Journal of Microbiological Methods 110 (2015) 78–84

Contents lists available at ScienceDirect

Journal of Microbiological Methods

journal homepage: www.elsevier.com/locate/jmicmeth

Comparison of four rapid diagnostic tests, ELISA, microscopy and PCR for
the detection of Giardia lamblia, Cryptosporidium spp. and Entamoeba
histolytica in feces
Dorien Van den Bossche ⁎, Lieselotte Cnops, Jacob Verschueren, Marjan Van Esbroeck
Department of Clinical Sciences, Institute of Tropical Medicine, Antwerp, Belgium

a r t i c l e i n f o a b s t r a c t

Article history: Purpose: Microscopy is the diagnostic reference standard for the detection of parasites, but it is labor-intensive
Received 7 December 2014 and requires experience. Rapid diagnostic tests (RDTs) can provide an alternative to microscopy.
Received in revised form 19 January 2015 Methods: RDTs from four different manufacturers were compared to enzyme-linked immunosorbent assay
Accepted 20 January 2015 (ELISA), microscopy and/or parasite-specific real-time PCR: ImmunoCardSTAT!®CGE (Meridian Bioscience Inc.,
Available online 20 January 2015
Cincinnati, Ohio, USA) (A), Crypto/Giardia Duo-Strip (Coris Bioconcepts, Gembloux, Belgium) (B), RIDA®QUICK
Cryptosporidium/Giardia/Entamoeba Combi (R-BioPharm, Darmstadt, Germany) (C) and Giardia/Cryptosporidium
Keywords:
Giardia
Quik Chek (Techlab Inc., Blacksburg, Virginia, USA) (D).
Cryptosporidium Results: Thirty frozen samples were analyzed retrospectively. For Giardia lamblia (n = 12) and Cryptosporidium
Entamoeba (n = 12) sensitivities ranged from 58% (B), over 83% (A, C) to 100% (D) and from 92% (B) to 100% (A, C, D),
Rapid diagnostic tests respectively. Specificity for both G. lamblia and Cryptosporidium was 100% for all RDT brands. Sensitivity for
Entamoeba histolytica (n = 5) was 100%, while specificity reached 80% (A) to 88% (C). In a prospective study,
fresh samples were tested. For G. lamblia (n = 30), sensitivity ranged from 66% (B), over 79% (A) and 83%
(C) to 100% (D) and specificity varied between 94% (D) and 100% (A, B, C). For Cryptosporidium (n = 3), sensitiv-
ity was 100% for all brands except (B) (67%) and specificities were 95% (A, B), 98% (C) and 100% (D). E. histolytica
(n = 1) was detected by both (A) and (C), while specificity was 81% and 87% respectively.
Conclusion: RDTs can be a valuable tool when microscopic expertise is poor and in remote and outbreak settings
where other techniques are often not available and rapid diagnosis is required.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction
Giardia lamblia and Cryptosporidium spp. are both protozoan para-
sites which can be present without symptoms or cause diarrhea and
abdominal discomfort with weight loss and malabsorption. Entamoeba
histolytica is a unicellular parasite responsible for intestinal and hepatic
amoebiasis and occasionally affects other organs. Clinical symptoms of
intestinal amoebiasis range from colitis to dysentery or an ameboma,
but can be asymptomatic as well. These three parasites can lead to
human infection via fecal–oral transmission of the cysts through con-
taminated food and water and person-to-person contact. They are com-
mon in both developed and developing countries, but with an increased
risk in the latter due to poor sanitation standards (Dillingham et al.,
2002; Ali and Hill, 2003).
A variety of methods for diagnosis of all three parasites is available.
G. lamblia cysts or trophozoites can be detected by microscopic, immu-
nological and molecular methods in stool samples. The same techniques
⁎ Corresponding author at: Central Laboratory of Clinical Biology, Department of Clinical
Sciences, Institute of Tropical Medicine, Kronenburgstraat 43/3, 2000 Antwerp, Belgium.
E-mail address: dorienvdb@hotmail.com (D. Van den Bossche).
can be applied for Cryptosporidium spp. but oocysts cannot be positively
identified in wet mounts and an additional staining, like acid fast Ziehl–
Neelsen or carbol-fuchsin staining, is necessary. Microscopic examina-
tion of E. histolytica cysts does not allow one to make a distinction
with Entamoeba dispar. Molecular investigation is indispensable as the
latter is considered non-pathogenic as opposed to E. histolytica. Molecu-
lar techniques are sensitive and specific, but not easily accessible and
expensive. Microscopy is time-consuming, labor-intensive, relies on
the technician's experience and three stool samples may be required
to increase sensitivity. Direct fluorescent antibody (DFA) is an immuno-
logical method which allows the visualization of the whole parasite
through fluorescence. Enzymatic immunoassays (EIA) permit an objec-
tive result by the obtained optical densities (ODs). Still, these tests
require more than an hour to generate a result and are optimally used
in settings with a lot of samples allowing one to test the samples in
batch. Rapid diagnostic tests (RDTs) are increasingly popular tools as
they provide a solution to overcome these disadvantages. RDTs are
immunochromatographic lateral-flow tests which allow the detection
of antigens of one or more protozoan parasites in a single test format,
are easy to perform and to interpret and can be used in settings with
poor resources. In this study, the performance of four commercial

http://dx.doi.org/10.1016/j.mimet.2015.01.016
0167-7012/© 2015 Elsevier B.V. All rights reserved.
 D. Van den Bossche et al. / Journal of Microbiological Methods 110 (2015) 78–84
RDTs was compared to routine diagnostic methods for the detection of
G. lamblia, Cryptosporidium spp. and E. histolytica.
2. Material and methods
2.1. Routine diagnostic methods
Analyses were performed by examining stool samples collected
from patients presenting at the outpatient clinic of the Institute of Trop-
ical Medicine (ITM), Antwerp, Belgium or stool samples that were sub-
mitted to the Central Laboratory of Clinical Biology (CLKB) of ITM for
diagnosis of parasitic infections. Microscopic detection of ova and
cysts was performed by the examination of direct smears with saline
and wet mounts after formalin-ether concentration (Loughlin and
Spitz, 1949). A carbol-fuchsin staining (Heine, 1982; Potters and Van
Esbroeck, 2010) was performed on formalin-ether concentrates for
the detection of Cryptosporidium. Freshly collected samples were fixed
with a sodium acetate, acetic acid and formaldehyde (SAF) solution
within 20 min and examined by microscopy after iron-hematoxylin
Kinyoun staining. Copro-antigen enzyme-linked immunosorbent assays
(ELISAs) were performed by the G. lamblia ProSpecT ELISA Micro-
plate assay, Cryptosporidium ProSpecT ELISA Microplate assay and
E. histolytica ProSpecT ELISA Microplate assay which detects both
E. histolytica and E. dispar (Remel, Lenexa, Kansas, USA). A parasite-
specific real-time polymerase chain reaction (PCR) to differentiate
between E. histolytica and E. dispar (Cnops and Van Esbroeck, 2010), to
detect G. lamblia (adapted from Verweij et al., 2003), or Cryptosporidium
hominis and Cryptosporidium parvum (adapted from Hadfield et al.,
2011) was performed on all samples positive for the corresponding
parasite by microscopy and/or ELISA, real-time PCR. The primer
and probe sequences were used as described before (Cnops, Verweij,
Hadfield), while the extraction method and PCR conditions were
slightly adapted. Briefly, primers and probes were purchased from
Integrated DNA Technologies (IDT, Belgium, Leuven). The lib13 target
of C. hominis and C. parvum was detected in a duplex reaction. DNA
extraction of samples of the retrospective study was performed with
the QIAamp DNA stool kit (Qiagen Benelux, Venlo, The Netherlands)
(Cnops and Van Esbroeck, 2010). Stool samples used in the prospective
study were incubated in cobas® PCR media buffer (Roche Diagnos-
tics, Vilvoorde, Belgium) and thoroughly mixed for 20 min with the
Hulamixer (Invitrogen, Merelbeke, Belgium) before overnight storage at
− 20 °C (Cnops and Van Esbroeck, 2010). Prior to automated DNA
extraction by the MagNA Pure LC 2.0 with the MagNA Pure LC Total
Nucleic Acid High Performance kit (Roche), samples were heated for
10 min at 95 °C, centrifuged and 500 μL of fecal suspension incubated
for 10 min at 56 °C after the addition of 4% polyvinylpolypyrrolidone
(PVPP) and proteinase K (L. Cnops, K. Demeulemeester, E. Van
Gintelenberg and M. Van Esbroeck, presented at the 8th European
Meeting on Molecular Diagnostics, Scheveningen, The Netherlands,
2–4 Oct 2013). All PCRs were run on a SmartCycler II (Cepheid Benelux,
Belgium) in a 25 μL reaction volume with 1 × Hotstar Taq mastermix
(Qiagen, Hilden, Germany) using 5 μL of DNA. Each sample was
tested for efficient extraction and inhibition of the PCR by an exoge-
nous extraction control (PhHV-1), and in each PCR a positive and
negative control were tested to control the PCR process (Cnops and
Van Esbroeck, 2010).
2.2. Rapid diagnostic tests
Four immunochromatographic RDT assays were evaluated
and performed according to the manufacturer's instructions.
ImmunoCardSTAT!®CGE (Meridian Bioscience Inc., Cincinnati,
Ohio, USA), Crypto/Giardia Duo-Strip (Coris Bioconcepts, Gembloux,
Belgium), RIDA®QUICK Cryptosporidium/Giardia/Entamoeba Combi
(R-BioPharm, Darmstadt, Germany) and Giardia/Cryptosporidium Quik
Chek (Techlab Inc., Blacksburg, Virginia, USA). All four brands are able
79
to identify G. lamblia and Cryptosporidium, and ImmunoCardSTAT!®CGE
and RIDA®QUICK Combi additionally detect E. histolytica. Specifications
of the different RDTs with regard to specific identification of protozoa
and sample storage are summarized in Table 1.
2.3. Retrospective study
The following 30 samples were selected for testing by all RDT
brands: 6 G. lamblia, 8 C. hominis, 2 C. parvum, 3 E. histolytica, 3
E. dispar, 1 Entamoeba hartmanni (included because initially identified
as E. histolytica/dispar), 1 G. lamblia/C. hominis, 1 C. hominis/E. dispar, 2
E. histolytica/G. lamblia and 3 E. dispar/G. lamblia.
The samples were kept frozen at −20 °C for a period ranging from
1 month to 5 years. Parasite detection was confirmed by PCR in 27/30
samples. E. histolytica/dispar PCR was negative in one sample containing
E. hartmanni cysts. For 2 Giardia samples, of which one was positive by
both microscopy and ELISA and one by microscopy (not tested by
ELISA), insufficient material was available for confirmation by PCR.
Each sample was tested with all RDT brands on the same day, except
for ImmunoCardSTAT!®CGE, which was tested two months later due
to the unavailability of this new test at study start. Two technicians
independently interpreted the result of the test lines as negative,
weaker than, equal to or stronger than the control line. In the case
of differences in qualitative interpretation (positive versus negative)
between both interpreters, the worst possible scenario (false positive or
false negative) was recorded as the final result.
2.4. Prospective study
In a prospective study, fresh samples collected during routine
work-up between April–August 2013 were included if enough material
was available to perform all tests. Ten non-fixed negative samples and
60 non-fixed samples in which G. lamblia, Cryptosporidium and/or
E. histolytica/dispar, were detected by either ELISA and/or microscopy,
were selected for RDT testing. Parasite-specific PCR was performed on
every sample positive by either microscopy or ELISA. In three cases
(three Giardia positives) not enough sample was available. In case
only E. histolytica/dispar was found by microscopy and/or ELISA,
only ImmunoCardSTAT!®CGE and RIDA®QUICK were performed.
Each sample was tested with all RDTs on the same day, except for
ImmunoCardSTAT!®CGE for the reasons explained above.
Test results were interpreted independently by two lab technicians
as described for the retrospective study. Specimens were considered
true G. lamblia or Cryptosporidium positives if they were positive by
PCR (or positive by microscopy and ELISA in case PCR analysis could
not be performed). Specimens were considered true E. histolytica posi-
tives if a positive microscopic and/or ELISA result was confirmed by
the E. histolytica specific PCR. Specimens were considered as true
negative when both microscopy and ELISA were negative for G. lamblia,
Cryptosporidium or E. histolytica.
2.5. Statistical analysis
Analyse-it Software (Leeds, England) was used to calculate 95% con-
fidence intervals (95% CI) for sensitivity and specificity. Kruskal–Wallis
test with Bonferroni correction allowed multiple pairwise comparison
of the ELISA ODs classified according to test interpretation as true posi-
tive, true negative and false negative.
3. Results
3.1. Retrospective study
Sensitivities and specificities of all four RDTs in the retrospective
study setting are summarized in Table 2. The specificity was 100.0%
for G. lamblia and Cryptosporidium, and ranged from 80.0% to 88.0% for
80 D. Van den Bossche et al. / Journal of Microbiological Methods 110 (2015) 78–84

Table 1
Specifications of four immunochromatographic RDT's according to manufacturer's package inserts.

RDT Identification specifications Sample storage preferences: Temperature (time) Freezing allowed? Fixated samples allowed?

ImmunoCard STAT!®CGE (Meridian) C. parvum 2–8 °C (48 h) Yes⁎ No 10% formalin, SAF, PVA, etc.
G. lamblia
E. histolytica
Duo-Strip (Coris) C. parvum 2–8 °C (24 h) Yes No formalin
G.lamblia
RIDA®QUICK (R-BioPharm) C. parvum Not specified Not specified Not specified
G. lamblia
E. histolytica (sensu lato)
Quik Chek (Techlab) Cryptosporidium spp. 2–8 °C (72 h) Yes (b90 days)⁎ No PVA
G. lamblia
⁎ The package insert mentions to avoid freezing and thawing. SAF: sodium acetate, acetic acid and formaldehyde; PVA: polyvinyl alcohol fixative.

E. histolytica. Sensitivities ranged from 58.3% to 100.0% for G. lamblia,
from 91.7% to 100.0% for Cryptosporidium and were 100.0% for both
RDTs able to identify E. histolytica.
3.2. Prospective study
Out of 70 samples, 36 (51%) were produced at the outpatient clinic
of the ITM, whereas 18 (26%) and 16 (23%) samples were sent by mail
to the ITM by the patient or via other laboratories, respectively. Samples
were tested between 0 and 13 days upon production. No difference was
seen between freshly collected samples and samples older than one day
for neither of the RDTs (data not shown).
Ten microscopy and ELISA negative samples (in 2 Blastocystis
hominis was found) were analyzed by all four RDT brands. Nine out of
10 samples tested negative by all RDT's and one sample tested positive
for E. histolytica/dispar by ImmunoCardSTAT!®CGE and RIDA®QUICK
Combi. The latter result was confirmed negative for E. histolytica/dispar
by PCR.
Additionally, 60 samples positive by microscopy and/or ELISA for
G. lamblia (n = 29), E. histolytica/dispar (n = 24), Cryptosporidium spp.
(n = 4) and G. lamblia + E. histolytica/dispar (n = 3) were tested by
RDT. Out of 32 G. lamblia included samples, 30 G. lamblia true positives
were found of which 27 were confirmed by PCR and three were positive
by both microscopy and ELISA. Two samples were ELISA false positive
results. Out of 27 E. histolytica/dispar included samples, only one was
confirmed as E. histolytica, 25 as E. dispar and one was negative by PCR.
Out of four Cryptosporidium included samples, three true positives
were identified by PCR as C. hominis (n = 2) or C. parvum (n = 1).
One sample was weakly positive by ELISA but negative by microscopy.
Upon repeat of the ELISA, according to prescribed routine lab proce-
dure, this sample tested negative, which was confirmed by a nega-
tive PCR. Other parasites found in these 60 samples were B. hominis
(n = 17), Endolimax nana (n = 12), Entamoeba coli (n = 9),
Chilomastix mesnili (n = 3), Iodamoeba butschlii (n = 1), E. hartmanni
(n = 1), Ascaris lumbricoïdes (n = 1) and Trichuris trichiura (n = 1).
All 60 samples were tested by RIDA®QUICK Combi, 58 samples by
ImmunoCardSTAT!®CGE and 35 samples by Duo-Strip and Quik Chek.
Due to the delay in testing for ImmunoCardSTAT!®CGE, an insufficient
amount of stool was left for 2 samples (1 G. lamblia and 1 E. histolytica/
dispar). For one G. lamblia + E. histolytica/dispar mixed infection not
enough sample was available and only the RDTs detecting E. histolytica/
dispar were performed.
Sensitivity and specificity of microscopy, ELISA's and RDTs are sum-
marized in Table 3. Specificity was 100.0% for G. lamblia for all tests
except for Quik Chek and ProspecT ELISA which generated false positive
results in one and two samples, respectively. Sensitivities for G. lamblia
ranged from 65.5% to 100.0%. ODs in the ELISA can be used semi-
quantitatively to give an estimation on the parasite load in the speci-
men. Because of the variation in sensitivity for G. lamblia between the
four RDTs, ODs were plotted according to interpretation (Fig. 1). ODs
of samples that tested false negative by ImmunoCardSTAT!®CGE and
RIDA®QUICK Combi were generally lower when compared to true pos-
itives, whereas ODs of samples that tested false negative by Duo-Strip
were more overlapping with ODs of true positives. Overall, Kruskal–
Wallis analysis showed a significant difference between all groups
within each RDT (data not shown). One false positive result found
by Quik Chek was weakly positive by ELISA as well but was not con-
firmed by PCR. No correlation between band intensity and ODs
existed.
Sensitivity was 100.0% for Cryptosporidium for all tests, except for Duo-
strip which showed a sensitivity of 66.7%, and specificities varied be-
tween 95.2% and 100.0%. The single positive E. histolytica sample was
picked up by both tests and specificities for this parasite ranged from
65.2% to 87.0%. For RIDA®QUICK Combi and ImmunoCardSTAT!®CGE,
respectively, 9/25 and 12/24 samples tested false positive of which 8/9
(88.9%) and 11/12 (91.7%) were identified as E. dispar.

Table 2
Sensitivity and specificity of four commercial RDTs of frozen samples selected retrospectively on positivity for G. lamblia, Cryptosporidium and E. histolytica.

G. lamblia Cryptosporidium E. histolytica

RDT Sens. Spec. Sens. Spec. Sens. Spec.

ImmunoCard STAT!®CGE (Meridian) 83.3% 100.0% 100.0% 100.0% 100.0% 80.0%

(51.6–97.9%) (81.5–100%) (73.5–100%) (81.5–100%) (47.8–100%) (59.3–93.2%)
(10/12) (18/18) (12/12) (18/18) (5/5) (20/25)
Duo-Strip (Coris) 58.3% 100.0% 91.7% 100.0% NA NA
(27.7–84.8%) (81.5–100%) (61.5–99.8%) (81.5–100%)
(7/12) (18/18) (11/12) (18/18)
RIDA®QUICK 83.3% 100.0% 100.0% 100.0% 100.0% 88.0%
(R-BioPharm) (51.6–97.9%) (81.5–100%) (73.5–100%) (81.5–100%) (47.8–100%) (68.8–97.5%)
(10/12) (18/18) (12/12) (18/18) (5/5) (22/25)
Quik Chek (Techlab) 100.0% 100.0% 100.0% 100.0% NA NA
(73.5–100%) (81.5–100%) (73.5–100%) (81.5–100%)
(12/12) (18/18) (12/12) (18/18)

NA: Not applicable.

Between brackets are 95% confidence intervals and the number of strains defined as true positives divided by the sum of true positives and false negatives for sensitivity and the number of
strains defined as true negatives divided by the sum of true negatives and false positives for specificity.
 D. Van den Bossche et al. / Journal of Microbiological Methods 110 (2015) 78–84 81

Table 3
Sensitivity and specificity of four commercial RDTs of fresh samples selected prospectively on positivity for G. lamblia, Cryptosporidium and E. histolytica by either microscopy and/or ELISA.

G. lamblia Cryptosporidium E. histolytica

RDT Sens. Spec. Sens. Spec. Sens. Spec.

ImmunoCard STAT!®CGE (Meridian) 79.3% 100.0% 100.0% 95.4% 100.0% 80.6%

(60.3–92.0%) (91.0–100%) (29.2–100%) (87.1–99.0%) (2.5–100%) (74.3–92.6%)
(23/29) (39/39) (3/3) (62/65) (1/1) (54/67)
Duo-Strip (Coris) 65.5% 100.0% 66.7% 95.2% NA NA
(45.7–82.1%) (79.4–100%) (9.4–99.2%) (83.8–99.4%)
(19/29) (16/16) (2/3) (40/42)
RIDA®QUICK (R-BioPharm) 83.3% 100.0% 100.0% 98.5% 100.0% 87.0%
(65.3–94.4%) (91.2–100%) (29.2–100%) (92.0–100%) (2.5–100%) (76.7–93.9%)
(25/30) (40/40) (3/3) (66/67) (1/1) (60/69)
Quik Chek (Techlab) 100.0% 93.8% 100.0% 100.0% NA NA
(88.1–100%) (69.8–99.8%) (29.2–100%) (91.6–100%)
(29/29) (15/16) (3/3) (42/42)
Microscopy 90.0% 100.0% 100.0% 100.0% ⁎ 65.2%
(73.5–97.9%) (91.2–100%) (29.2–100%) (94.6–100%) (52.8–76.3%)
(27/30) (40/40) (3/3) (67/67) (45/69)
ProspecT ELISA 100.0% 95.0% 100.0% 100.0% 100.0% 68.1%
(88.4–100%) (83.1–99.4%) (29.2–100%) (94.6–100%) (2.5–100%) (55.8–78.8%)
(30/30) (38/40) (3/3) (67/67) (1/1) (47/69)

NA: Not applicable.

Between brackets are 95% confidence intervals and the number of strains defined as true positives divided by the sum of true positives and false negatives for sensitivity and the number of
strains defined as true negatives divided by the sum of true negatives and false positives for specificity.
⁎ Too little sample was available for enrichment. Direct microscopic examination was negative.

Fig. 1. Optical densities (ODs) of the G. lamblia ProspecT ELISA according to classification as true positives (TP), true negatives (TN) and false negatives (FN) made by the four different RDTs
when compared to microscopy, ELISA and PCR. One false positive for Quick Check (Techlab) showed an OD of 0.129.
82 D. Van den Bossche et al. / Journal of Microbiological Methods 110 (2015) 78–84

3.3. Added value of RDT testing in comparison to microscopy
To evaluate the potential role of RDTs in a routine diagnostic lab
compared to microscopy, data from the retrospective and prospective
studies were analyzed together and are illustrated in Fig. 2. The kappa
score for the detection of G. lamblia, which demonstrates the level of
agreement between RDT and microscopy, was 0.80 (95%CI 0.68–0.92),
0.57 (0.39–0.75), 0.83 (0.71–0.94) and 0.89 (0.79–0.99) for
ImmunoCardSTAT!®CGE, Duo-Strip, RIDA®QUICK Combi and Quik
Chek respectively. Kappa-scores for Cryptosporidium were 0.88
(0.74–1.01), 0.89 (0.74–1.04), 0.96 (0.87–1.04) and 1.00 (1.00–1.00).
ImmunoCardSTAT!®CGE and RIDA®QUICK Combi had kappa values of
0.62 (0.46–0.79) and 0.44 (0.26–0.62) when compared to microscopic
results.
3.4. Ease-of-use of RDTs
Line intensities were interpreted compared to the intensity of the
control line for each RDT. Small differences in intensities between inter-
preters were recorded for 12 (12%), 8 (11%), 16 (16%) and 10 (13%)
samples tested by ImmunoCardSTAT!®CGE, Duo-Strip, RIDA®QUICK
Combi and Quik Chek respectively. Disagreements in presence or
absence were registered for 3 and 4 E. histolytica/dispar samples
by ImmunoCardSTAT!®CGE (3 E. dispar) and RIDA®QUICK Combi (3
E. dispar and 1 PCR negative sample) respectively and for 1 G. lamblia
positive sample by both Duo-Strip and Quik Chek. One test performed
by ImmunoCardSTAT!®CGE was interpreted as invalid due to the ab-
sence of the control line by one technician, and another test had to be
repeated because sample flow was hindered.
The ease-of-use was scored by lab technicians according to the num-
ber of steps required and difficulty of test execution and interpretation.
Duo-Strip was very quick and easy to perform with preparation of a sus-
pension of fecal sample into dilution buffer followed by immediate test
application, whereas ImmunoCardSTAT!®CGE and RIDA®QUICK Combi
required extra time for the sedimentation of the fecal suspension. Test
execution was more difficult for Quik Chek with the addition of wash
buffer and substrate following sample application. Ease of test inter-
pretation was comparable for all RDTs, however the verification of
the correct color of the reactive lines rendered interpretation of
ImmunoCardSTAT!®CGE and RIDA®QUICK Combi rendered interpreta-
tion more difficult.
4. Discussion
Performance characteristics for RDTs are widely described in litera-
ture but they are difficult to compare because of the variability in
study methodology, population investigated and what is used as a gold-
en standard. Microscopy is often used for comparison but its sensitivity
is influenced by the number of stool samples examined and the
microscopist's experience. The reference standard in this study was a
combination of microscopy and ELISA, confirmed by PCR. A possible
drawback of the study is that all RDTs used had the same lot number
and lot dependent variation was not evaluated, though theoretically
this should be small.
Sensitivities for G. lamblia reported in literature are comparable to
results found in this study, ranging from 44% to 100%, with mainly strips
from Coris showing lower sensitivities compared to the other RDTs
(Sharp et al., 2001; Garcia et al., 2003; Johnston et al., 2003; Oster
et al., 2006; Weitzel et al., 2006; Abdel Hammeed et al., 2008; Strand
et al., 2008; Goni et al., 2012; Minak et al., 2012; Alexander et al.,
2013; Duffy et al., 2013; Hawash, 2014; Ignatius et al., 2014). Parasite
load is expected to be the most important factor influencing the sensi-
tivity of RDTs. Previous studies have shown a relationship between
the ELISA reaction intensity and Giardia cyst concentration (Addiss
et al., 1991; Vidal and Catapani, 2005). Using the ODs of the ProspecT
ELISA as a semi-quantitative measurement of parasite load in the

Fig. 2. Comparison of RDT and microscopy results for G. lamblia, Cryptosporidium and E. histolytica. Black box: number of samples only detected by RDT. Gray box: number of samples only
detected by microscopy. White box: number of samples detected by both RDT and microscopy. A. ImmunoCard STAT!®CGE (Meridian). B. Duo-Strip (Coris). C. RIDA®QUICK (R-BioPharm).
D. Quick Check (Techlab). (The two samples positive for E. histolytica only by RDT were not tested by microscopy due to insufficient amount of samples.).
 D. Van den Bossche et al. / Journal of Microbiological Methods 110 (2015) 78–84
sample, it can be expected that samples with low ODs are most easily
missed and classified as false negatives by RDTs. This was true for
ImmunoCardSTAT!®CGE and RIDA®QUICK Combi. Duo-Strip however
was unable to detect 4 G. lamblia positives with higher ODs. This could
indicate a problem with the antibody used rather than a limited sensi-
tivity due to lower parasite loads. False negative copro-antigen tests
have been associated with low parasite densities by others as well (Ali
and Hill, 2003; Garcia et al., 2003; Johnston et al., 2003). Overall, speci-
ficity for G. lamblia was very good reaching 100% for 3 out of 4 RDTs,
which is comparable to results reported in literature (Sharp et al.,
2001; Garcia et al., 2003; Johnston et al., 2003; Oster et al., 2006;
Weitzel et al., 2006; Abdel Hammeed et al., 2008; Strand et al., 2008;
Goni et al., 2012; Minak et al., 2012; Alexander et al., 2013; Duffy
et al., 2013; Hawash, 2014; Ignatius et al., 2014). In this study only
two samples were found to test false positive, one by Quik Chek and
ProspecT ELISA and one only by ProspecT ELISA. The first sample was
only interpreted as positive with the Quik Chek RDT by one out of two
technicians. ODs of both samples were just above the cut-off (0.129
and 0.198). No other parasites were found in these samples which
could explain the false positive results. Specificities described in litera-
ture for ProspecT Giardia ELISA range between 98% and 100%, which is
slightly higher than reported here (Aldeen et al., 1998; Hanson and
Cartwright, 2001).
Cryptosporidium sensitivities of RDTs in literature range from
47% to 100% (Sharp et al., 2001; Llorente et al., 2002; Garcia et al.,
2003; Johnston et al., 2003; Alexander et al., 2013; Weitzel et al.,
2006; Abdel Hammeed et al., 2008; Agnamey et al., 2011; El-Moamly
and El-Sweify, 2012; Goni et al., 2012; Minak et al., 2012; Hawash,
2014). In this study only the two major human-infecting species
of Cryptosporidium (C. parvum and C. hominis) were tested. Others
have illustrated that the sensitivity of Quik Chek (Alexander et al.,
2013), ImmunoCardSTAT!®CGE and RIDA®QUICK Combi is compara-
ble for both species, and that it is lower for the other Cryptosporidium
species (Agnamey et al., 2011). Agnamey et al. (2011) also found a
lower sensitivity for Cryptosporidium when tested by Duo-Strip com-
pared to the other RDTs however without taking into account the
infecting species. We were unable to make conclusions on the influ-
ence of low parasite load due to the small amount of positive samples
found, but Johnston et al. reported that Cryptosporidium detection by
ImmunoCardSTAT!®CGE was even more affected by parasite load than
G. lamblia. Specificities for Cryptosporidium detection in literature range
from 97% to 100% which concurs to our results, though in our study
slightly lower specificities were obtained for ImmunoCardSTAT!®CGE
and Duo-strip. Cross-reactivity with Cryptosporidium antibodies used
in the RDTs has been reported previously (Cronquist, 2004; Robinson
et al., 2010). It is currently unclear what causes the false positive reac-
tions. Katanik et al. (2001) suggested that the presence of blood might
be responsible, though further research is needed.
Literature on the sensitivity and specificity of RDTs for the detection
of E. histolytica is scarce. Two studies compared Triage Micro Parasite
Panel (Triage) to microscopy and ELISA without differentiation between
E. histolytica and E. dispar, reporting sensitivities ranging from 68%
to 83% and specificities of 100% (Pillai and Kain, 1999; Sharp et al.,
2001). The sensitivity and specificity of the first RDT prototype for
specific E. histolytica detection, supplied by Techlab, Inc. and available
since 2006 (Leo et al., 2006) were 97% and 100% compared to an
E. histolytica specific ELISA respectively. Only one study evaluated
RIDA®QUICK Combi and revealed a sensitivity and specificity of 62%
and 96% for E. histolytica respectively (Goni et al., 2012), which differs
from the results found here but can be explained by the high number
of E. dispar samples in our study which substantially influences specific-
ity. Results on the performance of ImmunoCardSTAT!®CGE have not
been reported before as this was a newly developed RDT upon evalua-
tion in this study. Results were comparable between RIDA®QUICK
Combi and ImmunoCardSTAT!®CGE with cross-reactivity with E. dispar
being the main problem. When taking results from both the retrospective
83
and prospective part together, 13 (41%) and 17 (53%) out of 32 E. dispar
positives caused a positive reaction for RIDA®QUICK Combi and
ImmunoCardSTAT!®CGE respectively. One sample yielded a false posi-
tive result by both RDTs the reason of which remains unknown. Albeit,
for RIDA®QUICK Combi one technician interpreted this sample as a true
negative.
The decision to implement a RDT in routine practice and the choice
of RDT brand depends on many variables such as the population tested,
the availability of microscopic expertise, ease-of-use and cost-
effectiveness. Not only the antibody used in the test itself, parasite
load and correct homogenization of the sample influences sensitivities
of RDTs but also the population investigated. Prevalence in our laborato-
ry in a study performed in 2005–2006 was 4.7%, 0.5% and 0.3% for
G. lamblia, Cryptosporidium and E. histolytica respectively (ten Hove
et al., 2009) and has not changed substantially (unpublished data).
Microscopy of a stool sample allows the identification of all present
several parasites instead of searching for one or a few specific patho-
gens, however it lacks specificity and is labor-intensive. In areas and
situations where the prevalence of Giardia, Cryptosporidium and
E. histolytica is high and/or resources are poor, a RDT instead of micros-
copy can be useful, provided that other enteric pathogens are scarce and
RDTs have adequate performance characteristics. RDTs may prove valu-
able as point-of-care tests in remote regions, however test performance
should be re-evaluated when they are performed in endemic settings by
local laboratory personnel. Conversely, a RDT as a rapid screening
method prior to microscopic examination can be used in settings
where different enteric pathogens are present but overall prevalence
is low. In the case of a negative result and/or ongoing symptoms follow-
ing adequate treatment a microscopic examination should be required
to exclude the presence of other pathogens. Reduction in microscopic
skills is a possible disadvantage of this algorithm and differentiation be-
tween E. histolytica/dispar by PCR remains necessary due to poor speci-
ficity of RDTs. Costs of RDT are considerably higher than microscopy, but
labor costs will be considerably less due to the rapid turn-around-time
and ease-of-use. Sample conservation and sample collection with
different fixatives should be taken into consideration upon the imple-
mentation of a RDT as these can differ between brands substantially.
Manufacturers may provide stronger rules on conservation than neces-
sary, as was shown from the good results from the retrospective part of
this study with samples still found positive when kept frozen up to five
years after collection and comparable sensitivities in the retrospective
and prospective study. In this study none of the samples analyzed was
fixed so no conclusions could be made on the role of fixation.
In conclusion, sensitivities of the evaluated RDTs are excellent for
Cryptosporidium and E. histolytica, but variable for G. lamblia. Specific-
ities for all RDTs were excellent for G. lamblia and Cryptosporidium, but
differentiation between E. histolytica/dispar by PCR remains necessary.
We have illustrated that RDTs can be a valuable tool to replace micros-
copy, especially when microscopic expertise is poor. RDTs provide a
rapid screening method in developed countries where microscopy
may be a complementary technique for the detection of these and
other parasites or in remote and outbreak settings where other tech-
niques are often not available and rapid diagnosis is required.
Conflict of interest
The authors declare that they have no conflict of interest.
Acknowledgments
The authors would like to thank Hilde Cox, Idzi Potters and Henk
Vereecken for their work in the lab and the companies Meridian Biosci-
ence Inc. Belgium, Coris Bioconcepts Belgium, ApDia nv Belgium and
Alere Health Belgium for providing the test kits.
84 D. Van den Bossche et al. / Journal of Microbiological Methods 110 (2015) 78–84
References
Abdel Hammeed, D.M., Elwakil, H.S., Ahmed, M.A., 2008. A single-step immuno-
chromatographic lateral-flow assay for detection of Giardia lamblia and Cryptosporidium
parvum antigens in human fecal samples. J. Egypt. Soc. Parasitol. 36, 797–804.
Addiss, D.G., Mathews, H.M., Stewart, J.M., Wahlquist, S.P., Williams, R.M., Finton, R.J.,
Spencer, H.C., Juranek, D.D., 1991. Evaluation of a commercially available enzyme-
linked immunosorbent assay for Giardia lamblia antigen in stool. J. Clin. Microbiol.
29, 1137–1142.
Agnamey, P., Sarfati, C., Pinel, C., Rabodoniriina, M., Kapel, N., Dutoit, E., Garnaud, C., Diouf,
M., Garin, J.F., Totet, A., Derouin, F., 2011. Evaluation of four commercial rapid
immunochromatographic assays for detection of Cryptosporidium antigens in stool
samples: a blind multicenter trial. J. Clin. Microbiol. 49, 1605–1607.
Aldeen, W.E., Carroll, K., Robison, A., Morrison, M., Hale, D., 1998. Comparison of nine
commercially available enzyme-linked immunosorbent assays for detection of
Giardia lamblia in fecal specimens. J. Clin. Microbiol. 36, 1338–1340.
Alexander, C.L., Niebel, M., Jones, B., 2013. The rapid detection of Cryptosporidium and
Giardia species in clinical stools using the Quik Chek immunoassay. Parasitol. Int.
62, 552–553.
Ali, S.A., Hill, D.R., 2003. Giardia intestinalis. Curr. Opin. Infect. Dis. 16, 453–460.
Cnops, L., Van Esbroeck, M., 2010. Freezing of stool samples improves real-time PCR detection
of Entamoeba dispar and Entamoeba histolytica. J. Microbiol. Meth. 80, 310–312.
Cronquist, A., 2004. Manufacturer's recall of rapid cartridge assay kits on the basis of
false-positive Cryptosporidium antigen tests, Colorado 2004. MMWR 53, 198-198.
Dillingham, R.A., Lima, A.A., Guerrant, R.L., 2002. Cryptosporidiosis: epidemiology and
impact. Microbes Infect. 4, 1059–1066.
Duffy, T.L., Montenegro-Bethancourt, G., Solomons, N.W., Belosevic, M., Clandinin, M.T.,
2013. Prevalence of Giardiasis in children attending semi-urban daycare centres in
Guatemala and comparison of 3 Giardia detection tests. J. Health Popul. Nutr. 31,
290–293.
El-Moamly, A.A., El-Sweify, M.A., 2012. ImmunoCard STAT! cartridge antigen detection
assay compared to microplate enzyme immunoassay and modified Kinyoun's acid-
fast staining technique for detection of Cryptosporidium in fecal samples. Parasitol.
Res. 110, 1037–1041.
Garcia, L.S., Shimizu, R.Y., Novak, S., Carroll, M., Chan, F., 2003. Commercial assay for
detection of Giardia lamblia and Cryptosporidium parvum antigens in human
fecal specimens by rapid solid-phase qualitative immunochromatography. J. Clin.
Microbiol. 41, 209–212.
Goni, P., Martin, B., Villacampa, M., Garcia, A., Seral, C., Castillo, F.J., Clavel, A., 2012. Eval-
uation of an immunochromatographic dip strip test for simultaneous detection of
Cryptosporidium spp., Giardia duodenalis and Entamoeba histolytica antigens in
human faecal samples. Eur. J. Clin. Microbiol. Infect. Dis. 31, 2077–2082.
Hadfield, S.J., Robinson, G., Elwin, K., Chalmers, R.M., 2011. Detection and differentiation
of Cryptosporidium spp. in human clinical samples by use of real-time PCR. J. Clin.
Microbiol. 49, 918–924.
Hanson, K.L., Cartwright, C.P., 2001. Use of an enzyme immunoassay does not eliminate
the need to analyze multiple stool specimens for sensitive detection of Giardia
lamblia. J. Clin. Microbiol. 39, 474–477.
Hawash, Y., 2014. Evaluation of an immunoassay-based algorithm for screening and iden-
tification of Giardia and Cryptosporidium antigens in human faecal specimens from
Saudi Arabia. J. Parasitol. Res. http://dx.doi.org/10.1155/2014/213745.
Heine, J., 1982. A simple technic for the demonstration of cryptosporidia in feces.
Zentralbl. Veterinarmed. B 29, 324–327.
Ignatius, R., Gahutu, J.B., Klotz, C., Musemakweri, A., Aebischer, T., Mockenhaupt, F.P., 2014.
Detection of Giardia duodenalis assemblage A and B isolates by immunochromatography
in stool samples from Rwandan children. Clin. Microbiol. Infect. http://dx.doi.org/10.
1111/1469-0691-12596.
Johnston, S.P., Ballard, M.M., Beach, M.J., Causer, L., Wilkins, P.P., 2003. Evaluation of three
commercial assays for detection of Giardia and Cryptosporidium organisms in fecal
specimens. J. Clin. Microbiol. 41, 623–626.
Katanik, M.T., Schneider, S.K., Rosenblatt, J.E., Hall, G.S., Procop, G.W., 2001. Evaluation of
ColorPAC Giardia/Cryptosporidium rapid assay and ProSpecT Giardia/Cryptosporidium
microplate assay for detection of Giardia and Cryptosporidium in fecal specimens.
J. Clin. Microbiol. 39, 4523–4525.
Leo, M., Hasue, R., Kabir, M., Roy, S., Lahlou, R.M., Mondal, D., Tannich, E., Petri, J.W.A.,
2006. Evaluation of Entamoeba histolytica antigen and antibody point-of-care tests
for the rapid diagnosis of amebiasis. J. Clin. Microbiol. 44, 4569–4571.
Llorente, M.T., Clavel, A., Varea, M., Olivera, S., Castillo, F.J., Sahaqun, J., Rubio, M.C., Gomez-
Lus, R., 2002. Evaluation of an immunochromatographic dip-strip test for the detec-
tion of Cryptosporidium oocysts in stool specimens. Eur. J. Clin. Microbiol. Infect. Dis.
21, 624–625.
Loughlin, E.H., Spitz, S.H., 1949. Diagnosis of helminthiasis. J. Am. Med. Assoc. 139,
997–1000.
Minak, J., Kabir, M., Mahmud, I., Liu, Y., Liu, L., Haque, R., Petri Jr., W.A., 2012. Evaluation of
rapid antigen point-of-care tests for detection of Giardia and Cryptosporidium species
in human fecal specimens. J. Clin. Microbiol. 50, 154–156.
Oster, N., Gehrig-Feistel, H., Jung, H., Kammer, J., McLean, J.E., Lanzer, M., 2006. Evaluation
of the immunochromatographic Coris Giardia-strip test for rapid diagnosis of Giardia
lamblia. Eur. J. Clin. Microbiol. Infect. Dis. 25, 112–115.
Pillai, D.R., Kain, K.C., 1999. Immunochromatographic strip-based detection of Entamoeba
histolytica—E. disp ar and Giardia lamblia coproantigen. J. Clin. Microbiol. 37,
3017–3019.
Potters, I., Van Esbroeck, M., 2010. Negative staining technique of Heine for the detec-
tion of Cryptosporidium spp.: a fast and simple screening technique. Open Parasit.
J. 4, 1–4.
Robinson, T.J., Cebelinski, E.A., Taylor, C., Smith, K.E., 2010. Evaluation of the positive
predictive value of rapid assays used by clinical laboratories in Minnesota for the
diagnosis of cryptosporidiosis. Clin. Infect. Dis. 50, 53–55.
Sharp, S.E., Suarez, C.A., Duran, Y., Poppiti, R.J., 2001. Evaluation of the Triage Micro Para-
site Panel for detection of Giardia lamblia, Entamoeba histolytica/Entamoeba dispar,
and Cryptosporidium parvum in patient stool specimens. J. Clin. Microbiol. 39,
332–334.
Strand, E.A., Robertson, L.J., Hanevik, K., Alvsvag, J.O., Morch, K., Langeland, N., 2008.
Sensitivity of Giardia antigen test in persistent giardiasis following an extensive
outbreak. Clin. Microbiol. Infect. 14, 1069–1071.
ten Hove, R.J., Van Esbroeck, M., Vervoort, T., Van den Ende, J., Van Lieshout, L., Verweij, J.J.,
2009. Molecular diagnostics of intestinal parasites in returning travellers. Eur. J. Clin.
Microbiol. Infect. Dis. 28, 1045–1053.
Verweij, J.J., Schinkel, J., Laeijendecker, D., van Rooyen, M.A.A., van Lieshout, L., Polderman,
A.M., 2003. Real-time PCR for the detection of Giardia lamblia. Mol. Cell. Probes 17,
223–225.
Vidal, A.M.B., Catapani, W.R., 2005. Enzyme-linked immunosorbent assay (ELISA)
immunoassaying versus microscopy: advantages and drawbacks for diagnosing
giardiasis. Sao Paulo Med. J. 123, 282–285.
Weitzel, T., Dittrich, S., Möhl, I., Adusu, E., Jelinek, T., 2006. Evaluation of seven commercial
antigen detection tests for Giardia and Cryptosporidium in stool samples. Clin.
Microbiol. Infect. 12, 656–659.