To determine whether the detection of human IgG bound to Schistosoma haematobium eggs from filtered urine could be used as a rapid diagnostic test (RDT-Sh).
To determine whether the detection of human IgG bound to Schistosoma haematobium eggs from filtered urine could be used as a rapid diagnostic test (RDT-Sh).
We filtered 160 urine samples from children in the Kwale District of Kenya to isolate S. haematobium eggs and used anti-human IgG antibody conjugated to horseradish peroxidase to bind to the human IgG attached to the eggs. We then added 3,3′5,5′-tetramethylbenzidine base (TMB), which turns blue in the presence of horseradish peroxidase to detect the S. haematobium eggs. The RDT-Sh was compared in a blinded manner to urine microscopy.
The RDT-Sh was positive in 89% of urine samples containing >1 egg/10 ml (58/65 samples) and 97% of urine samples containing >11 eggs/10 ml urine (35/36 samples) seen by microscopy. The RDT-Sh was negative 79% of the time when no eggs were seen on urine microscopy, but because up to three times more urine was used for the RDT-Sh, there were likely cases in which eggs were on the RDT-Sh filter but not detected by microscopy. We used latent class analysis incorporating urine microscopy, haematuria, proteinuria and RDT-Sh results to determine an overall 97% sensitivity and 78% specificity for RDT-Sh, 96% and 81% for urine microscopy, 71% and 98% for microscopic haematuria and 46% and 89% for proteinuria, respectively.
The RDT-Sh is quick, inexpensive and easy to perform in the field for the diagnosis of S. haematobium.
Déterminer si la détection d'IgG humaines fixées aux œufs de Schistosoma haematobium, à partir d'urine filtrée peut être utilisée comme un test de diagnostic rapide (RDT-Sh).
Nous avons filtré 160 échantillons d'urine d'enfants dans le district de Kwale au Kenya pour isoler les œufs de S. haematobium et avons utilisé des anticorps IgG humains conjugués à la peroxydase de raifort pour les fixer aux IgG humaines attachées aux œufs. Nous avons ensuite ajouté du 3,3′–5,5′-tétraméthylbenzidine base (TMB), qui vire au bleu en présence de la peroxydase de raifort, pour détecter les œufs de S. haematobium. Le test RDT-Sh a été comparé à l'aveugle à la microscopie d'urine.
Le test RDT-Sh a été positif dans 89% des échantillons d'urine contenant > 1 œuf/10 ml (58/65 échantillons) et dans 97% des échantillons d'urine contenant > 11 œufs/10 ml d'urine (35/36 échantillons) vus par la microscopie. Le test RDT-Sh a été négatif dans 79% des cas où des œufs n'ont pas été observés par la microscopie d'urine, mais parce qu'une quantité jusqu’à trois fois plus d'urine a été utilisée dans le test RDT-Sh, il y a eu des cas probables où des œufs étaient sur le filtre RDT-Sh mais n'ont pas été détectés par la microscopie. Nous avons utilisé l'analyse de classe latente incorporant les résultats de la microscopique des urines, d'hématurie, de protéinurie et du test RDT-Sh pour déterminer une sensibilité globale de 97% et une spécificité de 78% pour le test RDT-Sh, 96% et 81% pour la microscopie d'urine, 71% et 98% pour l'hématurie microscopique et 46% et 89% pour la protéinurie, respectivement.
Le test RDT-Sh est rapide, peu coûteux et facile à réaliser dans le domaine du diagnostic de S. haematobium.
Determinar si la detección de la IgG humana, unida a huevos de Schistosoma haematobium filtrados de muestras de orina, puede utilizarse como una prueba diagnóstica rápida (PDR-Sh).
Filtramos 160 muestras de orina de niños del distrito de Kwale, en Kenia, para aislar los huevos de S. haematobium y utilizamos anticuerpos anti-IgG humana, conjugados con peroxidasa del rábano, para que se uniesen a la IgG humana unida a los huevos. Después añadimos 3,3′–5,5′-tetrametilbenzidina base (TMB), que se transforma en azul en presencia de la peroxidasa del rábano para detectar los huevos de S. haematobium. La PDR-Sh se comparó de en un ensayo ciego con la microscopía de la orina.
La PDR-Sh fue positiva en un 89% de las muestras de orina que contenían > 1 huevo/10 ml (58/65 muestras) y un 97% de las muestras de orina contenían > 11 huevos/10 ml orina (35/36 muestras) observadas mediante microscopía. La PDR-Sh fue negativa en un 79% de las veces en las que no se observaban huevos mediante microscopía, pero debido a que se utilizó tres veces más orina para la PDR-Sh, podría haber casos en los que los huevos estaban en el filtro de la PDR-Sh, pero que no se detectaron mediante microscopía. Utilizamos un análisis de clases latentes incorporando los resultados de la observación por microscopía de la orina, la hematuria, la proteinuria y del PDR-Sh, y se determinó una sensibilidad total del 97% y una especificidad del 78% para la PDR -Sh, 96% y 81% para la microscopía de la orina, 71% y 98% para la microscopía de hematuria, y 46% y 89% para la proteinuria, respectivamente.
La PDR-Sh es rápida, barata y fácil de realizar en el campo para el diagnóstico de S. haematobium.
Schistosoma haematobium is the most prevalent of the six Schistosoma species that infect humans. The other five Schistosoma species — S. mansoni, S. japonicum, S. mekongi, S. guineensis and S. intercalatum — generally cause gastrointestinal pathology, whereas S. haematobium causes genitourinary disease (WHO 2010). Schistosomiasis is endemic in Africa with hundreds of millions of people at risk for disease and millions affected (Gryseels et al. 2006; Utzinger et al. 2009). Among human parasitic diseases, schistosomiasis is second only to malaria in terms of socioeconomic and public health consequences (Gryseels et al. 2006). S. haematobium eggs are highly immunogenic and cause the majority of the disease's morbidity and mortality (Gryseels et al. 2006). The eggs can cause haematuria, calcified bladder, fibrosis of the neck of the bladder, hydroureter, hydronephrosis and bladder cancer (Gryseels et al. 2006). S. haematobium eggs can be found in both the human male and female reproductive organs, and people with mild worm burdens can have minimal or no symptoms (Gryseels et al. 2006). According to the European Association of Urology (EAU), a light S. haematobium infection is defined as <100 eggs/10 ml urine, while a severe infection is >400 eggs/10 ml (Bichler et al. 2006). WHO classifies a light infection as <50 eggs/10 ml urine and a heavy infection as 50 or more eggs/10 ml (WHO 2010).
Schistosoma haematobium eggs are excreted in the urine covered in human IgM and IgG (Koech et al. 1984). Urine from people infected with S. haematobium have protein precipitates containing IgG, IgM, IgA, IgE and complement C3 (Bosompem et al. 1996). Humans initially generate an IgM immune response to adult S. haematobium worms and eggs, followed by IgA, IgG and IgE (Mutapi et al. 1998; Woolhouse et al. 2000; Osada et al. 2003). Human IgG and IgE levels correlate with egg burden (Mutapi et al. 1998; Woolhouse et al. 2000; Osada et al. 2003). IgG levels rise dramatically after children are reinfected with S. haematobium (Mutapi et al. 1998).
The ‘gold-standard’ test for diagnosing S. haematobium is by the visualisation of eggs under the microscope, typically by centrifuging or filtering urine to concentrate the eggs (Gryseels et al. 2006). Urine microscopy is highly specific if eggs are identified; however, intermittent egg shedding and a low egg count can negatively affect the sensitivity. Because urine microscopy requires a microscope, a trained technician and the time and resources to read the slides, it poses challenges, especially in the field and during large-scale screenings of urine for S. haematobium.
Haematuria, and to a lesser extent proteinuria, is associated with S. haematobium infection (Houmsou et al. 2011). Urine dipsticks for microscopic haematuria and proteinuria are being used to guide treatment and public health interventions for S. haematobium (Houmsou et al. 2011). Urine dipsticks are cheap, and haematuria in a S. haematobium endemic area is highly specific for infection (Houmsou et al. 2011). However, haematuria has only modest sensitivity for diagnosing the disease and will miss light infections.
The societal costs and benefits of presumptive treatment of at-risk people for possible schistosomiasis have been compared with treatment based on the results of a potential rapid diagnostic test (RDT) (Stothard 2009). It was determined that a RDT costing less than US $ 3 would provide financial incentive to make treatment decisions based on the results of the RDT and one costing US $ 5 would be cost-neutral (Stothard 2009). The purpose of our study was to determine whether the detection of human IgG in filtered urine could be useful as a RDT for the detection of S. haematobium eggs in an endemic area.
The study received institutional review board (IRB) approval by Eastern Virginia Medical School (PR-149-INT) and the Kenya Medical Research Institute (KEMRI/RES/7/3/1) and was conducted in Kwale District on the coast of Kenya in September 2011. Study investigators met with parents, school administrators, teachers and government administrators prior to data collection for discussion and to obtain verbal consent. Written informed consent was given on the day of the study. After obtaining written consent, urine was collected between 10:00 am and 2:00 pm from 177 schoolchildren that stated they had seen blood in their urine recently. 160 schoolchildren were between the ages of 8 and 17 years and qualified for use in data analysis. Of the 17 study subjects who were not included in the analysis, 16 had incomplete data and one person turned out not to fulfil the age criterion. All study subjects received 40 mg/kg of praziquantel at the conclusion of the study regardless of the urine microscopy or rapid diagnostic test for Schistosoma haematobium (RDT-Sh) results as directed by the Kenyan Ministry of Health. There were no adverse events during the study.
Urine from each participant was noted to be yellow/clear, turbid or grossly bloody. The urine was then analysed for protein and blood using Rapid Response urine dipsticks (BTNX Inc., Markham, On, Canada). From each urine sample, two 10-ml aliquots were removed and each filtered using Nuclepore™ polycarbonate 12-micron filter paper (Whatman Ltd., Maidstone, Kent, UK) and examined with a microscope to determine the number of S. haematobium eggs in each sample. Egg counts ranged from 0.5 egg/10 ml (one egg was seen on one aliquot and no eggs on the other aliquot from the same urine sample) to more than 1000 eggs/10 ml urine. The remainder of the urine sample was analysed by the RDT-Sh in the field. Urine microscopy was completed within 48 h of slide preparation in the laboratory.
To prepare the antibody used in the RDT-Sh, we diluted anti-human IgG conjugated to horseradish peroxidase (Promega Corp., Madison, WI, USA) to 400 ng/ml using unsterile commercially available bottled drinking water. The antibody was diluted on the day of the experiment and kept at room temperature. We mixed 0.25 ml of the diluted anti-human IgG antibody into each urine sample.
The RDT-Sh filters were prepared in advance. A standard sized hole-punch was used to create paper discs of Whatman GE Healthcare Life Sciences grade 113 wet strengthened cellulose paper with a 30 micron pore size (Whatman GE Healthcare Life Sciences, Piscataway, NJ, USA). The paper discs were glued to one end of a Cole–Parmer polycarbonate female–female luer lock coupling device (Cole-Parmer, Vernon Hills, IL, USA) (Figure 1).
Up to 60 ml of urine (containing the anti-human IgG–conjugated antibody) was drawn up into a syringe and attached to the open end of the female luer lock coupling device (opposite to the end containing the glued filter paper) and the urine filtered (Figure 2). S. haematobium eggs remained trapped on the inner aspect of the filter paper. The empty 60-ml syringe was then removed. A syringe containing 3,3′5,5′-tetramethylbenzidine base (TMB) was then attached to the open female luer lock coupling device and approximately 0.25 ml of TMB was pushed through the filter. TMB turns from a pale yellow liquid to blue in the presence of horseradish peroxidase; therefore, the RDT-Sh was negative if the filter paper remained white and was considered positive if the filter paper turned blue (Figure 3). RDT-Sh filters that turned blue within 60 s were considered to be positive.
JMS and SB performed the RDT-Sh and were blinded to the urine microscopy results but not to the colour of the urine or the urine dipstick results. Urine microscopy was performed in the laboratory without knowledge of the urine colour, urine dipstick results or RDT-Sh results. Urine microscopy results were compared with RDT-Sh results after all data collection was completed.
Within 60 s, most of the samples having S. haematobium eggs on microscopy were positive by the RDT-Sh and, for heavy infections, the RDT-Sh was usually instantly positive. The results are summarised in Table 1. The RDT-Sh was positive in 46% (6/13) of urine samples containing 0.5–1 egg/10 ml urine, in 89% of urine samples containing >1 egg/10 ml (58/65 samples) and in 97% of urine samples containing >11 eggs/10 ml urine (35/36 samples). The RDT-Sh was negative in 79% of cases where no eggs were seen on urine microscopy. However, up to three times more urine was used for the RDT-Sh so there may have been eggs on the RDT-Sh filter but not seen by microscopy.
|Urine microscopy (number of subjects)||RDT-Sh test||Total|
|Negative (at 20 min)||1–20 min positive||30–60 s positive||<30 s positive||Instant positive|
|0 eggs/10 ml (82)||44||21||6||3||8||51% (82/160)|
|0.5–1 eggs/10 ml (13)||3||4||1||3||2||8% (13/160)|
|1.5–5 eggs/10 ml (17)||4||1||3||2||7||11% (17/160)|
|5.5–10.5 eggs/10 ml (12)||1||0||1||1||9||7.5% (12/160)|
|11 to >1000 eggs/10 ml (36)||0||1||1||4||30||22.5% (36/160)|
We estimated the number of eggs present on the RDT-Sh filter by taking the number of eggs per ml of urine seen on urine microscopy and multiplying it by the volume of urine filtered for the RDT-Sh; these results are summarised in Table 2. The RDT-Sh was less likely to be positive when very few eggs were found in the filter.
|Estimated number of eggs on the RDT-Sh filter (2.5 to >5500 eggs)||Percentage of RDT-Sh positive in 60 s||Total number of samples analysed|
|2.5–7.5 eggsa||46% (6/13)a||13|
|8–20 eggs||69% (9/13)||13|
|21–99 eggs||92% (21/23)||23|
|100 to >1000 eggs||96% (24/25)||25|
The RDT-Sh filter was not directly evaluated microscopically for the presence of eggs, so we could only assume that the average number of eggs seen on microscopy approximated the number of eggs that were on the RDT-Sh filter. However, different volumes of urine were used for the RDT-Sh and microscopy, so when few eggs were seen by microscopy, it is possible that no eggs were present on the RDT-Sh filter. The corollary is also possible that when no eggs were seen on microscopy, there were still eggs present on the RDT-Sh filter. Latent class analysis (LCA) was employed utilising microscopy, microscopic haematuria, proteinuria and RDT-Sh results in the analysis. LCA gives the RDT-Sh an overall 97% (95% confidence interval (CI): 91–100%) sensitivity and 78% (95% CI: 67–89%) specificity when the RDT-Sh is read at 1 min. The LCA reports sensitivities and specificities for urine microscopy, microscopic haematuria and proteinuria as 96% and 81%, 71% and 98%, and 46% and 89%, respectively (Table 3).
|Sensitivity (95% CI)||Specificity (95% CI)|
|RDT-Sh||97% (91–100)||78% (67–89)|
|Urine microscopy||96% (90–100)||81% (70–92)|
|Microscopic haematuria||71% (55–87)||98% (94–100)|
|Proteinuria||46% (31–60)||89% (83–95)|
A few of the urine samples contained visible sediment and obvious foreign bodies such as dirt and grass, clogging the RDT-Sh filter 39 of 160 times (24%). In some cases, this reduced the amount of urine that could be filtered. The filter broke during the filtering process 15 of 160 times (9%), although this did not necessarily affect the results. In general, gross blood or cloudy urine resulted in lower RDT-Sh filtration volumes.
The ability of microscopic haematuria, proteinuria and the RDT-Sh to detect S. haematobium infection compared to urine microscopy is summarised in Table 4. Three study subjects had microscopic haematuria on urine dipstick and no S. haematobium eggs seen on microscopy; one subject was a 10-year-old girl who also had a negative RDT-Sh. The other two subjects were 9- and 16-year-old boys who both had positive RDT-Sh results within 1 min. The presence of glucose, nitrates, ketones, bilirubin, leucocyte esterase, urobilinogen, or a certain pH or specific gravity on the urine dipstick was not correlated with the presence of S. haematobium eggs seen on microscopy. We retrospectively analysed our data using the following algorithm summarised in Figure 4:
|Gross haematuria present (number of study subjects)||Positive for microscopic haematuria (number of study subjects)||Positive for proteinuria (number of study subjects)||RDT-Sh positive in 1 min (number of study subjects)|
|0 eggs on microscopy||0% (0/82)||4% (3/82)||23% (19/82)||21% (17/82)|
|0.5–10.5 eggs on microscopy||5% (2/42)||31% (13/42)||37% (15/41)||69% (29/42)|
|11 to >1000 eggs on microscopy||28% (10/36)||81% (29/39)||69% (27/39)||97% (35/36)|
This algorithm would correctly identify as positive 85% (66 of 78 children) who had eggs (0.5 to >1000 eggs/10 ml) seen on urine microscopy, and identify as negative 78% (64 of 82 children) who did not have eggs seen on microscopy. If all samples with no haematuria and a negative RDT-Sh had the RDT-Sh filter examined microscopically for the presence of eggs, and those filters with eggs on microscopy treated as positive, the algorithm would have 100% sensitivity and 84% specificity. It would require that 76 filters be examined microscopically (to identify the 12 samples in our data set with no haematuria, a negative RDT-Sh and eggs seen on urine microscopy) compared with 320 filters needing urine microscopy if each of the original 160 urine specimens needed two 10-ml urine aliquots filtered and examined. This algorithm could be especially helpful in field settings where all but the lightest S. haematobium infections would be quickly and easily detected and treated. All samples negative for both haematuria and the RDT-Sh could have their RDT-Sh filter examined microscopically at a later time.
Human IgG coats S. haematobium eggs in the urine (Koech et al. 1984). These immune complexes have complicated the past use of enzyme-linked immunosorbent assay (ELISA) tests that attempted to detect egg antigens (Uga et al. 1989). Anti-human IgG–conjugated antibodies are cheap, easy to find and commercially available, whereas specific antibodies to S. haematobium egg antigens are not. This preliminary study was designed to determine whether the detection of human IgG in filtered urine was able to diagnose S. haematobium infection in an endemic area.
As the RDT-Sh detects the immunoglobulin attached to the S. haematobium eggs in the urine, the RDT-Sh will not correctly diagnose an infection when there are no eggs present in the urine. The number of S. haematobium eggs excreted into the urine in S. haematobium infected children varies throughout the day as well as day-to-day (Doehring et al. 1983).
It is theoretically possible that concurrent genitourinary tract infections, cystitis, casts in the urine, Enterobius vermicularis and protein complexes in the urine could cause false positives with the RDT-Sh. The ammonia, salts, water, proteins, carbohydrates, red blood cells (6–8 mm), white blood cells (10 mm), trichomoniasis (10 mm), crystals, yeast (10–20 mm) and bacteria (<3 mm) found in some urine samples should all pass through the RDT-Sh 30-mm filter. Theoretical potential RDT-Sh false-negative results could occur from the filter breaking or leaking, a low volume of urine being filtered, few eggs on the RDT filter or eggs bursting under the pressure of filtration. It is also possible that some eggs are excreted with little or no IgG attached or that the amount of anti-human IgG–conjugated antibody bound to the IgG on the eggs on the filter is not sufficient to cause a colour change when exposed to TMB.
Basic modifications to the RDT-Sh may further improve the test. The addition of a 30-mm filter inside of a commercially available filtering device will allow for microscopic counting of the eggs. Adding a synthetic larger pore pre-filter will remove sediment and debris. Some synthetic filters also bind non-specific proteins, and this could reduce the number of false-positive results with the RDT-Sh. Adding a water wash after the urine is filtered onto the RDT-Sh filter, but before the addition of TMB, may also reduce the number of false-positive RDT-Sh results, because it will remove anti-human IgG–conjugated antibody not bound to S. haematobium eggs. Adding more anti-human IgG–conjugated antibody to urine samples may help the filter turn positive faster when fewer eggs are present and improve the sensitivity of the RDT-Sh in light infections. Lastly, and most importantly, using an entire urine sample for the RDT-Sh will add as many eggs as possible onto the filter and should further improve the sensitivity of the test, especially in light infections.
Various methods for the diagnosis of S. haematobium have been recently evaluated using latent class analysis, including urine antigen detection of S. haematobium complexed to C3 (98% sensitive and 35% specific), the detection of anti-schistosome IgG from dried blood spots (48% sensitive and 57% specific), ultrasound examination (74% sensitive and 65% specific) and urine microscopy (93% sensitive and 98% specific) (Koukounari et al. 2009). There have been conflicting reports on the sensitivity and specificity of schistosome circulating cathodic antigen (CCA), but it is thought to be relatively insensitive for detecting S. haematobium (Stothard 2009). The test characteristics of the RDT-Sh compare favourably with these other methods of diagnosing S. haematobium, with the additional benefit that the test is rapid, cheap and easy to perform in the field.
The RDT-Sh was created with commercially available products and only costs approximately US $ 0.45 per reaction. This estimates the costs for the antibody, TMB, filter paper and luer coupling devices, but does not include the syringes that were reused during the experiment. The urine dipsticks cost approximately US $ 0.15 per strip.
Children living in S. haematobium endemic areas receive praziquantel for the treatment of schistosomiasis as part of organised large-scale drug administration campaigns. A cheap and easy-to-use RDT with sufficient sensitivity for diagnosing S. haematobium infection would be useful for targeting drug treatment only towards infected individuals; monitoring S. haematobium resistance to praziquantel; improving epidemiological mapping and monitoring of disease prevalence; and help guide public health interventions and prevention strategies. Current successful S. haematobium epidemiological mapping strategies include school-based questionnaires about self-reported haematuria as well as teacher-initiated testing of student's urine for blood (Brooker et al. 2009). The WHO currently advocates repeated stool and/or urine examinations of at least 50 people within a defined ecological zone to quantify the severity and types of infecting Schistosoma species (Brooker et al. 2009).
The mapping and treatment of neglected tropical diseases (NTDs) is being facilitated by the use of accurate rapid diagnostic tests that exist for malaria, onchocerciasis, visceral leishmaniasis and lymphatic filariasis (Brooker et al. 2009). There exists a urine antibody test to detect CCA, a protein regurgitated by adult Schistosoma worms. The CCA has good sensitivity and specificity for S. mansoni, but not for S. haematobium (Brooker et al. 2009). Concurrent use of CCA and the RDT-Sh on a single urine sample may allow for non-invasive, accurate and rapid mapping of schistosomiasis throughout Africa where S. mansoni and S. haematobium frequently co-exist. Additionally, after successful treatment with praziquantel, both the CCA and RDT-Sh should both become negative. This will allow for monitoring of praziquantel efficacy, an increasing public health concern especially because there are few acceptable treatment alternatives for schistosomiasis.
Haematuria has good specificity but only modest sensitivity for diagnosing S. haematobium infection in endemic areas, especially in people with light infections. The RDT-Sh has superior sensitivity to haematuria for detecting S. haematobium infection. Performing the RDT-Sh only on those samples negative for haematuria would be an additional strategy useful for rapidly detecting more S. haematobium infected individuals in a non-clinical setting than microscopic haematuria alone. Additionally, if only filters negative for both haematuria and the RDT-Sh are examined microscopically for the presence of S. haematobium eggs, it could substantially reduce the overall number of filters that would need to be evaluated microscopically and allow for the detection of nearly 100% of those infected individuals with eggs in the urine.
We have shown that the RDT-Sh can detect S. haematobium eggs by detecting the presence of human IgG on filtered urine. The RDT is inexpensive, quick, easy to perform and can be performed in the field. The RDT-Sh has excellent sensitivity, especially with larger numbers of eggs in the urine.
We thank the Department of Emergency Medicine, Eastern Virginia Medical School, Norfolk, VA USA for funding the research. We thank Dr. Simon Brooker in the Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, London. Lastly, we thank Charles S. Mwandawiro and Sammy M. Njenga of the Kenya Medical Research Institute, Nairobi, Kenya. Eastern Virginia Medical School has submitted a non-provisional Patent Convention Treaty application for the RDT-Sh.