Laboratory diagnosis of amebiasis in a sample of students from southeastern Brazil and a comparison of microscopy with enzyme-linked immunosorbent assay for screening of infections with Entamoeba sp .

introduction: Epidemiological studies on amebiasis have been reassessed since Entamoeba histolytica and E. dispar were fi rst recognized as distinct species. Because the morphological similarity of these species renders microscopic diagnosis unreliable, additional tools are required to discriminate between Entamoeba species. The objectives of our study were to compare microscopy with ELISA kit (IVD®) results, to diagnose E. histolytica infection, and to determine the prevalence of amebiasis in a sample of students from southeastern Brazil. Methods: In this study, diagnosis was based on microscopy due to its capacity for revealing potential cysts/trophozoites and on two commercial kits for antigen detection in stool samples. Results: For 1,403 samples collected from students aged 6 to 14 years who were living in Divinópolis, Minas Gerais, Brazil, microscopy underestimated the number of individuals infected with E. histolytica/E. dispar (5.7% prevalence) compared with the ELISA kit (IVD®)-based diagnoses (15.7% for E. histolytica/E. dispar). A comparison of the ELISA (IVD®) and light microscopy results returned a 20% sensitivity, 97% specifi city, low positive predictive value, and high negative predictive value for microscopy. An ELISA kit (TechLab®) that was specifi c for E. histolytica detected a 3.1% (43/1403) prevalence for E. histolytica infection. Conclusions: The ELISA kit (IVD®) can be used as an alternative screening tool. The high prevalence of E. histolytica infection detected in this study warrants the implementation of actions directed toward health promotion and preventive measures.

Amebiasis is a human infection caused by Entamoeba histolytica, a protozoan of cosmopolitan distribution, with or without clinical manifestations 1 .Infection by the species Entamoeba dispar is approximately 10 times more common than infection by E. histolytica 2 .Given the morphological similarity of these species, diagnosis based on light microscopy can yield either under-or overestimation of infection rates, leading to unnecessary treatment 3 .The sensitivity of microscopy ranges from 5% to 60%, and its specifi city ranges from 10% to 50% 4 .
Due to the invasive behavior of E. histolytica and the noninvasive nature of E. dispar, coupled with the inability of microscopy to distinguish between the species, the World Health Organization (WHO) recommends that diagnoses attained by microscopy be recorded as "E.histolytica/E.dispar" 1 .
In 1997, the WHO also advocated procedures capable of ensuring differentiation between these species so that treatment is restricted to confirmed cases of E. histolytica infection.Biochemical, immunological, and molecular biology methods are now capable of differentiating between Entamoeba species 5 .Among these methods, tests for antigen detection in stool samples are advantageous in terms of speed, accuracy, and reliability 3,5 .
The objectives of our study were to compare the parasitological examination of stools with ELISA kit (IVD®) results as a screening test for the diagnosis of infections by Entamoeba sp., to diagnose E. histolytica using an enzyme immunoassay for the detection of a specifi c antigen, and to determine the prevalence of amebiasis in a sample of students from southeastern Brazil.

MeTHODs
This cross-sectional epidemiological study with a stratifi edsampling design included a total of 1,403 male and female students aged 6 to 14 years who attended 15 public schools in Divinópolis county, State of Minas Gerais, Brazil.The subjects lived in urban neighborhoods and rural communities, thus representing all of the county's 11 geographical areas.In the   Students whose parents or guardians agreed to fi ll in a questionnaire and to sign a consent form were given a collection cup with no preservatives.The samples (one per student) were transported on ice to the Universidade Federal de São João del-Rei (UFSJ) Laboratory of Immunology and Parasitology, prepared on the day of collection, and processed using the Hoffmann-Pons-Janer (HPJ, or Lutz) method 7 .To increase the likelihood of parasite detection, four qualifi ed professional examined each sample (100% of fi elds read).An aliquot of each sample was stored at -20°C for later coproantigen testing using an E. histolytica/E.dispar ELISA (Enzyme-linked immunosorbent assay) kit (IVD® Research, Carlsbad, CA, USA) for the in vitro detection (but not discrimination) of E. histolytica and E. dispar.According to the manufacturer's instructions, the immunoassay is based on the interaction of monoclonal antibodies conjugate with peroxidase that bind to antigens of E. dispar and E. histolytica, and the reaction is revealed by the addition of a substrate containing tetramethylbenzidine and peroxide.The kit has a sensitivity of 88% and a specifi city of 100% 8 .In comparison, the E. histolytica II kit (TechLab®, Blacksburg, VA, USA) is an immunoassay based on the interaction of monoclonal antibodies with the single antigenic determinant adhesin present at the galactose affi nity E. histolytica.The kit has a sensitivity of 96.9% and a specifi city of 100% 9 .All tests were run and interpreted according to the manufacturer's instructions.
The data were encoded and processed using Statistical Package for the Social Sciences (SPSS) software, version 19.0,American University in Cairo -Department of University Academic Computing Technologies (UACT).The Chi-squared test was used to compare proportions, and the adopted signifi cance level was 5% (p-value < 0.05).To compare the microscopy test and the ELISA kit (IVD®), the sensitivity and specifi city, positive predictive value (PPV), and negative predictive value (NPV) were computed, assuming that the ELISA kit (IVD®) can adequately serve as the gold standard, using a dichotomous approach 10 .

Ethical considerations
The investigation was approved by the research ethics committee (opinion 56/2009) and was performed from February 2010 to October 2011.
Of the 1,403 samples, 52% (728/1403) were from females, and 48% (675/1403) were from males.The ages and genders of the subjects were evenly distributed.A signifi cant association (p-value = 0.01) was observed for E. histolytica with females but with not males.
E. histolytica infection was detected in all age groups, with the highest number of cases in individuals aged > 9 and ≤ 12 years (Figure 1).When the study population was segmented by age range, no signifi cant association was observed for the age groups.
Divinópolis county is composed of 11 so-called planning regions, 2 of which are rural and 9 of which are urban.Data on the prevalence of E. histolytica have emerged from studies on helminths and protozoans that were based on microscopic investigation of cysts and/or trophozoites in clinical specimens.Until relatively recently, E. histolytica and E. dispar were not differentiated, and infection with either of the two species was referred to as amebiasis, resulting in an overestimation of the true prevalence 11 .
However, the procedure for distinction has been reevaluated since Clark and Diamond 12 demonstrated that E. histolytica and E. dispar are distinct species that, despite morphological similarities, differ in pathogenicity.The genus Entamoeba contains many morphologically similar species, several of which, including E. histolytica, E. dispar, E. moshkovskii, E. polecki, and E. hartmanni, can be found in human stools 13,14 .
Therefore, the epidemiology of amebiasis is confusing, mainly because of the recently appreciated distinctions between E. histolytica, E. dispar, and E. moshkovskii 11 .Because light microscopy does not effi ciently identify E. histolytica, other tools have been developed to differentiate between Entamoeba species, including polymerase chain reaction (PCR), isoenzyme analysis, and ELISA, which are well suited for estimations of true prevalence and the treatment of patients 5 .Malatyali et al. 15 advocated the use of sensitive and effective tests, such as ELISA, for antigen detection in stools.
Several epidemiological studies have been conducted to estimate the incidence and prevalence of amebiasis by testing commercially available antigens from various manufacturers 11 .
Our study distinguished E. histolytica from other amebae and strengthened existing epidemiological data.Moreover, the study evaluated the ELISA kit (IVD®) as a screening test.
In the present investigation, microscopy underestimated the number of subjects infected with E. histolytica/E.dispar (5.7%), in contrast to the E. histolytica/E.dispar ELISA kit (IVD®) (15.7%).Whereas the ELISA (IVD®) was positive for 220 of the patients who had E. histolytica/E.dispar cysts in their stools, only 45 samples were positive in both tests.Additionally, 175 samples with negative results by direct microscopy were positive in the ELISA (IVD®) antigen detection test.This difference may be attributed to the quantity of the pathogen in the samples.Stools with a low number of cysts may be negative by direct microscopic examination but may yield positive results using ELISA 3 .
A comparison of the samples by light microscopy and ELISA (IVD®) revealed a low sensitivity (20%) and a high specifi city (97%) for light microscopy.The high NPV of 87% reduced the likelihood of false-negative results, yet the low PPV of 56% rendered the test unreliable.However, positivity on microscopy does not rule out the possibility that 44% of the samples are negative.Delialioglu et al. 3 reported that microscopy provided 53.8% sensitivity and 94% specifi city, with 78% PPV and 17% NPV, relative to an ELISA kit (Ridascreen Entamoeba, R-Biopharm AG, Darmstadt, Germany).In another study, compared with an ELISA triage kit (ProSpecT EIA, Alexon Inc., Sunnyvale, CA, USA), microscopy was more specifi c (92.1%) but less sensitive (68.4%) 16 .The ELISA kit (Alexon-Trend, Inc., Sunnyvale, CA, USA) had a sensitivity of 54.5% and a specifi city of 94%.If matched with culture and microscopy, the sensitivity of direct microscopic examination was 66%, and the specifi city was 83.7%.However, the results should be confi rmed with a larger number of fecal samples 17 .
Considering these data, the low sensitivity of microscopy may have been infl uenced by the collection of a single stool sample per student.According to Ravdin 18 , the examination of three separate stool specimens is required to attain 90% sensitivity, and a single examination identifi es only 40% to 60% of infections.If feces were collected more than once and were fi xed in preservatives, a higher prevalence of E. histolytica/E.dispar would be expected 11 .These data suggest that the ELISA (IVD®) can be used as a screen for the immediate testing of stools.The performance of antigen detection assays suggests that they may be considered as reference standards for the detection of E. histolytica and E. dispar 5,19 .
However, microscopy should still be considered as a screening method for the detection of Entamoeba found in human stools, despite the fact that this technique cannot differentiate between E. histolytica, E. moshkovskii, and E. dispar, although E. polecki, E. coli, and E. hartmanni can be differentiated morphologically from E. histolytica 20 .
One of the problems with screening kits is that they cannot differentiate between the amebae.However, the ELISA kit (TechLab®) is commercially available for the specifi c, direct detection of an E. histolytica antigen in stool specimens 5,19 .In the 1,403 samples subjected to ELISA (TechLab®), the prevalence of E. histolytica was 3.1%.Haque et al. 21, based on isoenzyme analysis of 202 samples from symptomatic individuals seen at the International Center for Diarrheal Disease Research in Dhaka, Bangladesh, obtained 52 culture-positive results using the E. histolytica II ELISA kit (TechLab®), a method that ensured faster diagnosis than when using isoenzyme analysis and achieved a higher sensitivity and specificity than did microscopy.The E. histolytica II ELISA (TechLab®) correctly identifi ed 21 of 22 cases of E. histolytica infection and 28 of 30 for E. dispar cases.Haque et al. 19 , examining 2000 samples using two TechLab® kits (the E. histolytica II ELISA kit and an Entamoeba ELISA kit), reported prevalence rates of 4.2% for E. histolytica and 6.5% for E. dispar in children aged 1 to 14 years who were living in the vicinity of Dhaka and presenting with diarrhea.In contrast, in asymptomatic children, the percentages were 1% for E. histolytica and 7% for E. dispar.Following the same strategy, Nesbitt et al. 22 examined 842 samples from Kilimanjaro, Tanzania, and detected prevalence values of 1% for E. histolytica and 7.3% for E. dispar.The high prevalence of E. histolytica infection detected in the present study warrants the implementation of actions directed toward health promotion and preventive measures.
The present results also corroborate previous fi ndings 23 that indicated that females are more prone than males to E. histolytica infection.In Brazil, amebiasis rates are highest in the northern region of the country, where both intestinal and extraintestinal forms of the disease exist, with serious public health implications 24 .In Belém, the capital city of the northern state of Pará, a prevalence rate of 29.35% has been reported using the E. histolytica II ELISA (TechLab®) 25 .The highest reported E. histolytica prevalence rates in Brazil that were detected using the E. histolytica II kit (TechLab®) were 36.6%(30/82) for stool samples from the state of Rondônia in Ariquemes and 19.4% (26/134) for stools taken from residents of Monte Negro 26 .
In Pernambuco state, in northeastern Brazil, Dourado et al. 27 detected only E. dispar, whereas in Macaparana county, within the same state, all samples investigated by Pinheiro et al. 28 tested negative in an E. histolytica-specifi c ELISA (TechLab®) and positive for E. dispar using a molecular biology method.
In southeastern Brazil, using light microscopy, Santos et al. 29 detected a 21% prevalence of the E. histolytica/E.dispar complex in urban and rural areas of Rio de Janeiro State, yet only two samples tested positive for E. histolytica by PCR and E. histolytica II ELISA (TechLab®).However, in São Leopoldo, within the southern State of Rio Grande do Sul, Tomé and Tavares 30 found no cases of E. histolytica infection using the E. histolytica II ELISA (TechLab®).
Light microscopy has several limitations when applied to the diagnosis of amebiasis, given factors such as examiner experience and similarities between Entamoeba cysts 3 , which can increase the likelihood of false-positive results 4 .Despite the low cost of light microscopy compared with culture, isoenzyme analysis, antigen detection, and PCR, the method's dependence on subjective diagnosis limits its reliability 3 .Therefore, microscopy is not appropriate for either rapid disease diagnosis or prevalence studies.
According Ngui et al. 31 , molecular techniques are indeed promising tools for epidemiological studies, particularly in discriminating the pathogenic from the non-pathogenic species of Entamoeba.E. moshkovskii, another morphologically indistinguishable human parasitic Entamoeba, has not been mentioned, nor has it been considered a contributor to prevalence fi gures in endemic areas 11 .Molecular techniques that can differentiate all studied species of Entamoeba, including E. moshkovskii, in human specimens have already been reported in Italy, Bangladesh, India, Australia, Turkey, Iran, and Malaysia [31][32][33][34][35][36] .
It is necessary to use new techniques to differentiate Entamoeba diagnoses 37 and to establish a readily available and cost-effective test for the specifi c diagnosis of amebiasis caused by E. histolytica in public laboratories 26 .
Diagnostic methods that are more sensitive and specifi c than light microscopy are required to establish the true distributions of E. histolytica and to reduce the rates of unnecessary treatment, thereby discouraging the development of drug resistance, precluding the risks of side effects, and reducing the costs of hospitalization.The present fi ndings demonstrate that the ELISA kit (IVD®) can be used as an alternative screening tool.In addition, this assay could be utilized by personnel who do not have extensive training in manual parasitological methods.The determination of the true prevalence of E. histolytica infection among students from southeastern Brazil is very crucial, as this information will lead to a better understanding of the public health problem and will help outline measures for controlling amebiasis.

TABLE 1 -
Comparison of samples by light microscopy and ELISA (E.histolytica/E.dispar ).