Giardiasis is a zoonotic disease that affects at least 280 million people worldwide each year [1]. It is one of the main causes of nonviral diarrhoea in industrialized countries [2] and it is associated with additional digestive disorders in children and adults including abdominal pain, nausea. Giardiasis can develop into a chronic condition, and disease can be aggravated in immunocompromised hosts. Infection with the causative intestinal protozoa Giardia intestinalis is commonly associated with unsafe drinking water, poor hygiene and sanitation, poverty and warm climate [3,4]. Correspondingly, the prevalence of G. intestinalis is estimated at 2-7% in developed countries and up to 20-30% in low- and middle-income countries (LMICs) [5]. Owing to their common faecal-oral mode of transmission G. intestinalis is often found in coinfection with other intestinal protozoa [6].

Studies among school-aged children across Africa are scarce; however, those available reports a G. intestinalis prevalence’s of 16% in urban and 24% in rural areas of Morocco [7] and 11.7% in a rural setting in southern Ethiopia [8]. From Côte D'Ivoire a 17.3% and 13.9% prevalence was demonstrated in the Man area [9] and the region of Agboville, respectively [6]. G. intestinalis prevalence in school children in Burkina Faso is currently described by several smaller studies reporting on local numbers, while the national prevalence unknown. One such study, conducted in 2014, has shown G. intestinalis prevalence’s of 13.3%, 12.5% and 9.8% among children aged 6-15 years from three different schools of the Central region [10]. Another setting revealed a prevalence of 43.7% in patients aged 5 months to 72 years suffering from gastroenteritis visiting the Saint Camille Hospital of Ouagadougou [11]. A retrospective study of parasitological aetiology assessments of gastroenteritis patients from the same hospital showed an infection rate of 24.8% [12].

Metronidazole that has been on the market since 1959 [13] is the main drug used to treat giardiasis. However, a growing resistance of G. intestinalis to this drug has been reported [13] and no new drug is available to efficiently treat the disease. To minimize the extent of resistance development it is strongly recommended to treat only diagnostically confirmed cases while avoiding blind or presumptive treatments [14]. There is currently no gold standard for the diagnosis of giardiasis [15]. However, the most common approach is the microscopic examination of trophozoïtes or cysts of G. intestinalis in stools samples [16]. Challenges of this microscopy-based diagnosis include low sensitivity (around 60% for single testing due to intermittent faecal excretion of the parasite) and a high dependence on the skills and experience of the technician [16]. Previous efforts on identifying alternative means of G. intestinalis diagnosis include the detection of parasite antibodies by direct immunofluorescence and immune-enzymatic methods (e.g. ELISA). However, as with microscopy, these techniques require a highly experienced technician along with specialized equipment [17]. A biochemical assay was also tested for the potential diagnosis of giardiasis, based on the detection of volatile organic compounds that derive from the protozoan metabolism, in faecal samples. However, this technique is not feasible in routine laboratory, because it also requires sophisticated equipment [18]. Strain characterization using molecular biology techniques such as real-time Polymerase Chain Reaction (PCR) has proven to be highly sensitive and specific for G. intestinalis [19], yet the use of PCR in parasitology is limited in African laboratories compared with its application in bacteriology or virology [19]. Many studies have focused on the use of PCR to test samples assessed as negative by microscopy, but very few were dedicated to the verification of positive cases of G. intestinalis in its cystic form. However, there is a particular need for the verification of microscopy-positive samples owing to the difficulty of differentiating G. intestinalis cysts from non-pathogenic intestinal protozoa and Morel spores when the technician is not highly trained [20].

The objective of this study was to assess real-time PCR to detect G. intestinalis for the first time in microscopy-positive stools samples from healthy schoolchildren in the Centre-Ouest and Plateau Central regions of Burkina Faso.

Study area and population

This work was conducted in the framework of the “Vegetable go to School: improving nutrition through agricultural diversification (VgtS).” project in Burkina Faso [21]. Four hundred and forty-one schoolchildren aged 8-14 years were randomly selected from four schools in the Centre Ouest region (Douré A, Goundi B, Ipendo A and Tio A) and four schools of the Plateau Central region (Loumbila A, Linoghin A, Tangzougou and Wavoussé).

These regions have a Soudano-Sahelian type of climate, and consist predominantly of agricultural areas and small farms. Questionnaires were administered to schoolchildren, their teachers and parents in order to collect the socio-demographics characteristics of the schoolchildren, and the socio-economic and WASH conditions in the corresponding schools and households. The nutritional status of the schoolchildren was assessed by measuring anthropometric parameters on the first day of the stool collection.

Collection of microscopy-positive and microscopy-negative stool samples

Two fresh morning stool samples were collected from each child on two consecutive days, into a clean and dry pot with an identification label. The stool samples were subsequently transported in coolers or thermostats containing ice boxes to the IRSS's Microbiology laboratory for parasitological analysis. The transport was conducted in the shortest possible time (in any case less than two hours) to avoid a temperature rise which might lyse the protozoan trophozoïtes. First, macroscopic examination of the stool was conducted to document the consistency and the potential presence of blood, phlegm, mucus, pus and macroscopic parasites. Second, the stool samples were analysed by direct microscopic examination of fresh sample and of sample concentrated by the formalin-ether sedimentation method of Ritchie [22], to detect intestinal protozoa and helminths (roundworm eggs are commonly destroyed by this technique). A total of 147 stool samples (corresponding to all 94 microscopy-positive and 53 microscopynegative for G. intestinalis ) were stored at -80°C prior real-time PCR analysis. As in many other previous molecular studies, we have been working on cysts that are more resistant than trophozoites [23].

Quality control of parasitological analysis was performed on 10% of the samples and conducted by a well-trained technician from the Teaching University Hospital Pediatric Charles de Gaulle, Ouagadougou, Burkina Faso who re-analysed the stool samples using microscopic-based examination. They included a total of 43 stools samples including 9 positive and 34 negative stool samples to Giardia intestinalis , randomly selected.

DNA extraction of Giardia intestinalis

DNA extraction from stool samples was conducted using the DNASorb- B Laboratory kit of Sacace Biotechnologies^® (Como, Italy). Samples were thawed at room temperature and pre-treated as follows: a suspension of each sample was prepared by introducing 1.0 ml of saline and 0.1 g of the stool sample into a polypropylene tube. The suspension was vortexed and then centrifuged at 10 000 g for 5 min. A volume of 100 μl of the pellet was subsequently transferred into an Eppendorf tube containing 800 μl of PBS buffer 10% and the supernatant removed after centrifugation. The pellet was suspended with 300 μl of PBS buffer 10%. The resulting mixture was used for the DNA extraction which was performed according to the manufacturer's protocol (Sacace Biotechnologies^® Como, Italy). The extracted DNA was stored at -20°C prior to use.

Real-time PCR analysis

Real-time PCR was performed using thermal cycler 7500 Fast Real Time PCR System (Applied Biosystems/Life Technologie, California, US). The total volume of the reaction mix for the amplification was 25 μl consisting of 12 μl of Taq Man Universal Master Mix, 3 μl of the probe (0.25 μM), 1 μl of each sense and anti-sense primer (50 μM) (Applied Biosystem, California, USA), 3 μl of water (DNAse-RNase free) and 5 μl of extracted DNA. The gene of the small subunit ribosomal RNA (SSU rRNA) was amplified using a (5'-GAC GGA GGC ATT CAA CGG TT-3') forward primer and a (5'-TTG CCA TCC GCG GTG G-3') reverse primer. The amplicons were detected using the sequence FAM-CCC GCG GCG CTC GTC GCT AG-BHQ as probe. The DNA amplification program was set up as previously describes by Mejia et al. [24], as follows: after preheating at 60°C for 1 min, a cycle was performed at 95°C for 10 min followed by 45 cycles at 95°C during 15 s, 60°C for 1 min and a final extension at 60°C for 1 min.

The analysis was done using 7500 FAST software v2.0.6 (Applied Biosystems/Life Technologie California, USA).

Statistical analysis

Data were entered in using Excel 2013 software (Microsoft Corp, Redmond, WA) and analysed in SPSS Version 20 software (SPSS Inc., Chicago, USA). Pearson's Chi2-test was used to study the association between G. intestinalis (microscopy) as the dependent variable and the PCR results and the socio-demographics and socioeconomic outcomes as independent parameters. To explore the presence of G. intestinalis (as assessed by real-time PCR) and associations with age, sex, level at school, nutritional status of schoolchildren, educational level of the household head, presence of a latrine in the residence and availability of drinking water, a binary logistic regression was performed. The Odds Ratio (OR) was estimated with a 95% Confidence Interval (CI) and P-values were calculated. The results were considered significant for P<0.05. Considering microscopy as ‘gold standard’, the sensitivity and specificity of real-time PCR were assessed according the TDR Diagnostics Evaluation Expert Panel [25].

The study was conducted as part of the VgtS case study [21] which received the approval of the ‘Ethikkommission Nordwest-und Zentralschweiz’ in Switzerland (EKNZ, Reference no. 2014-161), the Ethics Committee for Health Research of Burkina Faso (Reference no. 2015-02-026) and of the Institutional Ethics Committee of CERBA/ LABIOGENE, University of Ouaga I Joseph Ki Zerbo. The study is registered with the clinical trial registry ISRCTN (Identifier: ISRCTN17968589).

Informed consent of each care-givers was obtained and children assented orally before starting the investigation. Study participation was voluntary and children could withdraw anytime without further obligation. All infected schoolchildren received free treatment according to national guidelines.

Characteristics of the study population

The 147 schoolchildren enrolled for the real-time PCR study had an average age of 11 years (range 8 to 15 years), and 57.1% of them were male. Among the 94 schoolchildren with microscopy-positive results, 59.6% were male, 37.2% were malnourished, 40.4% had a latrine in their residence and 51.1% were in Grade 4. Three of the children with confirmed giardiasis (3.2%) had diarrhoea in the last two weeks preceding the survey. Among the 53 schoolchildren with microscopynegative results, 52.8% were male, 24.5% were malnourished, 62.3% had a latrine in their residence and 75.5% were in class of Grade 4.

The nutritional status of the 147 children was evaluated using the body mass index (BMI) for age as defined by the World Health Organisation (WHO) in 2007 [26]. According to the corresponding WHO score 32.7% (48/147) of the children were per definition underweight or malnourished represented by 6/19 (31.6%), 4/13 (30.8%), 6/14 (42.9%) and 11/24 (45.8%) schoolchildren from the schools in Doure A, Goundi B, Ipendo A and A Tio in the Centre Ouest Region and 7/19 (36.8%), 4/19 (21.1%), 5/21 (23.8%) and 5/18 (27.8%) schoolchildren from the schools in Linoghin A, Loumbila A, Tangzougou and Wavoussé A in the Plateau Central region. No statistically significant association was found between malnutrition and school. The binary logistic regression showed that schoolchildren without access to a latrine in their residence had a higher prevalence of G. intestinalis than those which did have access (OR 1.7; 95% CI 0.8– 3.8). This trend was, however, not significant.

Detection of Giardia intestinalis by direct microscopic examination of stool samples

According to the results of the microscopy-based diagnosis of stool samples from 441 schoolchildren, 120 children were found to harbour G. intestinalis (i.e. cystic forms or trophozoïtes), which represents an overall prevalence of 27.2%. Among them, 94 samples showed cystic forms of G. intestinalis while 26 samples contained vegetative forms.

The prevalence of G. intestinalis was 33.9%, 27.8%, 23.5% and 30.2% in the schools of Doure A, Goundi B, Ipendo A and Tio A in the Centre Ouest region and 20.0%, 30.9%, 19.6% and 23.2% in the schools of Linoghin A, Loumbila A, Tangzougou and Wavoussé in the Plateau Central region.

As previously reported, despite a varying intensity, G. intestinalis infection showed no statistically significant difference between schools (P=0.2) or between regions (P=0.1) [27]. Co- and multiple infections were significantly more frequent in boys (86%) than girls (77.1%) (P=0.038).

The rate of concordance between our results and those of the teaching hospital was very high, ie 95.3% (41/43). Indeed, of the 43 samples of which 09 positive and 34 Negative to Giardia intestinalis that were subjected to this external quality control, 1 was negative and another positive respectively.

G. intestinalis and Entamoeba histolytica/dispar showed the strongest association (P=0.025) (Table 1) followed by G. intestinalis and Entamoeba coli (34%). No significant association was found between the presence of multiple infections and underweight in schoolchildren (P=0.585).

Table 1: Coexistence of Giardia int 461 estinalis and other frequent parasites (significant^*).

The educational state of the head of a household seemed to influence the infection prevalence among schoolchildren. A large proportion of the infected children (81.4%) were associated with households where the head had not received any formal schooling. A proportion of 12.4% of infected children belonged to parents that had received primary education and 6.2% were associated with parents that had completed secondary school. However, no statistically significant difference was found. A significant association was demonstrated for G. intestinalis presence and availability of a latrine at the household (P=0.002) which was the case in less than half of all households (39/97). Most of the households assessed (87/97) had access to drinking water and no difference was observed in the infection prevalence of schoolchildren with and without this access.

Giardia intestinalis detection by real-time PCR

The results of the DNA amplification of Giardia intestinalis cysts by real-time PCR are given in the form of fluorescence intensity curves expressed as a function of the number of cycles (Figure 1).

Figure 1: Graph of real-time PCR amplification (ΔRn: Amplitude of the emitted fluorescence intensity.

Real-Time PCR confirmed the presence of G. intestinalis DNA in 76.6% (72/94) of the microscopy-positive cases. The sensitivity of realtime PCR ranged from 69.2% to 94.1% in the Centre Ouest region and 58.3% to 85.7% in the Plateau Central region (Table 2). Of the 53 microscopy-negative samples for G. intestinalis , 51 were confirmed as negative using real-time PCR. The overall specificity of the real-time PCR was 96.2% ranging from 95.4 % in the Centre-Ouest region to 96.7% in Plateau Central region (Table 2). No statistically significant difference was observed between a PCR positive status and schools of the two regions. However, a significant difference (P=0.032) was observed for G. intestinalis prevalence according to region, with Centre-Ouest having a higher rate. Furthermore, no significant association was observed between G. intestinalis positive samples, as assessed by real-time PCR, and age, sex, nutritional status, the school attended, the level of parent education, and availability of a latrine at home or availability of safe drinking water at home.

Table 2: Sensitivity and specificity results of the real-time PCR method to detect Giardia intestinalis (Overall sensitivity: 76.6; Overall specificity: 96.2).

In low and middle income countries, the estimation of the prevalence of parasitic diseases is often complicated by the absence of reliable statistical data owing to a lack of effective diagnostic tools. Despite being considered a major cause of parasite-related diarrhoea worldwide [28] and being included in the WHO ‘Neglected Diseases Initiative’ in 2004 [3], national prevalence of G. intestinalis are still not available in many African countries.

The aim of this study was to evaluate the performance of real-time PCR in detecting G. intestinalis in microscopy-positive stool samples for the first time in Burkina Faso. The overall observed prevalence of G. intestinalis among school-aged children was 27.2% which is in line with previous estimates between 20% and 30% in developing countries [5]. The children infected with G. intestinalis showed no apparent clinical signs of the disease which might be explained by a high proportion of cysts compared to trophozoïtes.

Given that microscopy is deemed the best available diagnostic method for G. intestinalis infection at the moment, the sensitivity found in previous studies for PCR-based detection of G. intestinalis differs wildly and ranges from 13-100%. In the current study, we have found a specificity of PCR ranging from 91-100% across the eight different schools.

High sensitivity of a PCR-based approach has been shown previously by Schuurman et al. who detected the presence of a single cyst by real-time PCR [29].

Within a study conducted in the Netherlands for instance, 100% of 20 microscopy-positive stools were confirmed positive by multiplex PCR [30]. These authors had reported 10/34 real-time PCR positive on microscopy-negative samples.

Using real time PCR, Verweij et al. detected 102 positive samples among 104 microscopy-positive samples that contained visible G. lamblia cysts, resulting in a sensitivity of 98.1% [19]. In Bangladesh Ng et al. reported 85% detection rate of G. intestinalis using multiplex real-time PCR from patients with diarrhoea [31]. Also, among subjects from Timor and Cambodia, Llewellyn et al. observed a detection rate of 65.7% (46/70) using multiplex real-time PCR [32]. In Morocco, from 84 samples only 11 were PCR positive (13.1%) using semi nested PCR targeting a sequence of glutamate gene dehydrogenase (gdh), and only 9 were positive (10.7%) to the nested PCR targeting the 18S rRNA gene [7].

Thus, the sensitivity of real-time PCR in our study was in line with those of previous studies. We showed an overall sensitivity of 76.6% ranging from 58.3% in the school of Linoghin A to 94.1% in the school of Doure A. The overall specificity of the real-time PCR was 96.2% ranging from 95.4 % in the Centre-Ouest region to 96.7% in Plateau Central region.

We also found 2 real-time positive cases among the 53 microscopynegative stool samples. Here, considering microscopy as ‘gold standard’, these unmatched cases could be taken for false positive. It is possible that we have overestimated the microscopy positive cases of G. intestinalis in our study, including potential false positives.

The lack of PCR detection of the 22 microscopic positive samples suggests an absence of target DNA and could be explained by lysis of cysts followed by a possible degradation of the parasite DNA by the DNase contained in faeces during sample storage [33]. Indeed, no preservatives have been added to the samples during the storage period of storage at -20°C in contrast to other study protocols [33,34]. In addition, PCR-based DNA amplification can be inhibited by the decomposition products of haemoglobin in stool such as bilirubin, bile acids and inorganic ions [33,35]. In our study, we have not provided special purification cysts prior to DNA extraction, while other authors advocate such a procedure [36]. Indeed, in such a context where the likely rate of PCR inhibition is high (23.4%), the optimization of the DNA extraction protocol to eliminate inhibitors was shown to increase yield [30,37]. However, the low concentration of G. intestinalis in stool samples could also be one reason for non-amplification. Even if realtime PCR can detect a single cyst [29], it is still necessary that the cyst is contained in the stool sample 0.1 g collected for DNA extraction. Sometimes, we have identified only a few cysts of G. intestinalis but only after the concentration of the stool by the formalin-ether technique. Regarding the real-time PCR on G. intestinalis , Bertrand et al. showed that for an initial concentration of 180 cysts per 200 μl of preparation, only 33% of positive responses were observed [38,39]. Indeed, a relatively low sensitivity (93.4%) was observed by targeting sequences of the small subunit gene of rRNA [40] as in our study.

PCR has major benefits to significantly increase the speed, the sensitivity and specificity for detection of microorganisms. It is, however, sensitive to a large number of inhibitory compounds which can lead to an underestimation of cysts and false negative results.

For the first time in Burkina Faso, real-time PCR, not yet routinely available in African laboratories, was used to detect G. intestinalis from cysts. Our data show that real-time PCR can be of great help in the diagnosis of giardiasis. However, it needs to be further optimized to become a more sensitive tool for G. intestinalis diagnosis in low income settings.

We found a strong association between infection of G. intestinalis and E. histolytica/dispar . In future studies, it would be interesting to use a multiplex PCR to simultaneously detect these protozoa on the same sample.

None declared

SD, TK, SE, PO, JU, PO, GC and JS conceived the study, TK, TRC, AZ, GT, BS performed the experiments, SD, JS and GC contributed for reagents and materials of the experiments, SD, TK, SE, JS analyzed the data and wrote the manuscript. All the authors reviewed drafts and approved the final version of the manuscript.

We are grateful to the SDC, Berne, Switzerland. This work is part of the VGtS project supported by the Swiss Agency for Development and Cooperation. We also thank the West African Economic and Monetary Union for financial support through the program PACER2 to CERBA/ LABIOGENE.

We thank schoolchildren and their parents, schools teachers, educational and health authorities, in the Plateau Central and Centre Ouest regions for their active participation in the study.