The Novodiag® Stool parasites assay, an innovative high-plex technique for fast detection of protozoa, helminths and microsporidia in stool samples: a retrospective and prospective study.

OBJECTIVES: We provide the first evaluation of the CE-IVD marked Novodiag® stool parasites assay (NVD), allowing rapid and high-plex detection of 26 distinct targets, encompassing protozoans, helminths and microsporidia in stool samples.
METHODS: A total of 254 samples (n = 205 patients) were prospectively processed by the NVD and our routine procedure (RP). Performances of the NVD were compared with RP. Samples only positive by the NVD assay were investigated by external PCR assays. Sensitivity and specificity (Se/Sp) and time from sample receipt to results were determined for each method. The NVD was also evaluated against 77 additional samples positive for a wide range of parasites.
RESULTS: Overall positivity rate was 16.9% for RP compared with 34% using the NVD assay, and 164 samples (66%) were negative by both methods. Only 30 positive samples (12%) showed full concordance between RP and NVD. Fifty-three discordant samples were sent for external investigations. Except for Giardia intestinalis and Trichuris spp., higher Se was observed for the NVD assay for Blastocystis spp. (100% vs. 63%), Dientamoeba fragilis (100% vs. 0%), Schistosoma spp. (100% vs. 17%), and Enterobius vermicularis (100% vs. 67%) but roughly similar to RP for the remaining parasites tested. False-positive results were identified for Blastocystis spp., G. intestinalis, and Trichuris spp. using the NVD assay. The NVD mostly provides a diagnosis on the day of sample receipt compared with a mean of three days with RP.
CONCLUSIONS: Besides some limitations, the NVD is a new diagnostic strategy allowing rapid and high-plex detection of gastrointestinal parasites from unpreserved stools. © S. Hartuis et al., published by EDP Sciences, 2022.

Introduction

Besides bacteria and viruses, various parasites and microsporidia account for a significant part of gastrointestinal diseases worldwide [30]. Prevalence of these parasites varies greatly between countries, with higher prevalence usually being reported in tropical areas and mainly being related to poor sanitation and hygiene, and elevated contamination of water resources or soils. In high-resource countries, these diseases are dominated by protozoan parasites and microsporidia, although their respective prevalence is probably underestimated as they are thought to be underdiagnosed. Yet, in many countries, routine microscopy remains the gold standard for the diagnosis of intestinal parasite diseases. Besides the diagnosis of gastrointestinal diseases, parasite stool screening is also part of the general screening of stool donors in the context of fecal microbiota transplantation [5]. However, it exhibits several limitations like being resource-consuming and displaying limited performances [18]. In addition, it is usually recommended to process multiple samples to ensure sufficient sensitivity. In this context, it is not surprising that alternative approaches relying on antigen detection or DNA-based techniques have been developed, first in-house PCRs then commercial assays. A large range of commercial assays is now available on the market, including syndromic panels or addressing specifically the issue of gastrointestinal parasites, primarily protozoans. These assays have been the subject of many evaluation studies or reviews [20, 31].

The first commercially available multiplex PCR assay specifically developed for the detection of both helminths and microsporidia has recently been evaluated [3]. Combined with a previously developed assay targeting several protozoan species, once automated on a DNA extraction platform, it allows high-plex detection of 15 gastrointestinal pathogens [2, 3]. Despite this improvement, some major pathogens are not included in this panel, including Schistosoma spp., and therefore cannot be diagnosed. In addition, Enterocytozoon and Encephalitozoon species, could not be differentiated, illustrating the need for additional techniques to allow their identification as they require distinct therapeutic strategies.

Recently, a novel commercial and CE-IVD marked assay (Novodiag® Stool Parasites, Mobidiag, Espoo, Finland), was approved. This technique relies on cutting-edge technology integrating DNA purification, PCR amplification and microarray hybridization/detection into a single, lab-on-chip cartridge. It allows for on-demand testing and high-plex detection of 26 distinct targets encompassing protozoans, helminths including nematodes and trematodes, and two microsporidian genera, from unpreserved stools. Thanks to fast processing of the samples, which includes mechanical disruption and ready-to-use cartridges, results are obtained within 90 min. Here we provide the first clinical evaluation of this new assay.

Materials and methods

Ethics

This study has been approved by our local ethic committee and registered under reference TS005-BIO.2019_6.

Prospective study

Two hundred and fifty-four unpreserved stool samples received at the Laboratory of Parasitology and Medical Mycology, Nantes University Hospital, France were processed using the Novodiag® Stool Parasites assay (NVD), according to the manufacturer’s instructions (see Table S1 for the list of parasites included in the panel). In parallel, the same samples were simultaneously analyzed according to our routine procedures (RP, considered here the reference method), blindly to the results of the NVD assay (see Fig. 1). The routine procedure, which was performed by skilled microscopists, systematically included microscopic examination of the fresh unpreserved stool samples before and after a concentration step (Para-Selles® Plus, Biosynex, Illkirch-Graffenstaden, France). Depending on the context (clinical symptoms, history of travel abroad, eosinophilia, etc.), additional techniques were performed: a further additional but different concentration step (Para-Selles® Plus), modified Ziehl–Neelsen staining for coccidiosis, two species-specific in-house single plex real-time PCR assays targeting Enterocytozoon bieneusi and Encephalitozoon intestinalis/E. hellem [6], and agar plate culture according to Arakaki for strongyloidiasis [1]. Samples that were positive only by the NVD were subjected to previously published PCRs at two expert laboratories. These external investigations were conducted at Clermont-Ferrand University Hospital, France, for Blastocystis spp. and D. fragilis [25, 27] and at Elisabeth-TweeSteden Hospital, the Netherlands, for the remaining parasites [15]. Samples negative by the NVD but positive by RP were defined as false-negative samples of the NVD. Conversely, samples positive by the NVD assay but negative by both RP and the external assay were considered false-positive samples of the NVD. Finally, samples negative by RP but positive by the NVD assay and the external assay, were considered false-negative by RP. Sensitivity (Se) and specificity (Sp) were then determined for each diagnostic strategy. The time to complete sample analysis (interval between sample receipt at our laboratory and final results) was also collected for each sample and for each diagnostic strategy. Samples “invalid” by the NVD (i.e., internal control not detected) were excluded from the analysis. A flowchart of the design of the study is given in Figure 1.

Figure 1

Study flow chart.

Retrospective study

In addition to the prospective evaluation study, 77 additional frozen stool samples, positive for a wide array of parasites and collected at our hospital and five other laboratories in France were also processed using the NVD assay. This additional set of samples included 21 distinct parasites, among which 14 were part of the diagnostic panel of the high-plex assay.

Results

Overall, 249 (n = 201 patients) of the 254 stool samples (n = 205 patients) included in the prospective study, could be analyzed (5 invalid samples with the NVD, 2%). The positivity rate was 21.3% for RP (53 of 249 samples) but 16.9% (42 of 249) when excluding parasites not targeted by the NVD assay (i.e., non-pathogenic amoeba species and flagellates). By comparison, the positivity rate of the NVD was 34.1% (85 of 249). One hundred and sixty-four samples (65.9%) were negative by both diagnostic strategies.

Distribution of the 11 different parasites identified in this prospective cohort, according to each diagnostic strategy, is given in Figure 2. When positive, most samples displayed a single parasite (72 of 85, 84.7% for NVD, compared with 38 of 40, 95% for RP). Thirty samples (12%) showed full concordance between both diagnostic approaches (i.e., perfect match between the parasites identified by RP and NVD). Nine samples (3.6%) showed partial concordance (i.e., incomplete match between RP and NVD), whereas 49 (19.7%) showed no concordance. Blastocystis spp. ranked first as they were identified in 27 samples by RP (10.8%) compared with 59 samples using the NVD assay (23.7%). A higher detection rate with the NVD assay was also noted for other parasites: Dientamoeba fragilis was identified in 10 stool samples compared with none by RP, and six samples were positive for Schistosoma spp. using the NVD compared with a single positive sample using RP. Two of the positive samples for Schistosoma by the NVD assay but negative by microscopy were from patients recently diagnosed with schistosomiasis. The complete list of samples sent for external investigations to resolve the discrepancies is given in Table S2.

Figure 2

Diversity of the parasites identified in the prospective cohort according to each diagnostic strategy (n = 249 samples). Note: Additional parasites not included in the Novodiag® panel were identified by microscopy: Entamoeba coli (n = 17), Endolimax nana (n = 18), and Chilomastix mesnili (n = 2).

In-depth investigations of the discrepancies by external PCRs revealed that all samples negative by RP but positive by the NVD assay for D. fragilis (n = 10), E. vermicularis (n = 1) and Schistosoma (n = 3) were positive by the external PCR assays (i.e., false-negative by RP). External investigations also revealed that microscopy failed to detect some Blastocystis spp. (n = 16). In some instances, external PCRs could not confirm the findings of the NVD assay (Table S2). These samples considered false positives of the NVD were observed for Blastocystis spp. (n = 16), G. intestinalis (n = 4) and Trichuris spp. (n = 7). Lastly, adjusted analytical performances for the NVD and RP are given in Table 1. Except for G. intestinalis and Trichuris spp., sensitivity was higher for the NVD, for Blastocystis spp. (100% vs. 62.8%), D. fragilis (100% vs. 0%), Schistosoma spp. (100% vs. 16.7%), and E. vermicularis (100% vs. 66.7%). Identical performances for both diagnostic strategies were observed for Cryptosporidium spp., Taenia saginata, and E. bieneusi. The time to complete sample analysis evaluated during the prospective study is outlined in Figure 3. Although a mean of 3 days was required using RP to obtain the complete results, due to multiple techniques, the NVD assay mostly provided a final diagnosis on the day of sample receipt.

Table 1

Comparative performances between routine procedure and the Novodiag® Stool Parasites assay on the prospective cohort (n = 249 samples).

Parasite of interest	Diagnostic strategy	Positive samples (n=)	Negative samples (n=)	False-positive samples* (n=)	False-negative samples* (n=)	Se (%)	Sp (%)
Blastocystis spp.	Routine	27	206	NA	16	62.8	NA
	Novodiag®	43	190	16	0	100	92.2
Dientamoeba fragilis	Routine	0	239	NA	10	0	NA
	Novodiag®	10	239	0	0	100	100
Giardia intestinalis	Routine	2	247	NA	0	100	NA
	Novodiag®	1	243	4	1	50	98.4
Cryptosporidium spp.	Routine	1	248	NA	0	100	NA
	Novodiag®	1	248	0	0	100	100
Schistosoma spp.a	Routine	1	243	NA	5	16.7	NA
	Novodiag®	6	243	0	0	100	100
Taenia saginata	Routine	1	248	NA	0	100	NA
	Novodiag®	1	248	0	0	100	100
Enterobius vermicularis	Routine	2	246	NA	1	66.7	NA
	Novodiag®	3	246	0	0	100	100
Trichuris spp.	Routine	1	248	NA	0	100	NA
	Novodiag®	0	241	7	1	0	97.2
Enterocytozoon bieneusi	Routine	6	243	NA	0	100	NA
	Novodiag®	6	243	0	0	100	100

NA: Not appropriate as routine procedure was considered the reference method (i.e., samples positive by RP were considered true positive).

False-negative or false-positive were determined based on external PCR investigations.

Except one identified to the genus level (Schistosoma spp.) all remaining Schistosoma-positive samples were identified as S. mansoni by the NVD and RP.

Figure 3

Time to final results according to each diagnostic strategy (routine procedure vs. Novodiag® Stool Parasites) during the prospective study (n = 249 samples): D: day from sample arrival at the clinical laboratory to final results.

To challenge the NVD on a larger number of parasites, this technique was further evaluated on a collection of frozen positive stool samples. In all, 74 of 77 samples (three invalid samples) could be analyzed (Table 2). All samples known to be positive for T. saginata (n = 4), Cystoisospora belli (n = 2), Entamoeba histolytica (n = 2), Ascaris lumbricoides (n = 1), Strongyloides stercoralis (n = 4) and hookworms (n = 4) were also positive by the NVD (100% sensitivity/specificity). An almost perfect agreement was observed for Cryptosporidium spp., G. intestinalis, E. vermicularis and E. bieneusi with only a few false-negative results by the NVD assay (Table 2). Again, the NVD assay identified parasites which had been missed by microscopy: Blastocystis spp. (n = 10), D. fragilis (n = 2), and S. mansoni (n = 2). Both false-negative S. mansoni samples by RP were from two patients recently diagnosed with schistosomiasis. The NVD assay was negative for a sample positive with a single S. mansoni egg by microscopy, obtained from a patient diagnosed and treated for schistosomiasis in 2017. Only four out of the 8 positive samples for Trichuris spp. were detected by the NVD, confirming its poor performance for this target. Finally, no cross-reactivity was evidenced with parasites not included in the high-plex panel (Table 2).

Table 2

Performances of the Novodiag® Stool Parasites assay on a collection of positive samples (n = 74 samples, retrospective study).

Parasite of interest	Positive samples according to routine procedure (n=)	Positive samples by the Novodiag® assay (n=)
Blastocystis spp.	16	26
Dientamoeba fragilis	0	2
Giardia intestinalis	14	12
Cryptosporidium spp.	8	7
Cystoisospora belli	2	2
Entamoeba histolytica	2	2
Schistosoma mansoni	9	10
Taenia spp.	4	4a
Ascaris lumbricoides/suum	1	1
Enterobius vermicularis	3	2
Strongyloides stercoralis	4	4
Hookworms	4	4b
Trichuris spp.	8	4
Enterocytozoon bieneusi	8	7
Other parasites*	42	0

All identified as T. saginata/Taenia asiatica by the Novodiag® assay.

Identified as N. americanus (n = 3), A. duodenale (n = 1) by the Novodiag assay®.

Other parasites not targeted by the Novodiag® assay included: Dicrocoelium dendriticum (n = 1), Entamoeba coli (n = 16), Endolimax nana (n = 12), Entamoeba dispar (n = 6), Entamoeba hartmanni (n = 3), Iodamoeba butschlii (n = 1) and unidentified flagellates (n = 4).

Discussion

The present study reports the first evaluation of the NVD assay compared with routine procedures, in real-life conditions. One strength of the prospective study was to include specific external PCR assays in order to resolve the discrepant results, a strategy which is often missing in similar studies but offers better evaluation of assay performances [3, 28, 29]. Most discrepancies were observed with Blastocystis spp., D. fragilis and S. mansoni, for which detection rates were higher using the NVD assay. Of note, the lack of D. fragilis detection by microscopy was expected as RP was performed on unpreserved stools and trichrome permanent staining is not performed in our laboratory. Except for Blastocystis spp., for which the improved detection rate also unmasked false-positive samples, the higher sensitivity of the NVD for D. fragilis and S. mansoni is in agreement with previous studies using PCR assays [3, 10, 13, 21, 25]. Another strength was to challenge the NVD assay on a collection of positive samples. This allowed us to consolidate our findings for Cryptosporidium spp., E. bieneusi and T. saginata, which were poorly represented in the prospective study. Likewise, it provides additional data for soil-transmitted helminths, E. histolytica and C. belli, for which the NVD assay performed well although the number of positive samples remains low. On the contrary, this set of samples confirmed the poor performance of the NVD for Trichuris spp., although the sample processing with the NVD includes a bead-beating step that is known to improve the detection of Trichuris spp. [4, 14]. The reduced performances of the NVD for G. intestinalis, observed during the prospective study, were unexpected as there is an overall consensus about the superiority of DNA detection over microscopy for this protozoa [2, 17, 21, 24]. Of note, better performances were observed during the retrospective study which included a higher number of Giardia-positive samples. These weak performances for the detection of Trichuris spp. and to a lesser extent G. intestinalis which are already known by the manufacturer (outlined in the technical note of the NVD assay), require specific optimizations.

Besides these limitations, the main advantages of this new assay include: (i) a high-plex format, allowing the detection of the 26 most common protozoa, helminths and microsporidia in a single run; (ii) a fast turnaround time allowing results within 90 min with few invalid results; (iii) higher performances for the detection of S. mansoni which is of major interest considering the notoriously poor sensitivity of microscopy [18]; (iv) its ability to provide species identification which can be useful for epidemiological purposes, as illustrated with hookworms and Taenia species eggs which cannot be identified to the species level using microscopy [9]. Here, except one identified as Ancylostoma duodenale, all hookworm-positive samples were Necator americanus. Further investigations are required to assess the ability of the NVD to identify the emerging species Ancylostoma ceylanicum [8]. The NVD assay also allows for the screening of underdiagnosed pathogens such as microsporidia and Cryptosporidium spp., which are often overlooked by physicians as their diagnosis requires specific techniques. Furthermore, its ability to discriminate E. bieneusi from Encephalitozoon species is of specific interest as distinct treatment regimens are required to treat these opportunistic fungi [11].

Because of the intermittent shedding of parasites, notably helminths eggs or larvae and the overall low sensitivity of microscopic methods, it is usually recommended to repeat parasite stool examination on different days, to ensure high sensitivity [31]. This strategy is controversial in populations with a low prevalence of the disease, as discussed elsewhere [7]. An alternative and cost-effective option is to limit the examination of a second specimen only when the first sample is negative and in symptomatic patients, with a third specimen for patients remaining negative despite a high index suspicion [23]. Our study was not designed to answer whether a single stool PCR performed similarly to RP on three consecutive samples, because only a few patients in our study had multiple samples available for analysis. Finally, a shortage of NVD reagents due to the COVID-19 pandemic, led us to stop this evaluation at some points of the prospective study.

Not all genera/species included in the panel of the NVD assay could be tested, which is a limitation of the present study. However, these missing parasites are the rarest identified in stool samples in European countries, which underlines the difficulty to assess the performances of multiplex panels in countries with a low prevalence of these diseases [12, 16, 19, 21, 22, 26].

Conclusion

To conclude, one major advantage of this high-plex assay is the rapid turnaround time which is explained by a reduced hands-on time together with the capability of the instrument to process four samples simultaneously, which makes it possible to reduce diagnostic delays. Importantly, although time to results is largely dependent on each laboratory organization and diagnostic strategies, compared with RP which requires multiple concentrations, staining methods as well as additional techniques in some cases, the NVD is simpler and therefore faster, providing a result within 90 min. The fully automated Novodiag® Stool Parasites assay is an alternative approach to labor-intensive microscopy-based methods especially in settings or countries with low parasite prevalence or personnel with reduced skills in parasitological diagnosis. Besides the diagnosis of gastrointestinal diseases, this technique also represents an attractive approach for rapid donor stool screening in the context of fecal microbiota transplantation [5]. This study provides better knowledge of the performances of the NVD in real-life conditions, which we hope will be of great help to determine its utility in the diagnostic algorithm of gastrointestinal parasites.

Conflict of interest

The authors declared no conflict of interest.

Funding

This study was supported by internal funding from Nantes University Hospital.

Supplementary material is available at https://www.parasite-journal.org/10.1051/parasite/2022026/olm.

Table S1. High-plex panel of the Novodiag® Stool Parasites assay.

Table S2. Results of external investigations for samples positive by Novodiag® assay and negative by the routine procedure (n = 53).