Integrated monitoring and evaluation and environmental risk factors for urogenital schistosomiasis and active trachoma in Burkina Faso before preventative chemotherapy using sentinel sites

Schistosomiasis and trachoma are among the most prevalent of the so-called "Neglected Tropical Diseases (NTDs)", an umbrella term that encompasses a group of parasitic, bacterial, and viral infections collectively imposing a similar disease burden to that of malaria and HIV [1, 2]. NTDs are widespread in sub-Saharan Africa [3], where they affect more than 500 million people [1]. These NTDs are both the result of, and a major contributor to the poverty of many rural and some disadvantaged urban populations in tropical regions of the world [4]. However, the fact that some NTDs are treatable with affordable (or donated), safe and effective drugs has encouraged the implementation of 'preventative' chemotherapy control programs. Although these drugs do not prevent such infections, they help prevent or minimize the burden of disease that would ensue if left untreated.

According to the World Health Organization (WHO), more than 206 million people are infected with parasites of the genus Schistosoma (of whom 120 million have symptoms and around 20 million show severe disease sequelae) [5]. Schistosomiasis is an indirectly (snail)-transmitted disease whose distribution is particularly sensitive to environmental changes, including changes of human origin. Transmission of the parasite is highly focal, with heterogeneity reflecting numerous human and snail host as well as environmental factors [6]. Identifying the broad scale patterns of schistosomiasis is crucial because ecological and climatic changes plus human migration may have led to changes in the prevalence and distribution of the disease in different parts of endemic countries [7]. The current mainstay for schistosomiasis control is to mitigate against the burden of the disease by controlling morbidity through chemotherapy with praziquantel (PZQ). It is recommended that mass drug administration (MDA) with PZQ be delivered to total at-risk populations defined where a survey shows prevalence of over 50% in school-aged children and to children aged 6-16 years where survey prevalence is between 10% and 50% in this group [8]. (Although the indicator group for schistosomiasis comprises school-aged children, pre-school-aged children can also be infected, the implications of which are considered below). Because schistosomiasis exhibits strong spatial heterogeneity at local levels, there is a need to identify and locate high-risk communities or schools that require MDA [7] so that cost-effective management of limited resources can be achieved.

Trachoma is the leading infectious cause of blindness worldwide [9]. Around 80 million people have active trachoma, an inflammatory early stage of the disease caused by recent infection with the ocular bacteria Chlamydia trachomatis, with the young children bearing the heaviest burden [10]. Repeat infection leads to scarring of the eyelid, which causes the eyelids to turn inwards and subsequently cause the eyelashes to scratch the cornea, ultimately leading to blindness (WHO estimates that 6 million people globally are blind from trachoma). The WHO endorses a four-pronged approach to eliminate blinding trachoma by the year 2020 known as SAFE: Surgery for trichiasis, Antibiotic against infection, Facial cleanliness, and Environmental sanitation [11]. Currently, Pfizer Inc. donates millions of doses of the antibiotic azithromycin (Zithromax, ZTX) for trachoma control. MDA with ZTX or tetracycline eye ointment is at present recommended for entire endemic communities to be conducted annually for three years wherever the prevalence of active trachoma in children of ages 1-9 years old at the district level is above 10% [11]. Trachoma is associated with individual and environmental risk factors as well as climatic conditions which may be important determinants of trachoma transmission (which in some cases can occur mechanically through flies) [12–15]. These may vary between settings and hence the need for studies aimed to identify risk factors relevant to specific environments. Indeed, understanding risk factors is essential in designing appropriate interventions for the 'F' and 'E' components of the SAFE strategy (e.g. water availability for Facial cleanliness and Environmental improvements) [16].

Based on the argument of co-endemicity of NTDs within countries [17], global advocacy for the amelioration or elimination of the disease burden imposed by these conditions has recently increased, with a renewed interest in the scaling-up of integrated NTD interventions in sub-Saharan Africa, mainly using preventative chemotherapy distributed through MDA [18]. More particularly, MDA is a relatively efficacious tool to control the infection and morbidity caused by the helminthiases and trachoma and such an approach offers opportunities for integration. There is also an interest in trying to deliver some of treatments through school-based programmes, and again schistosomiasis and trachoma lend themselves to this approach as the burden of infection is mostly concentrated among the young.

An important question in integrated NTD control relates to the spatial scale at which the programs should be integrated. This will depend on the geographical distribution and prevalence of both single and co-infections. Therefore, the principal aim of the current study was to identify whether combined survey designs based on sentinel sites are useful monitoring and evaluation (M&E) tools for the assessment of integrated schistosomiasis and active trachoma control programmes. A secondary aim was to establish the distribution of these two NTDs in these selected sentinel sites within Burkina Faso pre- and post-treatment (post-treatment data are not presented here). Finally, a post hoc analysis studied associations of urogenital schistosomiasis infection and active trachoma disease with demographic and environmental variables; such results should inform the construction of reliable disease maps [19] and mathematical modeling efforts of these two NTDs. In addition, these results should help highlight potential factors related to the physiology and behavior of mechanical vectors of these two NTDs which ultimately could inform policies aimed at reducing transmission of these diseases.

A total of 3,324 children aged 7 to 11 years were surveyed from the 21 sentinel sites (Figure 1). Observed prevalence values by sentinel site of single and dual S. haematobium and trachoma signs are illustrated in Table 2.

Figure 2 illustrates MDA recommendations for S. haematobium and trachoma signs for the 21 sentinel sites from observed data on the prevalence of the two diseases. Our data suggest that in only two sentinel sites would both PZQ and ZTX be required. In addition, Figure 2 shows that to the north (i.e. in the driest areas) of Burkina Faso the prevalence of S. haematobium infection was between 10% and 49% (i.e. medium) but that of trachoma was less than 10% (i.e. low).

Table 2 contains summary statistics for the interpolated and meteorological station-based values of the environmental variables. For instance, for altitude, minimum (min) and maximum (max) values were 259.0 and 440.8 MSL respectively; for precipitation, min and max values were 1.53 and 2.98 mm; for max temperature, min and max values were 34.0 and 37.3°C: for min temperature, min and max values were 21.8 and 23.3°C; for average temperature, min and max values were 27.5 and 29.5°C and finally for air pressure, min and max values were 961.0 and 978.5 mbars. The distributions of the interpolated values appear to be similar to the distributions of the original station-based data.

Through sentinel site surveillance defined at the school level, this study showed that urogenital schistosomiasis and trachoma surveys can be successfully implemented simultaneously. In addition, the prevalence of schistosomiasis infection and trachoma signs in these sentinel sites before integrated treatment in Burkina Faso was estimated. Results indicated that both of these diseases are prevalent in the country and they likely do constitute public health problems according to WHO recommended thresholds [15, 20, 34–36]. Hierarchical binomial logistic regressions identified, as may be predicted, these diseases to be independent (conditional on environmental factors), which further highlight the different mechanisms of their transmission. Treatment with both PZQ (for urogenital schistosomiasis) and ZTX (for trachoma) was only required, however, for two of the 21 sentinel sites surveyed here.

We chose sentinel site surveillance conducted in schools because in the case of schistosomiasis the selected methodology is ideal for the following reasons [37]: (i) schools are accessible and receptive; (ii) the highest prevalence levels of S. haematobium infection are found among school-age children; (iii) data collected in this age range may be used to evaluate not only if schistosomiasis threatens the health of school-aged children, but also if there is need for intervention in the community as a whole; (iv) children in intermediate grades (generally between ages 9-12) allow for the accompaniment of treatment impact over one to two years, before they leave school. In addition, we chose to recruit a longitudinal cohort of 7-11 years old at baseline for both logistical reasons (through integrating with the ongoing surveys) and as this would enable us to follow-up these same children during three years. Furthermore, our past experience with SCI data has demonstrated that morbidity is also detectable in young school-aged children, which suggests that children may become infected early in life [20, 38, 39]. As we were interested ultimately to monitor changes in the infection and to quantify the rate at which infections change, we suspect that changes in morbidity in the younger age groups of school children would be more sensitive to changes in the force of infection (i.e. per capita rate at which a host acquires new infections) by MDA. However, we recognize that if we had also included data from the even younger age groups (i.e. pre- school children) who did not receive schistosomiasis treatment in the monitoring programme, our understanding of the secular changes in transmission caused by a morbidity control programme could be improved, helping also differentiate between the impact of PZQ on the individual and the community [40].

School-based surveys might underestimate the prevalence of trachoma in the community as a whole and they certainly do not determine the priority for surgical intervention [25]. In addition, a key feature of trachoma epidemiology is that of the age-profile of infection prevalence, which increases to a peak at very young ages and declines at older ages (particularly in heavily infected communities). This feature would not be captured in study designs such as that used in the present study [16, 41–43]. Trachoma interventions using WHO thresholds were established for a different age group than the one analyzed here and thus we are aware that inference on this matter should be treated with some caution. Nevertheless, as the distribution of infection and disease will vary according to endemicity levels, the protocol applied in this study could be of use to hypoendemic trachoma areas of Burkina Faso if MDA treatment decisions have to be made based on results arising from similar study designs but not for planning where SAFE implementation is needed as this is based on district-level prevalence of TF among the preferred 1-9 year age group. We draw this conclusion because it has been observed in hypo- and meso-endemic trachoma areas the infection prevalence data come to a peak in ages slightly greater than 5 years [41].

In this school-based approach a significant proportion of the school-aged children may not attend school and so the practicality of this approach may be limited to areas where school attendance is not too low. In Burkina Faso the percentage of attendance for the primary schools (children aged 7-13 years old) is approximately 63% and thus we might have missed a significant proportion of S. haematobium infection and trachoma signs. Moreover, observer error might have compromised the sensitivity and specificity of clinical diagnosis, although the magnitude of this effect is likely to be smaller [44].

Environmental effects on the emergence and re-emergence of pathogens are well recognized and thus a secondary aim of this study was to characterize quantitatively the effect of specific environmental factors in addition to demographic factors on the two NTDs of interest here: urogenital schistosomiasis and trachoma.

Altitude, precipitation, temperature and air pressure were all important predictors of S. haematobium infection (Table 3) by simple hierarchical logistic regression modelling. Temperature and rainfall were important predictors of prevalence of S. haematobium infection also in Tanzania [45] and such statistical relationships are consistent with the known biology of freshwater snails, the intermediate hosts for schistosomes. S. haematobium is transmitted by snails from the genus Bulinus. In Burkina Faso B. truncatus and B. senegalensis occur in all endemic schistosomiasis areas of the country [35, 46] and their dynamics are known to be sensitive to temperature and rainfall [29]. However, multivariable analysis indicated only altitude to be negatively significantly associated with prevalence of S. haematobium infection.

There were significant negative associations of prevalence of trachoma signs with minimum temperature and air pressure (Table 4). The significant negative association between minimum temperature as well as air pressure and trachoma signs in our study, strengthens the argument that environmental factors, and particularly climatic conditions, may be important determinants of the physiology and behavior of Musca sorbens, a mechanical vector of trachoma in some settings [10, 13, 15]. Even if these are speculative, the following mechanisms can also be envisaged [14]: Preferred breeding sites of M. sorbens are fresh human faecal materials lying on the ground. High temperature and sunshine may induce rapid dryness of faecal materials, causing them to become inadequate breeding sites. Moreover, the lifespan of M. sorbens strongly depends on temperature, decreasing from 35 days at 24°C to less than 12 days at 32°C, yielding less frequent contacts of M. sorbens with children's faces, and in turn, lower contamination of M. sorbens with C. trachomatis [47, 48]. We did not find any significant association between age and trachoma prevalence and we suspect that this is due to the fact that pre-school aged children are not included in the current surveys. Conditional upon these predictors, these data are consistent with the two pathogens being independent.

As mentioned previously, the present study was designed to be as simple as possible in order to increase participation and to make it logistically manageable in settings with few resources. We therefore did not collect observational data, such as water-use behavior, evidence of latrine use and hygienic habits/states. Perhaps an investigation of the relationship between fly density, markers of socio-economic status, hygienic habits/states and altitude in these sentinel sites could be informative. Such facts may have led us to missing more subtle or complex relationships between risk factors for urogenital schistosomiasis and active trachoma [14] in the current study.

We believe that despite the potential limitations contained in terms of age groups selected for the assessment of S. haematobium infection and active trachoma, the fact that there was found very low prevalence of co-infections in the studied sentinel sites, carries important implications for integrated NTDs surveillance and control strategies. The presence of relatively few co-infections has clear consequences for the concept of integration of mapping, planning of drug delivery, and M&E of integrated control programs for these two NTDs. Results of the current study are in accordance with observations reported in the Plateau and Nasarawa States in Nigeria [17]. We agree with these authors' recommendation that current population-based district level surveys for the mapping of trachoma should not be replaced by school-based surveys. Finally, given that NTD interventions require baseline data for targeting treatment and evaluation, the benefits of including more than one disease indicator in overlapping age groups when using an integrated M&E protocol, warrants further study. In particular, different study designs would require further investigation. More research would also be needed to ascertain whether the findings presented here can be generalized to West Africa and beyond.

To our knowledge, very few previously published studies-three in Nigeria, Southern Sudan and Togo [17, 23, 49] and another more recent one offering a conceptual framework around these issues [50]-have discussed and tested different survey methodologies for M&E of different NTDs in integrated control programmes. The current study aims to bridge some of these gaps and stimulate further research around the methodologies for M&E for different NTDs by reporting on baseline findings from sentinel site surveillance of S. haematobium and trachoma signs with an integrated M&E protocol in Burkina Faso. We conclude that the current approach may be useful to monitor and evaluate integrated programs including trachoma, but not for planning where SAFE implementation should occur. In addition, we would recommend that current methodology to be modified to assess the younger age groups both for schistosomiasis and trachoma so that literature gets enriched from such data in the case of schistosomiasis, and at the same time key features of trachoma epidemiology can be captured.