JOURNAL OF CLINICAL MICROBIOLOGY, Sept. 1992, p. 2269-22740095-1137/92/092269-06$02.00/0

Evaluation of an Immunoenzymatic Assay Detecting SpecificAnti-Toxocara Immunoglobulin E for Diagnosis andPosttreatment Follow-Up of Human Toxocariasis

J.-F. MAGNAVAL,1* R. FABRE,2 P. MAURIERES,l J.-P. CHARLET,3 AND B. DE LARRARD4Laboratoire de Parasitologie, Centre Hospitalier Universitaire Purpan, F-31059 Toulouse Cedex, 1 Bioinova,F-78370 Plaisir,2 Service dI'nformatique et Statistiques Medicales, Centre Hospitalier Universitaire HotelDieu, F-31052 Toulouse Cedex,3 and Laboratoire de Biologie Medicale, F-31800 Saint Gaudens, 4 France

Received 18 February 1992/Accepted 8 June 1992

In order to complete the immunodiagnosis of human toxocaral disease, an immunoenzymatic assay withexcretory-secretory antigens from Toxocara canis larvae was developed for the detection of specific immuno-globulin E (sIgE enzyme-linked immunosorbent assay [ELISA]). The specificity of the assay was evaluated inpatients presenting with various allergic or helminthic diseases. The sensitivity was assessed in patientsexhibiting clinical and biological symptoms indicative of toxocariasis, serodiagnosis of which was made by theWestern blot (WB; immunoblot) procedure that used the same antigen as that used in the sIgE ELISA but thatdetected specific IgG. The value of the sIgE ELISA for posttreatment follow-up was tested in two groups ofpatients: one group was treated with diethylcarbamazine; the other group was not treated with DEC. Resultsshowed that the specificity and sensitivity of the sIgE ELISA were moderate. Thus, the sIgE ELISA appearedto be insufficient for properly ensuring the serodiagnosis of toxocariasis when it is used alone. However, sIgEELISA might be an interesting complementary method for the detection of specific IgG. It was the only assaythat was found to be positive in sera from some hypereosinophilic patients. sIgE ELISA values decreasedsignificantly among the patients treated with DEC, indicating that this test would be useful for posttreatmentfollow-up assessment.

Toxocariasis is a parasitic zoonosis caused by infestationof humans with larvae of Toxocara canis, the commonroundworm in dogs, and possibly, larvae of T. cati, acommon ascarid species of cats (16, 22). Two syndromeshave been identified: visceral larva migrans (3) and ocularlarva migrans (21). Results of recent studies suggest that thedisease frequently assumes the features of a syndrome thatcomprises chronic weakness, abdominal pain, various signsof allergy, and hypereosinophilia (11, 18, 30). Moreover,toxocariasis has been found to be a cofactor for epilepsy (1)and has been proposed as a possible etiology in variousneurologic syndromes (25).These data have been obtained by sensitive and specific

serodiagnostic methods by using excretory-secretory anti-gens from T. canis larvae (TES-Ag) (6). These methods arethe TES-Ag enzyme-linked immunosorbent assay (TESELISA) (7, 14) and, more recently, the Western blot (WB;immunoblot) procedure (20), both of which detect immuno-globulin G (IgG) specific for TES-Ag. However, clinicalsigns of allergy and increased levels of total IgE in manycases of toxocariasis (11, 13, 18) indicated that IgE specificfor TES-Ag (sIgE) would be present. Brunello et al. (5) andDesowitz et al. (8) detected IgE antibodies for larval extractsof T. canis in patients with toxocariasis and asthmaticchildren. In 1986, Genchi et al. (10) used a radioimmunoas-say with TES-Ag for the immunodiagnosis of toxocariasisand showed that this assay is valuable for the diagnosis ofocular larva migrans (9). In 1986, Oliver et al. (23) demon-strated that sIgE is detectable by ELISA.

In most countries, strict legal and technical regulations areapplied to radioimmunoassays that use isotopically labeled

* Corresponding author.

reagents. Thus, we perfected for the first time an ELISA forthe detection of sIgE (sIgE ELISA); we then assessed thespecificity of this test in sera from patients presenting withvarious allergic or helminthic diseases or containing immuneanti-blood group antibodies. Then, we studied the value ofsIgE ELISA for the serodiagnosis of toxocariasis in serafrom patients suspected of having the disease. The serodi-agnosis of this zoonosis was based on the results of a WBassay that detected IgG for TES-Ag. WB was chosen since itis well correlated with TES ELISA; moreover, it seemedmore specific. Banding patterns that included low-molecu-lar-weight bands appeared to be specific for toxocaral infes-tations; high-molecular-weight bands were found to bepresent at significant levels only when sera from patientswith various helminthic diseases were tested, 33% of whichyielded positive results by TES ELISA (20). Finally, theeffectiveness of sIgE ELISA for use in follow-up assess-ments of treated patients was evaluated.

MATERIALS AND METHODS

TES-Ag. Cultivation of T. canis larvae was carried out bythe prototype method of de Savigny (6) that was modified byBowman et al. (4). In culture vials, the larval density was1,000 larvae ml-', and the RPMI 1640 medium (Intermed,Venissieux, France) in the vials, supplemented with 1%glutamine, was renewed every week. The supernatants werefiltered on Whatman no. 1 filter paper and were frozen at-70°C. After thawing, the batches were pooled, dialyzed for48 h in distilled water (Visking dialysing tube C 75; PolyLabo, Strasbourg, France), and then freeze-dried in 5-mlvials. Before each pool was used, the protein titer wasmeasured by the method of Lowry et al. (17).

Positive and negative reference samples. A positive refer-

2269

Vol. 30, No. 9

on April 5, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

2270 MAGNAVAL ET AL.

ence serum pool was made from the sera of 30 patientspresenting with toxocariasis that was positively serodiag-nosed by WB, namely, a typical seven-band pattern (20). Anegative reference serum pool was made from the sera of 30children (age, 3 to 6 months) referred to the Laboratory ofParasitology by the Department of Paediatrics for toxoplas-mosis serology. Titers of total IgE in the positive andnegative reference pools were 2,760 and 3.1 kIU liter-1,respectively, as determined by radioimmunoassay in theDepartment of Nuclear Medicine, Centre Hospitalier Uni-versitaire Purpan. Evaluation of the reproducibility of thesIgE ELISA was done by testing positive and negativereference pools two times a week for 19 weeks.Human sera from patients with allergic diseases (group 1).

Sera were obtained from 41 patients attending the Depart-ment of Dermatology and Allergy for various symptoms ofallergy (asthma [n = 15], eczema [n = 16], or chronicurticaria [n = 10]). Group 1 patients exhibited sensitizationto various allergens (antibiotics, house dust mites, pollens,and foods) as shown by radioallergosorbent tests, patchtests, or prick tests. Only seven patients had atopic eczema.The mean total IgE value was 912.4 kIU liter-1 (95%confidence interval [CI], 555.1, 1,499.6; range, 10 to 11,000kIU liter-1).Human sera with immune anti-A and/or anti-B blood group

or heterophilic antibodies (group 2). Sera were obtained from12 patients with immune anti-A hemagglutinins, 3 patientswith both anti-A and anti-B immune antibodies, and 5patients with heterophilic anti-sheep erythrocyte antibodies.Human sera from patients with various helminthic diseases

(group 3). We collected in the Laboratory of Parasitology175 serum samples from patients with anisakiasis (n = 2);cystic hydatidosis (n = 7); fascioliasis (n = 5); variousfilariases (n = 22); a history of living in an African country,hypereosinophilia, and a negative parasitological checkupbut a positive serodiagnosis for filariasis by fluoroimmunoas-say (n = 22); various schistosomiases with a positive sero-diagnosis and/or direct evidence of parasite eggs (n = 24);strongyloidiasis with a positive stool examination (n = 28);hypereosinophilia with a negative parasitological checkupexcept positive serodiagnosis for strongyloidiasis (n = 9);and positive stool examination forAscaris lumbricoides (n =7), hookworms (n = 6), pinworms (n = 22), tapeworms (n =9), and whipworms (n = 12).With these sera, WB with TES-Ag was negative or dis-

played the high-molecular-weight banding pattern only.Human sera from patients presenting with a clinical and/or

a biological syndrome evocative of toxocariasis (group 4). Serawere collected from 150 patients attending the ConsultationOffice in the Laboratory of Parasitology. None of the group4 patients had ever lived outside of France; only 3 patientshad traveled overseas as tourists. For all patients repeatedstool examinations, including examinations by Baerman'smethod, were negative for intestinal helminthiasis, as wereserological checkups for the autochthonal parasitic diseasescystic hydatidosis, fascioliasis, and strongyloidiasis. A se-rodiagnostic test for trichinosis is no longer included in thepanel of tests for autochthonal parasitic diseases. From 1975to 1977, serodiagnostic tests for trichinosis were systemati-cally carried out on over 12,000 serum samples referred tothe Laboratory of Parasitology, but only two cases of coverttrichinosis were detected. For the three patients who hadtraveled overseas, parasitologic examinations of blood, skin,and urine were found to be negative, and serodiagnostic testsfor tropical diseases (i.e., amoebiasis, filariasis, malaria, andschistosomiasis) were also negative. In 138 patients, serodi-

agnostic tests for toxocariasis by WB exhibited a typicalseven-band pattern that included low-molecular-weightbands. For 12 patients, results of WB remained negative.The date of onset of toxocariasis was estimated. On

average, the duration of the disease before consultation atthe Laboratory of Parasitology was 4.8 months (95% CI, 4.1,5.6]). The main clinical features were chronic weakness(76%), abdominal pain (36.5%), cough (26.5%), and varioussigns of allergy, the most frequent of which was pruritus(40%). The clinical level of illness was quantified by using aclinical score system performed in the Laboratory of Para-sitology (mean, 7; 95% CI, 6.2, 8.2). The mean eosinophilcount was 784 cells ,ul-1 (95% CI, 676, 901), and the meanlevel of total IgE was 510 kIU liter-1 (95% CI, 399, 653).Human sera from patients included in a therapeutic trial

(group 5). Sera from 60 patients from group 4 were obtainedfor inclusion in a therapeutic trial. Twenty-eight patientswere treated with diethylcarbamazine (DEC), and 32 pa-tients were not treated with DEC. These two batches of serawere similar since, before treatment, no significant differ-ence was found (Mann and Whitney's test) between thevalues of dur'ation of disease before consultation, clinicalscore, eosinophil count, WB results, and total and specificIgE levels. According to epidemiologic findings (11), thefollowing advice was given to patients so that they couldavoid possible reinfestation: deworming of any pet threetimes a year, enclosing kitchen gardens, and careful hand-washing before meals. These patients were examined againfor a control checkup from 2 to 3 months after the firstconsultation.WB procedure. The WB procedure used in this study has

been described elsewhere (20). Briefly, electrophoresis ofTES-Ag was carried out in a gel containing 10% polyacryl-amide and 7.5 ,ug of TES-Ag per lane. Transfer (semidrytype) lasted for 45 min under 0.2 to 0.4 V cm-1. Forimmunodetection, sera were diluted at 1:100. For that pro-cedure, we used Auroprobe BL Plus and Intense BL kits(Amersham, Paris, France), in which anti-human IgG waslabeled with gold and the detection of immune complexeswas enhanced by a silver coprecipitation.

sIgE ELISA procedure. Ninety-six-well microtitrationpolystyrene plates (Immunoplate, Poly-Labo Block; Nunc,Strasbourg, France) were loaded with a TES-Ag solution (50,ul per well) containing 2.5 ,ug of protein ml- in a 0.1 Mcarbonate-bicarbonate buffer (pH 9.6). The plates weremaintained for 18 h at 25°C in a moist chamber. Then, theantigenic solution was discarded and the plates were washedthree times with phosphate-buffered saline (PBS)-Tween(0.05% Tween 20; pH 7.2) buffer; they were dried and finallystored at -70°C. All test reagents, including anti-human IgEconjugate labeled with 3-galactosidase, were part of thePhadezym Rast kit (Pharmacia, Bois d'Arcy, France). Serafrom patients and the negative reference pool were testedundiluted; positive reference pool sera were tested undilutedand at 1:10, 1:100, and 1:1,000 dilutions. All diluted andundiluted sera were tested in duplicate. Sera (50 ,ul per well)were incubated at 25°C for 3 h. The plates were washed threetimes with PBS-Tween (0.05% Tween 20; pH 7.2) buffer.Fifty microliters of the anti-IgE conjugate solution wasadded to each well, and the plates were maintained at 25°Cfor 16 h in a moist chamber. Then, they were washed threetimes with PBS-Tween buffer. A total of 150 ,ul of thesubstrate (o-nitrophenyl-0-D-galactopyranoside solution)was added to each well, and the plates were incubated at37°C for 2 h. Finally, the reaction was stopped by theaddition of 50 ,u of the blocking solution to each well. The

J. CLIN. MICROBIOL.

on April 5, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

IMMUNODIAGNOSIS OF TOXOCARIASIS 2271

TABLE 1. Assessment of the cross-reactivity of sIgE ELISA

No. of patients whose sera were infectedGroup with the following titers (TU liter-'):

.1 25 .10 .20 .50

1 (allergic patients) 41 13 13 5 1 0

2 (immune isohemaggluti- 20 7 1 0nins)

3 (helminthiasis) 175 89 54 36 23 10Anisakiasisa 2 0

Cystic hydatidosisb 7 3 3 1 1 1

Fascioliasisb 5 4 2 2 2 1

Filariasis caused by: 22 13 12 9 7 3Bancroftiasis 3 1 0Loiasis 4 3 3 1 0Mansonella perstans 6 5 5 5 4 1Mansonella strepto- 1 1 1 1 1 1

cercaOnchocerciasis 8 3 3 2 2 1

Occult filariasisb 22 12 6 3 2 1

Schistosomiasisa,b 24 12 7 3 1 0

Strongyloidiasisa 28 18 10 8 5 1

Strongyloidiasisb 9 4 2 1 0

Intestinal helminthiasisa 56 23 12 9 5 3Ascaridiasis 7 4 1 0Hookworms 6 2 1 1 0Pinworms 22 8 6 4 2 2Tapeworms 9 2 1 1 1 0Whipworms 12 7 3 3 2 1

Total 236 109 68 41 24 10

a Diagnosis by direct evidence of parasite.b Diagnosis by serology.

plates were read at 420 nm with a Multiskan photometer(Flow Laboratories, Puteaux, France).The following steps were done with software written in the

Laboratory of Parasitology. The software was compiled inGW-Basic. This software read the optical densities (ODs)from every pair of wells and calculated the means. When adifference of over 10% was found between the two ODscorresponding to the same pair of serum samples, the pair ofserum samples was discarded and tested again. The negativepool was read at first and its average OD at 420 nm (OD420)was substracted from the values of other sera. Then, anundiluted positive reference pool of sera and 1:10, 1:100, and1:1,000 dilutions of serum were read, and a positive refer-ence curve was constructed. Results for patient sera wereexpressed in units. According to Voller et al. (31), thismethod yields the second best result when various systemsthat give ratios for positive versus negative serum sampleswere tested. In Pharmacia's system of units, the positivereference serum was heterologous and contained IgE frompatients sensitized to birch tree flower allergens. Thus, wepreferred to use a homologous system; our positive refer-

20 *18.7

1614

12 ..% 10.

42-0- q

13.4

10.7

2~~~~~

o -_e17 c O,In oIn~~ ~Toxoca Units I I

FIG. 1. Repartitioning of sIgE ELISA values in sera from 150patients with toxocariasis.

ence pool contained 1,000 units of IgE for TES-Ag liter-'.This unit was named the Toxocara unit (TU). The softwarereported OD420 of patient sera on the positive referencecurve and converted them into TU liter-'. The technicalcutoff value was 1 TU liter-'.

Statistical analysis. Statistical analysis was conducted byusing an i386/387 computer with statistical software (North-west Analytical, Portland, Oreg.). All the calculations weremade with logarithmic values of data, so geometric meansare given.

RESULTS

Serological examinations. Thirty-eight tests of the undi-luted positive reference pool showed a mean OD420 of 0.992(95% CI, 0.968, 1.016; range, 0.898 to 1.173; variation, 1%).At a 1:1,000 dilution, i.e., 1 TU, the mean OD420 of thepositive reference pool was 0.050 (95% CI, 0.046, 0.054;range, 0.030 to 0.070; variation, 5%). The negative referencepool exhibited a mean OD420 of 0.003 (95% CI, 0.002, 0.004;range, 0.000 to 0.018; variation, 100%).Among sera from patients in group 1, 2, and 3, WB with

TES-Ag yielded negative results (all groups) or displayedonly high-molecular-weight banding patterns (group 3). Theresults of sIgE-ELISA are presented in Table 1. Sera fromgroups 1 to 3 exhibited very low values, especially sera fromallergic patients or those containing immune isohemaggluti-nins or heterophilic antibodies. Sera from patients in group 3had the highest sIgE values, but the maximum value did notexceed 100 TU liter-'.

Figure 1 shows sIgE results for group 4 patients. Sera from131 of 150 (81.3%) patients with clinical toxocariasis werereactive in the sIgE-ELISA at 21 TU liter-', while serafrom 138 (92%) patients were positive in the WB assay. Ahigh proportion (18.7%) of sera from patients found to bepositive by WB were negative by sIgE ELISA; sera from 12patients (8%) presenting with a negative WB result displayedhigh titers for total IgE (range, 181 to 37,162 kIU liter-') andsIgE (range, 276 to 25,000 TU liter-l).An assessment of the efficacy of the sIgE ELISA for the

follow-up of treated patients was carried out on the resultsfor group 5 patients, as was the evaluation of the efficacy ofDEC.

Results from statistical analysis. A study of the specificityand sensitivity of the sIgE ELISA was conducted on theresults for patients in groups 1 to 4. Five cutoff values weretested: 1, 5, 10, 20, and 50 TU liter-'. The first valuecorresponded to the technical threshold (Table 2).

VOL. 30, 1992

on April 5, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

2272 MAGNAVAL ET AL.

TABLE 2. Assessment of sensitivity and specificity ofsIgE ELISA

sIgE ELISA Percentcutoff value(TU liter-') Sensitivity Specificity

.1 81.3 53.8

.5 67.9 76.2

.10 55.9 82.6

.20 53.2 89.8250 41.9 95.8

The correlation of the sIgE ELISA with other parameters,i.e., clinical (duration of disease before consultation andclinical score) or biological (eosinophilia and total IgE level),was studied for the results for group 4 patients by usingSpearman's rank test and multivariate regression analysis.No positive or negative correlation was found between sIgEvalues and duration of disease before consultation, clinicalscore, or eosinophil count (Spearman's coefficient values[0.052, 0.014, and -0.199] not significant). A strong corre-lation between sIgE and total IgE was found (Spearman'scoefficient, 0.469; Z statistic, 6.16; P < 0.001).

In group 5 patients (therapeutic trial), at the controlcheckup significant variations were found for the clinicalscore (control group and treated patients), eosinophil count(control group and treated patients), total IgE (controlgroup), and sIgE (treated patients) (Table 3).

DISCUSSIONWhen it is used routinely and intensively, sIgE ELISA

appears to be a simple and easy-to-use method. The use ofundiluted sera avoided the tiresome procedure of massdilutions and probably contributed to the excellent reproduc-ibility of the method.The intrinsic specificity of the sIgE ELISA was good,

since this test reacted at a very low average level (mean, 2.1TU liter-', 95% CI, 1.5, 2.9) with sera from allergic patientswhen the mean value of total IgE in the sera was high (912.4kIU liter-'; 95% CI, 555.1, 1,499.6). Nonspecific adsorptionof total IgE onto the polystyrol solid phase did not occur.When Oliver et al. (23) tested 58 serum samples from atopicor asthmatic patients that exhibited total IgE levels of.1,000 kIU liter-', they also found that the sIgE ELISAwas nonreactive. Thus, the good statistical agreement foundbetween sIgE and total IgE levels in patients in group 4 maybe due to the marked stimulation of IgE synthesis that hasbeen shown in many patients with helminthiases, includingtoxocariasis (13, 15, 24).

In 1983, Smith et al. (27) demonstrated that TES-Agcontains blood group-like antigens; they stated that sera with

immune isohemagglutinins could cross-react and thus yieldfalse-positive results in the serodiagnosis of toxocariasiswhen sera were tested by TES ELISA (26). In 1985, Glick-man and Schantz (12) showed that sera containing high titersof such haemagglutinins were not reactive when tested byTES ELISA, and we corroborated this point when weassessed the WB procedure for the serodiagnosis of toxo-cariasis (20). In the present study, neither sIgE ELISA wasaffected by such immune hemagglutinins.

Testing of sera from patients presenting with varioushelminthic diseases showed that the sIgE ELISA reacted ata low mean level (3.1 TU liter-'; 95% CI, 2.6, 3.8), although23.8% of these sera had .5 TU liter-'. Those positivitiescould be due to concurrent minor toxocaral infestations,because that zoonosis is widespread. In a previous study,examination of sera from such patients by WB revealed that6.7% (8 of 118 serum samples) displayed typical patterns(20).

Analysis of the results from group 4 patients, showed thata majority of the patients (52.1%) exhibited moderate values(from 1 to 100 TU liter-') and a high proportion of negativeresults (18.7%) was found. These two factors produced amoderate mean titer for the sIgE ELISA (49.8 TU liter-1).Genchi et al. (9) have hypothesized that the synthesis of sIgEremains at a low level during the acute phase of toxocaralinfestation. Statistical analysis of data for group 4 patients,which indicated that there is no relationship between thesIgE level and parameters such as the estimated duration ofthe disease, the clinical score, and/or eosinophilia, does notagree with the hypothesis of Genchi et al. (9). The synthesisof sIgE for a given parasite is probably genetically regulated,although this point remains unclear (24).The wide range of sIgE ELISA values in 138 patients with

toxocariasis showing a positive WB result suggested that thetwo methods correlated poorly. Since the antigens that weretransferred to the nitrocellulose sheet might be different fromthose that adhered to the ELISA plate, we studied thecorrelation between the detection of specific IgG by the TESELISA and that by sIgE ELISA. TES ELISA values wereavailable for 43 patients in group 3 and for 13 patients ingroup 4. These sera were tested by TES ELISA in a previousstudy (20). The results for the 43 serum samples frompatients in group 3 showed a lack of correlation between thetwo methods (Spearman's coefficient, 0.19, Z statistic, 1.22).Among the 13 serum samples from patients in group 4, areverse quantitative correlation was found (Spearman's co-efficient, patients in -0.475, Z statistic, -1.634; P = 0.05).Moreover, using the data published by Genchi et al. (9,

10), we assessed the quantitative correlation between theTES ELISA (IgG) and radioimmunoassay with TES-Ag(sIgE) for sera from 21 patients; Spearman's rank test was

TABLE 3. Statistical analysis of the kinetics of various parameters in patients in group 5'

Eosinophilia (celis 1-J) Total IgE (kIU liter-) sIgE ELISA titer (TU liter-1)Consultation DEC-treated Untreated DEC-treated Untreated DEC-treated Untreated

patients patients patients patients patients patients

First 1,005 803 574 840 68 157Second 287 415 527 724 48 79Wilcoxon's signed rank test 2.75 (P < 0.001) 2.75 (P = 0.006) NSb 2.2 (P = 0.03) 2.46 (P = 0.01) NS

(Z statistic)a Values are means.b NS, not significant.

J. CLIN. MICROBIOL.

on April 5, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

IMMUNODIAGNOSIS OF TOXOCARIASIS 2273

not significant (Spearman's coefficient, 0.078; Z statistic,0.35).From these findings it can be inferred that an assay that

detects sIgE should not be used alone, but it appears to be agood complement to a method that detects IgG for TES-Ag.In fact, sera from 12 of 150 patients in group 4 were stronglypositive for TES-Ag by sIgE ELISA only.

Prophylactic measures must be used in association withdrug therapy for toxocariasis. When treated and controlpatients exhibited clinical and biological disturbances (onaverage, 4.6 months [95% CI, 3, 4.4] and 3.6 months [95%CI, 3.4, 7] after the presumed onset of disease, respectively),a significant decrease in the clinical scores and eosinophilcounts was found in both groups 2 months after the firstconsultation. A remarkable point is the significant decreasein sIgE ELISA only in the DEC-treated group, indicatingthat for posttreatment follow-up assessment, the sIgEELISA could be more valuable than ELISAs that detectIgG. The levels of specific IgG did not vary significantly, andthis result is in agreement with the results of previous studiesin two groups of children, in which Bass et al. (2) assessedthe efficacy of thiabendazole versus that of placebo andchecked the TES ELISA results 1 year later. Only a slightdecrease in the average TES ELISA values was found insera from the two groups.For a long time it remained a question whether anthel-

minthics are effective for the treatment of toxocariasis.Previous reports describing the results of therapeutic trialswith various benzimidazoles showed that these drugs did notaffect the kinetics of biological parameters (2, 19, 28). In thisstudy, the results for group 5 patients proved the efficacy ofDEC, especially its effect on the sIgE ELISA level. More-over, when the values of the parameters were comparedamong the two groups at the time of consultation forcontrols, DEC was found to be more effective (P = 0.03 byMann and Whitney's one-tailed test; Z statistic, 1.9) oneosinophilia, a biological parameter that appears to becorrelated with the presence of viable larvae in body tissues(29).According to the results of this study, in the Laboratory of

Parasitology the WB procedure remains the first-line assayfor the serodiagnosis of human toxocariasis. A serum samplefound to be negative for helminthic diseases by our panel ofserodiagnostic tests and by WB is tested by sIgE ELISA. Adiscrepancy (i.e., negative WB and positive sIgE ELISA), ifany, is interpreted in three ways when the patient's casehistory is not available. When sIgE ELISA values rangefrom 1 to 100 TU liter-', physicians are told to ask forrepeated stool examination, including an examination by theBaerman's method, plus other direct examinations when thepatient lived in or was a native of a tropical country. Whenthe sIgE titer lies between 100 and 500 TU liter-', a secondserodiagnosis is carried out, when possible, 1 month later.Patient sera that exhibit titers of more than 500 TU liter-'are classified as toxocariasis cases.

ACKNOWLEDGMENTS

We thank F. Giordano (Department of Dermatology, CentreHospitalier Universitaire Purpan, Toulouse), Y. Marty (RegionalBlood Bank, Toulouse), J. Fabre (Department of Nuclear Medicine,Centre Hospitalier Universitaire Purpan, Toulouse), and D. Hosford(Institut Henri-Beaufour, Paris) for help. We also thank A. Galasso,J. Jezequel, M.-J. Touchard, M. Vidal, G. Carriere, E. Dubly, andA. Thuries for technical assistance.

REFERENCES1. Arpino, C., G. C. Gattinara, D. Piergfli and P. Curatolo. 1990.

Toxocara infection and epilepsy in children. A case-controlstudy. Epilepsia 31:33-36.

2. Bass, J. L., K. A. Mehta, L. T. Glickman, R. Blocker, and B. M.Eppes. 1987. Asymptomatic toxocariasis in children: a prospec-tive study and a treatment trial. Clin. Pediatr. 26:441-446.

3. Beaver, P. C. 1956. Larva migrans. Exp. Parasitol. 5:587-621.4. Bowman, D. D., M. Mika-Grieve, and R. B. Grieve. 1987.

Circulating excretory-secretory antigen level and specific anti-body responses in mice infected with Toxocara canis. Am. J.Trop. Med. Hyg. 36:75-82.

5. Brunello, F., C. Genchi, and P. Falagiani. 1982. Detection oflarva-specific IgE in human toxocariasis. Trans. R. Soc. Trop.Med. Hyg. 77:279-280.

6. De Savigny, D. H. 1975. In vitro maintenance of Totxocara canislarvae and a simple method for the production of Taxocara ESantigens for use in serodiagnosis tests for visceral larva migrans.J. Parasitol. 61:781-782.

7. De Savigny, D. H., A. Voller, and A. W. Woodruff. 1979.Toxocariasis: serodiagnosis by enzyme immuno-assay. J. Clin.Pathol. 32:284-288.

8. Desowitz, R. S., R. Rudoy, and J. W. Barnwell. 1981. Antibodiesto canine helminth parasites in asthmatic and nonasthmaticchildren. Int. Arch. Allergy Appl. Immunol. 65:361-366.

9. Genchi, C., P. Falagiani, G. Riva, M. Tinelli, F. Brunello, M.Boero, and M. Almaviva. 1988. IgE and IgG antibodies inToxocara canis infection. A clinical evaluation. Ann. Allergy61:43-46.

10. Genchi, C., M. Tinelli, F. Brunello, and P. Falagiani. 1986.Serodiagnosis of ocular toxocariasis: a comparison of specificIgE and IgG. Trans. R. Soc. Trop. Med. Hyg. 80:993-994.

11. Glickman, L. T., J.-F. Magnaval, L. M. Domanski, F. S. Shofer,S. S. Lauria, B. Gottstein, and B. Brochier. 1987. Visceral larvamigrans in French adults: a new disease syndrome? Am. J.Epidemiol. 125:1019-1034.

12. Glickman, L. T., and P. M. Schantz. 1985. Do Toxocara canislarval antigens used in enzyme-linked immunosorbent assay forvisceral larva migrans cross-react with AB isohaemagglutininsand give false positive results? Z. Parasitenkd. 71:395-400.

13. Hogarth-Scott, R. S., S. G. 0. Johansson, and H. Bennich. 1969.Antibodies to T. canis in the sera of visceral larva migranspatients: the significance of raised levels of IgE. Clin. Exp.Immunol. 5:619-625.

14. Jacquier, P., B. Gottstein, Y. Stingelin and J. Eckert. 1991.Immunodiagnosis of toxocarosis in humans: evaluation of a newenzyme-linked immunosorbent assay. J. Clin. Microbiol. 29:1831-1835.

15. Jarrett, E. E. E., and H. R. P. Miller. 1982. Production andactivities of IgE in helminths infections. Prog. Allergy 31:178-223.

16. Kennedy, M. W., R. M. Maizels, M. Meghji, L. Young, F.Quereshi, and H. V. Smith. 1987. Species-specific and commonepitopes on the secreted and surface antigens of T. canis and T.cati. Parasite Immunol. 70:522-529.

17. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall.1951. Protein measurement with the Folin phenol reagent. J.Biol. Chem. 193:265-275.

18. Magnaval, J.-F. 1987. Elements nouveaux dans la s6meiologiedes larva migrans viscerales. Presse Med. 16:151-154.

19. Magnaval, J.-F., and J.-P. Charlet. 1987. Efficacit6 comparee duthiabendazole et du mebendazole dans le traitement de latoxocarose. Therapie 42:541-544.

20. Magnaval, J.-F., R. Fabre, P. Maurieres, J.-P. Charlet, and B.de Larrard. 1991. Application of the Western-blotting procedurefor the immunodiagnosis of human toxocariasis. Parasitol. Res.77:697-702.

21. Molk, R. 1983. Ocular toxocariasis. A review of the literature.Ann. Ophthalmol. 15:216-231.

22. Nichols, R. L. 1956. The etiology of visceral larva migrans.Diagnostic morphology of infective second-stage larvae. J.Parasitol. 42:349-362.

23. Oliver, S. T., G. Couslaud, H. V. Smith, and R. W. A. Gird-

VOL. 30, 1992

on April 5, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/

2274 MAGNAVAL ET AL. J. CLIN. MICROBIOL.

wood. 1986. Antibody isotype reactivity, isotype specific sero-diagnosis and IgE antibody specificity in humans infected withToxocara canis larvae. Trans. R. Soc. Med. Hyg. 80:985.

24. Romagnani, S. 1990. Regulation and deregulation of human IgEsynthesis. Immunol. Today 11:316-320.

25. Schantz, P. M. 1989. Toxcocara larva migrans now. Am. J. Trop.Med. Hyg. 41(Suppl.):21-34.

26. Smith, H. V., A. Hainson, and R. W. A. Girdwood. 1987.Serodiagnosis of human toxocariasis. Trans. R. Soc. Trop.Med. Hyg. 81:516.

27. Smith, H. V., J. F. Kusel, and R. W. A. Girdwood. 1983. Theproduction of human A and B blood group-like substances by invitro-maintained second stage Toxocara canis larvae: theirpresence in the outer larva surfaces and in their excretions/

secretions. Clin. Exp. Immunol. 54:625-633.28. Sturchler, D., P. Schubarth, M. Gualzata, B. Gottstein, and A.

Oettli. 1989. Thiabendazole vs. albendazole in treatment oftoxocariasis. Ann. Trop. Med. Parasitol. 83:473-476.

29. Sugane, K. and T. Oshima. 1984. Induction of peripheral bloodeosinophilia in mice by excretory and secretory antigens ofTotxocara canis larvae. J. Helminthol. 58:143-147.

30. Taylor, M. R., C. T. Keane, P. O'Connor, E. Mulvihill, and C.Holland. The expanded spectrum of toxocaral disease. Lanceti:692-694.

31. Voller, A., D. E. Bidwell, and A. Bartlett. 1979. The enzyme-linked immunosorbent assay (ELISA), p. 32-33. Nuffield Lab-oratories of Comparative Medicine, The Zoological Society ofLondon, London.

on April 5, 2021 by guest

http://jcm.asm

.org/D

ownloaded from

http://jcm.asm.org/