High prevalence of Strongyloides stercoralis in school-aged children in a rural highland of north-western Ethiopia: the role of intensive diagnostic work-up

Soil-transmitted helminthiases (hookworms, Ascaris lumbricoides and Trichuris trichiura) are extremely prevalent in school-aged children living in poor sanitary conditions. Recent epidemiological data suggest that Strongyloides stercoralis is highly unreported. However, accurate data are essential for conducting interventions aimed at introducing control and elimination programmes. We conducted a cross-sectional survey of 396 randomly selected school-aged children in Amhara region in rural area in north-western Ethiopia, to assess the prevalence of S. stercoralis and other intestinal helminths. We examined stools using three techniques: conventional stool concentration; and two S. stercoralis-specific methods, i.e. the Baermann technique and polymerase chain reaction. The diagnostic accuracy of these three methods was then compared. There was an overall prevalence of helminths of 77.5%, with distribution differing according to school setting. Soil-transmitted helminths were recorded in 69.2%. Prevalence of S. stercoralis and hookworm infection was 20.7 and 54.5%, respectively, and co-infection was detected in 16.3% of cases. Schistosoma mansoni had a prevalence of 15.7%. Prevalence of S. stercoralis was shown 3.5% by the conventional method, 12.1% by the Baermann method, and 13.4% by PCR, which thus proved to be the most sensitive. Our results suggest that S. stercoralis could be overlooked and neglected in Ethiopia, if studies of soil-transmitted helminths rely on conventional diagnostic techniques alone. A combination of molecular and stool microscopy techniques yields a significantly higher prevalence. In view of the fact that current control policies for triggering drug administration are based on parasite prevalence levels, a comprehensive diagnostic approach should instead be applied to ensure comprehensive control of helminth infections.


Background
Soil-transmitted helminths (STH) affect more than 2 billion people world-wide [1] and rank among the most prevalent neglected tropical diseases (NTDs) in sub-Saharan Africa (SSA) [2]. Hookworm is widely distributed in both rural and urban areas. While Ascaris lumbricoides and Trichuris trichiura are irregularly distributed, they are mainly found in urban areas [3]. Significant efforts have been and are being made to control their impact in endemic countries, with the ultimate target being to eliminate morbidity by 2020 [4]. Preventive chemotherapy is the key strategy for achieving this goal [5]. The World Health Organization (WHO) recommends regular mass drug administration (MDA) among school-aged and preschool-aged children (SAC, pre-SAC) in areas where prevalence exceeds 20% [6]. In areas of co-endemicity, integrated MDA is the recommended approach, e.g. albendazole-praziquantel for controlling STH and schistosomiasis [4].
Strongyloidiasis, caused by S. stercoralis, is an NTD known to be endemic in tropical and subtropical areas worldwide, affecting up to 100 million people [7,8]. However, data on its epidemiology suggest that is highly underreported [9,10]. This fact is mainly linked to diagnostic challenges: traditional identification of eggs in stool has nil sensitivity for S. stercoralis, and a high number of infections thus go undetected [11,12]. Its diagnosis relies on identification of larvae in stool, with the two most appropriate methods being the Baermann, Harada Mori and Koga agar plate culture techniques, even though their sensitivity is not optimal [13]. While highly sensitive molecular techniques have recently been developed [14,15], there is still no gold standard. Accordingly, a combination of different techniques is the most useful approach for the detection of S. stercoralis larvae [7].
Ethiopia is the second largest country in SSA, with a population of almost 92 million people, 84% of which live in rural areas [16]. Preventable communicable diseases, e.g. pneumonia, diarrhoea, measles and rabies, are the country's major health problems [17]. Data on the burden of NTDs are incomplete, yet onchocerciasis, trachoma, lymphatic filariasis, STH, schistosomiasis, leishmaniasis and podoconiosis are all prioritized for intervention [18]. Ethiopia has the third highest prevalence of hookworm, the second highest prevalence of A. lumbricoides and fourth highest prevalence of T. trichiura in SSA [19]. According to different studies, the prevalence of S. stercoralis in the country over the last 25 years has ranged from 0.6% to as a high as 17.4%, with the highest reported rate is related to HIV-infected patients and hospital-based studies [20,21]. However, S. stercoralis is not included in control programs [18].
The objectives of this study were: (i) to assess the prevalence of S. stercoralis infection among SAC in a rural area of Ethiopia; and (ii) to explore the usefulness of a diagnostic approach based on the combination of two parasitological techniques (conventional stool concentration and the Baermann method) and a molecular technique.

Study area
The Amhara Region is a subtropical zone lying at a height of 1,900 m above sea level, in the north-western of Ethiopia. The region's capital, Bahir Dar (population approximately 250,000), is centrally situated at the southern tip of Lake Tana. There is a central urban nucleus and a surrounding rural area, in a radius of about 30 km from the city centre [22]. There are three distinct seasons, i.e. a rainy season, from June to September; spring, from mid-October until the end of December; and a dry season, during the reminder of the year. In the rural area there are 16 primary schools, which are structured in eight grades, with pupil's age ranging from 5 to 6 in grade one, to 13-14 years in grade eight.

Study design and study population
We conducted a cross-sectional study targeting primary schools in the rural area of Bahir Dar from October to November 2013. Adequate sample size was computed by assuming a maximum prevalence of S. stercoralis of 5% [23], a marginal error of 2.5%, a 95% confidence interval, and a non-response rate of 10%. The list of schools was provided by the Amhara Regional Education Bureau. Sampling was carried out by random multistage stratified cluster strategy, with the primary sampling units being schools. To improve accuracy in the sample design, the schools were randomly selected with probability proportional to size. A total of eight schools were selected as follows: two in the south-west of the study area (Achadir and Meshenti); two in the north-east (Zenzelema and Gedro); three in the south-east, (Sebatamit, Tisabay 1-4 and Tisabay 1-8), and one in the north-west (Yiganda), on a small lakeside peninsula ( Fig. 1). Secondly, we randomly selected children attending the first grades, i.e. one to four, in every school. The list of pupils was provided by the head teachers of the eight schools. Although a total of 46 pupils were selected from each school, 48 to 50 samples were collected in each case, yielding a final sample size of 396.

Laboratory procedures
Stool samples were collected in a labelled container, without any preservation method. Samples were sent to the laboratory at room temperature, and processed within a maximum of four hours after collection.

Parasitological examination
First, a formol ether concentration (FEC) of 0.5 g of stool was processed with the Bioparaprep MINI® system (Leti Diagnostics, Barcelona, Spain) according to the manufacturer's instructions; the design of the system is based on a modification of Ritchie's method, with filtering by the classical method been replaced by a filtrationconcentration process [24]. Secondly, we applied the Baermann technique based on the active migration of larvae from fresh stool samples, i.e. when faeces are suspended in water, the larvae move into the water, sink to the bottom and can be collected for identification [25]. Samples were incubated for 18 h at 26°C with activated charcoal before applying the Baermann technique. Every sample was checked three times by three different, experienced microscopists. Pathogenic helminth eggs/ larvae and intestinal protozoa cysts were recorded. Strongyloides stercoralis larvae were identified by the buccal cavity and the genital organs in rhabditiform larvae, and the posterior tip in filariform larvae.

DNA extraction
One aliquot of one g of stool sample was preserved in a container and stored at 4°C for DNA extraction. A previous concentration of the sample with saline solution (0.9%) was prepared using a Bioparaprep MINI® device. DNA extraction, from 180 to 200 mg of the concentrated sample, was performed with a QIAamp® DNA stool mini kit (Qiagen, Hilden, Germany), following the manufacturer's instructions. Previous procedures were

DNA amplification
Strongyloides stercoralis-specific primers targeting the 18S ribosomal subunit, as described by Verweij et al. [26]: were used (Primer F: 5′-GAA TTC CAA GTA AAC GTA AGT CAT TAG C-3′; Primer R: 5′-TGC CTC TGG ATA TTG CTC AGT TC-3′). The amount of DNA was estimated with a nano-photometer, to ensure a concentration of five ng/μl in a final volume of 25 μl. A qualitative real-time polymerase chain reaction (RT-PCR) assay was performed, using a SybrGreen format (Invitrogen, San Diego CA, USA), as described by Saugar et al. [15]. Purified genomic DNA from Strongyloides venezuelensis L3 was used as positive control. The samples were assayed in duplicate. A third sample including 10 ng of S. venezuelensis DNA was also included as internal inhibition control. No template controls were included in each run. Amplification and detection were carried out using a Corbett Rotor-Gene™ 6000 RT-PCR cycler (Qiagen Corbett, Hilden, Germany) and data analysis in a Rotor Gene™ 6000 Series software version 1.7.

Data and definition
We recorded participant's gender, age and school grade. Infection by S. stercoralis was defined as positive when at least one of the three techniques was positive, i.e. presence of larvae in one/both parasitological test and/ or positive PCR.

Statistical analysis
We performed a descriptive analysis of parasitic infections using frequency tables; proportion and 95% confidence interval (CI) were used for the categorical variables while mean and standard deviation (SD) were used for the quantitative variables. Differences in parasite prevalence were assessed by the Chi-square test or Fisher's exact test, and the association between infection and quantitative variables was analyzed by the t-test. The level of statistical significance was set at a value of P < 0.05. All statistical analyses were performed using the software package for statistical analysis, SPSS v.15.0.

Results
The children had a mean age of 9.74 years (range 5-18 years; SD 1.91); 43% were female and 57% male. The gender and age distribution was similar across the total sample, at each school, and for the four school grades.

Discussion
To our knowledge, 20.7% is the highest S. stercoralis prevalence figure reported in Ethiopia to date. The strength of our study lies in its use of a combination of parasitological and molecular techniques: 35% of the samples (29/82) were identified by parasitological techniques alone, and 37.8% (31/82) were identified by PCR. The final number of positive samples (82/396) relied on a combination of all three techniques. This approach, based on a combination of microscopic and molecular methods, had not been previously applied in the country. Among SAC, the highest previously reported prevalence was 3.4% in a study based on egg-detection techniques [23]. Moreover, our study area had been described as having a low prevalence of S. stercoralis, in that the most recent survey conducted in the Amhara Region reported a figure of 0.2% (five cases in 2,338 stool specimens) among children aged 2-15 years. What is more, FEC was the sole diagnostic technique used [27].
Integrated control programmes, e.g. for schistosomiasis and STH, have been launched in the country [28].  Our results indicate the need to include S. stercoralis. The prevalence of other STH could decline, while MDA with albendazole/mebendazole is ongoing but is unlikely to have an impact on S. stercoralis. Ivermectin is the treatment of choice [29,30]. In countries where this has been introduced for control/elimination of onchocerciasis and lymphatic filariasis, prevalence of S. stercoralis has declined over time [31,32]. The prevalence of hookworm in our sample is much higher than the prevalence reported by a local Amharabased survey involving 2,338 children (9.7%) [27] and the estimated national prevalence of 16% [19]. This is probably linked to the poor sanitary conditions in the study area [33]. Such conditions make for the coexistence of and co-infection by both helminth species, since S. stercoralis has a similar transmission pattern [34].
The high prevalence of A. lumbricoides and T. trichiura on the peninsula in the north-west is noteworthy. The burden posed by A. lumbricoides and T. trichiura in Ethiopia is reportedly higher than that of hookworm, with the national average being estimated at 37% for A. lumbricoides and 30% for T. trichiura [19]. Low variation in temperature and environmental conditions in an isolated area such as the peninsula has been previously linked to these helminths infections [35].
The present study showed that S. mansoni and hookworm co-infections are common in SSA [36]; in our sample, S. mansoni displayed a clear pattern along the course of the river. Previous studies report a high transmission rate in the Lake Tana area at the end of the rainy season, exactly when our study was carried out. The low prevalence at the school on the peninsula has been accounted for by the displacement of the snail population -and the associated schistosome infectionstowards the river in the rainy season, following the water flow [37].
The study has a number of limitations: (i) Using a single faecal sample may underestimate the prevalence of parasitic infections and fails to detect larvae in up to 70% [13]; analyses of samples over consecutive days would help to narrow this gap; (ii) S. stercoralis multiplies within the human host, establishing long-lasting infections [38]. Prevalence rates are often lower in children than in adults [39], and it can therefore be assumed that prevalence in the area could be much higher; (iii) PCR was the most sensitive technique in our analysis, while the amount of stool for the DNA extraction was one g; for the Baermann technique we used a minimum of 30 g. Intermittent elimination of larvae in stools, a low parasitic load, as well as other factors that can make DNA extraction heterogeneous must thus be explored.
(iv) Applying PCR on the larvae isolate collected by the Baermann's method would increase the PCR detection rate; also would confirm the detected larvae as of S. stercoralis, as well as the co-infection with hookworm; however, this was not applied in our study; (v) The higher prevalence in the eastern areas could be related with the proximity to the water, since the larvae thrive in warm wet soil. Even so, more local/behavioural factors should be analyzed.

Conclusions
Our findings highlight the need for more studies to focus on the desirability of using a combination of methods for diagnosis of S. stercoralis. The information on the prevalence of infection rates suggests that S. stercoralis is highly underreported in SSA, owing to the fact that many studies often use low-sensitivity diagnostic methods, e.g. in our study, only 3.5% of the infections were diagnosed by FEC, whereas a figure of 20.7% was reached when techniques focusing on larvae detection were used in combination. National health policies in endemic countries must be supported by accurate ascertainment of STH epidemiology, including S. stercoralis and its related morbidity. Furthermore, our study data also support the possibility of routinely using a combination of albendazole and ivermectin during preventive chemotherapy campaigns targeting STH.