Abstract
BIOSITE Triage was 68.3% sensitive and 100% specific for the detection of Entamoeba histolytica-E. dispar (n = 71) compared to Alexon-Trend’s ProSpecT test (reference standard) using fresh-frozen stool. Neither test is able to distinguish E. histolytica from E. dispar. Triage was 83.3% sensitive and 100% specific compared to microscopy (formalin-ether concentrates and permanent stains) for the detection of Giardia lamblia.
Entamoeba histolytica, Giardia lamblia, and Cryptosporidium parvum are three of the major causes of protozoan-induced diarrheal disease (2, 3, 13). E. histolytica is responsible for approximately 100,000 deaths worldwide each year, making it second only to malaria as a cause of mortality due to a protozoan parasite (13). G. lamblia is among the most commonly reported parasitic infections in the United States, and on a global scale, giardiasis is responsible for approximately 100 million infections annually (7). Infection with the coccidian parasite C. parvum is usually self-limiting in immunocompetent individuals, but can be chronic and potentially life-threatening in the immunocompromised host (3). Contamination of municipal drinking water with C. parvum resulted in over 400,000 infections in Milwaukee during 1993 (8). Traditionally, the laboratory diagnosis of E. histolytica, G. lamblia, or C. parvum infection has relied upon the microscopic examination of fresh or fixed stool. However, microscopic diagnosis has several limitations, and recent studies have reported that stool antigen immunoassays equal (G. lamblia) or surpass (E. histolytica-E. dispar) microscopic detection of these pathogens (1, 6, 9–11).
Since multiple protozoal infections may coexist, a stool enzyme immunoassay capable of simultaneously detecting E. histolytica-E. dispar complex, G. lamblia, and C. parvum has recently been developed (Triage Parasite Panel; Biosite Diagnostics, San Diego, Calif.). Triage is a single immunochromatographic (IC) strip coated with monoclonal antibodies specific for 29-kDa surface antigen (E. histolytica-E. dispar), alpha-1-giardin (G. lamblia), and protein disulfide isomerase (C. parvum) (12). In a previous study which assessed all available stool antigen enzyme-linked immunosorbent assay (ELISA) kits (Techlab’s Entamoeba test and E. histolytica test, Merlin Diagnostika’s Optimun S, and Alexon-Trend’s ProSpecT), we showed that ProSpecT is the most sensitive and specific stool antigen ELISA currently available for the detection of the E. histolytica-E. dispar complex (11). In addition, ProSpecT outperformed microscopy (formalin-ether concentration and permanently stained smears) carried out in either community laboratories or referral centers (11). In this study, we evaluated the performance of Triage compared blindly to those of (i) ProSpecT for the detection of E. histolytica-E. dispar antigen in patient specimens as well as in mock reconstitution experiments and (ii) microscopy (formalin-ether concentration and permanently stained smears) performed at a referral center for the detection of G. lamblia in patient specimens.
Patients presenting between 1993 and 1998 to the Tropical Disease Unit of The Toronto Hospital with gastrointestinal symptoms (with diarrhea defined as ≥3 loose bowel movements/day which conform to the container, abdominal pain, nausea, weight loss, or bloody stool) or risk factors for E. histolytica-E. dispar, G. lamblia, or C. parvum infection (travel to the tropics within 6 months, men who have sex with men, or immigrants from the tropics or subtropics within the previous 2 years) were eligible for inclusion in this study. Verbal informed consent was obtained from each patient, and the study was approved by the Ethical Review Committee of The Toronto Hospital. Subjects were requested to provide fresh stool samples (within 1 h of passage) for microscopy and ELISA analysis. Stool specimens (minimum of two) were transported to the parasitology laboratory for routine evaluation of ova and parasite by microscopy (formalin-ether concentration and permanent stains [iron hematoxylin and modified acid fast]). Aliquots of fresh unpreserved stool were frozen (−20°C) for subsequent ELISA analysis.
Once thawed, samples were resuspended in the specimen dilution buffer (buffered protein solution–0.1% NaN3) provided in each kit. The assay procedure for both Triage and ProSpecT was carried out according to the manufacturer’s recommendations. A positive reaction in the Triage kit is identified by a qualitative colorimetric reaction when the antibody conjugate (alkaline phosphatase) reacts with the substrate (indoxyl phosphate), resulting in a dark blue-purple line on the IC strip. The ProSpecT ELISA was read at λ450 with a microplate reader (Thermomax; Molecular Devices Corp., Sunnyvale, Calif.). The ELISA plate format of the ProSpecT allows for multiple tests (up to 96 per ELISA plate), which reduces the assay time per sample. The Triage kit had the advantage of being completely self-contained and can be stored at room temperature. Triage also contains internal positive and negative controls for each test strip.
The results of stool antigen detection by Triage and ProSpecT are summarized in Table 1. Triage was 100% specific and 68.3% sensitive compared to ProSpecT (used as the reference standard) for the detection of the E. histolytica-E. dispar complex. ProSpecT was used as the comparative test in this study, since this assay had the highest specificity (98%) and sensitivity (100%) in a previous study (n = 112) that compared all available ELISA kits for E. histolytica-E. dispar detection with microscopy (11). Based on these performance characteristics, ProSpecT was chosen as the reference standard. The lower sensitivity of the Triage IC strip was explained in part by a fourfold difference in the limit of detection for E. histolytica-E. dispar trophozoites. HM1:IMSS trophozoites grown axenically in YI-S medium were serially diluted in specimen dilution buffer and subjected to both the Triage and ProSpecT assays (5). The results of these reconstitution experiments are indicated in Table 2. ProSpecT was able to detect E. histolytica-E. dispar antigen at 250 trophozoites per ml, whereas Triage required >1,000 trophozoites per ml for an unequivocal positive signal. Although Triage is a qualitative colorimetric test, we found that band intensity correlated well with trophozoite number independent of the strip reader (Table 2). At 500 (before filtration) or 1,000 (after filtration) trophozoites per ml, band intensity was positive but weak and therefore prone to subjective reader error. No association has been established between parasite burden (number of cysts or trophozoites shed in stool) and clinical severity. Molecular epidemiological data indicate that E. histolytica is associated with invasive disease, while the genetically distinct species E. dispar results in asymptomatic infection (4, 11).
TABLE 1.
Results of Triage versus Alexon-Trend ProSpecT for the detection of E. histolytica-E. dispar
| Alexon-Trend ProSpecT ELISA result | No. of samples with BIOSITE Triage IC strip result
|
||
|---|---|---|---|
| Positive | Negative | Total | |
| Positive | 28 | 13 | 41 |
| Negative | 0 | 30 | 30 |
| Total | 28 | 43 | 71 |
TABLE 2.
Limit of detection of E. histolytica-E. dispar by BIOSITE Triage IC strip and Alexon-Trend ProSpecT ELISA
| No. of HM1:IMSS trophozoites/ml | OD in Alexon-Trend ProSpecTa | BIOSITE Triage IC strip resultb
|
|
|---|---|---|---|
| Before filtration | After filtration | ||
| 105 | 1.481 (+) | +++ | ++ |
| 104 | 0.883 (+) | ++ | + |
| 103 | 0.264 (+) | + | +/− |
| 500 | 0.253 (+) | +/− | − |
| 250 | 0.169 (+) | − | − |
| 100 | 0.088 (−) | − | − |
| 10 | 0.066 (−) | − | − |
An optical density (OD) of >0.100 was regarded as positive (+) for ProSpecT in this assay run.
Filtration of stool specimens resuspended in specimen dilution buffer (protein-buffered solution containing 0.1% NaN3) was carried out with the filter devices provided in the Triage kit according to the manufacturer’s recommendations (Biosite Diagnostics). Band intensity with the Triage kit was measured visually by two readers and scored as follows: +++, very strong; ++, strong; +, medium; +/− weak; − none.
Comparison of Triage with microscopy for the detection of G. lamblia is summarized in Table 3. A minimum of two stools per patient were examined by microscopy (formalin-ether concentration and permanent stains). Aliquoted fresh-frozen specimens were subsequently tested by IC strip. Triage was 83.3% sensitive and 100% specific compared to microscopy performed blindly at our referral center (n = 71). Triage was able to detect two mixed infections containing both E. histolytica-E. dispar and G. lamblia. No C. parvum infections were detected in these samples by either microscopy or Triage IC strip.
TABLE 3.
Results of Triage versus standard ovum and parasite microscopy for the detection of G. lamblia
| Microscopy result | No. of samples with BIOSITE Triage IC strip result
|
||
|---|---|---|---|
| Positive | Negative | Total | |
| Positive | 5 | 1 | 6 |
| Negative | 0 | 65 | 65 |
| Total | 5 | 66 | 71 |
In summary, the Triage IC strip is highly specific for the detection of E. histolytica-E. dispar complex. However, Triage is less sensitive (68.3%) than an alternative ELISA diagnostic kit (ProSpecT). The lower sensitivity of Triage may be due to its inability to detect E. histolytica-E. dispar antigen at or below 1,000 trophozoites per ml, especially following the required filtration of the sample. Triage has the advantage of being able to detect multiple protozoal pathogens in a single test. Compared to reference microscopy, Triage was 83.3% sensitive and 100% specific for the detection of G. lamblia. These results indicate that Triage may be a useful alternative system for the detection of multiple pathogens in stool. However, both of these tests are limited by an inability to distinguish pathogenic E. histolytica from nonpathogenic E. dispar and by the requirement for fresh or fresh-frozen stool.
Acknowledgments
Funding was provided by Medical Research Council (MRC) of Canada grant MT12665 to K.C.K. K.C.K. is supported by a career scientist award from the Ontario Ministry of Health. D.R.P. is funded by an MD/Ph.D. Studentship from the MRC.
We thank L. Wamsley (Somagen Diagnostics), M. Bergh (Alexon-Trend), and P. Cordero (Alexon-Trend) for providing the diagnostic kits used in this study.
REFERENCES
- 1.Addiss D G, Mathews H M, Stewart J M, Wahlquist S P, Williams R M, Finton R J, Spenser H S, Jurnnek D D. Evaluation of a commercially available enzyme-linked immunosorbent assay for Giardia lamblia antigen in stool. J Clin Microbiol. 1991;29:1137–1142. doi: 10.1128/jcm.29.6.1137-1142.1991. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Black R E, Dykes A C, Sinclair S P, Wells J G. Giardiasis in day-care centers: evidence of person-to-person transmission. Pediatrics. 1977;60:486–491. [PubMed] [Google Scholar]
- 3.Chapman P A. Cryptosporidiosis: recent trends in epidemiology, diagnosis, and treatment. Serodiagn Immunother Infect Dis. 1988;2:311–317. [Google Scholar]
- 4.Diamond L S, Clark C G. A redescription of Entamoeba histolytica Schaudinn, 1903 (Emended Walker, 1911) separating it from Entamoeba dispar Brumpt, 1925. J Eukaryot Microbiol. 1993;40:340–344. doi: 10.1111/j.1550-7408.1993.tb04926.x. [DOI] [PubMed] [Google Scholar]
- 5.Diamond L S, Clark C G, Cunnick C C. YI-S, a casein-free medium for axenic cultivation of Entamoeba histolytica, related Entamoeba, Giardia intestinalis and Trichomonas vaginalis. J Eukaryot Microbiol. 1995;42:277–278. doi: 10.1111/j.1550-7408.1995.tb01579.x. [DOI] [PubMed] [Google Scholar]
- 6.Haque R, Ali I K M, Akther S, Petri W A., Jr Comparison of PCR, isoenzyme analysis, and antigen detection for diagnosis of Entamoeba histolytica infection. J Clin Microbiol. 1998;36:449–452. doi: 10.1128/jcm.36.2.449-452.1998. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kappus K, Juranek D. Giardia in the well. JAMA. 1988;259:1810. [Google Scholar]
- 8.MacKenzie W R, Hoxie N J, Proctor M E, et al. A massive outbreak in Milwaukee of cryptosporidium infection transmitted through the public water supply. N Engl J Med. 1994;331:161–167. doi: 10.1056/NEJM199407213310304. [DOI] [PubMed] [Google Scholar]
- 9.Mirelman D, Nuchamowitz Y, Stolarsky T. Comparison of use of enzyme-linked immunosorbent assay-based kits and PCR amplification of rRNA genes for simultaneous detection of Entamoeba histolytica and E. dispar. J Clin Microbiol. 1997;35:2405–2407. doi: 10.1128/jcm.35.9.2405-2407.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Newman R D, Jaeger K L, Wuhib T, Lima A A M, Guerrant R L, Sears C L. Evaluation of an antigen capture enzyme-linked immunosorbent assay for detection of Cryptosporidium oocytes. J Clin Microbiol. 1993;31:2080–2084. doi: 10.1128/jcm.31.8.2080-2084.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Pillai, D. R., J. S. Keystone, D. C. Sheppard, J. D. MacLean, D. W. MacPherson, and K. C. Kain.Entamoeba histolytica and Entamoeba dispar: epidemiology, and comparison of diagnostic methods in a non-endemic setting. Clin. Infect. Dis., in press. [DOI] [PubMed]
- 12.Reed S L, Flores B M, Batzer M A, Stein M A, Stroeher V L, Carlton J E, Diedrich D L, Torian B E. Molecular and cellular characterization of the 29-kilodalton peripheral membrane protein of Entamoeba histolytica: differentiation between pathogenic and nonpathogenic isolates. Infect Immun. 1992;60:542–549. doi: 10.1128/iai.60.2.542-549.1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Walsh J A. Problems in recognition and diagnosis of amebiasis: estimation of the global magnitude of morbidity and mortality. Rev Infect Dis. 1986;8:228–238. doi: 10.1093/clinids/8.2.228. [DOI] [PubMed] [Google Scholar]