Assessing the marginal error in diagnosis and cure of Schistosoma mansoni in areas of low endemicity using Percoll and PCR techniques

Authors


Corresponding Author A. F. Allam, Medical Research Institute, Alexandria University, 165 El-Horeya Avenue, Alexandria, Egypt. E-mail: amalalam2005@yahoo.com

Summary

Objective  The Kato–Katz technique misses a proportion of low-intensity schistosomiasis cases. We set out to estimate the rate of false negative results after Kato–Katz.

Method  Comparison with the more sensitive Percoll technique and polymerase chain reaction (PCR).

Results  Amongst Kato–Katz negative Schistosoma mansoni specimens in a hypoendemic area, Percoll detected 11% positive cases and PCR detected 23%. After treatment, the cure rate was 97% by Kato–Katz, 90% by Percoll and 71% by PCR. Results of Kato–Katz and Percoll showed moderate agreement, all cases positive by Kato–Katz were positive by Percoll. However, PCR showed some negative results amongst cases considered sure positives.

Conclusion  The Percoll technique seems to be the technique of choice for diagnosis of S. mansoni in low intensity areas and after treatment.

Abstract

Objectif:  La technique Kato-Katz rate une proportion de cas de schistosomiase à faible intensité. Nous avons entrepris d’estimer le taux de résultats faux négatifs avec Kato-Katz.

Méthode:  Comparaison avec la technique Percoll plus sensible et la polymérisation en chaîne (PCR).

Résultats:  Parmi spécimens de S. mansoni négatifs par Kato-Katz dans une région hypoendémique, le Percoll a détecté 11% de cas positifs et la PCR 23%. Après traitement, les taux de guérison étaient de 97% selon Kato-Katz, 90% selon le Percoll et 71% selon la PCR. Les résultats de Kato-Katz et Percoll montraient une concordance modérée, tous les cas positifs par la technique de Kato-Katz étaient positifs par le Percoll. Cependant, la PCR a révélé des résultats négatifs parmi les cas trouvés positifs.

Conclusion:  La technique Percoll semble être la technique de choix pour le diagnostic de S. mansoni dans les zones de faible intensité et après le traitement.

Abstract

Objetivo:  Con la técnica de Kato-Katz se pierde una proporción de casos de esquistosomiasis de baja intensidad. Hemos estimado la tasa de falsos negativos después de Kato-Katz.

Método:  Comparación con técnicas más sensibles: la del Percoll y la reacción en cadena de la polimerasa (PCR).

Resultados:  En un área híper endémica, entre los especímenes S. mansoni negativos por Kato-Katz, el Percoll detectó un 11% de los casos positivos y la PCR un 23%. Después de recibir tratamiento, la tasa de curación era del 97% por Kato-Katz, 90% por Percoll y 71% por PCR. Los resultados de Kato-Katz y Percoll muestran una concordancia moderada; todos los casos positivos por Kato-Katz fueron positivos por Percoll. Sin embargo, la PCR mostró algunos resultados negativos entre casos considerados como positivos seguros.

Conclusión:  La técnica del Percoll parece ser la técnica de preferencia para el diagnóstico de S. mansoni en áreas de baja intensidad y después del tratamiento.

Introduction

Throughout its recorded history Egypt has been plagued by schistosomiasis caused by snail-transmitted parasites of the genus Schistosoma, which inhabit the human vasculature. But in recent years the epidemiological features of Schistosoma mansoni regarding prevalence, intensity and morbidity have profoundly changed thanks to sustained control efforts that have rendered some highly endemic areas low endemic (WHO 1994).

Notwithstanding the large number of tests for diagnosing schistosomiasis, few have withstood tests of effectiveness, reproducibility, cross reactivity and predictive values (Rabello et al. 2002). Consistent diagnosis of schistosomiasis still depends on coprological demonstration of Schistosoma eggs in faecal samples. The Kato–Katz technique is currently the method of choice for parasitological diagnosis of S. mansoni (Katz et al. 1972), despite the fact that its sensitivity is less appropriate for low endemic areas, post-treatment, and for determination of incidence (De Vlas & Gryseels 1992; Engels et al. 1997). If control is the objective, more sensitive methods must be used (Gryseels et al. 1991). Eberl et al. (2002) established and validated Percoll, a novel separation technique based upon the greater density of viable Schistosoma eggs relative to faecal material. Comparative analysis demonstrated that this method always gives better and more consistent estimates than conventional Kato–Katz, and efficiently detects low intensity infections (<100 eggs/g) (Doenhoff et al. 1993).

Antibody detection methods have been evaluated as adjuncts to faecal examinations (Doenhoff et al. 2004). However, the existence of cross-reactivity and slow reduction of specific antibody levels after treatment constituted great disadvantages of the immunological methods (Pontes et al. 2002). One possible solution to this problem could be the search for circulating antigens rather than antibodies (Hassan et al. 1994). Circulating cathodic antigen (CCA) and circulating anodic antigen (CAA) in serum and urine were reported as specific and sensitive markers of presence and intensity of S. mansoni infection. They were considered field applicable screening tests (Polman et al. 1998).

Polymerase chain reaction (PCR) was described for the first time by Pontes et al. (2002) for detecting S. mansoni DNA in human faecal samples. It was acclaimed as an outstandingly sensitive and specific diagnostic tool. PCR can detect as few as 2.4 eggs per gram of faeces, which makes it ten times more sensitive than the Kato–Katz examination. Its specificity was demonstrated by the absence of amplification when DNA samples from other helminths commonly found in the same endemic areas as S. mansoni were used as templates. Thus this DNA amplification assay may constitute an unprecedented alternative diagnostic technique in low endemic areas and in light infection intensity (Hamburguer et al. 1991).

We used these more sensitive techniques (Percoll density centrifugation and PCR) to assess the true prevalence and intensity of S. mansoni before and after treatment in areas of low endemicity, and estimated the marginal error after Kato–Katz examination.

Material and methods

The study was conducted in the Abis area near Alexandria: villages Abis 4 and Abis 8, which had low S. mansoni prevalence and intensity.

Study design

A total of 995 school children aged 9–12 years participated in the study. After informed consent of the school authority and the children’s parents, we visited schools on two consecutive days. On the first day, properly labelled plastic cups for stool collection were distributed; on the next day, the containers were collected and returned to the laboratory, where the samples were examined by Kato–Katz technique (41.7 mg/slide, three slides for each sample).

After the survey, stool samples were collected again from 100 randomly chosen students who had tested negative by Kato–Katz. These samples were processed by Percoll, part of the samples were stored at −70 °C for the PCR study.

A total of 156 cases positive for S. mansoni (56 cases were obtained from the present survey and 100 cases from a survey undertaken in nearby areas) received praziquantel as a single oral dose of 40 mg/kg of body weight administered on an empty stomach. One month later, 105 of these cases submitted stool samples, all of which were examined by Kato–Katz and Percoll; 70 were analysed by PCR.

Percoll technique

A sample of 250 mg fresh faeces was suspended in 3 ml phosphate buffer saline (PBS) and layered on top of 3 ml of 0.9% Nacl/60% Percoll (Amersham Pharmacia, Sigma, Germany) solution. After centrifugation for 15 min at 330 g, the supernatant (which contains contaminating debris) was discarded and the pellet was resuspended in 1 ml PBS. Larger particles were excluded by passing the suspension through a 150 μm mesh sieve into a collecting tube. The eggs in the flow-through were pelleted at 30 g for 1 min. The pellet was resuspended and transferred into a watch glass and stained with a drop of saturated malachite green. Finally the sediment was distributed into microscope slides and examined by the bright field microscopy (100x) (Dalton et al. 1997; Eberl et al. 2002). To establish and perfect the Percoll technique, 50 cases diagnosed positive after Kato–Katz were studied by Percoll and positivity and intensity of infection compared.

PCR

For the PCR assay, S. mansoni egg DNA was extracted from stool by QIAamp® DNA stool mini Kit (QIAGEN, GmbH, Hilden, Germany). As recommended by the manufacturer, 2 ml of stool lysis buffer were added to 220 mg of stool, followed by vortexing till homogenisation. The suspension was heated at 70 °C for 5 min followed by centrifugation to pellet the stool particles. A tablet of inhibitex X was added to the supernatant to allow inhibitors to adsorb to the inhibitex matrix, followed by centrifugation to pellet the inhibitors and stool particles. The supernatant was treated with proteinase K, lysis buffer and incubated at 70 °C for 10 min. Ethanol was added, followed by transfer to the QIAamp (GmbH) spin column and washing with 500 μl of buffer AW1 and AW2, then elution with elation buffer.

The eluate was used for PCR amplification. Primers were designed to amplify the 121 bp tandem repeat DNA sequence of S. mansoni as described by Hamburguer et al. (1991). Faecal samples obtained from uninfected school children were included in each run as negative control to detect any cross contamination. The amplification was conducted in a 25 μl reaction mixture using 2 X master mixture (Fermentas). The reaction mixtures were heated at 94 °C for 5 min, then subjected to 35 cycles of 95 °C for denaturation for 30 s, at 55 °C for 30 s for annealing and at 72 °C for 60 s for extension followed by the final extension for 10 min at 72 °C in a DNA thermal cycler (9600 Perkin Elmer, GmbH). Amplification products were analysed by agarose gel electrophoresis in Tris acetate EDTA buffer pH 8.0 and stained with ethidium bromide. The size marker used was a 50 bp ladder (Hamburguer et al. 1991; Pontes et al. 2002).

Sample size

To assess agreement between Kato and Percoll, the sample size was determined by using McNemar test with the following input data: level of significance equal to 0.05, β error = 0.2. The resulting sample size was 154. Accordingly, 100 negative cases before treatment and 105 after treatment, (i.e. 205 children) were included in this study.

Statistical analysis

spss-11 was used to edit and analyse the data, the t-test to compare egg counts between independent groups; chi-square to test for association. Kappa index (K) was used for measuring the agreement between the results of any two diagnostic tests applied to the same stool samples. (K < 0.2 poor agreement, K 0.2–0.4 fair agreement, K 0.41–0.6 moderate agreement, K 0.61–0.8 good agreement, K 0.81–1.0 very good agreement) (Altman 1992).

Results

Survey of schoolchildren

The prevalence of S. mansoni in Abis 4 (n = 429) was 4.9% and in Abis 8 (n = 566) it was 6.18%. The overall prevalence in the two villages (995) was 5.72%. Children aged 11 years showed the highest infection rate (15.2%). Males had a significantly higher infection rate than females (8.05%vs. 3.6%, P < 0.01). 84% had low intensity (≤96 eggs/g), 10.7% had moderate intensity (120–480 eggs/g) and 5.4% had high intensity (>480 egg/g) infections.

Quantitatively the egg counts differed by Kato–Katz and Percoll. At very low-intensity Percoll detected significantly more eggs than Kato–Katz, whilst Kato–Katz method gave significantly higher egg counts at higher intensities of infection (Table 1).

Table 1.   Comparison of egg counts as diagnosed by Kato and Percoll in 50 positive cases
Intensity (eggs/g)Kato mean ± SDPercoll mean ± SD T P
  1. *Number of cases.

  2. †Statistically significant.

8–24 (20)*9 ± 312 ± 71.760.05†
25–96 (15)*51 ± 3359 ± 500.520.46
> 96 (15)*123 ± 10109 ± 182.630.012†

Amongst 100 samples negative by Kato–Katz, Percoll detected 11 cases (Table 2). Nine of these (80%) had an intensity of four eggs/g stool (i.e. one egg was detected in 250 mg after Percoll). The stool of 77 of these negative cases was studied by PCR; 18 (23.4%) tested positive.

Table 2.   Results of Percoll and PCR techniques in diagnosis of Schistosoma mansoni amongst negative cases after Kato–Katz method
No. examined S. mansoni Positive no.%
Percoll1001111
PCR771823.37

Comparing the results of the 77 cases analysed by PCR with their results by Percoll revealed that only two patients gave concording positive results; seven cases positive by Percoll were negative by PCR (Table 3). Statistically no agreement was detected (K = 0.009, P > 0.05).

Table 3.   Agreement of the Percoll and PCR for the detection of Schistosoma mansoni cases negative after Kato–Katz technique
 PercollTotal
PositiveNegative
  1. Kappa index = 0.009, P > 0.05 no agreement.

PCR
 Positive21618
 Negative75259
Total96877

Table 4 and Figure 1 show that Kato–Katz, Percoll and PCR gave cure rates of 97%, 90.4% and 71.4% respectively. All cases found positive by Kato–Katz were positive by Percoll. However, four cases found positive by both Kato–Katz and Percoll gave negative results by PCR.

Table 4.   Observed cure rates in children infected with Schistosoma mansoni after one dose of praziquantel, examined by Kato–Katz, Percoll and PCR
TechniqueNo. examined casesNo. cured individualNo. uncured% cure rate
Kato–Katz105102397.14
Percoll105951090.47
PCR70502071.4
Figure 1.

 Cure rates in children infected with Schistosoma mansoni after single dose of praziquantel diagnosed by PCR, Percoll and Kato–Katz.

Comparing egg counts by Kato–Katz and Percoll showed that cases with very low intensity by Percoll were missed by Kato–Katz, whilst cases with relatively high intensity by Percoll showed a high egg count by Kato–Katz. Tables 5–7 present the agreement of cure rates of S. mansoni cases one month after praziquantel treatment examined by Kato–Katz, Percoll and PCR.

Table 5.   Agreement of cure rates by Kato–Katz and Percoll*
 Kato–KatzTotal
NegativePositive
  1. *Kappa index = 0.478, P > 0.05 moderate agreement.

Percoll
 Negative64064
 Positive426
Total68270
Table 6.   Agreement of cure rates by Kato and PCR*
 KatoTotal
NegativePositive
  1. *Kappa index = 0.04, P > 0.05 no agreement.

PCR
 Negative49150
 Positive19120
Total68270
Table 7.   Agreement of cure rates by Percoll and PCR*
 PCRTotal
NegativePositive
  1. *Kappa index = 0.025, P > 0.05 no agreement.

Percoll
 Negative461864
 Positive426
Total502070

Discussion

Two Kato slides obtained from a single specimen are routinely used for diagnosis of schistosomiasis in the field, yet even examination of three slides still misses a proportion of positives particularly in low prevalence and intensity areas (Sleigh et al. 1982; Barreto et al. 1990). In an attempt to surpass this diagnostic limitation, more sensitive techniques (Percoll and PCR) were used in this study to detect false negative S.mansoni cases after Kato–Katz method.

The most impressive finding in this study is the detection of a marginal error of 11% (95% CI 5–17%) using Percoll compared with Kato–Katz. 80% of cases missed by Kato–Katz passed 4 epg, which is a very low egg count. This indicates the efficiency of the Percoll technique in detecting cases with low egg counts compared with Kato–Katz.

The efficiency of Percoll is based on the fact that it depends on 250 mg faeces; one egg in the sediment would mean 4 epg, whilst three Kato slides can diagnose up to 8 epg when only one egg is detected in one of the three slides examined. Moreover, detection of one to two eggs per faecal smear could be missed more easily within the faecal debris of Kato–Katz than the clear fields of the Percoll technique. Similar findings were reported by Eberl et al. (2002), who found that the triple Kato–Katz examination on consecutive samples still missed 7.4% of all patients, whereas low egg numbers were detected efficiently only by Percoll technique.

Concerning the performance of Kato–Katz and Percoll according to intensity of infection, Eberl et al. (2002) reported that Percoll always gave higher egg count estimates than the conventional Kato–Katz technique. This does not agree with the results of our study, which revealed lower egg counts by Percoll amongst cases with >100 epg. Knight et al. (1976) reported that all concentration methods lose some eggs in the procedural steps, but clear field allows better visualisation of ova. Moreover, Katz et al. (1970) proved that straining of stool through a wire mesh in the thick-smear technique allowed passage of all ova whilst removing the large particles, resulting in concentration of the eggs. This explains the greater number of eggs in Kato than Percoll at higher intensities.

PCR detected more positive cases than both Kato–Katz and Percoll. However, seven patients with eggs in their stools (diagnosed by Percoll) were missed by PCR and considered negative. Thus variation in sensitivity either way occurred. Similar results were detected by Rabello et al. (2002) and Pontes et al. (2003); they reported that amongst 194 patients studied, no eggs were found in the faeces of 16 individuals with a positive PCR, whereas two patients with eggs in their stools (diagnosed by Kato) were PCR negative.

We compared and evaluated the assessment of cure rates by Kato–Katz, Percoll and PCR. Kato–Katz gave the highest cure rate (97%), followed by Percoll (90%) and then PCR (71%). Percoll was more accurate than Kato–Katz in estimation of cure rate. On the other hand, PCR failed to detect four positive cases (one detected by Kato and Percoll and three by Percoll only). There is no reason to believe that these cases correspond to false positive parasitologic results, as Kato and Percoll methods are 100% specific (Turner et al. 2004). Pontes et al. (2003) reported that missed cases after PCR were certainly misdiagnosed by the DNA amplification assay due to many factors such as: inhibition of the amplification reaction by faecal compounds and/or DNA degradation during transportation from the field, variation in egg output and uneven distribution in faeces. In addition, Rabello et al. (2002) reported that PCR demands a more sophisticated laboratory, more complex operational efforts and high cost. Even by using multiplex real-time PCR, Ten Hove et al. (2008) reported discrepancies between microscopy and PCR analysis (both ways) in samples with very low egg counts or low parasite DNA concentration.

Agreement between Kato–Katz and Percoll was moderate: all positives diagnosed by Kato–Katz were positive by Percoll technique. There was no significant agreement between PCR and Kato nor PCR and Percoll. Thus it can be concluded that in surveys conducted in low intensity areas using Kato–Katz method, we should consider a marginal error of 11% as proved by Percoll technique.

Using the conventional Kato–Katz a proportion of young people with low prevalence and low intensity of infection can be easily missed. They constitute a potential source of transmission. This added to an increase in snail population due to neglect of control may lead to gradual flare up of schistosomiasis. So, continuous surveillance of the level of endemicity in cases with low intensity is necessary.

Our results indicate that Percoll is the technique of choice for diagnosis of S.mansoni in low intensity conditions and post-treatment. However, it is not cost-effective for surveys in Egypt. PCR revealed more very low intensity cases, yet it missed a small proportion of cases confirmed positives by coprological examination. Accordingly there must be additional diagnostic tests to verify the results obtained by PCR.

Acknowledgements

The authors acknowledge the TDR, WHO, for funding and supporting this work.

Ancillary