Study design
This was a single-center, retrospective, diagnostic accuracy study. Sera were retrieved from the “Tropica Biobank” of the Department of Infectious Tropical diseases and Microbiology (DITM) of IRCCS Sacro Cuore Don Calabria hospital, Negrar, Verona, Italy. Results are reported following the STARD (Standards for Reporting Diagnostic Accuracy) guidelines [14].
Participants and inclusion/exclusion criteria
All available sera in the Tropica Biobank were considered eligible. The inclusion criteria were: (i) serum collected from immigrants from S. stercoralis endemic countries (i.e. individuals from Africa, Latin America, South-East Asia and Western Pacific regions); and (ii) serum with a matched result for APC and/or PCR assay for S. stercoralis (i.e. the latter test/s done on the same day or ± 30 days from the collection of the serum), or routine coproparasitology after formol–ether concentration if positive for S. stercoralis larvae. The exclusion criteria were: (i) unavailable/insufficient quantity of serum; (ii) sera from patients who received treatment with ivermectin in the previous 6 months; and (iii) sera from patients for whom only routine coproparasitology was performed after formol-ether concentration, if negative for S. stercoralis larvae.
A list of eligible samples was generated from the electronic database of the biobank, and all consecutive samples meeting the given criteria were included.
Test methods
Fecal tests
Both APC and the PCR assay for S. stercoralis are routinely used at the study site. The procedure for APC follows a modified Koga agar plate method. The cultures are incubated up to 6 days and visually inspected on day 2, 4 and 6. On day 6, cultures are flooded with saline and the sediment examined by microscopy. In a previous retrospective study, the sensitivity of the method was 45.4% (95% CI: 30.4–61.1%) [15].
The PCR is a real-time assay based on the method published by Verweij et al. [16]. Briefly, approximately 200 mg of unfixed feces was suspended in 200 µl of phosphate-buffered saline containing 2% polyvinylpolypyrolidone (Sigma-Aldrich Merck KGaA, Darmstadt, Germany) and frozen overnight at − 20 °C. After thawing, the samples were boiled for 10 min at 95 °C. The DNA was then extracted using Magnapure LC.2 extractor (F. Hoffmann-La Roche AG, Basel, Switzerland) following the protocol DNA_I_Blood_Cells_High performance_II, using the DNA isolation kit I (F. Hoffmann-La Roche AG), with a final elution volume of 100 µl. The real-time assay, targeting the small-subunit rRNA gene sequence of S. stercoralis, was performed as described. Appropriate positive and negative controls, as well as controls for PCR inhibitors and amplification quality using PhHV-1 control DNA in the same sample reaction, were included in each run. Thermocycling consisted of 40 cycles, and cycle threshold values < 40 were considered positive. The reactions, detection and data analysis were performed with the CFX96 detection system (Bio-Rad, Hercules, CA, USA). In a previous retrospective study the method demonstrated a sensitivity of 56.8% (95% CI: 41.0–71.6%) [15].
Serological tests
The following serological tests were used:
-
i.
The new Strongy Detect ™ IgG ELISA and IgG4 ELISA (InBios International, Inc., Seattle, WA, USA). Positive and negative control samples are provided in the kit. The tests were performed according to the manufacturer’s instructions. Interpretation of the results (positive/negative) was based on the analysis of the receiver operating characteristics (ROC) curve, as detailed below.
-
ii.
An in-house IFAT. This IFAT is routinely implemented at DITM for screening and individual diagnosis of strongyloidiasis. A positive result is defined as a titer ≥ 1:20.
-
iii.
The Strongyloides ratti IgG ELISA (Bordier ELISA; Bordier Affinity Products SA, Crissier, Switzerland). At DITM, the test is performed as per the manufacturer’s instructions. As the cut-off varies between runs, a normalized optical density (OD) ratio is used to compare the results obtained in different sessions. A ratio ≥ 1 defines positive results. The test is widely available and deployed for routine screening and diagnostic activities across Europe.
Test procedures
The new InBios ELISA assays were performed by staff of the DITM parasitology laboratory, which is a regional reference laboratory for parasitic infections. Sera were re-coded by staff not directly involved in the laboratory procedures. The laboratory staff were blinded to the results of any fecal and serological tests previously performed. The samples were tested shortly after thawing. For sera already tested using the Bordier ELISA and IFAT at the time of routine analysis, only the new InBios ELISA tests were performed. Conversely, in case results for either routine seroassay were missing, the Bordier ELISA and/or IFAT were carried out as needed.
Data on concomitant infections, assessed as per routine procedure in the parasitology laboratory, were retrieved from the electronic clinical records of DITM. These included routine microscopy after formol–ether concentration of feces for intestinal helminths and ELISA serology for filariasis and schistosomiasis (both assays from Bordier Affinity Products SA).
Data analysis
Reference standards and sera classification
The performances of the new InBios ELISA tests were first assessed against the results of fecal tests only (primary reference standard). For this endpoint, S. stercoralis infection was defined based on at least one positive fecal test among standard copromicroscopy of formol–ether concentrated feces, or on positive APC or PCR results.
As the included fecal tests are virtually 100% specific but lack sensitivity, a composite reference standard (CRS) based on the combination of the results of fecal and serological tests was also used to classify patients, as recommended for the evaluation of diagnostic tests in the absence of a gold standard [17]. Patients were classified as positive for S. stercoralis infection if they were: (i) positive on at least one fecal test among standard copromicroscopy of formol–ether-concentrated feces, APC or PCR; or (ii) negative on all fecal tests performed, including at least APC or PCR, but concordantly positive by Bordier ELISA and IFAT; or (iii) negative on all fecal tests performed, including at least APC or PCR, and positive only in one routine serology test but at a high titer (≥ 1:160 for IFAT, ≥ 2 for Bordier ELISA). These threshold values were chosen based on the results of a previous study [6]. Negativity for S. stercoralis infection was defined in the presence of negative fecal (all performed, including at least APC or PCR) and routine serology (Bordier ELISA and IFAT) tests. Finally, doubtful infection status (indeterminate cases) was defined when there was a negative fecal test (all performed, including at least APC or PCR) and positivity in only one of the two routine seroassays, at a low titer (< 1:160 for IFAT or < 2 for Bordier ELISA).
Sample size calculation
For sample size calculation, it was assumed that the new InBios ELISAs based on recombinant antigens NIE and SsIR would not have lower sensitivity and specificity than the existing NIE ELISA (75.4 and 89.5%, respectively [6]). For sensitivity, a sample size of 246 was estimated to be required based on a two-sided 95% CI of 0.10 width and test sensitivity of 0.80. For specificity, a sample size of 151 was calculated, assuming a specificity of 0.89 with a 95% CI with width of 0.10. Calculations were carried out using Power Analysis and Sample Size Software (PASS) version 2019 (NCSS Statistical Software, Kaysville, UT USA).
Statistical analysis
The optimal cut-off for the two index ELISA assays was obtained using the maximum sum of sensitivity and specificity (Youden’s J statistic), determined across all cut-off points of the ROC curve for each endpoint. Cut-off points obtained minimizing the distance to the upper-left corner of the ROC plot (D0-1) were also calculated. Individual samples analyzed with the InBios ELISA tests with OD exceeding the reading limit of the ELISA reader were assigned an OD = 5. Areas under the ROC curve (AUC) were compared by the DeLong test.
Tests results were displayed in contingency tables from which sensitivity and specificity were calculated against the two reference standards: fecal test (primary reference standard) and CRS. When applying the CRS, results obtained from sera of patients with doubtful infection status (i.e. indeterminate cases with negative fecal tests and positive serology at “low titer” in only one of the routine IFAT and Bordier ELISA tests) were first excluded from the analysis, then included while classifying them as either positive or negative.
The association between serology results of the index ELISA assays and presence of other intestinal helminth infections, schistosomiasis or filariasis was assessed using the Chi-square or Fisher Exact tests. Differences between sensitivities and specificities of the two tests were assessed using the McNemar test. Estimations were reported together with the 95% CI. Data analysis was performed using SAS software version 9.4 (SAS Institute, Cary, NC, USA).