Evaluation of the multi-test device for immediate hypersensitivity skin testing

Evahmtion of the Multi-Test device for e hypersensitivtiy skin testbg Robert B. Berkowitz, MD, David G. Tinkelman, MD, Cheryl Lutz, MT, Angela Crummie, RN, and Karen Smith, RN Atlanta, CU. Test-to-test reproducibility, user variability, and the effect of positive reactions on adjacent negative sites were evaluated for the Multi-Test skin testing device. Twenty-Jive subjects had skin tests with 24 histamine (1 mglml), four glycerosaline controls, and four “d@’ controls on two testings 7 days apart. To assess reproducibility from large and smaller histamine reactions andlor their effect on adjacent negative controls, 10 of the 25 subjects were retested once with the same testing format, but histamine at 10 mglml was substituted. To determine whether IUIX( allergen reactions affect adjacent negative controls differently than histamine reactions, 24 additional patients were retested on arms and back with negative controls adjacent to allergens to which they had prior 3 + to 4 + skin test reactions. Conclusions from 2688 skin tests on 49 patients are as follows: large (mean >I0 mm) histamine reactions reproduced better than smaller (mean <7 mm) responses-coeficients of variation were 12.3 and 21.4 respectively. A 14% user variability occurred when comparing mean wheal sizes from histamine produced b! each nurse and 1.2% when comparing their coefJicients of variation. Neither small histamine reactions nor large reactions from histamine and allergens affected adjucent negative controi\ We conclude that Multi-Test is a highly reliable skin testing technique that provides good reproducibility of results and low user variability. (.I ALLERGY CLIN IMMLJNOL 1992;90:9794?. ! Key words: Multi-Test, ullergens, site spacing

skin testing, variabiliry.

Immediate hypersensitivity skin testing is the standard clinical procedure for identifying allergen-specific IgE in allergic patients. Results from skin testing, a relatively specific and sensitive bioassay, are influenced by the quality of extracts used and the precision of the testing method used. General agreement exists that standardization of allergenic extracts is an essential goal COimprove diagnosis and treatment in allergic patients. Equally important for optimal diagnostic value of the skin test is a method of administering extracts that can be used with a high degree of accuracy. Relatively few studies of skin testing devices have been conducted to evaluate their test-to-test reproducibility and degree of user variability. The data that are available have originated predominantly from academic institutional settings where individually ad-

From the Atlanta Allergy and Immunology Research Foundation and the Atlanta Allergy Clinic, Atlanta. Supported by Lincoln Diagnostics and Center Laboratories. Received for publication Oct. 17, 1991. Revised April 21, 1992. Accepted for publication May 6, 1992. Reprint requests: R. B. Berkowitz, MD, Atlanta Allergy Clinic, 6667 Vernon Woods Dr.. N.E., Suite A12, Atlanta, GA, 30328.

l/l/42085

reproducibiliq.

standardization,

histamine

Abbreviation used CV: Coefficient of variation

ministered single prick or single puncture type instruments have been evaluated. Multi-Test (Lincoln Diagnostics, Decatur, ill.), a plastic device consisting of eight test heads attached to a rigid handle, administers eight skin tests simuftaneously via multiple puncture. The purpose of this study was to evaluate its precision and diagnostic accuracy under conditions of office practice. MATERIAL AND METHODS Subjects and skin testing In the first part of this study 2.5 healthy subjects. tive males and 20 females, ages 14 to 49 years (mean, 33 years), had skin tests on two occasions I week apart performed by two registered nurses (A.C. and K.S.) who were t.rained to use Multi-Test according to the manufacturer’s recommendations. Each subject received 24 glycerinstted hista@ne ( I mg / ml), four glycerosaline controls, and four “dry” controls during each skin testing session. Test sites wet divided between arms and back, and each glycerosaline control was adjacent to three histamine sites. All skin tests were per979

980

Berkowitz

J ALLERGY CLIN IMMUNOL DECEMBER 1992

et al.

FIG.

1. Multi-Test

formed between 8:00 AM and noon. Dry controls, points devoid of test or control substance, were used to detect dermatographism. During the first skin testing session one nurse administered the tests and the other read and recorded results. Roles were reversed on the second testing day. Histamine reactions were read at 10 minutes in keeping with the histamine manufacturer’s recommended reading time of 8 to 10 minutes after administration. Negative and dry controls were read at 20 minutes, which is consistent with the extract manufacturer’s recommended reading time of 15 to 20 minutes for negative controls and extracts. Wheal sizes were calculated as the average of the longest diameter and its midpoint orthogonal diameter, traced with a fine-tipped felt pen and transfered by tape to permanent records. After an interval of more than 7 days from the time the first and second skin testing sessions were completed in which histamine at 1 mgiml was used, 10 of the same 25 subjects, two males and eight females ages 14 to 49 years (mean, 36 years), were retested once by nurse A.C. using the same testing format, but with a lo-fold increase in histamine dose ( 10 mg / ml). Results from high-dose histamine and controls were read and recorded as described. The highdose histamine was used to determine whether larger reactions from it affected negative control sites differently than reactions from low-dose histamine, and to compare test-to-test reproducibility from high- and low-dose histamine. The 10 mg/ml histamine dose was not used in assessing user variability. In the second phase of this study 24 additional patients with 3 + and 4 + skin test history, 12 males and 12 females, ranging in age from 17 to 50 (mean, 31 years), were tested once on the arms and back by nurse A.C. using those allergens to which each patient had established prior skin test

skin

testing

device.

positivity. The allergens included mite and pollens. Each patient was tested with four Multi-Test devices each containing two allergens, in matched pa&, that were used on test heads 1 and 8 and 3 and 6 so that each of four negative control sites per device was either adjacent to an allergen site or between two allergen sites. Each patient received a total of 16 allergen tests and 16 tests with negative control solution. Results from allergens and negative controls were read by nurse A.C. at 20 minutes and again traced and transferred by tape to permanent records. Wheal sizes were calculated from the average of the longest diameter and its midpoint orthogonal diameter by use of the central portion of the wheal, but pseudopodia, when present, was traced and recorded. All testing was completed between 8:00 AM and noon. This portion of the study was designed primarily to determine whether large reactions from allergens affected adjacent negative control sites differently than reactions from histamine at two concentrations, 1 mg/ml and 10 mg / ml. Since each patient had established 3 + and 4 + skin test history to each of the two allergens used per patient, the frequency at which Multi-Test produced 3 + and 4 + reactions was recorded. In the second testing phase, no attempt was made to assess reproducibility of 3 + and 4 + reactions from matched pairs of allergens at adjacent sites because in some cases a difference of only 1 mm in diameter would have caused a reaction to vary by one class. Test sites for histamine, allergens, and controls were divided on the patient’s right and left sides, both for the volar surface of the forearms and for the upper and lower back. Because Multi-Test has eight fixed heads attached to a rigid handle (Fig. l), test sites were selected that provided adequate subcutaneous tissue and relatively flat surfaces to ensure uniform penetration of points on each test head. Each

VOLUME 90 NUMBER 6, PART 1

FM;. 2. Test site locations trol, and dry control.

for

histamine,

glycerosaline

con-

test site for Multi-Test requires an area 6.4 cm by 3.5 cm. Test head locations for histamine phosphate, allergen 1 and allergen 2, negative controls, and dry controls are shown in Figs. 2 and 3. The points on each test head were loaded uniformly by capillary action, with a Dip ‘N Touch liquid dispenseras directed by the manufacturer’s package insert. None of the 49 subjectswas receiving medication known to suppressimmediate hypersensitivity skin test reactions, The study protocols were approved by an institutional review board, and eachsubject read and signed an informed consent form before being tested. Material The Multi-Test device is injection-molded from acrylic. It has nine fine points on each test head that are arranged in an area 2 mm by 2 mm and that are 1.9 mm long. The longitudinal spacing between test sites is 20 mm, and the lateral spacing is 30 mm. The Dip ‘N Touch liquid dispenser is also injection-molded from acrylic and consists of an elongated solid shaft with an open cylinder at the end. The Dip ‘N Touch unit fits into a standard 5 ml vial as a replacement for the dropper. Test substancesfill the Dip ‘N Touch cylinders and are then delivered to the Multi-Test points. Histamine phosphate in 50% glycerosaline was used at concentrations of 1 mgiml and IO mg/ml histamine base. Diagnostic allergenic extracts, pollens (grasses,trees, and weeds) and mite (Dermatophagoides farinae), in 50% glycerosalinewere used at concentrationsof 1: 20 (wt / vol) and 10,000 AU/ml, respectively. The negative control solution was 50% glycerosaline. Statistical analysis The statisticalanalysiswas performed with useof a Systat 3.0 program on a Northgate 386 (Northgate Computer Systems, Inc.. Eden Prairie, Minn.) computer. Paired-differ-

FIG. 3. Test glycerosaline

site locations controls.

for

allergen

1, allergen

2, and

ence t tests were used to determine the significance of differences between wheal sizeson the right and left sides ot the body, on the arms and back, and for the twu nurses. An independent samples f test was used to compare the mean wheal sizesfor the upper and lower back. A one,-wayanalysis of variance (ANOVA) test was used to compare the mean wheal sizesat the negative control sues for the three evaluations. A 0.05 level of significance was used for all tests and 95% confidence for all interval estimates. A coefficient of variation (CV) was calculated for each patient and for each nurse. CV, a measure of how much reaction sizes deviate from the mean, is a valuable determinant of precision in any assay procedure including skin testing and is expressedas a percentage of the arithmetic mean. The level of precision is inversely proportional to the CV value: that is. low CV-high precision. hrgh C’V-loa precision. RESULTS Histamine at a concentration of 1 mg/ml produced wheal reactions with a mean diameter of 6.7 mm. The mean CV for these reactions was 21.4 (range, 6 to 90). Ninety-six percent of the CVs were between 6 and 39, whereas 4% were between 44 and 90 Histamine at a concentration of 10 mg/ml elicited wheal responses with a mean diameter of 10.1 mm and a mean CV of 12.3 (range, 7 to 17) (Tables 1 and Hf. When calculating intrasubject test-to-test reproducibility from skin testing with histamine, reaction sizes on each subject were compared with means for each patient. When evaluating user variability with histamine, mean reaction sizes produced by each nurse were compared, as were their CVs. The mean reaction size produced from all histamine ( 1 mg/ml) sites on

982

J ALLERGY CLIN IMMUNOL DECEMBER 1992

Berkowitz et al.

III. Distribution of wheal sizes compared with mean for nurses A.C. and KS. (testing with histamine 1 mg/ml)

I. Distribution of wheal sizes compared with mean from 25 subjects tested twice with histamine 1 mg/ml

TABLE

TABLE

NO.

Deviation mean

from

?5% r 10% +- 15% -c 20% ?25% ? 30% Outside 4 30% CV, 21.4. Mean wheal

size,

test sites

306 275 212 124 103 58 122

No.

Percent test sites

Deviation from mean

Nurse A.C. ?5% ? 10% k 15% k 20% k 25% It 30% Outside 4 30% Nurse K.S. +-5% k 10% 2 15% + 20% k 25% ‘- 30% Outside k 30%

25.5% 23.0% 17.7% 10.3% 8.6% 4.9% 10.0%

6.7 mm.

II. Distribution of wheal sizes compared with mean from 10 of 25 Table I subjects retested once with histamine 10 mg/ml

TABLE

No. Deviation from mean

t5% + 10% + 15% * 20% 5 25% k 30% CV, 12.3. Mean wheal

size,

test sites

101 65 38 17 13 6

Percent test sites

42.1% 27.1% 15.8% 7.1% 5.4% 2.5%

10.1 mm.

25 patients was 7.2 mm. by A.C. and 6.2 mm by K.S. A difference of 1 mm is statistically significant (p value 0.0005) but constitutes 86% agreement. The CVs from test results produced by A.C. and K.S. were 21.9 and 20.7, respectively, showing user variability of 1.2%. A high degree of agreement between users is shown in Table III where distribution of wheal sizes is compared with the mean for each nurse. When evaluating skin test reactions elicited from histamine at concentrations of 1 mg/ ml and 10 mg/ ml, no statistically significant differences occurred from either concentration when comparing mean reaction sizes from each on both arms, both sides of the back, upper and lower back, or when comparing all sites from both arms to all back sites (p values ranged from 0.463 to 0.966). When similar comparisons of testing sites were made for allergens, only those reactions from the arms (mean, 13 mm)

Nurse Nurse

A.C., KS.,

test sites

Percent test sites

154 128 108 64 61 26 59

25.7% 21.3% 18.0% 10.7% 10.2% 4.3% 9.8%

152 147 103 61 47 27 63

25.3% 24.5% 17.2% 10.2% 7.8% 4.5% 10.5%

CV, 21.9; mean wheal size, 7.2 mm. CV, 20.7; mean wheal size, 6.2 mm.

when compared with those from the back (mean, 14.6 mm) revealed a statistically significant size difference (p value 0.002). It is significant that neither small nor large positive wheal reactions affected adjacent negative controls. As mean wheal reaction sizes for low-dose histamine, high-dose histamine, and allergens increased, the mean wheal reaction sizes for adjacent negative controls remained essentially constant (Table IV). A reaction of less than a 3 mm wheal at a histamine or allergen site was considered false negative, and a wheal reaction equal to or greater than 5 mm at a negative control was considered false positive. When applying these standards, consistent with Multi-Test package insert instructions, to the subjects tested with histamine the following results were recorded: from histamine 1 mg/ml, 2% false negative and 7% false positive; from histamine 10 mg/ml, no false negatives and 3.8% false positive. When grading reactions from the subjects tested with allergen with established 3 + and 4 + skin test history according to Multi-Test package insert instructions, these results were observed from Multi-Test: all reactions on the arms were 3 + or 4 + , whereas 96% of the back reactions were 3 + or 4 + ; 2.8% of the reactions on the back were 2 -t , 1% were 1 + , and one response (0.3%) were negative. A whea17 mm to 10 mm in diameter was graded 3 + ,

Evaluation

VOLUME 90 NUMBER 6. PART 1

TABLE

IV.

Results

from

negative

controls

adjacent

Negative mean (mm)

Low-dose histamine High-dose histamine Allergens

0.564 0.544 0.526

and a wheal larger than 10 mm or pronounced pseudopodia was graded a 4 + . The incidence of falsepositive reactions at negative control sites was 3.9%. If the level of positivity is lowered from a wheal 5 mm in diameter to a wheal 3 mm in diameter, the false-positive calculations for subjects receiving lowdose and high-dose histamine and patients tested with allergen become 10.8%, 8.8%, and 10.4%, respectively. Eighty-eight percent of all negative responses showed no whealing. Wheals 2 to 4 mm in diameter occurred at 6.6% of the negative control sites. Whealing reactions of 5 mm, 6 mm, and 7 mm occurred at 2.9%, 1.7%, and 0.8%, respectively, of the negative control sites.

DISCUSSION Published findings on test-to-test reproducibility and user variability are limited, and there are no generally established standards for grading the precision of skin testing procedures. Aas’ reported that some individuals can regularly produce skin test responses within ? 10% of the mean for most reactions, but not for all. Aas further stated that results from other users consistently vary from the mean more than + 30%. Aas et al.’ suggest that for optimal results, reactions should be reproduced within +20% of the mean. An important factor affecting reproducibility is wheal size. Taudorf et a1.3 found the CV between 30 to 60 for wheals smaller than 4.4 mm in diameter, and a CV between 20 to 30 for wheals with diameters larger than 4.4 mm. Our results from histamine responses with a mean wheal size less than 7 mm showed a CV of 21.4, and results showed a CV of 12.3 from histamine reactions with a mean wheal size greater than 10 mm. For the smaller histamine reactions the CV of 21.4 was 1.4% higher than the CV of 20 recommended by the Nordic Society of Allergology,’ whereas the CV of 12.3 from the larger histamine reactions was 7.7% lower. When using half-log dilutions of histamine in concentrations up to 5 mg/ml, Engler et al.,4 in a preliminary report, showed a CV of 24.4 for Multi-Test

to positive Standard deviation

1.628 1.256 1.581

of the Multi-rest

devi~:e

983

reactions Positive mean (mm)

6.7 10.1 13.8

Standard deviation

: .“‘%I .i..3? 3; :‘i . .. . -. .----

wheals, which is comparable to our findings from histamine reactions with a mean wheal size less than 7 mm. Mahan et al.’ in a preliminary report comparing Multi-Test and the Morrow Brown needle In patients with known skin test sensitivity to specific allergens, showed a 0.91 coefficient of correlation from MultiTest wheal reactions. Although expressed as coefficient of correlation, the reproducibility of Multi-Test results from the study by Mahan et al.’ wds comparable in diagnostic precision to that gained in our study from histamine reactions with a mean wheal size greater than 10 mm. Adinoff et a1.6 in a study with histamine at 10 mg/ml, found a CV of 39.5 for Multi-Test wheals compared with our CV of 12.3 from the same histamine concentration. This difference in findings between the two studies could be attributable to study design and to some degree the different methods used to calculate CV. We used wheal diameters, whereas Adinoff et al.h measured wheal area. Adinoff et ai.” rotated Multi-Test sites over the entire surface of the back in a manner that could have resulted in MultiTest being used over an area on the back that is not optimal for uniform puncture of points from all test heads. Nevertheless, we note that 11 of the f 5 subjects in the study by Adinoff et a1.6 showed reproducibility of results from Multi-Test wheals that was as good as, or better than, the reproducibility ;)f results (CV 22 to 28) from other techniques. Murphree and Kniker’ first characterized the sensitivity and specificity from Multi-Test in 1979. Since then other investigators have also reported reliable levels of diagnostic accuracy from the Multi-Test technique .5. ‘-I1 When Menardo et a1.8 compared skin test results from Multi-Test with those gained fi+om a modified prick test, the Morrow Brown needle. and the intradermal technique in diagnosis of mite allergy, specific IgE antibody levels correlated Mter with Multi-Test results than with results from ;iny of the other three skin testing procedures. Another area of concern with skin testing is the level of false-positive and false-negative results. Adinoff et al.h when using the Multi-Test with histamine

984

Berkowitz

J ALLERGY CLIN IMMUNOL DECEMBER 1992

et al.

at 10 mg/ ml, reported 1% false-negative and 49% false-positive results, with a wheal size of 3 mm or greater at a negative control site as the definition of a false-positive test. When we used the same concentration of histamine, our results from Multi-Test yielded no false-negative results and 3.8% false-positive results with a wheal 5 mm or greater in size to define a false-positive test. When we applied the 3 mm criterion to our results, the false-positive rate was 8.8%, with wheal sizes ranging from 3 mm to 5 mm. It should be noted that the Multi-Test package insert states that wheals up to 5 mm in diameter can occur at negative control sites in some patients, and it stresses reading all results against the negative control. The large difference in false-positive findings from our study and those from the work of Adinoff et a1.6 does not appear attributable to the different levels used to define false-positive reactions. A possible explanation for the difference in false-positive rates is that Adinoff et a1.6 used older Multi-Test units on which the points were approximately 0.5 mm longer than those on the current devices used in our study. The longer points could have caused more pricking of the dermis and, consequently, a greater degree of irritant response. Adinoff et a1.6suggest that the spacing between test heads on Multi-Test may have contributed to the falsepositive rate they reported. We found no evidence that large wheal reactions from either histamine or allergens caused false-positive results at adjacent negative control sites. Garibaldi and Slavin,‘* when evaluating negative controls adjacent to large allergen reactions and histamine responses from Multi-Test, found no evidence that test head spacing influenced results at negative controls. The findings of Garibaldi and Slavin are consistent with those reported by Mahan et al.’ A point of interest regarding false-positive results in our study is the fact that more false-positive reactions were recorded from the lower concentration of histamine than from the higher concentration. The reason was undetermined. An important observation from our study is that Multi-Test affords the user greater protection against blood-borne diseases than other skin testing techniques. Its construction and method of application are such that the user’s fingers and hands are not in close proximity to any bleeding that might occur on the patient’s skin. In addition, the design of Multi-Test and its method of administration greatly reduce the risk of inadvertent pricking of the user’s skin. Several investigators have used Multi-Test for re-

search purposes and have reported satisfactory precision of results from it, as well as other advantages.‘3-‘7 Our goal was to evaluate Multi-Test under conditions of office practice. We found the Multi-Test technique of skin testing to be rapid, convenient, and well accepted by patients. Reproducibility of test results from Multi-Test was high, and variability of results between users low. We concluded that the MultiTest technique of skin testing provided sufficient precision, sensitivity, and specificity to be classified as a reliable in-office diagnostic procedure. We Lincoln oratories

thank Randal Beck, PhD, Diagnostics for Multi-Test for

allergens

and histamine

for statistical analysis, devices, and Center Labphosphate.

REFERENCES 1. Aas K. Some variables in skin prick testing. Allergy 1980;35:250-2. 2. Aas K, Backman A, Weeke B, Belin L. Standardization of allergen extracts with appropriate methods: the combined use of skin prick testing and radioallergosorbent tests. Allergy 1978;33:130-7. 3. Taudorf E, Malling HJ, Lauresen LC, Lanner A, Weeke B. Reproducibility of histamine skin prick test. Allergy 1985;40:344-9. 4. Engler DB, DeJamatt AC, Sim TC, et al. Comparison of sensitivity and precision of four skin test devices [Abstract]. J ALLERGY CLIN IMMUNOL 1991;87(suppl):329. 5. Mahan CA, Spector SL, Siegel SC, Katz RM, Rachelefsky GS, Rohr AS. Validity and reproducibility of Multi-Test Skin Test Device (Poster). Presented at the Forty-seventh Annual Meeting of the American College of Allergy and Immunology, San Francisco, California, Nov. 10-14, 1990. 6. Adinoff AD, Rosloniec DM, McCall LL, Nelson HS. A comparison of six epicutaneous devices in the performance of immediate hypersensitivity skin testing. J ALLERGY CLIN IMMUNOL 1989;84: 168-74. I. Murphree JT, Kniker WT. Correlation of immediate skin test responses to antigens introduced by Multi-Test and intracutaneous routes. AM Allergey 1979;43:279-85. 8. Menardo JL, Bousquet J, Michel FB. Comparison of three prick test methods with the intradermal test and with the RAST in the diagnosis of mite allergy. Ann Allergy 1982;48:235-9. 9. Grater WC, Do&horn R, Boggs P, Don RL. Report of the American College of Allergists’ Committee on Standardization of Allergenic Extracts. Ann Allergy 1982;49:49-54. 10. Zeiger RS. Special considerations in the approach to asthma in infancy and early childhood. J Asthma 1983;20:341-59. 11. Basaran M, Hein EW. Evaluation of a Multi-Test device for allergy testing [Abstract]. J ALLERGY CLM IMMUNOL 1988;81:171. 12. Garibaldi E, Slavin RG. Positive Multi-Test reactions do not cause false positive reactions at adjacent sites. Ann Allergy 1990;65:481-4. 13. Zeiger RS, Heller S, Mellon MH, et al. Effect of combined maternal and infant food-allergen avoidance on development of atopy in early infancy: a randomized study. J ALLERGY CLIN IMMUNOL 1989;84:72-89. 14. Sullivan TJ. Pharmacologic modulation of the whealing re-

VOLUME 90 NUMBER 6. PART 1

sponse to histamine in human skin: identification of doxepin as a potent in vivo inhibitor. J ALLERGY CLIN IMMUNOL 1982;69:260-7. 15. Rao KS, Menon PK, Hilman BC, Sebastian CS, Baimsfather L. Duration of the suppressive effect of tricyclic antidepresSants on histamine-induced wheal-and-flare reactions in human skin. J ALLERGY CLIN IMMUNOL 1988;82:752-7.

Evaluation

of the

Multi-

i‘e~~ C!WKX

16. Lin RY, Erlich ER, Don PC. Skin prick test -c\ponse\ ttr codeine, histamine, and ragweed utilizing the Multitext del’ice Ann Allergy 1990;65:222-6. 17. Hamburger RN, Berger WE, Quiwa NB. ‘Temrzas \‘. Ca~lllab R, Miller SP. Skin testing compared with in vitro resting tor screening allergic patients. Ann Allergy I L)Y I :ir- i 3 ? -7

Comparison of the sensitivity four skin test devices

and precision

of

David B. Engler, MD, Alan C. DeJamatt, MD, Tommy C. Sim, MD, Jana L. Lee, RN, and J. Andrew Grant, MD Galveston and Houston, Te.ra.r Twenv volunteers were skin tested with seven concentrations of histamine phosphate and a glycerosaline control to determine the relative sensitivity and precision qf,four skin tear tieviczs Greer Pen (GP), Greer DermaPIK (DP), Center Multi-Test (MT), and Morrow Brown need/t8 (MB). The end points of the study were (1) wheal and flare response of’ each device, with a dose-response curve, (2) the time required to apply euch set of eight tests, and (3) the volunteers’ subjective assessment of each device. On a diRerent day. 10 oj. the volunteers tt’crttested to determine the precision of each device. Dose-response curves for hay-log dilurictns c?/ histamine phosphate were produced with a glycerosaline control. The DP and GP induced when1 and flare responses discernible from that of the giycerosaline control at a lower concentration of histamine phosphate than the MB and MT. The DP took a shorter time 60 uppi\ eight samples than any other device. The MB was preferred by the most vo1unteer.c. but an! device tested on the upper half of the back was usually prqferred over that tested on the /owe, half. When 5 mglml histamine phosphate was used, coej$cients of variation ,for euch dcI~ic.e demonstrated that for wheals the precision of the DP, GP, and MT was similar {mearr. L’i . I’;ic 23.190, and 24.590, respectively). The MB was larger (meun, 5Y.0LTc). Forjiares, the prec.ision of GP and DP was similar (mean, 22.090 and 23.57~~ respectivei!), with the MT and MB larger imean, 35.5% and 58.270, respectively). Repeating the analysis for the MB with 10 rn~> ml histamine phosphate, we found a mean coeflcient of variation of 16.5% for wheals anti lV.6% for,pares. We conclude that the DP is equal to the GP in precision and sensitivirv,for wlrtwl cm9 flare, but it has the advantage of allowing allergy skin testing to be performed more quickie. The MT has similar precision for wheals but has less precision for,pares. The MB de\-ice !,A more precise only with higher concentrations of histamine phosphate than those t.outirlel\, usgbri in the clinical practice of allergy. (J ALLERGY CLIN IMMKWX 1992;90:985-93.1 Key words: Skin testing, allergy testing, histamine

From the Division of Allergy and Immunology, Department of Internal Medicine, The University of Texas Medical Branch at Galveston. Supported by a grant from Greer Laboratories, Lenoir, N.C. Received for publication May 23, 1991. Revised Jan. 2. 1992. Accepted for publication Feb. 7, 1992. Reprint requests: David B. Engler, MD, The Allergy Clinic, 7707 Fannin, Suite 100, Houston. TX 77054-1918.

111142084

For more than a century, skin testing has been the standard procedure for diagnosing inhalant allergies. Most authors recommend use of positive and negative controls with the testing. ” ’ Recently several new devices have become available to perform immediate skin testing. They are claimed to be disposable, easy to use, comfortable, and more accurate. We examined the sensitivity, time necessary to apply material. volunteers’ preference of the devices, and the precision of four of these devices. 985