Evaluation of an Indirect Immunofluorescence Assay for Strongyloidiasis as a Tool for Diagnosis and Follow-Up

The diagnostic accuracy of an indirect immunofluorescence antibodytest (IFAT) for Strongyloides stercoralis at different serumantibody titers was evaluated. To assess diagnostic sensitivity,sera from 156 patients with known strongyloidiasis were collected.Negative control sera were obtained from a composite group of427 subjects (blood donors and hospitalized patients). Withan area under the receiver-operating characteristic plot of0.98, the IFAT showed a high level of diagnostic accuracy forstrongyloidiasis. An antibody titer of $≥$ 1:20, with 97% sensitivityand 98% specificity, was identified as the diagnostic thresholdwith the best overall performance. Cross-reactions were evaluatedwith 41 additional samples from patients with other known helminthinfections, and the IFAT detected low-titer positivity in onlyone subject with filariasis. A positive IFAT result at an antibodydilution of $≥$ 1:80 was virtually 100% specific, with 71% sensitivity.To test the usefulness of the IFAT as a monitoring tool, thechanges in specific-antibody titers after treatment in a groupof 155 patients were evaluated. Seroreversion or a decreasein antibody titer of twofold or more was observed in 60% ofthe patients. Response to treatment was directly correlatedto the initial antibody titer, and a baseline titer of $≥$ 1:80was identified as the best predictor of response. In conclusion,a positive IFAT result at an antibody dilution of $≥$ 1:20 is theoptimal cutoff for screening. A titer of $≥$ 1:80, with virtuallyno false-positive result, is a reliable cutoff for a serologicalassessment of treatment efficacy and for inclusion in clinicaltrials.

INTRODUCTION

Top

Introduction

Strongyloidiasis is a chronic, soil-transmitted helminthiasiswith a worldwide distribution. Current estimates indicate thatit affects 30 to 100 million people, but accurate prevalencedata are lacking. Although strongyloidiasis is endemic in manytropical and subtropical countries (sub-Saharan Africa, LatinAmerica, Southeast Asia, and Northern Australia), foci of lowendemicity are also reported in temperate climates, such asthe Mediterranean coast, Eastern Europe, and the southeasternUnited States (1, 3, 8, 18, 20, 23, 26, 38, 39, 42, 47, 48,50, 52, 53). Due to its unique autoinfective cycle, Strongyloidesstercoralis may persist in the host for indefinite periods (21).Most infected individuals are asymptomatic or may present aspecificand intermittent clinical symptoms and unexplained eosinophilia(14, 38, 43). However, in cases of immunosuppression, strongyloidiasismay become a disseminated, life-threatening disease (7, 16,29, 30, 54). Unfortunately, its diagnosis remains a challenge.Given the irregular and often low larval output in uncomplicatedinfections, the sensitivity of direct stool examination is verypoor (12, 27, 59). Furthermore, currently available antiparasiticdrugs do not always eradicate the infection (27). Consequently,the risk of failing to diagnose the disease is compounded bythe difficult determination of treatment efficacy. Thus, anenzyme-linked immunosorbent assay (ELISA) and an indirect immunofluorescenceantibody test (IFAT) have been developed, with a wide rangeof reported sensitivities and specificities (19, 40, 55, 56,57, 58, 60). Assessing the accuracies of serological assayshas been limited primarily by the absence of a gold standard.Therefore, the utility of immunoassays both for diagnosis andfor follow-up remains controversial (38, 43, 57).

The aim of the present study was to assess the sensitivity andspecificity of an IFAT for S. stercoralis at different antibodytiters in order to determine the best cutoff both for a clinicaldecision and for enrollment in clinical trials. In addition,to test the usefulness of serology as a monitoring tool, changesin specific-antibody titers were measured after treatment ina group of 155 patients over a 4-month follow-up period.

MATERIALS AND METHODS

Top

Materials and Methods

Samples used to assess diagnostic sensitivity. Sera were collected from 156 patients in whose stool specimensS. stercoralis larvae had been detected.

Samples used to assess diagnostic specificity. Negative control sera were obtained from a group of 427 subjects,comprising 229 healthy Italian blood donors and 198 patientsadmitted to the hospital in medical and pediatric departmentsfor any reason. None of these subjects had a history of travelto high-risk areas or any other known epidemiologic risk factorfor S. stercoralis infection. None of them had an eosinophilcount of more than 300/µl or complained of any cutaneous,gastrointestinal, or respiratory symptom indicative of strongyloidiasis.

Samples used to assess cross-reactivity. The sera used to evaluate cross-reactions were obtained from41 patients with the following other parasitologically provenhelminth infections: filariasis (11 cases), schistosomiasis(9 cases), hookworm infections (7 cases), toxocariasis (5 cases),ascariasis (5 cases), trichuriasis (3 cases), and Strongyloidesfuelleborni infection (1 case). All of them had multiple (atleast five) stool specimens negative for S. stercoralis upondirect examination.

IFAT for S. stercoralis. Immunoglobulin G antibodies against S. stercoralis were detectedby an IFAT developed and validated at the Centre for TropicalDiseases, Sacro Cuore Hospital, Negrar, Verona, Italy. Briefly,for antigen preparation, intact S. stercoralis filariform larvaewere obtained from a positive charcoal fecal culture (15). Byusing the Baermann technique (5), all the liquid containingthe vital larvae was collected and centrifuged at 4°C. Thesediment was washed five times with phosphate-buffered saline(PBS), pH 7.2, at 4°C. It was then exposed to acetone for30 min at 4°C and washed again five times with PBS. Tenreaction fields of 8 mm in diameter, with waterproof surroundings,were prepared on a microscope glass coverslip (thickness, 0.13to 0.16 mm; 24 by 60 mm; Menzel-Glaser). The material containingthe intact filariform larvae was then applied to the reactionfields, with at least 10 larvae/field. Each reaction field wascovered with a 12-µm-pore-size filter of the same diameter.Twenty-five microliters of diluted serum sample was appliedas an overlay onto the filter of each field. Negative and positivereference control serum samples of 25 µl at a 1/20 dilutionwere added. The coverslip was then incubated in a moist chamberat 37°C for 30 min. After the coverslips were rinsed initiallyfor 15 min and then for 10 min with PBS, 25 µl of 50-times-dilutedfluorescein isothiocyanate-conjugated anti-human immunoglobulinG (bioMérieux Italia SpA, Rome, Italy) containing 0.2%Evans blue (bioMérieux Italia SpA) was added to eachfield. The coverslip was again incubated at 37°C for 30min. After being washed twice in PBS for 10 min, the specimenswere embedded in Fluoroprep (bioMérieux Italia SpA) andapplied upside-down onto a microscope slide. The light fromthe bound, fluorescein-conjugated antibodies was detected byan experienced microbiologist, blinded to clinical data, througha Leitz Laborlux S fluorescence microscope with a 20x objective.A positive reaction was indicated by the defined and completefluorescence of larvae on the slide. Results were expressedas the greatest dilution in serum that gave a positive result.

Patients. Patients with a diagnosis of strongyloidiasis from October 1993to March 2005 were identified from the database of the Centrefor Tropical Diseases. They were included in the analysis ifthey met at least one of the following criteria: a positivemicroscopic stool examination, a positive fecal culture, and/ora positive S. stercoralis IFAT result with an antibody dilutionof $≥$ 1:20. Demographic and clinical data, including history oftravel or immigration, geographical areas of long- or short-termresidence, presence or absence of eosinophilia (defined as aneosinophil count of $≥$ 300/µl), and clinical signs and symptoms,were reviewed.

Treatment regimens. All patients received one of the two alternative treatment regimensconsidered to be the most effective for strongyloidiasis atthe time of their admission: thiabendazole at 25 mg/kg of bodyweight twice a day orally for 2 days (74 patients) or ivermectinat 200 µg/kg orally in a single dose (81 patients).

Assessment of therapeutic efficacy. Patients were considered to be responsive to treatment if theirpositive baseline IFAT result reverted to negative and/or showeda decrease in antibody titer of twofold or more within 4 monthsof the initial treatment. For subjects with positive directstool examinations (and/or cultures) upon recruitment, negativeresults after treatment were also mandatory as a criterion forcure.

Statistical analysis. Diagnostic accuracy was assessed as the area under the receiver-operatingcharacteristic (ROC) plot. Diagnosis obtained by direct stoolexamination was the external gold standard. Sensitivity, specificity,and positive and negative likelihood ratios were calculatedfor the 1:20, 1:40, 1:80, and 1:160 serum antibody titer (31,63).

For the population of patients treated for strongyloidiasis,baseline variables potentially associated with the serologicalresponse to treatment were evaluated. The distribution of continuousvariables was reported as the median and the interquartile range(25th to 75th percentiles). Categorical variables were presentedas numbers and percentages. The comparison between the responsiveand nonresponsive groups was carried out using Student's t testor the Mann-Whitney U test for continuous variables. Qualitativedata were compared by using the chi-square test with Yates'corrections and the Fisher exact test when necessary. The binarylogistic regression model was used to assess the predictorsof serological response to pharmacological treatment. A valueof P of less than or equal to 0.05 was considered significant.

RESULTS

Top

Results

Decision cutoff. The complete spectrum of sensitivity and specificity combinationsfor the different serum antibody titers in the IFAT for S. stercoraliswas represented by the ROC plot (Fig. 1). The area under theROC plot of 0.98 (95% confidence interval, 0.97 to 0.99; P =0.000001) showed the high diagnostic accuracy of this immunoassay.An antibody titer of $≥$ 1:20, which represents the serum antibodytiter closest to the upper left corner of the ROC plot, maximizedthe sensitivity-specificity index (Fig. 1). The estimated diagnosticaccuracy of IFAT at different serum antibody titer is reportedin Table 1. The best overall performance of IFAT was found ata titer of $≥$ 1:20, with 97% sensitivity and 98% specificity. Ata titer of $≥$ 1:80, the test showed the maximum level of specificity,99.8%, still with an acceptable level of sensitivity (71%).Among the group of 41 patients with other known helminth infections,only one positive IFAT result was observed, at an antibody dilutionof 1:20, in a patient with Mansonella perstans infection.

View larger version (17K):
[in this window]
[in a new window]

FIG. 1. ROC curve for the IFAT test at different antibody dilutions.

View this table:
[in this window]
[in a new window]

TABLE 1. Estimated diagnostic accuracy of Strongyloides IFAT at different serum antibody titers

Serological follow-up. From the database search, 360 patients with a diagnosis of strongyloidiasiswere identified. One hundred fifty-five had a confirmed serologicaldiagnosis and a serological follow-up and were thus includedin the analysis. Patient characteristics are shown in Table2. Ninety-nine patients (64%) had a history of travel or residenceoutside Italy, mostly in areas where strongyloidiasis is highlyendemic, including Africa (61%), Asia (20%), and South America(16%). Fifty-six subjects (36%) had never traveled outside thecountry. More than 80% of the subjects reported symptoms, mostlycutaneous and gastrointestinal disturbances.

View this table:
[in this window]
[in a new window]

TABLE 2. Main characteristics of the study population (155 patients)

Clinical adverse effects of treatment were observed in 74 patients(48%). The most important were dizziness (63%), nausea (32%),and vomiting (20%), all with higher frequencies in the grouptreated with thiabendazole (P = 0.0001). In no case did theside effects determine discontinuation of treatment.

Among the whole study population, at the 4-month serologicalfollow-up, 93/155 (60%) subjects were found to be responsiveto treatment, with either seroreversion (36%) or a decreasein antibody titers of twofold or more (24%). Among the nonresponsivepatients, 23% of the subjects had unvaried or increased titersand 17% had a reduction by one titer (i.e., a reduction by oneconsecutive serum dilution).

Response to treatment was directly correlated to the baselineantibody titer (Spearman correlation, 0.25; P = 0.002). Amongthe group of patients with an initial titer of $≥$ 1:80, 54/74 (73%)patients met the criterion for cure. In contrast, among thegroup of patients with an initial titer of <1:80, the eradicationrate was 39/82 (48%; P = 0.002).

The binary logistic regression model identified a baseline titerof $≥$ 1:80 as the best predictor of the response to treatment (P= 0.001) (Table 3).

View this table:
[in this window]
[in a new window]

TABLE 3. Multivariate analysis of predictors of response to treatment^a

DISCUSSION

Top

Discussion

Recent reports of several cases of fatal disseminated strongyloidiasisin developed countries highlight the problem of a disease whichis underestimated due both to the low level of clinical suspicionand to the insensitivity of direct diagnostic tools (4, 6, 11,12, 24, 34, 35, 44, 45, 49, 51, 59).

The low sensitivity of conventional stool examination may wellexplain the underestimation of the disease prevalence on theone hand and the wide range of reported drug efficacies on theother.

Therefore, more-sensitive immunodiagnostic techniques have beenanalyzed. Most studies have evaluated the accuracy of ELISA,with reported sensitivities ranging from 80% to 97% and specificitiesranging from 29% to 99% (9, 13, 17, 26, 33, 36, 55, 58), whilelittle data are available on the IFAT (22, 33). The IFAT istechnically more complex than ELISA and requires skillful personnelboth for antigen preparation and for reading. Compared to ELISA,the IFAT has the advantage of giving a quantitative result througha precise determination of the specific-antibody titer. Thisis particularly useful for follow-up. Recent studies reporteda decrease in the antibody titer after successful treatment(28, 32, 38, 43, 46). As suggested by several authors, the opticaldensity from an ELISA may be used as a proxy for the antibodytiter (32, 38, 43), but the IFAT gives an objective determinationof the titer through real serum antibody dilutions.

Our results show that both the sensitivity and the specificityof the IFAT are very high and that the cross-reactivity rate,in subjects with other nematode infections, is negligible.

The ROC plot (Fig. 1) supported evidence of the high accuracyof the IFAT. The best overall performance of the IFAT was foundat a titer of $≥$ 1:20. At a titer of $≥$ 1:80, the test showed themaximum level of specificity, still with an acceptable levelof sensitivity.

In agreement with those from previous studies (38, 56), theseresults show that the immunoassay has much higher sensitivity,even at higher titers, than direct smear or agar plate culture.

Because in clinical practice the optimal serological cutofffor screening for a potentially fatal disease should correspondto the highest sensitivity, a positive titer of $≥$ 1:20 was foundto be the best cutoff for screening for this infection. Allsubjects with a positive IFAT result at this dilution shouldbe treated, regardless of the results of direct tests.

In contrast, the case definition for enrollment in a clinicaltrial should correspond to the highest specificity, as falsepositives could give rise to a misleading interpretation ofthe results, with underestimation of the drug efficacy (12).A positive IFAT result at a dilution of $≥$ 1:80, virtually 100%specific, with acceptable sensitivity, was identified as theoptimal cutoff for enrollment in clinical trials.

A limited number of studies have evaluated the use of serologyas a tool for monitoring therapeutic efficacy. Most (28, 32,37, 38, 43, 46) reported a decline in antibody titer within6 to 12 months. However, results are heterogeneous and difficultto compare due to variable and nonstandardized laboratory techniquesand follow-up timings (26).

In the present study, after a $≥$ 4-month follow-up, seroreversionor a decrease in antibody titer of twofold or more was observedin 60% of patients. These results provide further evidence thatserology is a useful tool to assess treatment efficacy (28,32, 36, 43, 46).

However, this response rate was lower than expected. Previousstudies with the same drug regimens (2, 10, 14, 25, 41, 43,61, 62) reported eradication rates ranging from 85% to 100%.Given the absence of a reference diagnostic strategy, thesestudies employed a variety of follow-up protocols. Moreover,all randomized trials used insensitive stool examination techniquesto assess therapeutic efficacy, with a consequent overestimationof the anthelmintic effects of the drugs (2, 10, 14, 41, 61,62).

A direct correlation between the diagnostic titer and the responseto treatment as defined above was found. The eradication rateamong the group of patients with baseline titers of $≥$ 1:80 wassignificantly higher than that achieved among the group of patientswith diagnostic titers of <1:80 (73% versus 48%; P = 0.002).This difference may be explained by the suboptimal specificityof serology at lower titers: the inclusion of patients withfalse-positive results upon recruitment could account for anunderestimation of therapeutic efficacy. A baseline serum antibodytiter of $≥$ 1:80 was the best predictor of response to pharmacologicaltreatment.

A limitation of this study is the scarce availability of theStrongyloides IFAT, currently used only by research laboratories.Our promising results suggest the need for a wider disseminationof this serological assay, both as a diagnostic and as a monitoringtool.

However, more studies are needed in order to identify the mostappropriate follow-up timing. In conclusion, the IFAT is a sensibleand specific diagnostic tool for strongyloidiasis. A titer of $≥$ 1:20 is the optimal cutoff for diagnosis. A titer of $≥$ 1:80 isa reliable cutoff for a serological assessment of treatmentefficacy and for inclusion in clinical trials.

FOOTNOTES

* Corresponding author. Mailing address: Centre for Tropical Diseases, Sacro Cuore Hospital, via Sempreboni 5, 37024, Negrar, Verona, Italy. Phone: (0039) 045-601-3326. Fax: (0039) 045-601-3694. E-mail: zeno.bisoffi{at}sacrocuore.it.

Published ahead of print on 29 November 2006.

REFERENCES

Top