- Email: [email protected]
Allergic skin testing: A clinical investigation
Allergic skin testing: investigation
A clinical
Wayne E. Imber, A.B.* St. Louis, MO.
Forty patients were administered a total of 2,395 skin tests in order to assess the effects oj method, concentration, diluent, manufacturer, and mixtures on patient reaction to inhalant extract solutions. The prick method was found to be superior to the scratch method. The +ttracutaneous technique exhibited the greatest degree of specificity. Patients reacted at a higher level to higher concentrations of extract solution, descending to a lower level of reaction to lesser concentrations of extract solution. Patient reactions were variable to aqueous vs glycerin .solutions. Clear diflerences in the potency of extruct solutions produced by manufacturers A. B, and C were evident. Specifically, the products of manufacturer C were least potent. The results illustrate the possibility of a dilution factor existing in mixtures of extract solutions, producing a decrease in their potency when compared to the potency of their individually administered components. A technique with the potential for wide acceptance and clinical applicability is proposed.
The establishment of the diagnosis of clinical respiratory allergy is a multifactorial process involving the assessment and integration of the patient’s personal and familial history, presenting symptoms and their seasonal variation, environment, occupation, and the results of physical examination as well as specific diagnostic procedures. Direct skin testing is the most widely accepted of the in vivo procedures, because of its high reliability, easy accessibility, and relatively low risk as compared to ophthalmic, nasal, and bronchial provocative testing and the PrausnitzKiistner passive transfer method. The diagnostic and practical value of in vitro procedures such as the radioallergosotbent test (RAST) is still under investigation . Presently, methods of allergic skin testing are remarkably diversified. The three most widely accepted methods are the prick, scratch, and intracutaneous
From the Department of internal Medicine, Section of Allergy and Immunology, St. Louis University School of Medicine. Supported in part by a scholarship from the Allergy Foundation of America. Received for publication Dec. 17, 1976. Accepted for publication May 5, 1977. Reprint requests to: Wayne E. Imber, St. Louis University School of Medicine, 1402 S. Grand Blvd., St. Louis, MO. 63104. *Summa cum laude, Phi Beta Kappa graduate of Washington University and presently a medical student at St. Louis University School of Medicine.
techniques, the latter being the most sensitive, and consequently presenting the greatest risk of a systemic anaphylactic reaction. The concentration of antigenic extracts in clinical use may vary from 1:5 weight/volume (w/v)’ to 1: 2,000,OOOw/v,~ although for the scratch and prick methods 1: 10, 1:20, and 1:50 w/v are most commonly used, as are 1: 1,000, 1: 5,000 and 1: 10,000 for the intracutaneous method. Aqueous and glycerin diluents are widely employed for the prick and scratch methods, while aqueous solutions are used almost exclusively for intracutaneous testing because of the nonspecific irritant properties of glycerin upon injection.” There are several manufacturers of allergenic extracts in the United States, from which three were chosen for a comparative study of the clinical potency of their products. All three were informed of the study and kindly contributed the necessary extract solutions, Mixtures containing two or more antigens in a single testing solution are used in many offices and clinics as screening solutions in the initial skin testing regimen. A savings to the patient of trauma, time, and money is the motivation here. However, according to Lewis and Imber,4 the levels of clinical potency of mixtures and those of their individual components may differ. Method, concentration, diluent, manufacturer, and mixtures are the five variables involved in this study. Each was studied while all other variables were Vol. 60, No. 1, pp. 47-55
48 lmber
TABLE I. Allergenic
J. ALLERGY
extract
solutions
ConcenAllergen
Fungi: Alfernaria Alternaria Alternaria Alternaria Alternaria Alternaria Alternaria Alternaria Alternaria Helminthosporium Helminthosporium Hormodendrum Hormodendrum Mix-Alternaria, Helmin. thosporium, Hormodendrum Grasses: Timothy Timothy Timothy Timothy Timothy Timothy Timothy Timothy Timothy Sweet vernal Sweet vernal Mix-timothy, sweet vernal Trees: Sycamore Sycamore Sycamore Sycamore Sycamore
CLIN. IMMUNOL. JULY 1977
tration
Method
Manufacturer
Diluent
Date of pollen collection
Date of manufacture
Expiration date
1:lO 1:20 1:50 1:20 1:50 1:20 1:20 I:20 1 : 10,000 1:20 1:50 1:20 1:50 1:20
Prick Prick Prick Prick Prick Prick Prick Scratch Intracutaneous Prick Prick Prick Prick Prick
Aqueous Aqueous Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin
A A A A A B C A A A A A A A
l/3/76 l/3/76 l/3/76 9123175 9123175 1l/75 1976 9123175 l/3/76 1219175 1219175 5128175 5128175 -
318176 318176 318176 1218175 1218175 5176 l/29/76 1218175 318176 l/12/76 l/12/75 619175 619175 -
l/23/78 l/23/78 l/23/78 617179 617179 7112177 6129179 617179 l/23/78 617179 617179 617179 617179 -
1: 10 1:20 1:50 1:20 1:50 1:20 1:20 1:20 1: 10,000 1:20 1:50 1:20
Prick Prick Prick Prick Prick Prick Prick Scratch Intracutaneous Prick Prick Prick
Aqueous Aqueous Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin Aqueous Glycerin Glycerin Glycerin
A A A A A B C A A A A A
1974 1974 1974 1974 1974 9173 1974 1974 1974 1973 1973 -
1I7176 l/7/76 l/7/76 4117175 4117175 3176 8122175 4117175 4117175 l/27/75 l/27/75 -
l/7/78 l/7/78 l/7/78 617179 617179 7112177 6129179 617179 617179 617179 617179 -
Prick Prick Prick Prick Prick
Aqueous Aqueous Aqueous Glycerin Glycerin
A A A A A
1975 1975 1975 1974 1974
l/5/76 l/5/76 l/5/76 1O/24/74 1O/24/74
l/7/78 l/7/78 l/7/78 617179 617179
1: 10 1:20 1:50 1:20 1:50
held constant. All analyses, statistics, and conclusions are based upon within-patient assessment, with the patient as a self-matched control. METHODS Patients Patients were selected from the allergy clinic at the St. Louis University School of Medicine Ambulatory Care Center, St. Louis, Missouri. All had been skin-tested previous to this study and were found to be skin-reactive to pollen as well as exhibiting symptomatic allergic rhinitis or asthma. After explaining the procedure, written informed consent was obtained. All patients participated as volunteers and all skin testing was done in a single visit. Patients were instructed to discontinue antihistamine therapy 24 hr before
their appointment with us because of the reported suppressive effect of antihistamine on skin test results53 6 All other pharmacologic therapy continued without interruption.3 A total of 60 skin tests was administered to each of the 40 patients. The mean age of the volunteers was 33 yr, with a range of 18 to 61 yr.
Extracts Extract preparations were kindly contributed to this study by three major United States manufacturers, at 1: 10 w/v in aqueous solution, and 1: 20 w/v in 50% glycerin solution. Western water hemp was supplied at 1: 40 concentration in glycerin solution. Further dilution was performed with physiologic buffered saline when necessary. The date of pollen collection, extract manufacture, and expiration for each solution is included in Table I.
VOLUME NUMBER
TABLE
Allergic
60 1
skin testing
49
I. Cont’d Concentration
Allergen
Trees: Sycamore Sycamore Sycamore Sycamore Box elder Box elder Shellbark hickory Shellbark hickory Black willow Black willow Mix-sycamore, box elder, shellbark hickory, black willow Weeds: Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Western water hemp Western water hemp Lamb’+quarters Lamb’s-quarters Mix-giant ragweed, western water hemp, lamb’s-quarters Control Control Control Control
Skin testing
Method
I:1
100% 1:l 100%
Date of pollen collection
Date of manufacture
Expiration date
Glycerin Glycerin Glycerin Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin Glycerin Glycerin
B C A A A A A A A A A
S/75 1975 1974 1975 1974 1974 1972 1972 1972 I972
5176 I l/10/75 1O/24/74 1I5176 11I6174 1l/6/74 1l/7/72 1l/7/72 6127172 6127172 -
7112177 6129179 617179 l/7/78 1I/6/77 1l/6/77 I l/7/78 1l/7/78 6127178 6127178 -
Prick Aqueous Prick Aqueous Prick Aqueous Prick Glycerin Prick Glycerin Prick Glycerin Prick Glycerin Scratch Glycerin Intracutaneous Aqueous Prick Glycerin Prick Glycerin Prick Glycerin Prick Glycerin Prick Glycerin
A A A A A B C A A A A A A A
1975 1975 1975 1974 1974 10/75 1975 1974 1975 1973 1973 9174 9174 -
I/19/76 l/19/76 l/19/76 3124175 3124175 3176 1O/27/75 3124175 I/19/76 711175 711175 515175 515175 -
l/23/78 l/23/78 l/23/78 I I7178 l/7/78 7112177 6129179 l/7/78 I/23/78 617179 617179 617179 617179 -
Prick Prick Scratch Intracutaneous
-
Prick 1:20 Prick I:20 Scratch 1:20 I : lO,@OO Intracutaneous Prick 1:20 Prick I :50 Prick 1:20 Prick 1:50 Prick 1:20 Prick 1:50 Prick 1:20
1: 10 1:20 1:so 1:20 I :50 1:20 1:20 1:20 I : 10,000 1:40 I:50 1:20 I:50 I:20
Menufacturer
Diluent
Glycerin Aqueous Glycerin Aqueous
methods
Three skin testing methods were employed: prick, scratch, and intracutaneous. The prick and scratch tests were performed on the patient’s back; the intracutaneous tests on the volar forearm. Each set of tests was interpreted 15 min after administration. The extract solutions were randomly arranged on the patient’s backs and on the data sheets in order to eliminate any intrinsic bias in the interpretation of the results. Disposable Monolet lancets were used for the prick method. A drop of extract solution was placed on the skin through which the prick was made.’ The Von Pirquet scarifier, supplied by Hollister-Stier Laboratories, was used for the scratch method. The skin was slightly abraded with the scarifier followed by the addition of a drop of extract solution upon the area. All scratch
-
-
-
solutions were 1: 20 w/v in 50% glycerin solution donated by Company A (Table I). Using a 25- or 26-gauge disposable needle and 1 cc syringe, 0.01-0.02 ml of extract solution was injected into the superficial layers of the skin for the intracutaneous technique. All intracutaneous test solutions were 1: 10,000 w/v in aqueous solution, donated by Company A (Table I). Physiologic buffered saline and/or 50% glycerin control solutions were used for each technique where appropriate (Table 1). The skin tests were interpreted using the O-4+ scale described by Patterson. I 0 No reaction or a reaction no different from the control l+ Erythema no larger than a nickel (21 mm) in diameter
50
lmber
J. ALLERGY
TABLE II. Mean values and Kendall tau p correlation coefficients between administered at 1 : 20 and 1: 50 w/v concentration in glycerin solution Kendall tau /3 correlation coefficients 1:20with 1:50tauP (N)
Fungi: Alternaria Helminthosporium Hormodendrum
patient
CLIN. IMMUNOL. JULY 1977
reaction
to extracts
Mean values I:20 (X 2 SD)
1:50 (i7 2 SD)
1.30 t 1.60
1.58 ?I 1.57
0.88 ? 1.24
1.18 r 1.30
0.48 + 0.93
0.55 k 0.96
2.50 + 1.55
2.48 k 1.54
2.05 5 1.63
1.93 2 1.53
1.00 & 1.32
1.05 + 1.30
0.3248 (31) 0.5083 (29) 0.2858 (30)
1.65 5 1.55
1.48 ” 1.38
1.55 ? 1.38
0.93 -t 1.14
1.20 5 1.18
1.18 + 1.11
0.5582 (31) 0.0750 (27) 0.4845 (30)
2.05 k 1.68
1.58 2 1.60
0.85 k 0.95*
1.00 2 1.20
1.30 * 1.22
1.05 + 1.38
1.45 t 1.39t
1.33 2 1.33
0.5996 (25) 0.0615 (24) -0.2174
(18) Grasses: Timothy Sweet vernal Trees: Sycamore
0.7383 (35) 0.5570 (35) 0.5667
(22) Hickory Box elder Willow Weeds: Giant ragweed Western water hemp Lamb’s-quarters Mean values
0.4061 t
*Western water hemp was administered at 1:40 and 1: 50 w/v concentrations. t Western water hemp was excluded from the Kendall tau p and 1: 20 mean calculations
2f Erythema larger than a nickel in diameter 3+ Erythema and wheal formation without pseudopods 4+ Erythema and wheal formation with pseudopods All procedures were performed and interpreted by the author.
Mixtures The individual allergenic extracts from Company A at I : 20 w/v in glycerin solution were combined by the author in equal parts for each mixture and administered using the prick test (Table I).
Concentrations Diluents Alternaria, timothy, sycamore, and giant ragweed extracts from Company A were administered using the prick test at I : 10, 1:20, and 1:50 w/v in aqueous solution and at I : 20 and I : 50 w/v in glycerin solution (Table I).
Aqueous solutions ofAlternaria, timothy, sycamore, and giant ragweed at 1: 10, 1: 20, and 1: 50 w/v concentrations, as well as glycerin solutions of the same four antigens at 1: 20 and I : 50 w/v, were administered using the prick test. All were contributed by Company A (Table I).
Manufacturers
Data
The three manufacturers will be referred to as Companies A, B, and C. Alternaria, timothy, sycamore, and giant ragweed glycerin extracts from the three companies were administered at 1: 20 w/v using the prick test (Table I).
Skin test results were transferred to computer cards and subsequently analyzed, using the Statistical Package for the Social Sciences (SPSS),s on the IBM/360 computer. Subprogram NONPAR CORR was used in obtaining Kendall
VOLUME NUMBER
Allergic
60 1
III. Mean values and Kendall tau /3 correlation administered at 1 : 10, 1 : 20, and 1 : 50 w/v concentration
TABLE
Kendall
tau p correlation
coefficients between in aqueous solution
patient
skin testing
reaction
51
to extracts
coefficients Mean values
l:lOwith1:50 tau p WI
l:lOwith1:20 tau j? (N)
1:20with1:50 tau p (N)
1:lO (2 rt SD)
1:20 (g k SD)
I:50 (x f SD)
1.70 ? 1.73
1.73 k 1.69
1.38 k 1.53
0.3687 (27)
0.43 17 (25)
0.6743
Timothy
0.6869
0.7020
0.6623
2.30 -t 1.62
2.23 t
1.72
1.95 k 1.57
Sycamore
(33) 0.5390
(34) 0.4916
(33) 0.6222
1.25 + I.21
1.08 t
1.31
I.13 2 1.20
Giant ragweed
(29) 0.4814
(31) 0.4664
(25) 0.4565
2.05 k 1.68
1.70 L 1.62
1.59 + 1.52
(31)
(31)
~29)
0.5190
0.5229
0.6038
1.83 ? 1.56
1.69 -I- 1.59
I.51 t
Alternariu
Mean values
TABLE IV. Mean values administered
in aqueous
(26)
and Kendall tau p correlation and glycerin solutions
coefficients
between
patient
reaction
1.46
to extracts
Kendall tau 6 correlation coefficients I:20 Aqueous with glycerin taup (N)
I:50 Aqueous with glycerin tau p
Alternaria
0.4384
0.5499
(26)
Timothy
(27) 0.6839 (34)
Sycamore
Giant ragweed Mean values
0.2642 (24) 0.6034 (32) 0.4975
(N)
0.6061 (34) 0.2572
Mean values 1:50
1:20 Aqueous (r[ f SD)
Glycerin (R f SD)
Aqueous (x + SD)
Glycerin (it 2 SD)
1.73 k 1.69
1.30 + 1.60
1.38 5 1.53
1.58 +- 1.57
2.23 2 1.72
2.50 -c I.55
1.95 2 1.57
2.48 ri 1.54
1.08 k 1.31
1.00 ? 1.32
1.13 i 1.20
1.05 f 1.30
1.70 5 1.62
2.05 3- 1.68
1.59 t
I.58 k 1.60
1.69 5 1.59
1.71 2 1.54
1.51 rfr 1.46
(26) 0.4887
1.52
(27) 0.4755
tau /3 correlation coefficients. Subprogram PEARSON CORR, statistical option I, provided the mean patient reactivity and N values. All Kendall tau j3 correlation coefficients were basedon the population exhibiting a 1+ or greaterreaction to at least one of the two antigens involved in eachof the calculations, in order to avoid the occurrence of false-positive results. The correlation coefficients calculated for the mixture analyses were based on the population exhibiting a 1+ or greaterreaction to either the 1: 20 w/v or 1: 50 w/v concentration of the particular antigen under consideration. The range of Kendall tau j3 correlation coefficients is - 1.O (a perfect negative correlation) to + 1.O(a perfect positive correlation). The mean patient reactivity (“mean values”) to each antigen extract was calculated for the entire 40-patient population reacting 0 to 4+ with the exception of those involving the intracutaneousmethod, in which one patient
1.67 +- 1.50
was excluded becauseof a systemicreaction to hyposensitization therapy on the day of testing. RESULTS Concentration The degree of correlation between 1: 20 and 1: 50 glycerin extracts was quite variable, as is illustrated in Table II. Timothy produced the highest concentration correlation (7 = 0.7383) as well as the highest mean patient reactivity levels (2.50 * 1.55 for 1:20; 2.48 + 1.54 for I : 50). Alternaria, sweet vernal, sycamore, and giant ragweed are similar in concentration correlation, with Kendall tau p values ranging from T = 0.5570 for sweet vernal to 7 = 0.5996 for Alternaria. Box elder (7 = 0.5083) and Lamb’squarters (7 = 0.4845) follow in their degree of correlation. Helminthosporium, Hormondendrum, hick-
52
lmber
J. ALLERGY
TABLE V. Mean values administered
and Kendall tau p correlation coefficients between patient reaction from manufacturers A, 6, and C at 1 : 20 w/v concentration in glycerin solution Kendall tau /3 correlation coefficients
Alternaria Timothy Sycamore Giant ragweed Mean values
to extracts
Mean values
A with B taup INI
A with C tau p IN)
B with C taup (N)
0.5620 (23) 0.7426 (36) 0.4634 (30) 0.5788 (35)
0.4228 (23) 0.5592 (37) 0.1958 (25) 0.4468 (30)
0.6012 (23) 0.5169 (33) 0.2371 (30) 0.4865 (36)
0.5867
0.4062
0.4604
(X AD, N = 40
B (X ” SD) N=40
1.30 2 1.60
1.53 + 1.69
1.23 r 1.49
2.50 k 1.55
2.30 2 1.65
1.55 2 1.40
1.00 k 1.32
1.50 2 1.32
0.80 4 1.02
2.05 k 1.68
2.65 -r 1.44
1.65 +- 1.56
1.71 k 1.54
2.00 2 1.53
1.31 2 1.37
TABLE VI. Mean values administered
CLIN. IMMUNOL. JULY 1977
and Kendall tau p correlation coefficients between using the prick,* scratch, and intracutaneous methods
patient
(ii AD, N=40
reaction
to extracts
Mean values Prick with intracutaneous tauj? (N)
Alternaria
0.5229
Prick with scratch tau p (N)
Scratch with intracutaneous tau /3 INI
lntracutaneous (x k SD) N = 39
Prick (x 2 SD) N=40
Scratch (x k SD) N = 40
1.30 + 1.60
2.03 -e 1.69
1.95 r 1.89
2.50 -c 1.55
2.63 k 1.64
2.46 ” 1.70
1.00 + 1.32
1.80 + 1.52
1.33 f 1.53
Timothy
0.5383
0.3085 (33) 0.6811
Sycamore
(35) 0.1122
(37) 0.6335
0.2243 (34) 0.3534 (37) 0.3241
Giant ragweed
(29) 0.5043
(31) 0.5535
(32) 0.4761
2.05 k 1.68
2.30 r 1.65
2.15 +- 1.94
Control solutions
(29) 0.2332
(35) 0.3303
(34) 0.1042
0.25 k 0.67
0.80 + 1.14
0.05 2 0.22
Mean valuest
(39) 0.4194
(40) 0.5442
(39) 0.3445
1.71 c 1.54
2.19 k 1.63
1.97 2 1.77
(26)
*Results of the prick method administeredat 1: 20 w/v concentrationin glycerin solution were employed in the above calculations. tMean value calculations do not include the control solutions.
ory, willow, and western water hemp all exhibit low correlation coefficients. The results of patient reaction to the aqueous solutions are seen in Table III. Once again, timothy exhibits the highest correlation coefficients and patient reactivity levels. Among the three concentration comparisons, the 1:20 with 1: 50 correlation coefficients are greatest for the four species administered. This trend holds for Alternariu and sycamore individually as well. However, for timothy, 1: 10 with 1: 50 produced the highest correlation coefficients with 7 = 0.7020. 1: 10 with 1: 20 produced the greatest of giant ragweed’s correlation coefficients (7 = 0.4814). High standard deviation values lead to difficult interpretation of the mean patient
reactivity levels. However, a high (1: 10) to low (1: 50) trend can be detected for each of the species administered. Diluent The results of comparisons between aqueous and glycerin solutions are illustrated in Table IV. Immediately apparent are the low Kendall tau p values for sycamore at both concentrations. One possible explanation is provided by the dates of manufacture in Table I. The aqueous sycamore extract solution was manufactured on 1I5 176, the glycerin extract solution on 10124174. Perhaps the longer storage of the glycerin extract solution resulted in a loss of antigen potency. The low mean patient reactivity levels may
VOLUME NUMBER
Allergic
60 1
also be indicative of potency loss. The remaining correlation Icoefficients are variable, with timothy exhibiting the highest values, ragweed intermediate values, and 141ternaria slightly lower values. The correlation between the 50% glycerin control solution and the saline control was 7 = 0.4600, where N = 40 and p < 0.001. The mean patient reactivity levels, which might be expected to reflect the irritant properties of glycerin if present, show little evidence of this. However, considering the: high levels of standard deviation, the possibility of irritation cannot be ruled out. Manufacturer The results of patient response to manufacturers A, B , and C are summarized in Table V. The correlation coefficients and mean patient reactivity levels indicate a closer association between the potency of the extracts of Companies A and B than between either A or B and Company C. In addition, the mean correlation coefficients reveal the extracts of Company C as the least potent among the three. Those of Company B predominate in three out of the four species administered, the exception being timothy, where Company A elicited the highest level of patient reaction.
Mixtures Ideal analysis of the correlation of mixtures with their components and possible dilution effects would entail the calculation of Kendall’s tau fi for the subpopulation reactive only to one component of the mixture, so that the degree of patient reaction to the mixture would be solely a reflection of the patient’s sensitivity to that particular component. The population of 40 patients under present consideration is not adequate for this type of analysis. Since the mixtures were composed of 1:20 w/v extract solutions, it would be assumed that the degree of correlation between patient reaction to the mixtures and patient reaction to their individual components administered separately at the same concentration
53
TABLE VII. Kendall between individual
tau p correlation coefficients patient reaction to mixtures and their components* Mixture with 1:20 wlv component tau p (N)
Mixture with 1: 50 w/v component tau /3
(N)
Grasses:
Timothy
0.7515
Sweet vernal
(351 0.6107 (35)
0.6597 (35) 0.5878 (35)
0.5714 (25) 0.3112 (18) 0.7029 (24)
0.4361 (25) 0.3419 (18) 0.1950 (24)
0.7114 (31) 0.3496 (30) 0.2298 (27) 1:40 w/v
0.6400 (31) 0.6021 (30) 0.3955
Sycamore
0.3766
0.2045
Hickory
0.3887 (30) 0.4628 (27) 0.1932 (29)
Fungi: Alternaria Hormodendrum Helminothosporium Weeds:
Giant ragweed Lamb’s-quarters Western water hemp?
(27)
Trees:
Method Table VI contains the results of the prick, scratch, and intracutaneous methods. The highest correlations between any two of the three methods considered are seen for the prick and scratch methods in all species except Akrnaria, where prick with intracutaneous predominates. The least degree of correlation occurs between the scratch and intracutaneous methods. Mean patient reactivity levels are greatest for the scratch test in all species administered. Significantly, the aforementioned patterns hold for the controls as well.
skin testing
(21) Box elder
Willow
(21) 0.4543 (30) 0.4808 (27) 0.4180 (29)
*The subpopulationusedfor each calculation consistedof those patientsreacting I+ or greaterto at least one of the two component concentrations. tWestem water hemp was administeredat I :40 and I : 50 w/v concentrations.
would be high. However, if a dilution factor exists and affects mixture potency, it would be expected that the degree of correlation between patient response to the mixture and patient response to each of the mixture’s individual components administered separately at a lower wlv concentration would exceed the correlation between patient response to the mixture and patient response to each of its components administered at a concentration equal to that of each component in the mixture. Therefore, a greater degree of correlation in this study would be expected between patient reaction to each mixture and patient reaction to the individual glycerin extract components of that mixture administered separately at 1: 20 w/v concentration, assuming a mixture potency dilution factor is
54
lmber
nonexistent. Alternatively, assuming the existence of a mixture potency dilution factor, a greater degree of correlation between patient reaction to each mixture and patient reaction to the individual components of that mixture administered separately at 1:SO w/v concentration would be expected. In addition, it would be expected that as the number of components in the mixture increases, the resultant higher dilution factors would increase the likelihood of greater correlation coefficients between patient reaction to each mixture and patient reaction to each of the individual components of that mixture administered separately at 1 : 50 w/v concentration. Referring to Table VII, it will be seen that the two-component grass mixture, containing timothy and sweet vernal, exhibits greater Kendall tau p correlation coefficients with each of its components administered at 1: 20 w/v concentration. The mold mixture, containing the three components Alternaria, Hormodendrum, and Helminthosporium, exhibits higher correlations with Alternaria and Helminthosporium administered at 1: 20 w/v concentration. Hormodendrum, however, correlates at a higher level with the mixture when it was administered at 1: 50 w/v concentration. The weed mixture, also containing three components, i.e., ragweed, lamb’s-quarters, and western water hemp, shows better correlation with ragweed administered at 1:20 w/v concentration. Lamb’s-quarters and western water hemp, on the other hand, correlate better with the mixture when they were administered at 1: 50 w/v concentration. Particularly interesting is the tree mixture, containing four components, three of which (hickory, box elder, and willow) show greater Kendall tau p correlation coefficients with the mixture when they were administered at 1: 50 w/v concentration as compared to their administration at 1: 20 w/v concentration. The 1: 20 w/v components and their mixtures show a descent in their overall level of correlation as the number of components in the mixtures increase, i.e., the mean Kendall tau p correlation coefficient of the grasses (T = 0.6841), with two components, is greater than that of the fungi (5 = 0.5285) and weeds (7 = 0.4303), each with three components, while patient reaction to the tree mixture, consisting of four components, exhibits the least overall degree of correlation with patient reaction to its component extract solutions (T = 0.3553). The overall correlation coefficients between the 1: 50 w/v solutions and the mixtures exhibit a similar but more variable pattern. These data suggest that if in fact a dilution factor is present, the likelihood of a resultant significant dilutional effect on the level of patient skin reaction
J. ALLERGY
CLIN. IMMUNOL. JULY 1977
would increase as the number of components in the mixture increases. These results are inconclusive, however, and require further investigation. DISCUSSION Some interesting patterns of correlation between the three levels of extract concentration employed in this study are elucidated in Tables II and III. It is evident by the mean Kendall tau p values that the 1: 20 and 1: 50 concentration levels exhibit the highest degree of correlation among the aqueous extracts, matched by a similarly high mean value among the same four extracts in glycerin solution (7 = 0.6157). Correlations of 1: 10 with 1: 20 were surprisingly low, considering the twofold difference between them relative to the 2.5- and 5-fold concentration correlations. The aberrant values in Table II may be attributed to the low patient population response to those antigens, as is evidenced by the low mean patient reactivity levels. In general, despite high levels of standard deviation, the mean patient reactivity levels allude to a logical pattern of higher patient response to higher extract concentrations, descending to a lower patient response to lower extract concentrations. Glycerin is known to cause nonspecific irritation upon intradermal injection.3 Except for sycamore, significant (p < 0.001) correlation coefficients were obtained between the aqueous and glycerin solutions studied (Table IV). Causation of potency differences between the sycamore extract solutions was discussed earlier. Differences between the mean patient reactivity levels are overshadowed by high standard deviation levels, and, as a result, no definite pattern emerges. It is clear from Table VI that the highest correlation between the three methods studied was between the prick and scratch techniques. Similarity in site and degree of penetration, i.e., mast cell exposure, may explain these results. Prick with intracutaneous and scratch with intracutaneous followed respectively in their general level of correlation, supporting similar results obtained by Indrajana and co-workers2 using mixed grass pollen extracts. It should be mentioned here that all 40 patients, upon administration of the scratch test, exhibited an irritative wheal-like reaction without pseudopods accompanied by a small amount of erythema. This reaction was elicited by the control solution (50% glycerin) as well as the 1: 20 extract solutions, producing what could naively be considered a 3+ response in nonreactive patients. In addition, when skin reactivity was present, as judged by the prick technique, the intensity of the reaction to the scratch method was masked by the inherent irritation present. Quantitative
VOLUME NUMBER
Allergic skin testing
60 1
reflections of these observations are evident in the high mean patient reactivity levels to the scratch SOlutions in Table VI, particularly the control solution. The prick method is less traumatic and more specific than the scratch as evidenced by the mean patient reactivity control values (0.25 +- 0.67 for glycerin prick; 0.60 5 1.30 for saline prick). The lack of irritation facilirates the interpretation of low levels of patient skin reactivity. A higher correlation with the intracutaneous method (7 = 0.4194 for prick; 7 = 0.3445 for scratch) also indicates the higher specificity of the prick as compared to the scratch method. The intracutaneous method proved to be the most specific method studied, as evidenced by the very low patient reactivity to the control solution (0.05 -+ 0.22). It is well known that the intracutaneous method carries the highest risk of systemic anaphylactic response among the three methods and is often employed only after screening with one of the techniques previously discussed. However, it is a very useful technique for pinpointing skin reactivity in a symptomatic patient exhibiting a questionable response to the seasonally appropriate inhalant prick or scratch test. According to Feinberg and Steinberg,g the potency of pollen shows distinct variation from season to season. Some of the extracts donated by the three manufacturers were not in routine production and were therefore prepared exclusively for this study. Baer and co-wofkers’O found up to a hundredfold variation of active allergen in two batches of ragweed extract labeled as having identical protein nitrogen unit (PNU) content. The same group found considerable variation between the antigen E content of extracts from different companies, as well as between different lots of the same company. It is clear then that the results presented earlier pertaining to the three companies studied represent a limited sample of their products an.d are not necessarily applicable as a quality indicator. The results of the patient reactions elicited by the mixtures as compared to those elicited by their components, although inconclusive, raise certain questions concerning the sensitivity of mixtures as a screening method. Some clinics use mixtures containing ten or twelve components in their initial skin testing regimens, followed by administration of the individual components contingent upon a positive patient response to the appropriate mixture. The magnitude of the dilution factor produced by such a mixture
55
quite possibly results in the oversight of skin reactivity in some allergic patients. Considering the current diversity in allergic skin testing throughout the United States and the world, a single method of skin testing which could be universally accepted and clinically employed would be most advantageous. As a step in that direction, based on the results of this study, the prick test administered at a concentration of 1: 20 w/v in aqueous solution is proposed. Hopefully, this study will lead to further clinical investigation into the methodology of allergic skin testing. 1 would like to express my gratitude to Raymond G. Slavin, M.D., Director of the Section of Allergy and Immunology, St. Louis University School of Medicine, for his
professional guidance and review of the manuscript and the enthusiastic assistance of his patients and staff. Many thanks to Greer Laboratories, Lenoir, N. C., Hollister-
Stier Laboratories, Spokane,Wash., and Center Laboratories, Port Washington, N. Y. for contributing the inhalant extract solutions, and Clement J. Sullivan, M.D.,
and David A. Specht, Ph.D., for their review of the manuscript. I also want to expressmy appreciationto Mrs. Joyce Edwards for her typing of the manuscript. REFERENCES
I. Sheldon,J. M., Lovell, R. G., and Mathews, K. P.: A manual of clinical allergy, ed. 2, Philadelphia, 1967. W. B. Saunders co., p. 56. 2. Indrajana, T., Spieksma,F. Th. M., and Voorhorst, R.: Comparative study of the intracutaneous,scratch and prick tests in allergy, Ann. Allergy 29~639, 1971. 3. Slavin, R. G.: Skin tests in the diagnosis of allergies of the immediate type, Med. Clin. North Am. 55~65, 1974. 4. Lewis, W. H., and Imber, W. E.: Allergy epidemiology in the St. Louis, Missouri area. II. Grasses, Ann. Allergy 35:42, 1975. 5. Arbesman, C. E., Loepf, Cl. F., and Miller, G. E.: Some antianaphylactic and antihistamine properties of N’pyridyl, N’bensyl, demethyethylenediamine monohydrochloride, J. ALLERGY 17:203, 1946.
6. Cook, T. J., Macqueen,D. M., Wittig, H. J., Thornby, J. I., Lantos, R. L., and Virtue, C. M.: Degreeand duration of skin test suppressionand side effects with antihistamine, J. ALLERGY51:71, 1973.
7. Patterson,R.: Allergic diseases:Diagnosis and management, Philadelphia, 1972, J. B. Lippincott Co., p. 77. 8. Nie, N., Hull, C. H., Jenkins. J., Steinbrenner,K., and Bent, D.: Statistical package for the social sciences, ed. 2. New York, 1975, McGraw-Hill Book Co. 9. Feinberg, S. M., and Steinberg, M. J.: Studies in pollen potency, J. ALLERGY 5: 19, 1933. 10. Baer, H., Godfrey, H., Maloney, C. J., Norman, P. S., and Lichtenstein, L. M.: The potency and antigen E content of commercially preparedragweedextracts, J. ALLERGY 45:347, 1970.
A clinical
Wayne E. Imber, A.B.* St. Louis, MO.
Forty patients were administered a total of 2,395 skin tests in order to assess the effects oj method, concentration, diluent, manufacturer, and mixtures on patient reaction to inhalant extract solutions. The prick method was found to be superior to the scratch method. The +ttracutaneous technique exhibited the greatest degree of specificity. Patients reacted at a higher level to higher concentrations of extract solution, descending to a lower level of reaction to lesser concentrations of extract solution. Patient reactions were variable to aqueous vs glycerin .solutions. Clear diflerences in the potency of extruct solutions produced by manufacturers A. B, and C were evident. Specifically, the products of manufacturer C were least potent. The results illustrate the possibility of a dilution factor existing in mixtures of extract solutions, producing a decrease in their potency when compared to the potency of their individually administered components. A technique with the potential for wide acceptance and clinical applicability is proposed.
The establishment of the diagnosis of clinical respiratory allergy is a multifactorial process involving the assessment and integration of the patient’s personal and familial history, presenting symptoms and their seasonal variation, environment, occupation, and the results of physical examination as well as specific diagnostic procedures. Direct skin testing is the most widely accepted of the in vivo procedures, because of its high reliability, easy accessibility, and relatively low risk as compared to ophthalmic, nasal, and bronchial provocative testing and the PrausnitzKiistner passive transfer method. The diagnostic and practical value of in vitro procedures such as the radioallergosotbent test (RAST) is still under investigation . Presently, methods of allergic skin testing are remarkably diversified. The three most widely accepted methods are the prick, scratch, and intracutaneous
From the Department of internal Medicine, Section of Allergy and Immunology, St. Louis University School of Medicine. Supported in part by a scholarship from the Allergy Foundation of America. Received for publication Dec. 17, 1976. Accepted for publication May 5, 1977. Reprint requests to: Wayne E. Imber, St. Louis University School of Medicine, 1402 S. Grand Blvd., St. Louis, MO. 63104. *Summa cum laude, Phi Beta Kappa graduate of Washington University and presently a medical student at St. Louis University School of Medicine.
techniques, the latter being the most sensitive, and consequently presenting the greatest risk of a systemic anaphylactic reaction. The concentration of antigenic extracts in clinical use may vary from 1:5 weight/volume (w/v)’ to 1: 2,000,OOOw/v,~ although for the scratch and prick methods 1: 10, 1:20, and 1:50 w/v are most commonly used, as are 1: 1,000, 1: 5,000 and 1: 10,000 for the intracutaneous method. Aqueous and glycerin diluents are widely employed for the prick and scratch methods, while aqueous solutions are used almost exclusively for intracutaneous testing because of the nonspecific irritant properties of glycerin upon injection.” There are several manufacturers of allergenic extracts in the United States, from which three were chosen for a comparative study of the clinical potency of their products. All three were informed of the study and kindly contributed the necessary extract solutions, Mixtures containing two or more antigens in a single testing solution are used in many offices and clinics as screening solutions in the initial skin testing regimen. A savings to the patient of trauma, time, and money is the motivation here. However, according to Lewis and Imber,4 the levels of clinical potency of mixtures and those of their individual components may differ. Method, concentration, diluent, manufacturer, and mixtures are the five variables involved in this study. Each was studied while all other variables were Vol. 60, No. 1, pp. 47-55
48 lmber
TABLE I. Allergenic
J. ALLERGY
extract
solutions
ConcenAllergen
Fungi: Alfernaria Alternaria Alternaria Alternaria Alternaria Alternaria Alternaria Alternaria Alternaria Helminthosporium Helminthosporium Hormodendrum Hormodendrum Mix-Alternaria, Helmin. thosporium, Hormodendrum Grasses: Timothy Timothy Timothy Timothy Timothy Timothy Timothy Timothy Timothy Sweet vernal Sweet vernal Mix-timothy, sweet vernal Trees: Sycamore Sycamore Sycamore Sycamore Sycamore
CLIN. IMMUNOL. JULY 1977
tration
Method
Manufacturer
Diluent
Date of pollen collection
Date of manufacture
Expiration date
1:lO 1:20 1:50 1:20 1:50 1:20 1:20 I:20 1 : 10,000 1:20 1:50 1:20 1:50 1:20
Prick Prick Prick Prick Prick Prick Prick Scratch Intracutaneous Prick Prick Prick Prick Prick
Aqueous Aqueous Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin
A A A A A B C A A A A A A A
l/3/76 l/3/76 l/3/76 9123175 9123175 1l/75 1976 9123175 l/3/76 1219175 1219175 5128175 5128175 -
318176 318176 318176 1218175 1218175 5176 l/29/76 1218175 318176 l/12/76 l/12/75 619175 619175 -
l/23/78 l/23/78 l/23/78 617179 617179 7112177 6129179 617179 l/23/78 617179 617179 617179 617179 -
1: 10 1:20 1:50 1:20 1:50 1:20 1:20 1:20 1: 10,000 1:20 1:50 1:20
Prick Prick Prick Prick Prick Prick Prick Scratch Intracutaneous Prick Prick Prick
Aqueous Aqueous Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin Aqueous Glycerin Glycerin Glycerin
A A A A A B C A A A A A
1974 1974 1974 1974 1974 9173 1974 1974 1974 1973 1973 -
1I7176 l/7/76 l/7/76 4117175 4117175 3176 8122175 4117175 4117175 l/27/75 l/27/75 -
l/7/78 l/7/78 l/7/78 617179 617179 7112177 6129179 617179 617179 617179 617179 -
Prick Prick Prick Prick Prick
Aqueous Aqueous Aqueous Glycerin Glycerin
A A A A A
1975 1975 1975 1974 1974
l/5/76 l/5/76 l/5/76 1O/24/74 1O/24/74
l/7/78 l/7/78 l/7/78 617179 617179
1: 10 1:20 1:50 1:20 1:50
held constant. All analyses, statistics, and conclusions are based upon within-patient assessment, with the patient as a self-matched control. METHODS Patients Patients were selected from the allergy clinic at the St. Louis University School of Medicine Ambulatory Care Center, St. Louis, Missouri. All had been skin-tested previous to this study and were found to be skin-reactive to pollen as well as exhibiting symptomatic allergic rhinitis or asthma. After explaining the procedure, written informed consent was obtained. All patients participated as volunteers and all skin testing was done in a single visit. Patients were instructed to discontinue antihistamine therapy 24 hr before
their appointment with us because of the reported suppressive effect of antihistamine on skin test results53 6 All other pharmacologic therapy continued without interruption.3 A total of 60 skin tests was administered to each of the 40 patients. The mean age of the volunteers was 33 yr, with a range of 18 to 61 yr.
Extracts Extract preparations were kindly contributed to this study by three major United States manufacturers, at 1: 10 w/v in aqueous solution, and 1: 20 w/v in 50% glycerin solution. Western water hemp was supplied at 1: 40 concentration in glycerin solution. Further dilution was performed with physiologic buffered saline when necessary. The date of pollen collection, extract manufacture, and expiration for each solution is included in Table I.
VOLUME NUMBER
TABLE
Allergic
60 1
skin testing
49
I. Cont’d Concentration
Allergen
Trees: Sycamore Sycamore Sycamore Sycamore Box elder Box elder Shellbark hickory Shellbark hickory Black willow Black willow Mix-sycamore, box elder, shellbark hickory, black willow Weeds: Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Giant ragweed Western water hemp Western water hemp Lamb’+quarters Lamb’s-quarters Mix-giant ragweed, western water hemp, lamb’s-quarters Control Control Control Control
Skin testing
Method
I:1
100% 1:l 100%
Date of pollen collection
Date of manufacture
Expiration date
Glycerin Glycerin Glycerin Aqueous Glycerin Glycerin Glycerin Glycerin Glycerin Glycerin Glycerin
B C A A A A A A A A A
S/75 1975 1974 1975 1974 1974 1972 1972 1972 I972
5176 I l/10/75 1O/24/74 1I5176 11I6174 1l/6/74 1l/7/72 1l/7/72 6127172 6127172 -
7112177 6129179 617179 l/7/78 1I/6/77 1l/6/77 I l/7/78 1l/7/78 6127178 6127178 -
Prick Aqueous Prick Aqueous Prick Aqueous Prick Glycerin Prick Glycerin Prick Glycerin Prick Glycerin Scratch Glycerin Intracutaneous Aqueous Prick Glycerin Prick Glycerin Prick Glycerin Prick Glycerin Prick Glycerin
A A A A A B C A A A A A A A
1975 1975 1975 1974 1974 10/75 1975 1974 1975 1973 1973 9174 9174 -
I/19/76 l/19/76 l/19/76 3124175 3124175 3176 1O/27/75 3124175 I/19/76 711175 711175 515175 515175 -
l/23/78 l/23/78 l/23/78 I I7178 l/7/78 7112177 6129179 l/7/78 I/23/78 617179 617179 617179 617179 -
Prick Prick Scratch Intracutaneous
-
Prick 1:20 Prick I:20 Scratch 1:20 I : lO,@OO Intracutaneous Prick 1:20 Prick I :50 Prick 1:20 Prick 1:50 Prick 1:20 Prick 1:50 Prick 1:20
1: 10 1:20 1:so 1:20 I :50 1:20 1:20 1:20 I : 10,000 1:40 I:50 1:20 I:50 I:20
Menufacturer
Diluent
Glycerin Aqueous Glycerin Aqueous
methods
Three skin testing methods were employed: prick, scratch, and intracutaneous. The prick and scratch tests were performed on the patient’s back; the intracutaneous tests on the volar forearm. Each set of tests was interpreted 15 min after administration. The extract solutions were randomly arranged on the patient’s backs and on the data sheets in order to eliminate any intrinsic bias in the interpretation of the results. Disposable Monolet lancets were used for the prick method. A drop of extract solution was placed on the skin through which the prick was made.’ The Von Pirquet scarifier, supplied by Hollister-Stier Laboratories, was used for the scratch method. The skin was slightly abraded with the scarifier followed by the addition of a drop of extract solution upon the area. All scratch
-
-
-
solutions were 1: 20 w/v in 50% glycerin solution donated by Company A (Table I). Using a 25- or 26-gauge disposable needle and 1 cc syringe, 0.01-0.02 ml of extract solution was injected into the superficial layers of the skin for the intracutaneous technique. All intracutaneous test solutions were 1: 10,000 w/v in aqueous solution, donated by Company A (Table I). Physiologic buffered saline and/or 50% glycerin control solutions were used for each technique where appropriate (Table 1). The skin tests were interpreted using the O-4+ scale described by Patterson. I 0 No reaction or a reaction no different from the control l+ Erythema no larger than a nickel (21 mm) in diameter
50
lmber
J. ALLERGY
TABLE II. Mean values and Kendall tau p correlation coefficients between administered at 1 : 20 and 1: 50 w/v concentration in glycerin solution Kendall tau /3 correlation coefficients 1:20with 1:50tauP (N)
Fungi: Alternaria Helminthosporium Hormodendrum
patient
CLIN. IMMUNOL. JULY 1977
reaction
to extracts
Mean values I:20 (X 2 SD)
1:50 (i7 2 SD)
1.30 t 1.60
1.58 ?I 1.57
0.88 ? 1.24
1.18 r 1.30
0.48 + 0.93
0.55 k 0.96
2.50 + 1.55
2.48 k 1.54
2.05 5 1.63
1.93 2 1.53
1.00 & 1.32
1.05 + 1.30
0.3248 (31) 0.5083 (29) 0.2858 (30)
1.65 5 1.55
1.48 ” 1.38
1.55 ? 1.38
0.93 -t 1.14
1.20 5 1.18
1.18 + 1.11
0.5582 (31) 0.0750 (27) 0.4845 (30)
2.05 k 1.68
1.58 2 1.60
0.85 k 0.95*
1.00 2 1.20
1.30 * 1.22
1.05 + 1.38
1.45 t 1.39t
1.33 2 1.33
0.5996 (25) 0.0615 (24) -0.2174
(18) Grasses: Timothy Sweet vernal Trees: Sycamore
0.7383 (35) 0.5570 (35) 0.5667
(22) Hickory Box elder Willow Weeds: Giant ragweed Western water hemp Lamb’s-quarters Mean values
0.4061 t
*Western water hemp was administered at 1:40 and 1: 50 w/v concentrations. t Western water hemp was excluded from the Kendall tau p and 1: 20 mean calculations
2f Erythema larger than a nickel in diameter 3+ Erythema and wheal formation without pseudopods 4+ Erythema and wheal formation with pseudopods All procedures were performed and interpreted by the author.
Mixtures The individual allergenic extracts from Company A at I : 20 w/v in glycerin solution were combined by the author in equal parts for each mixture and administered using the prick test (Table I).
Concentrations Diluents Alternaria, timothy, sycamore, and giant ragweed extracts from Company A were administered using the prick test at I : 10, 1:20, and 1:50 w/v in aqueous solution and at I : 20 and I : 50 w/v in glycerin solution (Table I).
Aqueous solutions ofAlternaria, timothy, sycamore, and giant ragweed at 1: 10, 1: 20, and 1: 50 w/v concentrations, as well as glycerin solutions of the same four antigens at 1: 20 and I : 50 w/v, were administered using the prick test. All were contributed by Company A (Table I).
Manufacturers
Data
The three manufacturers will be referred to as Companies A, B, and C. Alternaria, timothy, sycamore, and giant ragweed glycerin extracts from the three companies were administered at 1: 20 w/v using the prick test (Table I).
Skin test results were transferred to computer cards and subsequently analyzed, using the Statistical Package for the Social Sciences (SPSS),s on the IBM/360 computer. Subprogram NONPAR CORR was used in obtaining Kendall
VOLUME NUMBER
Allergic
60 1
III. Mean values and Kendall tau /3 correlation administered at 1 : 10, 1 : 20, and 1 : 50 w/v concentration
TABLE
Kendall
tau p correlation
coefficients between in aqueous solution
patient
skin testing
reaction
51
to extracts
coefficients Mean values
l:lOwith1:50 tau p WI
l:lOwith1:20 tau j? (N)
1:20with1:50 tau p (N)
1:lO (2 rt SD)
1:20 (g k SD)
I:50 (x f SD)
1.70 ? 1.73
1.73 k 1.69
1.38 k 1.53
0.3687 (27)
0.43 17 (25)
0.6743
Timothy
0.6869
0.7020
0.6623
2.30 -t 1.62
2.23 t
1.72
1.95 k 1.57
Sycamore
(33) 0.5390
(34) 0.4916
(33) 0.6222
1.25 + I.21
1.08 t
1.31
I.13 2 1.20
Giant ragweed
(29) 0.4814
(31) 0.4664
(25) 0.4565
2.05 k 1.68
1.70 L 1.62
1.59 + 1.52
(31)
(31)
~29)
0.5190
0.5229
0.6038
1.83 ? 1.56
1.69 -I- 1.59
I.51 t
Alternariu
Mean values
TABLE IV. Mean values administered
in aqueous
(26)
and Kendall tau p correlation and glycerin solutions
coefficients
between
patient
reaction
1.46
to extracts
Kendall tau 6 correlation coefficients I:20 Aqueous with glycerin taup (N)
I:50 Aqueous with glycerin tau p
Alternaria
0.4384
0.5499
(26)
Timothy
(27) 0.6839 (34)
Sycamore
Giant ragweed Mean values
0.2642 (24) 0.6034 (32) 0.4975
(N)
0.6061 (34) 0.2572
Mean values 1:50
1:20 Aqueous (r[ f SD)
Glycerin (R f SD)
Aqueous (x + SD)
Glycerin (it 2 SD)
1.73 k 1.69
1.30 + 1.60
1.38 5 1.53
1.58 +- 1.57
2.23 2 1.72
2.50 -c I.55
1.95 2 1.57
2.48 ri 1.54
1.08 k 1.31
1.00 ? 1.32
1.13 i 1.20
1.05 f 1.30
1.70 5 1.62
2.05 3- 1.68
1.59 t
I.58 k 1.60
1.69 5 1.59
1.71 2 1.54
1.51 rfr 1.46
(26) 0.4887
1.52
(27) 0.4755
tau /3 correlation coefficients. Subprogram PEARSON CORR, statistical option I, provided the mean patient reactivity and N values. All Kendall tau j3 correlation coefficients were basedon the population exhibiting a 1+ or greaterreaction to at least one of the two antigens involved in eachof the calculations, in order to avoid the occurrence of false-positive results. The correlation coefficients calculated for the mixture analyses were based on the population exhibiting a 1+ or greaterreaction to either the 1: 20 w/v or 1: 50 w/v concentration of the particular antigen under consideration. The range of Kendall tau j3 correlation coefficients is - 1.O (a perfect negative correlation) to + 1.O(a perfect positive correlation). The mean patient reactivity (“mean values”) to each antigen extract was calculated for the entire 40-patient population reacting 0 to 4+ with the exception of those involving the intracutaneousmethod, in which one patient
1.67 +- 1.50
was excluded becauseof a systemicreaction to hyposensitization therapy on the day of testing. RESULTS Concentration The degree of correlation between 1: 20 and 1: 50 glycerin extracts was quite variable, as is illustrated in Table II. Timothy produced the highest concentration correlation (7 = 0.7383) as well as the highest mean patient reactivity levels (2.50 * 1.55 for 1:20; 2.48 + 1.54 for I : 50). Alternaria, sweet vernal, sycamore, and giant ragweed are similar in concentration correlation, with Kendall tau p values ranging from T = 0.5570 for sweet vernal to 7 = 0.5996 for Alternaria. Box elder (7 = 0.5083) and Lamb’squarters (7 = 0.4845) follow in their degree of correlation. Helminthosporium, Hormondendrum, hick-
52
lmber
J. ALLERGY
TABLE V. Mean values administered
and Kendall tau p correlation coefficients between patient reaction from manufacturers A, 6, and C at 1 : 20 w/v concentration in glycerin solution Kendall tau /3 correlation coefficients
Alternaria Timothy Sycamore Giant ragweed Mean values
to extracts
Mean values
A with B taup INI
A with C tau p IN)
B with C taup (N)
0.5620 (23) 0.7426 (36) 0.4634 (30) 0.5788 (35)
0.4228 (23) 0.5592 (37) 0.1958 (25) 0.4468 (30)
0.6012 (23) 0.5169 (33) 0.2371 (30) 0.4865 (36)
0.5867
0.4062
0.4604
(X AD, N = 40
B (X ” SD) N=40
1.30 2 1.60
1.53 + 1.69
1.23 r 1.49
2.50 k 1.55
2.30 2 1.65
1.55 2 1.40
1.00 k 1.32
1.50 2 1.32
0.80 4 1.02
2.05 k 1.68
2.65 -r 1.44
1.65 +- 1.56
1.71 k 1.54
2.00 2 1.53
1.31 2 1.37
TABLE VI. Mean values administered
CLIN. IMMUNOL. JULY 1977
and Kendall tau p correlation coefficients between using the prick,* scratch, and intracutaneous methods
patient
(ii AD, N=40
reaction
to extracts
Mean values Prick with intracutaneous tauj? (N)
Alternaria
0.5229
Prick with scratch tau p (N)
Scratch with intracutaneous tau /3 INI
lntracutaneous (x k SD) N = 39
Prick (x 2 SD) N=40
Scratch (x k SD) N = 40
1.30 + 1.60
2.03 -e 1.69
1.95 r 1.89
2.50 -c 1.55
2.63 k 1.64
2.46 ” 1.70
1.00 + 1.32
1.80 + 1.52
1.33 f 1.53
Timothy
0.5383
0.3085 (33) 0.6811
Sycamore
(35) 0.1122
(37) 0.6335
0.2243 (34) 0.3534 (37) 0.3241
Giant ragweed
(29) 0.5043
(31) 0.5535
(32) 0.4761
2.05 k 1.68
2.30 r 1.65
2.15 +- 1.94
Control solutions
(29) 0.2332
(35) 0.3303
(34) 0.1042
0.25 k 0.67
0.80 + 1.14
0.05 2 0.22
Mean valuest
(39) 0.4194
(40) 0.5442
(39) 0.3445
1.71 c 1.54
2.19 k 1.63
1.97 2 1.77
(26)
*Results of the prick method administeredat 1: 20 w/v concentrationin glycerin solution were employed in the above calculations. tMean value calculations do not include the control solutions.
ory, willow, and western water hemp all exhibit low correlation coefficients. The results of patient reaction to the aqueous solutions are seen in Table III. Once again, timothy exhibits the highest correlation coefficients and patient reactivity levels. Among the three concentration comparisons, the 1:20 with 1: 50 correlation coefficients are greatest for the four species administered. This trend holds for Alternariu and sycamore individually as well. However, for timothy, 1: 10 with 1: 50 produced the highest correlation coefficients with 7 = 0.7020. 1: 10 with 1: 20 produced the greatest of giant ragweed’s correlation coefficients (7 = 0.4814). High standard deviation values lead to difficult interpretation of the mean patient
reactivity levels. However, a high (1: 10) to low (1: 50) trend can be detected for each of the species administered. Diluent The results of comparisons between aqueous and glycerin solutions are illustrated in Table IV. Immediately apparent are the low Kendall tau p values for sycamore at both concentrations. One possible explanation is provided by the dates of manufacture in Table I. The aqueous sycamore extract solution was manufactured on 1I5 176, the glycerin extract solution on 10124174. Perhaps the longer storage of the glycerin extract solution resulted in a loss of antigen potency. The low mean patient reactivity levels may
VOLUME NUMBER
Allergic
60 1
also be indicative of potency loss. The remaining correlation Icoefficients are variable, with timothy exhibiting the highest values, ragweed intermediate values, and 141ternaria slightly lower values. The correlation between the 50% glycerin control solution and the saline control was 7 = 0.4600, where N = 40 and p < 0.001. The mean patient reactivity levels, which might be expected to reflect the irritant properties of glycerin if present, show little evidence of this. However, considering the: high levels of standard deviation, the possibility of irritation cannot be ruled out. Manufacturer The results of patient response to manufacturers A, B , and C are summarized in Table V. The correlation coefficients and mean patient reactivity levels indicate a closer association between the potency of the extracts of Companies A and B than between either A or B and Company C. In addition, the mean correlation coefficients reveal the extracts of Company C as the least potent among the three. Those of Company B predominate in three out of the four species administered, the exception being timothy, where Company A elicited the highest level of patient reaction.
Mixtures Ideal analysis of the correlation of mixtures with their components and possible dilution effects would entail the calculation of Kendall’s tau fi for the subpopulation reactive only to one component of the mixture, so that the degree of patient reaction to the mixture would be solely a reflection of the patient’s sensitivity to that particular component. The population of 40 patients under present consideration is not adequate for this type of analysis. Since the mixtures were composed of 1:20 w/v extract solutions, it would be assumed that the degree of correlation between patient reaction to the mixtures and patient reaction to their individual components administered separately at the same concentration
53
TABLE VII. Kendall between individual
tau p correlation coefficients patient reaction to mixtures and their components* Mixture with 1:20 wlv component tau p (N)
Mixture with 1: 50 w/v component tau /3
(N)
Grasses:
Timothy
0.7515
Sweet vernal
(351 0.6107 (35)
0.6597 (35) 0.5878 (35)
0.5714 (25) 0.3112 (18) 0.7029 (24)
0.4361 (25) 0.3419 (18) 0.1950 (24)
0.7114 (31) 0.3496 (30) 0.2298 (27) 1:40 w/v
0.6400 (31) 0.6021 (30) 0.3955
Sycamore
0.3766
0.2045
Hickory
0.3887 (30) 0.4628 (27) 0.1932 (29)
Fungi: Alternaria Hormodendrum Helminothosporium Weeds:
Giant ragweed Lamb’s-quarters Western water hemp?
(27)
Trees:
Method Table VI contains the results of the prick, scratch, and intracutaneous methods. The highest correlations between any two of the three methods considered are seen for the prick and scratch methods in all species except Akrnaria, where prick with intracutaneous predominates. The least degree of correlation occurs between the scratch and intracutaneous methods. Mean patient reactivity levels are greatest for the scratch test in all species administered. Significantly, the aforementioned patterns hold for the controls as well.
skin testing
(21) Box elder
Willow
(21) 0.4543 (30) 0.4808 (27) 0.4180 (29)
*The subpopulationusedfor each calculation consistedof those patientsreacting I+ or greaterto at least one of the two component concentrations. tWestem water hemp was administeredat I :40 and I : 50 w/v concentrations.
would be high. However, if a dilution factor exists and affects mixture potency, it would be expected that the degree of correlation between patient response to the mixture and patient response to each of the mixture’s individual components administered separately at a lower wlv concentration would exceed the correlation between patient response to the mixture and patient response to each of its components administered at a concentration equal to that of each component in the mixture. Therefore, a greater degree of correlation in this study would be expected between patient reaction to each mixture and patient reaction to the individual glycerin extract components of that mixture administered separately at 1: 20 w/v concentration, assuming a mixture potency dilution factor is
54
lmber
nonexistent. Alternatively, assuming the existence of a mixture potency dilution factor, a greater degree of correlation between patient reaction to each mixture and patient reaction to the individual components of that mixture administered separately at 1:SO w/v concentration would be expected. In addition, it would be expected that as the number of components in the mixture increases, the resultant higher dilution factors would increase the likelihood of greater correlation coefficients between patient reaction to each mixture and patient reaction to each of the individual components of that mixture administered separately at 1 : 50 w/v concentration. Referring to Table VII, it will be seen that the two-component grass mixture, containing timothy and sweet vernal, exhibits greater Kendall tau p correlation coefficients with each of its components administered at 1: 20 w/v concentration. The mold mixture, containing the three components Alternaria, Hormodendrum, and Helminthosporium, exhibits higher correlations with Alternaria and Helminthosporium administered at 1: 20 w/v concentration. Hormodendrum, however, correlates at a higher level with the mixture when it was administered at 1: 50 w/v concentration. The weed mixture, also containing three components, i.e., ragweed, lamb’s-quarters, and western water hemp, shows better correlation with ragweed administered at 1:20 w/v concentration. Lamb’s-quarters and western water hemp, on the other hand, correlate better with the mixture when they were administered at 1: 50 w/v concentration. Particularly interesting is the tree mixture, containing four components, three of which (hickory, box elder, and willow) show greater Kendall tau p correlation coefficients with the mixture when they were administered at 1: 50 w/v concentration as compared to their administration at 1: 20 w/v concentration. The 1: 20 w/v components and their mixtures show a descent in their overall level of correlation as the number of components in the mixtures increase, i.e., the mean Kendall tau p correlation coefficient of the grasses (T = 0.6841), with two components, is greater than that of the fungi (5 = 0.5285) and weeds (7 = 0.4303), each with three components, while patient reaction to the tree mixture, consisting of four components, exhibits the least overall degree of correlation with patient reaction to its component extract solutions (T = 0.3553). The overall correlation coefficients between the 1: 50 w/v solutions and the mixtures exhibit a similar but more variable pattern. These data suggest that if in fact a dilution factor is present, the likelihood of a resultant significant dilutional effect on the level of patient skin reaction
J. ALLERGY
CLIN. IMMUNOL. JULY 1977
would increase as the number of components in the mixture increases. These results are inconclusive, however, and require further investigation. DISCUSSION Some interesting patterns of correlation between the three levels of extract concentration employed in this study are elucidated in Tables II and III. It is evident by the mean Kendall tau p values that the 1: 20 and 1: 50 concentration levels exhibit the highest degree of correlation among the aqueous extracts, matched by a similarly high mean value among the same four extracts in glycerin solution (7 = 0.6157). Correlations of 1: 10 with 1: 20 were surprisingly low, considering the twofold difference between them relative to the 2.5- and 5-fold concentration correlations. The aberrant values in Table II may be attributed to the low patient population response to those antigens, as is evidenced by the low mean patient reactivity levels. In general, despite high levels of standard deviation, the mean patient reactivity levels allude to a logical pattern of higher patient response to higher extract concentrations, descending to a lower patient response to lower extract concentrations. Glycerin is known to cause nonspecific irritation upon intradermal injection.3 Except for sycamore, significant (p < 0.001) correlation coefficients were obtained between the aqueous and glycerin solutions studied (Table IV). Causation of potency differences between the sycamore extract solutions was discussed earlier. Differences between the mean patient reactivity levels are overshadowed by high standard deviation levels, and, as a result, no definite pattern emerges. It is clear from Table VI that the highest correlation between the three methods studied was between the prick and scratch techniques. Similarity in site and degree of penetration, i.e., mast cell exposure, may explain these results. Prick with intracutaneous and scratch with intracutaneous followed respectively in their general level of correlation, supporting similar results obtained by Indrajana and co-workers2 using mixed grass pollen extracts. It should be mentioned here that all 40 patients, upon administration of the scratch test, exhibited an irritative wheal-like reaction without pseudopods accompanied by a small amount of erythema. This reaction was elicited by the control solution (50% glycerin) as well as the 1: 20 extract solutions, producing what could naively be considered a 3+ response in nonreactive patients. In addition, when skin reactivity was present, as judged by the prick technique, the intensity of the reaction to the scratch method was masked by the inherent irritation present. Quantitative
VOLUME NUMBER
Allergic skin testing
60 1
reflections of these observations are evident in the high mean patient reactivity levels to the scratch SOlutions in Table VI, particularly the control solution. The prick method is less traumatic and more specific than the scratch as evidenced by the mean patient reactivity control values (0.25 +- 0.67 for glycerin prick; 0.60 5 1.30 for saline prick). The lack of irritation facilirates the interpretation of low levels of patient skin reactivity. A higher correlation with the intracutaneous method (7 = 0.4194 for prick; 7 = 0.3445 for scratch) also indicates the higher specificity of the prick as compared to the scratch method. The intracutaneous method proved to be the most specific method studied, as evidenced by the very low patient reactivity to the control solution (0.05 -+ 0.22). It is well known that the intracutaneous method carries the highest risk of systemic anaphylactic response among the three methods and is often employed only after screening with one of the techniques previously discussed. However, it is a very useful technique for pinpointing skin reactivity in a symptomatic patient exhibiting a questionable response to the seasonally appropriate inhalant prick or scratch test. According to Feinberg and Steinberg,g the potency of pollen shows distinct variation from season to season. Some of the extracts donated by the three manufacturers were not in routine production and were therefore prepared exclusively for this study. Baer and co-wofkers’O found up to a hundredfold variation of active allergen in two batches of ragweed extract labeled as having identical protein nitrogen unit (PNU) content. The same group found considerable variation between the antigen E content of extracts from different companies, as well as between different lots of the same company. It is clear then that the results presented earlier pertaining to the three companies studied represent a limited sample of their products an.d are not necessarily applicable as a quality indicator. The results of the patient reactions elicited by the mixtures as compared to those elicited by their components, although inconclusive, raise certain questions concerning the sensitivity of mixtures as a screening method. Some clinics use mixtures containing ten or twelve components in their initial skin testing regimens, followed by administration of the individual components contingent upon a positive patient response to the appropriate mixture. The magnitude of the dilution factor produced by such a mixture
55
quite possibly results in the oversight of skin reactivity in some allergic patients. Considering the current diversity in allergic skin testing throughout the United States and the world, a single method of skin testing which could be universally accepted and clinically employed would be most advantageous. As a step in that direction, based on the results of this study, the prick test administered at a concentration of 1: 20 w/v in aqueous solution is proposed. Hopefully, this study will lead to further clinical investigation into the methodology of allergic skin testing. 1 would like to express my gratitude to Raymond G. Slavin, M.D., Director of the Section of Allergy and Immunology, St. Louis University School of Medicine, for his
professional guidance and review of the manuscript and the enthusiastic assistance of his patients and staff. Many thanks to Greer Laboratories, Lenoir, N. C., Hollister-
Stier Laboratories, Spokane,Wash., and Center Laboratories, Port Washington, N. Y. for contributing the inhalant extract solutions, and Clement J. Sullivan, M.D.,
and David A. Specht, Ph.D., for their review of the manuscript. I also want to expressmy appreciationto Mrs. Joyce Edwards for her typing of the manuscript. REFERENCES
I. Sheldon,J. M., Lovell, R. G., and Mathews, K. P.: A manual of clinical allergy, ed. 2, Philadelphia, 1967. W. B. Saunders co., p. 56. 2. Indrajana, T., Spieksma,F. Th. M., and Voorhorst, R.: Comparative study of the intracutaneous,scratch and prick tests in allergy, Ann. Allergy 29~639, 1971. 3. Slavin, R. G.: Skin tests in the diagnosis of allergies of the immediate type, Med. Clin. North Am. 55~65, 1974. 4. Lewis, W. H., and Imber, W. E.: Allergy epidemiology in the St. Louis, Missouri area. II. Grasses, Ann. Allergy 35:42, 1975. 5. Arbesman, C. E., Loepf, Cl. F., and Miller, G. E.: Some antianaphylactic and antihistamine properties of N’pyridyl, N’bensyl, demethyethylenediamine monohydrochloride, J. ALLERGY 17:203, 1946.
6. Cook, T. J., Macqueen,D. M., Wittig, H. J., Thornby, J. I., Lantos, R. L., and Virtue, C. M.: Degreeand duration of skin test suppressionand side effects with antihistamine, J. ALLERGY51:71, 1973.
7. Patterson,R.: Allergic diseases:Diagnosis and management, Philadelphia, 1972, J. B. Lippincott Co., p. 77. 8. Nie, N., Hull, C. H., Jenkins. J., Steinbrenner,K., and Bent, D.: Statistical package for the social sciences, ed. 2. New York, 1975, McGraw-Hill Book Co. 9. Feinberg, S. M., and Steinberg, M. J.: Studies in pollen potency, J. ALLERGY 5: 19, 1933. 10. Baer, H., Godfrey, H., Maloney, C. J., Norman, P. S., and Lichtenstein, L. M.: The potency and antigen E content of commercially preparedragweedextracts, J. ALLERGY 45:347, 1970.