testing.

Results. A list of the abbreviations used in all the tables and figures is presented in Table 2. Table 3 presents a comparison of the history-enumerated symptom complaints with those provoked on the OCFTs and the initial IPFTs. Obviously, multiple symptoms were frequently reported, especially on the OCFT. Some form of rhinitis was the most common pretest complaint and provoked symptom. Headache was a distant second. Other common complaints concerned the gastrointestinal system, the broncho pulmonary system, the throat, the conjunctiva, sensations of dizziness and fatigue.

Table 3. Comparison of pretest and test-provoked symptoms

Symptoms Pretest OCFT IPFT
Rhinitis 45 44 33
Headache 24 39 9
Gastro-intestinal 11 19 2
Broncho-pulmonary 11 7 5
Fatigue 9 28 9
Urticaria 6 3 2
Throat 6 15 6
Dizziness 4 11 9
Ear discomfort 3 -- 1
Skin itch/rash 2 9 2
Edema 2 2 5
Conjunctiva 2 11 6
Palpitations 2 3 3
Arthralgic 1 5 --
Hot Flash 1 4 3
Mood Change 1 -- --

The primary objective of this study was to assess the validity of the IPFT, that is, to determine the extent to which it detected food allergenicity. This was done by use of the OCFT as a standard and correlating OCFT vs. IPFT findings (Fig. 1). For the IPFT SK, the average validity coefficient for the five foods tested was 0.78, significant beyond the 0.01 level of confidence. Coefficients for the five foods ranged in value from 0.95 to 0.49, and all were significant beyond the 0.01 level of confidence. Validity of the initial IPFT PR was also significant beyond the 0.01 level of confidence overall, with the average of the five foods tested being 0.61. The individual values for the IPFT PR ranged from 0.50 to 0.69, varying from 0.05 to 0.01 level of confidence.

Figure 2 presents the IPFT individual food responses that agreed with their comparison OCFTs. The test results were similar, whether interpreted by skin response or symptom provocation, except for a consistently smaller percentage of positive responses when symptom provocation was required for positivity. Wheat was the most frequent positive responder – positive in 25 of the 37 patients by IPFT SK and in 21 of the patients by IPFT PR. Corn was second, followed by milk, potato, and beef. The IPFT responses to wheat, corn, and beef – regardless of the method of interpretation – were approximately 80% consistent with their companion OCFT results. White potato was approximately 70% consistent. Milk was the least consistent, at nearly 65%.

Another objective as to assess the reliability, or repeatability, of the IPFT. This was accomplished by correlating test vs. retest findings (Fig.3). The average correlation of the second and third IPFT SK, with the initial IPFT SK summed across the five foods tested, was 0.68 – significant beyond the 0.01 level of confidence. Individual food correlations ranged from 0.60 to 0.74, each significant beyond the 0.01 level of confidence. The reliability of the IPFT PR was less and more variable than that of the IPFT SK. The average reliability of the IPFT PR was 0.40 significant beyond the 0.05 level of confidence. The individual correlations ranged from 0.21 to 0.75, with three of the five correlations significant beyond the 0.05 level of confidence.

Consistency of the IPFT responses is shown in Fig. 4 by comparing the results of the #2 and #3 IPFT, SK and PR, with those of the initial IPFT. Because of an extract lot change for the two follow-up tests on 17 foods, 34 comparisons were deleted. The results of the remaining 336 comparisons are presented. Skin response efficiency was 83%. Provocation efficiency was 75% consistent. In the 89 instances when symptoms were persistently provoked, the same symptoms were provoked approximately one half of the time.

Table 4 details the consistency of the final endpoint or neutralizing dose. Recall, neither a final endpoint nor a neutralizing dose was determined after the first IPFT, but was determined after the second and third IPFT. The latter two responses were compared. Positive response to positive-within-one dilution response was included in the acceptably consistent category, but comprised only 16 of the 152 consistent comparisons. Overall consistency was shown to be 82.8%. There was no apparent effect on symptoms provocation during the third IPFT by the application of a neutralizing dose after the second IPFT 7 days earlier. The incidence of non-repeated symptom provocation during the third IPFT was no greater than during the second IPFT, when no previous final endpoint or neutralizing dose was delivered.

Table 4. Final dose consistency on comparison of the #2 with the #3 IPFT


Skin responses

Consistent – to –  – 74

+ to + same dilution

– 62 82.2%

+ to + within 1 dilution

– 16

Not consistent

+ to + over 1 dilution – 5

+ to – 

– 11 17.7%

– to +

– 17

Table 5 summarizes the qualities of all the various tests accomplished, with the OCFT being the designated standard. Formulas used were:

True positives

Sensitivity =True positives + false negatives X 100

True negative

Specificity =True negatives + false positives X 100

True positives + true negative

Efficiency = Grand total X 100

Sensitivity represents percent positivity in disease, whereas specificity represents percent negativity in the absence of disease. Low sensitivity of a test results in the occurrence of false-negative responses, whereas low specificity of a test results in false positive responses. Efficiency concerns overall test accuracy.

The data show the IPFT modalities to be accurate approximately 75% of the time. The IPFT SK shows a 20% greater propensity for false-positive responses than the IPFT PR. Any missed delayed response after the OCFT would tend to exaggerate this proclivity. The IPFT PR shows a 20% greater propensity for false negative responses than the IPFT SK.

During this study, a companion double-blind comparison between the OCFT results and those of in vivo Multi-Test prick testing and in vitro specific IgE, IgG, and the food enzyme linked immune assay test was accomplished. The detailed data will be presented in a separate article. Suffice here to state that the efficiency of these other test modalities was approximately 50%. Both studies suggest that skin response is not exclusively IgE-mediated.

Table 5. Test sensitivity, specificity and efficiency compared to OCFT standard.


Sensitivity Specificity Efficiency
IPFT SK 79.7% 72.4% 76.8%
IPFT PR 59.6% 92.1% 73.0%


Food hypersensitivity test modalities may be divided into three categories: diet tests, other in vivo tests, and in vitro tests. The classic 2-week elimination diet with individual re-introduction of food risks false-negative responses by allowing development of tolerance for the tested foods by their prolonged avoidance. Also, only so many foods can be withheld at a time for 2 weeks without nutritional supplementation. Repeated diet-trial periods are required for thorough evaluation. Prolonged avoidance of hidden foods, such as yeast, milk, soy, wheat, egg, and corn, is very difficult. Because of the impracticality of such testing, physician and patient compliance is often poor, and results disappointing.

Double-blinded feeding tests using capsules of unidentified foods is not a practical office test for delayed food hypersensitivity. The number of capsules required to provide the volume of the usual challenge feeding is prohibitive. Also, the purity of the test is sacrificed by the capsule itself. Capsules are gelatins of beef or pork.

Although the OCFT is the most practical and accurate of the diet procedures, it does result in occasional false interpretation, negative and positive, as do all diet test modalities. False-positive OCFTs most commonly result from overzealous reporting of potential symptoms, whereas false-negative responses occur from either stories non-reporting of occurring symptoms, the occurrence of unquestioned or unreported delayed symptoms, or failure to maintain a strict 4-day pretest abstinence from the tested food, with resultant masking of the food challenge symptoms. As the classic test the IPFT was designed to replace the OCFT serves well as a standard for comparison.

Single dilution food skin testing has been earlier disavowed as of marginal value by many physicians1,4,8, because the results did not appear to coincide with companion food challenges. The skin tests routinely used were either epicutaneous or no stronger that l:1000 weight: volume intracutaneous tests, much weaker than the maximally used IPFT test dose. Non-correlation of food challenge with skin positive responses should also not be surprising when the usual non-utilization of a 4-day per-challenge fast to allow sensitivity to maximize is considered.

This study determined the IPFT PR to be 73% efficient when compared to the OCFT. The IPFT PR had a specificity of 92% but was only 60% sensitive. The IPFT SK was determined to be 77% efficient, with a more balanced 72% specificity and 80% sensitivity. These figures indicate that dependence on symptom provocation for diagnosis will give very reliable positive responses, but may miss foods that need inclusion in the treatment program. Use of skin response for diagnosis will miss fewer "guilty" foods, but falsely implicate more "innocent" foods.

The IPFT SK is recommended as the standard technique for the following reasons: (1) the IPFT SK was the more valid and reliable: (2) the IPFT SK provides a more cautious approach for the highly sensitized patient, allowing sensitivity screening with weaker doses, without concern over inadvertent neutralization of a later provoking dose; (3) the IPFT SK obviates false-negative IPFT PR responses caused by unrecognized delayed symptom provocation; (4) the IPFT SK allows recognition of delayed skin responses; (5) the IPFT SK allows for symptom provocation to emphasize positivity; and (6) the IPFT SK list of positive responses probably includes the true offenders. In this circumstance, false-positive responses are more easily managed clinically that false-negative responses.


The IPFT is a practical and useful test for food hypersensitivity. It provides a neutralizing dose, if desired, for patient treatment. The IPFT SK provides the physician with more information for correlation with the patient’s history and diet-symptom diary. If the chosen treatment necessitates patient hardship or nutritional peril – a rare occurrence when neutralization treatment is used – OCFT double-testing or treatment trial periods can be accomplished. A successful treatment program that includes false-positive responders appears preferable to a program that fails because needed foods are not included. Use of the IPFT SK as the standard technique makes provocation of symptoms an incidental occurrence rather than a goal. The emphasis on provocation in the current name, intracutaneous provocative food test, become misleading. The more accurately descriptive names, "progressive dilution intracutaneous food test," is recommended as replacement.

This study was sponsored by the American Academy of Otolaryngic Allergy.

The Research Committee of the AAOA wishes to thank statistical consultants John L Fletcher, PhD (Physiological Psychology, University of Kentucky, 1955) and D. Theron Stimmel, PhD (Biostatistics, University of Michigan, 1963) for their review of the data.

The Research Committee also wishes to thank Meridian Bio-Medical Inc, for its cooperation.

Finally, a sincere thank you is extended to the involved staffs of all the participating physicians, especially Trina Whitehead, MLT, and Debbie Garcia, LVN, for their technical assistance, and Nancy Kennedy for her help in manuscript preparation.


To buy products for the chemically sensitive see
For more information on medical treatment see
For more articles on the relationship of health and disease to environmental factors, see the list of available articles and other information available here.