Imported fire ant (IFA) sensitivity: IFA-specific levels of serum IgG and IgG4

Imported fire ant (IFA) sensitivity: levels of serum IgG and lgG4 Margaret

A. Reed, MS, and Brian T. Butcher,

IFA-specific

PhD New Orleans, La.

ImportedJire ant (IFA) whole body extract (WBE) and venom (V)-specific IgG and IgG4 antibodies and specific IgE antibodies were evaluated in sera from 56 IFA-sensitive individuals (18 undergoing immunotherapy with IFAWBE and 38 individuals not being treated) and 44 nonatopic and atopic control subjects with no history of IFA allergy. Although there was no difSerence in the level of IFAWBE- or IFAV-specific IgG between treated and untreated patients, both groups had higher levels of IFAWBE- and IFAV-specific IgG (p =C0.05) than did control subjects. Patients receiving treatment tended to have higher levels of IFAWBE-speciJic IgG4 than did either untreated patients (0.05 < p CC0.10) or control subjects (p < 0.051. Levels of IFAV-specijic IgG4 were higher in treated patients than in control subjects (p < 0.05) but were not dtflerent between treated and untreated patients. Levels of IFAWBE- and IFAV-specific IgE antibodies did not differ between the two patient populations but were higher in both groups than in control subjects (p < 0.05). The ratio of IgG4 to IgE (G4lE) for both IFAWBE and IFAV was calculated for all groups. The ratios of IFAWBE- and IFAV-specijk G4lE were higher in treated patients and in control subjects as compared to nontreated patients (p < 0.05). IFAV-speciJic G4lE ratios were lower in treated patients than in control subjects, but IFAWBE-specijc G4 I E ratios were the same as ratios for control subjects. These jindings demonstrated that immunotherapy with IFAWBE tends to raise the level of IFAWBE-spec$c IgG4 and thus influence the ratio of IFAWBE G4lE. The observation that IFAV G4lE ratios were higher in treated patients as compared to untreated patients may reject an interaction between a modest, but nonsignificant, rise in IFAV IgG4 together with a modest, but nonsignificant, decrease in IFAVIgE. (JALLERGYCLINIMMUNOL 1990;85:737-42.)

Allergic reactions to IFA are a significant medical problem in the southeasternUnited States.Of the two speciesof’ IFA, Solenopsis invicta and S. richteri, the former is more common and is responsible for most allergic reactions. Up to 58% of the population in infested areas are stung yearly.’ As many as 16% experiencesometype of hypersensitivity reaction with approximately 1% requiring medical treatment.* Reactions range in severity from large local dermal lesions, which persist for several days, to severesystemic anaphylaxis, manifestedby hives, edema,shortnessof breath, fainting, and in somecases,death.3z4 For individuals who are diagnosed as IFA sensitive, immunotherapy is often initiated. Becauseof the difficulty in obtaining sufficient amounts of pure IFAV, IFAWBE is the only material currently available com-

Abbreviations used IFA: Imported fire ant WBE: Whole body extract Whole body extract IgG or IgG4 ratio WBE-G,G4: WBE-E: Whole body extract RAST percent binding Whole body extract IgG4 to IgE ratio WBE-G4/E: V: Venom V-G,G4: Venom IgG or IgG4 ratio V-E: VenomRASTpercentbinding V-G4/E: Venom IgG4 to IgE ratio KWANOVA: Kruskal-Wallis analysis of variance OD: Optical density PBS: Phosphate-buffered saline BSA: Bovine serum albumin RT: Room temperature (approximately

22” C)

From the Section of Clinical Immunology, Department of Medicine,

Tulane Univesity Medical Center, New Orleans, La. Supportedby National Institutes of Health Grant AI 21552. Received for publication April 21, 1989. Revised Sept. 5, 1989. Accepted for publication Sept. 11, 1989. Reprint requests: Margaret A. Reed, MS, New Orleans, LA 70112. l/1/18029

1700 Perdido St.,

mercially for desensitization. In studies of hypersensitivity

to other Hymenoptera,

venom was demon-

strated to contain the relevant allergens.5The venom was also superior to WBE for desensitization of allergic patienk6s ’ Levels of Hymenoptera-specificserum IgG and IgG4 antibodies were demonstratedto 737

738 Reed and Butcher

J. ALLERGY

TABLE I. Mean IFAWBE- and IFAV-specific Group

Treated Untreated Control subjects

IFAWBE-G

3.08 -+ 0.18 2.58 k 0.15 1.57 * 0.12

CLIN. IMMUNOL. APRIL 1990

IgG and lgG4 OD ratios + SEM IFAV-G

IFAWBE-G4

3.27 k 0.16 2.90 k 0.15 0.95 2 0.08

6.34 k 0.47 3.41 + 0.42 0.82 k 0.03

IFAV-G4

5.55 + 0.45 4.13 +- 0.42 0.51 2 0.02

n

18 38 44

rise over the progression of immunotherapy.s-10 The

IFA-specific IgG ELISA

increasein Hymenoptera-specificIgG and, in particular, IgG4 antibodies, has been proposed to be important in conferring protection against Hymenoptera anaphylaxis.“-I3 No studies of class-specific antibody responseto immunotherapy for IFA sensitivity have been reported. We have developed an ELISA to measure IFAWBE- and IFAV-specific IgG and IgG4 in sera from patients with lFA sensitivity, someof whom are undergoing immunotherapy.The results of thesestudies are reported here.

ELISA was performedaccordingto, and with the reagents supplied in, the ELISAmate kit manufacturedby Kirkegaard & Perry Laboratories, Gaithersburg, MD.,‘5 with modifications, as follows: For IFA-specific total IgG (all subclasses) antibodies, wells of Immulon II polystyrene microtiter plates (Dynatech Industries, Alexandria, Va.) were coatedwith 100p.1of a 50 kg/ml solution of either IFAWBE or IFAV in coating buffer (PBS, pH 8.6), and plates were incubatedovernight at 4” C. Coating antigen was discarded, wells were filled with 300 ~1 of PBS,pH 7.4, containing 1% BSA to block unbound sites, and plates were incubated for 1 hour at RT. Blocking solution was discarded,and 100 p.1of patient’s sera, diluted 1: 50 in PBS-BSA, was added to duplicate wells. A series of dilutions (1: 20, 1:50, 1:200, and 1: 500) of a pool of IFA RAST-positive (IFAWBE-E, 19.8; IFAV-E, 28.6) sera or a pool of IFA RAST-negative (IFAWBE-E, 0.8; IFAV-E, 0.9) sera were added to duplicate wells. Plates were incubated at RT for 1 hour. After removal of unreacted serum, plates were washedthree times with PBS containing 0.02% Tween 20. One hundred microliters of affinity-purified, peroxidaselabeledgoatantihumanIgG (1 Fg/ml) (Kirkegaard& Perry) in PBS-BSA was added to all wells, and plates were incubated 1 hour at RT. After removal of peroxidase conjugate, plates were washedthree times as above, but the last wash was left for 5 minutes. Color was developedby adding 100 p,l of 0.03% 2,2’-azinodi[3-ethylbenzthiazoline sulfonate] in citrate buffer containing 0.01% hydrogenperoxide. The reaction was stopped after 10 minutes by addition of 50 p.1of 1% of sodium dodecyl sulfate, and plates were read at 690 em. The resulting ODs were recorded, and the ratios of test OD to the IFA-negative semm pool OD (IFAWBE-G and IFAV-G) were calculated. These ratios were used for statistical analyses.

MATERIAL AND METHODS Study population Serawere obtained from 56 individuals with documented reactions to IFA sting attending allergy clinics in New Orleans and the surrounding area. Thirty-eight serawere from patients who had had reactions to IFA sting rangingin severity from large local to systemic allergic reactions, and, for various reasons,did not receive immunotherapy. Some, but not all, patients had been skin tested, and not all those patients skin testedhad a positive skin test. Detailed information of skin testsfor thesepatients is not available. Eighteen of these sera were from patients receiving immunotherapy for different lengths of time (mean,4.8 years, ranging from 0.5 to 14.0 years), with IFAWBE extracts from commercial suppliers. All 18 patients had had at least one documentedsystemicallergic reaction to IFA sting. All but one patient were skin test positive. (This patient was a minor and the parentsdeclined the skin test; RAST was performed for confirmation of sensitivity.) Allergic and demographic history were obtained by questionnaire. Control sera were from six individuals stung by IFA who denied reaction (five nonatopicandone atopic individual) andfrom 38 individuals residing in Ohio, where IFA are not indigenous, who had not been exposed to IFA (19 atopic and 19 nonatopic individuals) .

IFAWBE and IFAV reagents ReferenceIFAWBE was preparedby a methoddeveloped in our laboratory.14Dry weight and protein concentration were 35.0 mglrnl and 18 mg/ml, respectively. IFAWBE extracts were aliquoted and stored at -20” C until use. S. invictu IFAV, collected by an electroshockmethod, was a generous gift fromVespaLaboratories (SpringHills, Pa.). IFAV was supplied in sterile glass vials as a freeze-dried powder, with reported weight of 0.4 mg per vial.

IFA-specific lgG4 ELISA For IFA-specific IgG4 antibodies, Immulon II plateswere coated with IFAV or IFAWBE, incubated, and blocked as describedabove. Plates containing sera in duplicate of patients or control subjectswere incubated overnight at 4” C. Serawere removedand plates were washedthree times with PBS containing 0.02% Tween 20. One hundred microliters of mouse IgGl monoclonal antihuman IgG4 (1 Fgg/ml in PBS-BSA) was added, and plates were incubated at RT for 2 hours. The monoclonal antibody was discarded,andplates were washedthree times as describedabove. One hundred micmliters of peroxidase-conjugated,affinity-purified goat antimouseIgG (1 kg/ ml in PBS-BSA) was addedto wells

IFA IgG and lgG4 levels

VOLUME 85 NUMBER 4

TABLE II. Mean IFAWBE- and IFAV-specific ratios * SIEM Grwp

Treated Untreated Control subjects

IgE RAST percent IFAV-E

IFAWBE-04/E

IFAV-GQIE

n

7.41 ? 1.24 9.22 * 1.34 0.89 5 0.12

14.54 + 2.03 19.37 r 1.72

1.41 -c 0.24 0.89 rt 0.20

0.59 t 0.11 0.39 k 0.11

18 38

0.70 2 0.08

1.27 ” 0.08

0.89 ? 0.06

44

IgE RAST

Sera were tested for IFAWBE and IFAV-specific IgE by RAST (Kallestad Laboratories, Austin, Texas).16Briefly, IFAWBE (10 mg/ml of protein) or IFAV (0.1 mg/ml) was coupled to cyanogenbromide-activatedpaper disks. Disks were incubated overnight with 50 p.1of sera plus 50 t.~lof PBS, in dul)licate, at RT, and then washedthree times with 2.5 ml of physiologic saline for 10 minutes per wash to remove unreacted serum. One hundred microliters of “‘Ilabeled anti-IgE diluted 1: 1 with physiologic saline was added, and disks were incubated overnight at RT. Disks were washedas aboveto removeunreactedisotopeand then countedon a gammacounter to determineamount of bound radioactivity. Results were expressedas percent binding of total activity added. Statistical

analyses

Summary descriptive statistics and analysis of the data were performed with the Statistix program for the IBM.” Since most of the data were not normally distributed, nonparametricstatistics, which are applicable to both normally and nonnormally distributed sample populations, were done. Analyses included the rank-sum test and one-way KWANOVA.I8 Multiple comparison testsI were performed on the ANOVA of the ranksof the three groupsto determine which one of the three groups was different. RESULTS Study population Patient and control populations were divided into three groups: untreated patients (not currently receiving immunotherapy, n = 38), treated patients (receiving immunotherapy, n = 18), and all control subjects (n == 44). There were no statistical differences between nonexposed and exposed control subjects in any parameters analyzed. IFA-specific

-c SEM and G4/E

IFAWBE-E

and incubatedfor 1 hour at RT. The remaining washing and color-development steps were performed as above. ODs were recorded, and ratios of test OD to IFA-negative serum pool (IFAWBE-G4 and IFAV-G4) were calculated. These ratios were used for statistical analyses. IFA-specific

binding

739

IgG and lgG4

As an ndication of IgG and IgG4 levels relative to negative control pool, mean values for IFAWBE- and IFAV-specific IgG and IgG4 ratios were determined

(Table I). KWANOVA and multiple comparison of the ranks of the values indicated the following: levels of IFAWBE-G and IFAV-G did not differ between treated and untreated patients (p > 0.05), but both groups of patients had higher levels of both as compared to control subjects (p < 0.05). Treated patients tended to have higher levels of IFAWBE-G4 than did untreated patients (0.05 < p < O.lO), but both groups of patients had significantly higher levels than control subjects (p < 0.05). Levels of IFAV-G4 were not different between treated and untreated patients (p > 0.60) but were higher in untreated patients and treated patients than in control subjects (p < 0.05). Ranksum test demonstrated that in both the treated and untreated patients, IFAV-G was not different from IFAWBE-G (p > 0.05); however, IFAWBE-G was higher than IFAV-G in control subjects (p < 0.001). Treated and untreated patients’ levels of IFAV-G4 did not differ from levels of IFAWBE-G4 (p > 0.10); whereas, in control subjects, IFAWBE-G4 antibody values were higher than IFAV-G4 (p -=c0.001). Summaries of the statistical analyses are presented (see Tables III and IV). IFA-specific

IgE (RAST) and G4/E ratios

Mean values of IFAWBE- and IFAV-specific IgE (IFAWBE-E and IFAV-E) serum levels are presented in Table II). Untreated patients were not statistically different from treated patients for either IFAWBE-E or IFAV-E (p > 0.05), but both groups had higher levels than control subjects for both (p < 0.05). Levels of IFAV-E were higher than IFAWBE-E in both untreated and treated patients (p < O.Ol), but IFAWBE-E was higher than IFAV-E in control subjects (p < 0.05). A ratio of the IFA-specific IgG4 to IFAspecific IgE RAST percent binding (G4/E) was determined to establish a relationship between IgG4 and IgE antibody levels as a possible measure of efficacy of immunotherapy. Mean values of these ratios are also presented in Table II. Nonparameteric statistical analyses performed on these ratios indicated that treated patients had a significantly higher IFAWBEG4/ E and IFAV-G4/E than did untreated patients (p < 0.05). Control values of IFAV-G4/E were statistically higher than either treated or untreated pa-

740 Reed and Butcher

TABLE III. Summary

J. ALLERGY

of KWANOVA

and multiple

comparison

Treated vs untreated

Parameter

IFAWBE-G IFAV-G IFAWBE-G4 IFAV-G4 IFAWBE-E IFAV-E IFAWBE-G4/E IFAV-G4/E

tests*

Treated vs control subjects

No difference No difference No difference? No difference No difference No difference Treated higher Treatedhigher

CLIN. IMMUNOL. APRIL 1990

Untreated vs control subjects

Treatedhigher Treatedhigher Treatedhigher Treatedhigher Treated higher Treatedhigher No difference Control subjectshigher

Untreated higher Untreated higher Untreatedhigher Untreated higher Untreated higher Untreated higher Control subjectshigher Control subjectshigher

*All significant p values, CO.05. to.05 < p < 0.10.

TABLE IV. Summary levels among

of rank-sum

test of IFAWBE-specific

as compared

to IFAV-specific

antibody

groups*

Antibody

IgG IgG4 W IgG4/E

Untreated

Treated

No difference No difference IFAV higher IFAWBE higher

No difference No difference IFAV higher IFAWBE higher

Control subjects

IFAWBE higher IFAWBE higher IFAWBE higher IFAWBE higher

*See text p values.

tients (p < 0.05). Levels in control

subjects of

IFAWBE-G4/ E were higher than in untreated patients (p < 0.05)

but were not different

from levels in

treated patients (p > 0.05). Levels of IFAWBEG4/E was higher than IFAV-G4/E in all three groups (p < 0.05). Summariesof the statistical analysesof these parametersare found in Tables III and IV.

DISCUSSION These studies demonstrate that patients being treated for IFA sensitivity did not have significantly higher levels of IFAWBE- or IFAV-specific IgG than patients not treated. However, both groups of patients had higher levels of these antibodies than did control subjects.Patientsreceiving immunotherapy tendedto have higher levels of IFAWBE-specificIgG4 antibody levels than did patients not receiving immunotherapy, and IFAV-G4 levels were the same for both groups. IFAWBE- or IFAV-specific IgE levels did not differ between treated and untreated patients, but both groups had higher levels than did control subjects. In individuals receiving immunotherapy, IEAWBE IgG4/IgE ratios did not differ significantly from control subjectsbut were significantly higher than those of IFA-allergic patients not receiving immunotherapy. Treatedpatients had higher ratios of IFAV-G4/E than

untreated patients, but control subjects had higher levels of IFAV-G4/E than either treated or untreated patients. Our findings that IFAWBE-G levels did not differ between treated and untreated patients and that sera from control subjects also had significant IFAWBEspecific IgG antibodies may reflect cross-reactivity between IFAWBE components and other commonly

encountered antigens. IFAWBE have been demonstrated to contain as many as 29 antigenic components,14many of which may shareepitopeswith other antigens. Our inability

to demonstrate a significant

increase in IFAV-specific IgG antibodies of treated patients over patients who were not treated may be related to the IFAV content of IFAWBE. In other studies we have demonstrated that some commercial IFAWBEs have significant deficiencies in IFAV content and therefore may lack potency.19 The finding that IFAWBE-G4 levels tended to increasein patients undergoing immunotherapy is similar to results observed with other allergens, that is, housedust mite,” ryegrasspollen,” and beevenom.‘* These studies indicated that high levels of allergenspecific IgG4 have beendemonstratedto be associated with immunotherapy and may be important in conferring protection. It is interesting to note that levels

VOLUME NUMBER

IFA IgG and

85 4

of IFAV-G4.did not differ betweenthe treated and the untreated patients. This, too, may be due to the low or absent levels of IFAV in commercially available IFAWBE used for treatment. We have not yet evaluated the time course of the IgG4 response to immunotherapy nor of other IgG antibody subclasses.In studiesof patientsundergoing immunotherapyfor other Hymenoptera, such as bees, wasps, and hornets, a rise in venom-specific IgG1 occurred early in the course of immunotherapy, reacheda plateau, and then decreased,whereasIgG4 rose more slowly but stayedelevatedwith time, even after the maintenance dose was reached. Levels of venom-specific IgG2 and IgG3 were virtually undetectable.*‘.24It has been proposedthat this initial rise in IgGl , b’eforethe developmentof IgG4 antibodies, is important in conferring protection and that this early IgGl response may modulate venom-specific IgE production.25 Subsequently, only IgG4, which acts as blocking antibody, is necessaryto maintain protection. In previous studies we have indicated that level of IFAV-specific IgE is a better indicator of sensitivity than IFAWBE-specificIgE.26.27Thus, it would follow that production of IgG blocking antibodies against IFAV would be more useful in conferring protection against IFA stings than IFAWBE IgG antibodies. This is in keepmg with findings concerning other Hymenoptera antibody measurementsin which it has been demonstratedthat venom-specific rather than whole body-specific antibody measurementsprovide a better indication of the protection conferred by immunotherapy.‘, ’ We calculated the ratio of IFA-specific IgG4 to specific IgE based on the premise that immunotherapy increases allergen-specific IgG4 over time while it concurrently inhibits allergen-specific IgE production.24.25The significant increasein IFAWBEandIFAV G4/ E ratios in treatedpatients over untreatedpatients could be attributed to immunotherapy, indicating that there is IFAV present in some commercial preparations. However, the fact that control subjectshad significantly higher IFAV G4/E ratios than treated patients, whereas IFAWBE G4/E levels did not differ betweenthesetwo groups, may again indicate that the amount of IFAV contained in commercially available extracts may not be sufficient to increase levels of IFAV IgG4 and/or inhibit IFAV IgE antibody production. Further studiesrelating antibody subclassresponses to immunotherapy and to outcome of controlled IFA sting challenge or field sting will be necessaryto determine whether increasedlevels of IFAWBE-specific IgG or IgG4 antibodies are sufficient for protection

lgG4

levels

741

or whether it is necessary to insure that levels of IFAV-specific IgG and/or IgG4 antibodies are also increasedfor patients to be successfully immunized. It might also be useful to evaluatethe useof IFAWBEand/or IFAV-specific G4/E ratios as indicators of efficacy of immunotherapy. We thank Mr. M. Guralnick, Vespa Laboratories, Spring Mills, Pa., for the generous donation of IFAV; Mr. S. Sackett, New Orleans Mosquito Control, New Orleans, La., for identification of IFA species; and Ms. Dane11Watkins for excellent technical assistance. REFERENCES

1. deShazo RD, Ciriffing C, Kwan TH, Banks WA, Dvorak HF.

2. 3 4

5.

6.

7.

Dermal hypersensitivity reactions to imported fire ants. J ALLERGYCLIN IMMUNOL 19843743841-7. Triplett RF. The imported fire ant: health hazard or nuisance? South Med I 1976;69:258-9. Stablein JJ, Lackey RF. Adverse reactions to ant stings. Clin Res Allergy 1987;5:161-75. Rhoades RB, Stafford CT, James FK Jr, Bunker-Soler A, Impson LK. Survey of fatal anaphylactic reactions to imported fire ant stings. J ALLERGYCLIN IMMUNOL 1988;81:202. Hoffman DR. Allergens in Hymenoptera venom. XIV. IgE binding activities of venom proteins from three species of vespids. J ALLERGYCLANIMMUNOL 1985;75:64&10. Golden DBK, Langlois J, Valentine MD, Kagey-Sobotka A, Lichtenstein LM. Treatment failures with whole body extract therapy of insect allergy. JAMA 1981;246:2460-3. Hunt KJ, Valentine MD, Sobotka AK, Benton AW, Amodio FJ, Lichtenstein LM. A controlled trial of immunotherapy in insect hypersensitivity. N Engl J Med 1978;299: 157-61.

8. Grant JA, Goldblum RM, Rabr R, Thueson DO, Famam J, Gillaspy J. Enzyme-linked immunosorbcnt assay (ELISA) for immunoglobulin G antibodies against insect venoms. J ALLERGYCLIN IMMUNOL 1981;68:112-8. 9. Cheung NK, Blessing-Moore J, Reid MJ, Yang G. Honeybee venom-specific immunoglobulin G4 in honeybee sting-allergic patients and bee keepers. Ann Allergy 1983;50: 155-60. 10. Halpem GM. Recent applications of IgG4 in diagnosis and management of allergic diseae. Immunol Allergy Pratt 1986;8:386-96. 11. Golden DBK, Meyers DA, Kagey-Sobotka A, Valentine MD. Lichtenstein LM. Clinical relevance of the venom-specific immunoglobulin G antibody level during immunotherapy. J ALLERGYCLIN IMMUNOL 1982;69:489-93. 12. Nakagawa T, Miyamoto T. The role of IgG4 as blocking antibodies in asthmatics and beekeepers. Int Arch Allergy Appl Immunol 1985;77:204-5. 13. Urbanek R, Dold S. The key role of IgG4 subclass antibodies in the development of protection against allergic reactions to insect stings [English abstract]. Monatsschr Kinderheilkd 1986;134:536-40. 14. Butcher BT, Reed MA. Crossed immunoelectrophoretic studies of whole body extracts and venom from the imported fire ant Solenopsis invicru. J ALLERGYCLIN IMMUNOL 1988;81:33-40. 15. ELISAmate enzyme immunoassay for detection of antigen or antibody in microplates (package insert). Gathersburg, Md.: Kirkegaard & Perry Laboratories, Inc., 1985. 16. Allercoat RAST (package insert). Austin, Texas: Kallestad Laboratories, Inc., 1985. 17. Statistix: an interactive statistical analysis program for micro-

742

Reed and Butcher

computers(IBM version). St. Paul Minn.: NH analytical software, 1985. 18. Hollander M, Wolfe DA. Nonparametric statistical methods. New York: John Wiley & Sons, 1973. 19. Butcher BT, Reed MA. Evaluation of commercial imported fue ant extracts by crossed immunoelectrophoresisand radioallergosorbent test. J ALLERGY CLIN IMMUNOL 1988;82: 770-7. 20. Ohashi Y, Nakai Y, Kihara S, Nakagawa T, Miyamoto T. House dust mite-specific IgE, IgGl, and IgG4 antibodies in patients with perennial rhinitis. Ann Otol Rhino1 Laryngol 1987;96:434-7. 21. Moss RB, Hsu Y-P, Kwasnicki JM, Sullivan MM, Reid MJ. Isotypic and antigenic restriction of the blocking antibody responseto ryegrass pollen: correlation of rye group 1 antigenspecificIgG 1 with clinical response.J ALLERGYCLM IMMUNOL 1987;79:387-98. 22. Urhanek R, Kemeny DM, Richards D. Subclass of IgG antibee venom antibody produced during bee venom immunotherapy and its relationship to long-term protection from bee stings and following termination of venom immunotherapy. Clin Allergy 1986;16:317-22.

J. ALLERGY

CLIN. IMMUNOL. APRIL 1990

23. Malling HJ, Djurup R, SondergaardI, Weeke B. Clustered immunotherapy with yellow jacket venom: evaluation of the influence of time interval on in vivo and in vitro parameters. Allergy 1985;40;373-83. 24. Djurup R, Malling HJ, SondergaardI, WeekeB. The IgE and IgG subclassantibody responsein patients allergic to yellow jacket venom undergoing different regimens of venom immunotherapy.J ALLERGYCLIN IMMUNOL1985;76:46-55. 25. Djurup R. The subclassnature and clinical significanceof the IgG antibody responsein patientsundergoingallergen-specific immunotherapy.Allergy 1985;40:469-86. 26. Butcher BT, deShazoRD, Grtiz A, Reed MA. Superiority of Solenopsisinvictu venom to whole body extract in RAST for diagnosis of imported fire ant allergy. Int Arch Allergy Appl Immunol 1988;85:458-61. 27. Bahna S, Strimas J, Reed MA, Butcher BT. Imported fire ant allergy in young children: skin reactivity and serum IgE antibodies to venom and whole body extracts. J ALLERGYCLIN IMMUNOL1988:82:419-24.