INTRODUCTION
Schistosomiasis, caused by the parasitic blood flukes Schistosoma haematobium, S. mansoni, S. intercalatum, S. mekongi and S. japonicum, remains one of the most prevalent human infectious diseases. An estimated 207 million persons are currently infected worldwide, over 85% of which live in sub-Saharan Africa (Steinmann et al. Reference Steinmann, Keiser, Bos, Tanner and Utzinger2006). Invasion, migration and maturation of the parasite may initially induce an acute feverish disease phase (acute schistosomiasis), followed (up to years) later by a chronic disease phase (chronic schistosomiasis). The chronic phase can be further divided into early and late stages of disease. Each successive phase represents the outcome of the host immunological responses to schistosome egg deposition in tissues, which induce pathological changes, manifested as signs and symptoms. Acute intestinal schistosomiasis (S. mansoni and S. japonicum, as well as S. intercalatum and S. mekongi) may cause fever, abdominal pain, bloody diarrhoea and tender hepato-splenomegaly. In their chronic form, these intestinal species may cause obstruction to blood flow, liver fibrosis, splenomegaly, bleeding from oesophagal varices, haematemesis, ascites and, in the case of S. japonicum, liver cancer. Urinary schistosomiasis (S. haematobium) may manifest as dysuria and haematuria in both acute and chronic stages, and cause renal colic, hydronephrosis, renal failure and bladder cancer in the late stages of disease progression.
The potential severity, to the extent of mortality, of schistosomiasis-associated pathologies is well known. For example, hospital records over a 30-year period in the Lake and West Nile regions of Uganda have shown S. mansoni to be the leading cause of hospital deaths (Williams, Hayes and Smith, Reference Williams, Hayes and Smith1986), and more recently, 29% of deaths in a one year period were attributed to haematemesis or ascites likely to be due to infection with S. mansoni (Kabatereine et al. Reference Kabatereine, Tukahebwa, Kazibwe, Namwangye, Zaramba, Brooker, Stothard, Kamenka, Whawell, Webster and Fenwick2006). Yet such schistosome-associated mortality is only the tip of the iceberg in terms of the health impact upon those infected. According to the ‘Global Burden of Disease’, schistosomiasis caused the loss of 1·7 million disability-adjusted life years (DALYs) worldwide in 2001, of which 82% (1·4 million DALYs) were lost in sub-Saharan Africa alone. Schistosomiasis was the third largest cause of the tropical disease cluster burden (which does not include malaria) worldwide and caused one quarter of the tropical disease cluster burden in sub-Saharan Africa alone. Moreover, it is now generally accepted that such calculated DALY figures for schistosomiasis represent a serious underestimate and do not fully capture the true magnitude of the disease (King, Dickman and Tisch, Reference King, Dickman and Tisch2005). Indeed, more recent studies have used evidence-based associations between schistosomiasis and its known morbidity-related sequelae to project a revised disability weight of up 2–24% for the different forms of infection (including the most virulent S. japonicum), which indicates a 4–50-fold greater DALY burden for schistosomiasis than previously reported, with revised estimates on the scale of 3–70 million DALYs lost to schistosomiasis (reviewed in King and Dangerfield-Cha, Reference King and Dangerfield-Cha2008).
With the 2001 World Health Assembly resolution (54.19) and increasing affordability of broad spectrum, effective and safe single-dose anthelminthics, there has been renewed interest and commitment to implement national control programmes (NCPs) directed against schistosomiasis and intestinal worms. A central foundation of contemporary control is based upon the delivery of anthelminthics, often en masse (mass drug administration, MDA), to communities known, or suspected, to be at high-risk of infection, as well as promotion of accessibility to such drugs at ‘first-point-of-call’ health centres. Based upon this existing template and rationale, the ‘Schistosomiasis Control Initiative’ (SCI), established in 2002 with funding from the Bill and Melinda Gates Foundation, has assisted several selected sub-Saharan African countries to develop and implement schistosomiasis, together with soil-transmitted helminthiasis (STH), morbidity control programmes (Fenwick et al. Reference Fenwick2009; see this special issue). These initial SCI countries included Burkina Faso, Mali, Niger, Tanzania including Zanzibar, Uganda and Zambia, and more recently also Rwanda and Burundi. The strategy included the donation to the countries of praziquantel (PZQ) for MDA for schistosomiasis and co-administration, where appropriate, of albendazole for STH, combined with the provision of health education. A primary objective of the SCI-supported control programmes was to achieve, and hence also demonstrate, a quantifiable reduction in schistosome-associated morbidity as a consequence of such chemotherapeutic intervention. Inherent within such an objective, it is imperative to define and validate direct and indirect indicators of schistosome-related morbidity (Hatz et al. Reference Hatz, Savioli, Mayombana, Dhunputh, Kisumku and Tanner1990) and to characterize pre-treatment baseline morbidity levels within the at-risk populations accurately, so that any subsequent changes in morbidity due to chemotherapy can be accurately determined. Furthermore, identification of sensitive and specific indicators of schistosome-associated morbidity that can be logistically and financially implemented within such large-scale control programmes, as distinct from individual clinical or research-based settings, should also prove invaluable in assisting governments and authorities in identifying target populations for ongoing and future intervention (Wiest, Reference Wiest1996; Bergquist, Johansen and Utzinger, Reference Bergquist, Johansen and Utzinger2009). Evaluation of such markers may also be beneficial in terms of determining their predictive value during different stages of a schistosomiasis control programme (Campagne et al. Reference Campagne, Garba, Barkire, Vera, Sidiki and Chippaux2001), with markers ideally having the potential to predict recrudescence of morbidity after treatment (Richter, Reference Richter2003). Indeed, the more successful a control programme becomes, the more crucial the need for an accurate assessment of the epidemiological and clinical situation becomes (Bergquist et al. Reference Bergquist, Johansen and Utzinger2009). Nevertheless, throughout the history of schistosomiasis control programmes in general, the tendency has been to emphasize the drug treatment itself, whilst the importance of quality-assured diagnostic tests, for infection prevalence, intensity and/or associated morbidity, has often been neglected (Ridley, Reference Ridley2006).
To document such specific morbidity reduction may not be straightforward. Schistosome infection is not equivalent to schistosomiasis disease. Interactions between the parasite, the host and the environment determine the pace and severity of schistosomiasis disease progression (Blanton et al. Reference Blanton, Salam, Kariuki, Magak, Silva, Muchiri, Thiongo, Abdel-Meghid, Butterworth, Reis and Ouma2002; Ross et al. Reference Ross, Bartley, Sleigh, Olds, Li, Williams and McManus2002). There is currently no agreement regarding the list of sequelae that should be included in any schistosome morbidity assessment (Michaud, Gordon and Reich, Reference Michaud, Gordon and Reich2004). Assessing morbidity may be particularly difficult for S. mansoni infections in school-aged children, the main target group of the SCI and many other MDA programmes. In contrast to urinary schistosomiasis, where gross haematuria is an obvious early visual sign of morbidity, morbidity associated with intestinal schistosomiasis is often less obvious in all but the most severe or chronic cases. Even at patient presentation, the often cryptic symptomology of intestinal schistosomiasis confounds further interrogative or palpation methods of diagnosis. Furthermore, few, if any, of the clinical manifestations of schistosomiasis are specific, being potentially related to other causes, in particular intestinal nematode infections and malaria, which often co-exist with schistosomiasis.
Novel and improved ways of measuring morbidity associated with schistosomiasis, especially in school-aged children, and particularly after treatment when egg excretion declines, are therefore needed. Vennervald and colleagues (Reference Vennervald, Kahama and Reimert2000) set down seven points for consideration during the evaluation of a new marker which should ideally: (1) reflect morbidity as well as intensity of infection, thus providing additional information to egg counts; (2) vary little with respect to the time of day that the biological sample is collected; (3) give a reasonable estimate of the prevalence of disease using only one sample collected from individuals within the population under study; (4) measure a continuous variable; (5) be able to distinguish between early morbidity and later or more severe morbidity; (6) be non-invasive, and preferably measurable in urine; and (7) be assayed directly in the field. Such measures should also be as inexpensive as possible if they are to be realistically incorporated into any MDA programme.
Here we discuss a range of both traditional and novel morbidity measures performed within our SCI activities, evaluate their relative advantages and disadvantages for such large-scale MDA programmes, and provide recommendations for ongoing and future morbidity-related monitoring and evaluation activities.
INFECTION INTENSITY
For S. mansoni eggs, microscopic examination of faecal smears, using the standardized thick smear (41·7 mg) Kato-Katz (KK) method, remains the World Health Organization (WHO) accepted ‘gold standard’ for routine epidemiological surveillance. S. mansoni intensity of infection is classified according to WHO guidelines as either infection-negative, 1–99 eggs per gram faeces (EPG) as ‘light’ infection; 100–399 EPG as ‘medium’ infection and a ‘heavy’ infection as 400 or more EPG. For S. haematobium, filtration of urine and direct parasitological examination are used to determine infection status and intensity, where an absence of eggs is classified as infection-negative, 1–49 eggs per 10 ml of urine as ‘light’ infection and a ‘heavy’ infection is defined as 50 or more eggs present in a 10 ml urine sample (WHO, 1998).
SCI activities to date have generally reported dramatic reductions in mean parasite infection intensity following treatment implementation. For example, mass treatment with PZQ within Uganda has led, in relation to baseline, to a significant 70% decrease in the S. mansoni EPG counts after one year and 82% after two years of treatment (Kabatereine et al. Reference Kabatereine, Brooker, Koukounari, Kazibwe, Tukahebwa, Fleming, Zhang, Webster, Stothard and Fenwick2007). Within Burkina Faso, again in relation to baseline, there was on average an overall significant decrease in the S. haematobium egg counts by 99% one year post-treatment and 93% by two years post-treatment (Koukounari et al. Reference Koukounari, Gabrielli, Touré, Bosque-Oliva, Zhang, Donnelly, Fenwick and Webster2007; Touré et al. Reference Touré, Zhang, Bosque-Oliva, Ky, Ouedraogo, Koukounari, Gabrielli, Sellin, Webster and Fenwick2008). However, for S. mansoni infection in Burkina Faso (in Sud Ouest region only, and caution may be required as all S. mansoni sample sizes were small) observed arithmetic mean intensities did not show a decrease one year post-treatment, although they did show an 87% decrease by two years post-treatment (Koukounari et al. Reference Koukounari, Gabrielli, Touré, Bosque-Oliva, Zhang, Donnelly, Fenwick and Webster2007; Touré et al. Reference Touré, Zhang, Bosque-Oliva, Ky, Ouedraogo, Koukounari, Gabrielli, Sellin, Webster and Fenwick2008). In Niger, the observed overall mean intensity of S. haematobium infection was reduced by 83% after one year of mass treatment (SCI unpublished data) although in three schools little or no change in infection intensities following mass treatment has been reported (Tohon et al. Reference Tohon, Mainassara, Garba, Mahamane, Bosqué-Oliva, Maman-Laminou, Duchemin, Chanteau and Boisier2008).
However, one must acknowledge these KK and urine filtration measurements may not be reliable in terms of accurately measuring schistosome intensity (Booth et al. Reference Booth, Vounatsou, N'Goran, Tanner and Utzinger2003; Carabin et al. Reference Carabin, Balolong, Joseph, McGarvey, Johansen, Fernandez, Willingham and Olveda2005). Due in part to the overdispersed nature of schistosome eggs in stool and daily variation in excretion, the ideal protocol is to use replicate faecal/urine samples over several (ideally a minimum of three) consecutive days. Unfortunately such demands are rarely logistically or financially feasible within large-scale control programmes in comparison to smaller-scale research settings and SCI have attempted to use duplicate KK thick smears taken from either a single day's stool (Uganda) or single filtrations from a single day's urine sample or, where possible, over two days (Tanzania and Niger). One solution to this is to use statistical methods to improve accuracy in the parasitological estimates obtained. For example, Booth and colleagues (Reference Booth, Vounatsou, N'Goran, Tanner and Utzinger2003) used latent class models to estimate single and dual infections with S. mansoni and hookworm from 101 schoolchildren from a rural area of Côte d'Ivoire over five consecutive days and using five KK readings from each stool specimen. The cumulative prevalence of co-infections was estimated at approximately 10% after reading slides on the first day, and 57·0% after reading all 25 slides per person. The sensitivities of a single KK thick smear for detection of S. mansoni alone, hookworm alone, or S. mansoni-hookworm co-infection were 22%, 8% and 18%, respectively. A similar study in the Philippines, although using faecal samples of animals infected with S. japonicum, further attempted to determine the sensitivity and specificity of one, two, three, four or five consecutive days of stool samples. Results revealed a sensitivity of up to 96% when three or more stool samples were collected on three separate days and a specificity of at least 91%. In contrast, results based on only one stool sample underestimated the true prevalence of infection, and were associated with sensitivity values of less than 80% across all species (Carabin et al. Reference Carabin, Balolong, Joseph, McGarvey, Johansen, Fernandez, Willingham and Olveda2005). Similar SCI analyses from a sub-sample of children in Uganda indicated that the sensitivity of only one day's double KK smear was significantly lower than that obtained from two or more days sampling. Moreover, this sensitivity appeared to decrease further post-treatment, highlighting the need for more days of stool examination, and/or more sensitive alternative diagnostic assays, as the intensity of infection at the population level is reduced (P. H. L. Lamberton, unpublished data). Therefore, for future monitoring and evaluation, provided that logistical resources are available, the use of at least two or three stool samples collected on different days is recommended in order to achieve the necessary sensitivity and subsequent accuracy.
Another potential complication in terms of monitoring and evaluation of MDA is that, as discussed above, current infection intensity, as measured by egg counts, is not always associated with morbidity. Development of disease is related, in part, to the number of parasite eggs that become trapped in host tissues. Whilst a number of studies have revealed, for instance, significant associations between heavy (or higher-intensity) infection and risk for diarrhoea, anaemia, calorie undernutrition, hepatosplenomegaly, and/or structural urinary tract disease, the correlation is not perfect. Emerging evidence indicates that immune regulation of inflammatory response, independent of infectious burden, determines risk for some forms of schistosomiasis-associated morbidity (King et al. Reference King, Malhotra, Mungai, Wamachi, Kioko, Muchiri and Ouma2001). Therefore, ‘light’ infection intensities should not be overlooked or discounted as a cause for schistosome-associated morbidity. Consequently, additional direct and indirect measures of potential morbidity, in complement to prevalence and intensity data, including that of light infection intensities, must be incorporated into the monitoring and evaluation of any large-scale control programme.
QUESTIONNAIRES
Questionnaires concerning current or previous blood in the urine, as an alternative to urine examinations, have been documented as a useful, albeit potentially underestimating, reflection of S. haematobium prevalence (Lengeler, Utzinger and Tanner, Reference Lengeler, Utzinger and Tanner2002; van der Werf, Borsboom and de Vlas, Reference van der Werf, Borsboom and de Vlas2003). Such questionnaires appear to be generally well accepted and are low cost. Such potential rapid assessment procedures are certainly useful for identifying high-risk communities for MDA. Furthermore, several studies have also investigated whether the questionnaire approach could be adapted for diagnosing S. haematobium in individuals to identify morbidity and to determine if chemotherapy could be targeted more selectively (reviewed in Lengeler et al. Reference Lengeler, Utzinger and Tanner2002). Although questionnaires alone missed a significant proportion of infected children, most of those individuals missed had light infections, suggesting that S. haematobium blood-in-urine questionnaires may still be useful as indicators of more severe morbidity.
In contrast, questionnaires have generally not proved to be of consistent value to determine prevalence or morbidity due to S. mansoni (Booth et al. Reference Booth, Mayombana, Machibya, Masanja, Odermatt, Utzinger and Kilima1998; Utzinger et al. Reference Utzinger, N'Goran, Ossey, Booth, Traoré, Lohourignon, Allangba, Ahiba, Tanner and Lengeler2000; Lengeler et al. Reference Lengeler, Utzinger and Tanner2002). This may be explained, in part, by the variety of symptoms associated with S. mansoni that are also influenced by other infections, any fluctuations in symptomology over time and also by a lack of consistency noted within individuals over time, as observed in sub-samples of SCI cohorts (unpublished data), indicating a potentially high level of unreliable patient recall and/or interviewer bias. Nevertheless, Lengeler and colleagues (Reference Lengeler, Utzinger and Tanner2002) reanalysed previous studies and found that, whilst the diagnostic performance of reported and/or observed blood in stool had a low-to-moderate sensitivity (7–66%), it usually did have a moderate-to-high specificity (54–96%). Furthermore, modified questionnaires, taking into account only three yes/no questions (concerning frequent water contract, frequent weakness and frequent diarrhoea), have shown both a high sensitivity (86%) and specificity (98%) as part of screening of schoolchildren for S. japonicum-associated intestinal schistosomiasis in China (Zhou et al. Reference Zhou, Ross, Hartel, Sleigh, Williams, McManus, Luo, He and Li1998). Likewise, a more recent meta-analysis performed by King and colleagues (Reference King, Dickman and Tisch2005) did find that both S. mansoni and S. haematobium infections were significantly associated with a subjective history of diarrhoea, abdominal or genito-urinary pain and fatigue or exercise intolerance. Therefore, questionnaires have remained within SCI monitoring and evaluation, including that relating to potential side-effects of treatment.
ANTHROPOMETRIC MEASURES (HEIGHTS/WEIGHTS)
King et al.'s (Reference King, Dickman and Tisch2005) meta-analysis revealed that studies reporting measures of growth and nutrition indicated significant deficits in weight, weight for height, and skin fold thickness occurring with schistosome infection. The extent of these effects differed depending on whether analysis was performed according to infection prevalence, infection intensity or response to treatment, and no consistent effect was shown on linear growth (height). No such associations in any of these parameters have yet been consistently demonstrated within the current SCI data. For example, data arising by the national Burkinabé helminth control programme was examined to test if higher S. haematobium egg counts were associated with increased risks of wasting or stunting at baseline, adjusting for demographic characteristics and other potential predictors such as microhaematuria and anaemia status. The results did not suggest any significant association between the risk of undernutrition and S. haematobium infection intensity, with only age being revealed as a significant factor (Koukounari et al. Reference Koukounari, Gabrielli, Touré, Bosque-Oliva, Zhang, Donnelly, Fenwick and Webster2007). Younger children were less likely to be wasted or stunted than those aged 14 years or over, which could imply prior malnutrition in these older children as has previously been reported for Zanzibari boys (Stoltzfus et al. Reference Stoltzfus, Albonico, Tielsch, Chwaya and Savioli1997). Similar results were observed in data arising from the National Rwandese neglected tropical diseases (NTD) control programme. Random effects multivariate logistic regression models for the risk of an individual being wasted or stunted yielded significant positive odds ratios for older ages if compared with the youngest age group (i.e. 5–7 year old children). No other significant associations were found (SCI unpublished data). Further investigation into the utility and applications of these measures within such MDA programmes are continuing.
UMBILICAL CIRCUMFERENCE
While the visual signs of a distended or swollen abdomen can be clearly associated with advanced schistosomiasis morbidity, the occurrence, distribution and severity of this condition appeared to have not been previously investigated within the school-aged community where it likely develops, and thus served as a potentially interesting morbidity measure for consideration within SCI. A cross-sectional survey was conducted on 4354 schoolchildren across eight districts within Uganda with a longitudinal one-year follow-up of 2644 (60·7%) (Balen et al. Reference Balen, Stothard, Kabatereine, Tukahebwa, Kazibwe, Whawell, Webster, Utzinger and Fenwick2006). Distended abdomens, defined as an abdominal circumference ratio (ACR), were calculated as the umbilical circumference divided by the sternal circumference, where the greater the abdominal distension the greater the ACR from unity. An ACR greater than 1·05 was observed in 2·5% of the sampled children, several of whom presented with particularly severe distensions necessitating referral to local hospitals. ACR scores were highly overdispersed between districts and schools. Multivariate regression analysis revealed that S. mansoni infections accounted for only a small fraction of ACR variation, suggesting that either single point-prevalence and intensity measures failed to reflect this more chronically evolved morbidity and/or that other interacting factors, such as malnutrition and/or malaria, were involved. At the one-year follow-up, ACR scores showed an overall trend of regression towards the mean, potentially indicative of amelioration after chemotherapy, but geographic overdispersion still remained. Thus, the ACR appeared an interesting measure, but given the very low sensitivity and specificity for measuring schistosome-associated morbidity as part of large-scale MDA, was not continued over subsequent SCI follow-ups, although smaller scale studies may continue.
CLINICAL EXAMINATION
Palpation of the liver provides a crude measure of the prevalence of firm and enlarged livers in a population, an indication of population morbidity. However, although hepatomegaly and periportal fibrosis appear to be reduced with PZQ chemotherapy in general, some studies do not support a similar reduction in splenomegaly. In Burundi, for example, PZQ administration successfully reduced periportal fibrosis and hepatomegaly, but was associated with an increase in splenomegaly up to two years after treatment with PZQ (Doehring-Schwerdtfeger et al. Reference Doehring-Schwerdtfeger, Abdel-Rahim, Kardorff, Kaiser, Franke, Schlake, Richter, Elsheikh, Mohamed-Ali and Ehrich1992). Likewise in Sudan, the percentage of patients with hepatomegaly decreased significantly seven months and two years after PZQ adminstration but splenomegaly remained unchanged (Gryseels and Nkulikyinka, Reference Gryseels and Nkulikyinka1989). SCI results to date do nevertheless look convincing for both reduced hepatomegaly and splenomegaly following treatment. For example, in Uganda the percentage of children with a firm liver significantly decreased from 63% in 2003 to 3% in 2004 and 0·8% in 2005. While three (0·8%) children had a hard liver in 2003, by 2004 none had a hard liver. The prevalence of hardened spleens also decreased from 3% in 2003 to none in 2004 (Kabatereine et al. Reference Kabatereine, Brooker, Koukounari, Kazibwe, Tukahebwa, Fleming, Zhang, Webster, Stothard and Fenwick2007). Likewise in Tanzania the percentage of children with a firm or hard liver decreased from 4% and 0·13%, respectively at baseline to 1% and none at the one-year follow-up. The percentage of children with firm or hard spleen also decreased from 8% and 0·13%, respectively at baseline to 6% and 0·07% at one-year follow-up. Similar findings were also observed in Mali, where, for example, firm and hard livers at baseline were observed in approximately 3% and 0·5% of the children, but by the one-year follow-up only 0·7% of the children showed a firm liver and none were observed to have a hard liver (SCI, unpublished data).
Further analyses of SCI country-wide data, together with comparisons with other techniques such as ultrasonography (see below), are ongoing. Thus, whilst there is some controversy over the accuracy of this method as a reliable indicator of Schistosoma spp. periportal fibrotic morbidity (Gerspacher-Lara et al. Reference Gerspacher-Lara, Pinto-Silva, Serufo, Rayes, Drummond and Lambertucci1998; Lambertucci et al. Reference Lambertucci, Cota, Pinto-Silva, Serufo, Gerspacher-Lara, Costa Drummond, Antunes, Nobre and Rayes2001), its logistical utility and our results to date suggest that clinical palpation is an appropriate measure for continued application within SCI for the immediate future at least.
ULTRASONOGRAPHY
Ultrasonography (US) has replaced invasive techniques (i.e. hepatic biopsy and urography) as the ‘gold standard’ to assess the severity of the distribution of schistosomiasis disease in communities. Furthermore, population–based studies conducted during the 1990s using portable US, contributed an objective assessment of disease progression in communities. US, being non-invasive and yet capable of providing an accurate and permanent measure of change, is currently the diagnostic tool of choice for detecting pathological conditions associated with schistosomiasis. US is capable of detecting dilatation of the renal pelvis, bladder wall lesions and urinary tract pathology, liver fibrosis and enlargement, and dilatation of the portal veins for both S. haematobium and S. mansoni (Richter, Hatz and Häussinger, Reference Richter2003). Examinations can be carried out, not only in the hospital setting but also, due to portable US machines, in field-based studies, with the advantage of being relatively simple to perform and very well accepted by communities, as well as providing directly comparable images of the pathological changes.
In the attempt to objectively define and categorize the pathological changes associated with schistosomiasis, and to standardize the different scoring systems used previously in various endemic areas, successive “ultrasound consensus” meetings were held in Niamey, Niger in 1996 and Belo Horizonte, Brazil in 1997, leading to the revision of standardized scoring protocols and the development of the WHO-recommended US protocol (Niamey-Belo Horizonte protocol). Nevertheless, the prognostic features of individual US findings in different endemic situations, as well as whether or not US can be incorporated into a MDA programme in order to monitor morbidity, were still to be confirmed. SCI therefore assessed indicators of US-detectable morbidity due to infection with both S. haematobium and S. mansoni in the context of their large-scale control interventions targeting school-aged children.
In an initial study within Mali, morbidity indicators for both S. haematobium and S. mansoni infections were assessed and the appropriateness of the WHO coding guidelines for ultrasound in schistosomiasis in the context of MDA programmes evaluated (Koukounari et al. Reference Koukounari, Sacko, Gabrielli, Landouré, Dembelé, Clements, Whawell, Donnelly, Fenwick, Traoré and Webster2006a). Abdominal and urinary tract US was performed on 2247 and 2822 schoolchildren respectively from 29 randomly-selected schools in Mali before the implementation of MDA. It was observed that, whilst the WHO protocol proved useful for detection of S. haematobium pathology, it overestimated the risk of portal vein dilatation and left liver lobe enlargement associated with S. mansoni infection. This thus suggested that US was indeed a useful tool in large-scale control interventions, albeit it should be interpreted with some caution for S. mansoni infections. Accordingly, it has been incorporated, wherever logistically and financially possible, into all ongoing SCI monitoring and evaluation activities. For example, US measurements in Uganda were conducted on children in selected schools in Mayuge and Hoima districts. One hundred and eighty children, aged 8–13 years, had complete longitudinal ultrasound data during three years of study (i.e. baseline, follow-up year 1 and follow-up year 2). At baseline, 39% of the children had liver pattern B (early suggestive changes) that significantly decreased to 9% in 2004 and 2% in 2005. No schoolchild had patterns C to F (indicative of fibrosis). Dilated portal vein was recorded in 18% children in 2003, 2% in 2004 and 3% in 2005 (Kabatereine et al. Reference Kabatereine, Brooker, Koukounari, Kazibwe, Tukahebwa, Fleming, Zhang, Webster, Stothard and Fenwick2007). Such results thereby provide further support both for the utility of this morbidity measure and the impact MDA can achieve on reducing schistosome-associated morbidity. Thus the current findings, if combined with consideration of costs involved, which is of course a critical issue in the economically developing countries, lead to the recommendation of the inclusion of US examinations in the routine monitoring and evaluation activities of schistosomiasis control programmes where possible.
ANAEMIA PROFILES
Anaemia is a common problem throughout the world and of enormous public health concern in developing countries, but its aetiology is complex making the effect of any one factor difficult to assess. Malaria plays a key causative role for anaemia among young African children, although HIV infection, haemoglobinopathies, intestinal helminths, in particular that of hookworm and schistosomiasis, poor nutritional status and micronutrient deficiencies are also likely to make important additional contributions. Meta-analysis on previously published studies have, for example, indicated that individuals with S. mansoni infection manifest significantly greater haemoglobin (Hb) deficit than non-infected individuals, and specific anti-schistosomal therapy significantly improved anaemia (King et al. Reference King, Dickman and Tisch2005). Likewise, heavy S. haematobium infection has been reported to be significantly associated with lower Hb levels relative to light infection status (King et al. Reference King, Dickman and Tisch2005). Combined infections with S. mansoni and S. haematobium have also been reported to contribute to greater anaemia among heavy vs. light infections (Salih, Marshall and Radalowicz, Reference Salih, Marshall and Radalowicz1979).
An initial SCI study, using data arising from the Ugandan National Bilharzia Control Programme, evaluated the potential relationship between S. mansoni infection, anthelminthic chemotherapy, and anaemia (Koukounari et al. Reference Koukounari, Fenwick, Whawell, Kabatereine, Kazibwe, Tukahebwa, Stothard, Donnelly and Webster2006b). Hb levels in 2788 children, in relation to their schistosome and/or hookworm infection intensity category at baseline and one year after treatment, were analyzed. It was observed that children who were heavily infected with S. mansoni or hookworm had significantly lower Hb levels at baseline compared to those not infected. However, at least among those children anaemic at the baseline survey, a significant increase in Hb counts was found post-treatment. This suggested that intestinal schistosomiasis, together with hookworm, is indeed associated with anaemia, and such anaemia is reduced following chemotherapy. Another recent SCI study across two years within Burkina Faso indicated that anaemic children or children with more severe microhaematuria scores at baseline had higher S. haematobium infection intensities, suggesting an association between heavy intensities of S. haematobium infection and anaemia and haematuria. Moreover, there was evidence that heavy infections of urinary schistosomiasis were associated with lowered Hb counts and consequently to potential anaemia (Koukounari et al. Reference Koukounari, Gabrielli, Touré, Bosque-Oliva, Zhang, Donnelly, Fenwick and Webster2007). One may thereby conclude that anaemia is a logistically practical, inexpensive (via HeamaCue finger-prick blood assay) and valuable measure of schistosome-associated morbidity amenable to such large-scale control programmes monitoring and evaluation. There do remain, nevertheless, a few complicating issues. For example, there are inconsistencies across the published literature regarding the actual definition of ‘anaemia’ – varying among studies by region and age, from Hb concentration cut-offs of 88 to 144 g/l (King et al. Reference King, Dickman and Tisch2005). Moreover, the potential confounders of co-infections with malaria and hookworm, in particular in terms of attributing schistosome-specific-associated anaemia, can remain an issue in co-infected individuals. Ongoing SCI work, primarily within Zambia and Tanzania, where malaria infection data have been included, will help elucidate this issue further.
EXCRETED ALBUMIN IN URINE
Healthy kidneys produce physiologically appropriate urine, in that they retain essential blood components while excreting waste materials. Excess albumin in the urine (albuminuria) has been regarded for more than 40 years as a principal marker for renal (as well as cardiovascular and hypertension) disease, where more albumin appears in the urine when the kidneys are not functioning properly (Clavant, Osicka and Comper, Reference Clavant, Osicka and Comper2007). The persistent excretion of urine-albumin has also been proposed to have clinical significance in relation to the daily protein–energy balance children (Stephenson, Reference Stephenson1993).
Testing for albuminuria is often also referred to as ‘microalbuminuria testing’. A cross-sectional survey of 305 children from Unguja, Zanzibar, was performed to investigate the relationship between levels of excreted albumin in urine and S. haematobium infection status (Rollinson et al. Reference Rollinson, Klinger, Mgeni, Khamis and Stothard2005). Sensitivity analysis of urine-albumin values determined that a concentration of >40 mg/l, as measured with the HemoCue urine-albumin photometer (an assay initially developed for use in European first-point-of-care health centres for rapid detection and quantification of microalbuminuria as a biochemical marker of kidney dysfunction) had sensitivity, specificity, positive and negative predictive values of 0·90, 0·83, 0·86 and 0·89, respectively against ‘gold-standard’ parasitology. There was a clear association of reported pain upon micturition for children with elevated urine-albumin levels, with an odds ratio of 20. Reductions in albumin excretion were observed in a sub-sample of children after PZQ adminstration. These results suggest that a urine-albumin assay has promise as a biochemical marker of S. haematobium induced kidney and upper urinary tract-morbidity, and moreover a potential use as a rapid operational field diagnostic tool.
However, as with every potential assay, there are disadvantages. One major issue is that this technique does not appear to be appropriate for S. mansoni infections (J. R. Stothard and P. H. L. Lamberton, unpublished data). Even within S. haematobium infections, complicating factors, such as other uro-genital infections/dysfunctions or scarification practices, mean that not all raised albumin levels will be solely due to schistosome infections. Further potential drawbacks include the necessity for an electrical power supply, time-delays for each electronic reading, and a urine dilution series, if initial urine-albumin concentration is in excess of 200 mg/l, to produce a precise reading. Finally, the technique itself is also rather expensive for both the initial purchase of equipment and continued use of consumables. Therefore, this tool will continue to be incorporated into certain SCI monitoring and evaluation activities, but only where logistically and financially feasible.
EXCRETED BLOOD/HAEMOGLOBIN IN URINE
The presence of blood in urine, as discussed earlier, is a well-accepted marker of S. haematobium infection (Lwambo et al. Reference Lwambo, Savioli, Kisumku, Alawi and Bundy1997; Lengeler et al. Reference Lengeler, Utzinger and Tanner2002) and has led to the wide-spread use of both questionnaires and hemastix dipstick reagents across a range of studies including within SCI. For example, as part of the urinary schistosomiasis control programme on Zanzibar, a cross-sectional survey of 305 children from Unguja were examined to investigate the relationships between levels of excreted haemoglobin in urine and S. haematobium infection status (Rollinson et al. Reference Rollinson, Klinger, Mgeni, Khamis and Stothard2005). Urine specimens were visually inspected for red colour against a urine colour chart by the schoolchildren presenting the urine, a teacher at the respective school and a laboratory technician/investigator. Infection intensity as measured by egg counts was correlated with both self-assessed haematuria (rs=0·24, P=0·001) and teacher-assessed haematuria (rs=0·40, P<0·001). Hemastix dipsticks were also used to assess microhaematuria in 304 schoolchildren which successfully revealed a high prevalence of microhaematuria in contrast to very few specimens reported as red by visual inspection only. Detection and quantification of urine-Hb was also assessed using a HemoCue Plasma/Low Hb photometer, an assay originally developed to better estimate the associated blood loss and transfusion requirements of patients undergoing surgery through the rapid quantification of Hb lost in bodily fluids collected during operation. The urine-haemoglobin assay results clearly showed that quantification of excreted Hb is possible but that the assay is much less sensitive than the hemastix, which can detect as little as between 0·15–0·62 mg/l Hb. The hemastix assay soon saturates (+++) before 0·01 g/l values of Hb are obtained, which is approaching the limit of detection with the Plasma/Low Hb photometer (~0·1 g/l). Hence these results would indicate that hemastix remains the method of choice for the detection of microhaematuria. Furthermore, additional SCI studies, such as those within both Burkina Faso and Mali, have demonstrated that the hemastix test correlates well with US results and parasitological measures of S. haematobium infection intensities (Koukounari et al. Reference Koukounari, Fenwick, Whawell, Kabatereine, Kazibwe, Tukahebwa, Stothard, Donnelly and Webster2006b, Reference Kabatereine, Brooker, Koukounari, Kazibwe, Tukahebwa, Fleming, Zhang, Webster, Stothard and Fenwick2007).
Nevertheless, some disadvantages remain. As for the albumin measures discussed above, complicating factors such as menses or other uro-genital infections/dysfunctions may cause blood in urine independent of schistosome infections. It has also been noted, for example, that areas with high levels of iron in the soil (red soil), particularly on windy days with an increased risk of contamination of the sample with the soil, it is imperative to read the haemastix immediately, as contamination results in the stick reading non-representative maximum ‘haemoglobin’ levels (P. H. L. Lamberton, unpublished data). In addition, reporting error might be common within visual inspection data – for example, in the aforementioned Zanzibar study, 47 students (15·5%) reported visibly red urine while teachers recorded 81 (26·6%) urine samples to be visually red (Rollinson et al. Reference Rollinson, Klinger, Mgeni, Khamis and Stothard2005). Finally, in terms of actual schistosome-associated morbidity, whilst it appears intuitive that daily levels of associated blood loss could over time constitute a substantial amount of continual haemoglobin loss, no clear relationship between excreted urine-Hb and venous-Hb levels has been demonstrated.
COGNITIVE PERFORMANCE AND SCHOOL ATTENDANCE
Previous studies have concluded that impaired learning may be important only in heavy infections and there was no conclusive evidence that schistosomiasis reduced school attendance (Jamison et al. Reference Jamison, Mosley, Measham and Bobadilla1993), although more recent data available from large-scale studies of combined therapy have suggested that deworming improves school attendance (Miguel and Kremer, Reference Miguel and Kremer2001). King and colleagues' meta-analysis also revealed no consistent association between school performance and schistosomiasis infection (King et al. Reference King, Dickman and Tisch2005). One explanation for these, often conflicting, results may simply relate to the inherent difficulty in measuring such parameters accurately across difference socio-economic endemic situations, particularly for cognitive performance. Likewise, a lack of ‘uninfected’ control children, without schistosomes, STHs and/or a range of other infections such as malaria, provides serious constraints in terms of accurate evaluation of such morbidity measures. Nevertheless, as such potential indicators of schistosome morbidity would be of great fundamental and applied utility, particularly with respect to their impact on policy and advocacy, particularly with Ministry of Education within endemic countries, SCI felt it appropriate to, where possible, monitor and evaluate the potential impact of MDA on these parameters also.
SCI results to date on these measures remain inconclusive. For instance, if one were to assess school attendance through successful follow-up of the longitudinal cohorts, the results may not be promising. Follow-up rates in Uganda during four years of study and relative to baseline decreased continuously, from 62·7% at follow-up year 1, 41·2% at follow-up year 2 and 27·9% at follow-up year 3, although in Niger during two years of study and relatively to baseline the follow-up rate was estimated to be 87·5%. Such drop-out rates are likely to be explicable by a range of alternative factors, such as weather conditions, farming requirements or migration, rather than necessary schistosome infection status. Alternative measures such a class registers may thus provide slightly more accurate data, and analyses from which are underway.
DISCUSSION AND RECOMMENDATIONS
Schistosomiasis remains a significant but preventable chronic health problem across many parts of the world. Schistosoma infections are associated with persistent inflammation, causing a range of morbidities including, but not at all exclusive to, anaemia, chronic diarrhoea and undernutrition, which can be found with all levels of infection intensity (King et al. Reference King, Dickman and Tisch2005; King and Dangerfield-Cha, Reference King and Dangerfield-Cha2008). The present review, focusing primarily on results available from ongoing SCI activities in sub-Saharan Africa, but based within the context of already published studies from a range of countries, should help illustrate the substantial impact MDA activities can achieve on the infection intensity levels of schistosome and STH infections and thereby the subsequent health of endemic populations. The results gained to date should also help identify potential morbidity markers suitable for application in both ongoing and future large-scale MDA programmes. For instance, within the context of SCI data, the portable HemoCue proved an excellent minimally-invasive measure for monitoring Hb and potential anaemia profiles in the field over space and time. US also proved highly suitable for measuring schistosome-associated morbidity in the field, even in young children, and appeared to be particularly appropriate for S. haematobium infections where economically and logistically feasible. Some measures, such as Hb dipsticks and that of the portable HemoCue Albumin to measure excreted albumin in urine appeared very useful for S. haematobium at least, and further evaluations of the latter, including that for S. mansoni are underway. On the other hand, other potential morbidity markers evaluated here, such as that of the ACR profiles, tended not to show any consistent association with schistosome-specific infection status or intensity, nor improvement with chemotherapy, and may therefore not be currently recommended for inclusion with ongoing MDA programmes aimed at reducing host morbidity, although continued smaller-scale research on these measures are certainly still warranted.
Particularly in view of the current move towards a ‘rapid impact package’ of a combination of drugs for integrated NTD control (Hotez et al. Reference Hotez, Ottesen, Fenwick and Molyneux2006; Fenwick, Reference Fenwick2009), what remains to be considered is how such morbidity findings can or cannot be extrapolated into non-surveyed areas and countries. We will now continue to expand the application of these key morbidity measures to those previously unsurveyed and untreated populations and individuals. Likewise, there is undoubtedly a need for further research, monitoring and evaluation into additional potential schistosomiasis-associated morbidity indicators, for both severe and subtle cases. For instance, genital schistosomiasis of both men and women is an aspect of chronic schistosomiasis that is finally receiving more prominence, at least in the medical literature, if not yet within the public health sector (Feldmeier et al. Reference Feldmeier, Leutscher, Poggensee and Harms1999; Swai et al. Reference Swai, Poggensee, Mtweve and Krantz2006; King and Dangerfield-Cha, Reference King and Dangerfield-Cha2008). Similarly, the role and potential schistosomiasis-associated morbidity in even the youngest of infants is now coming to light and should be incorporated into future MDA and concurrent monitoring and evaluation (Odogwu et al. Reference Odogwu, Ramamurthy, Kabaterenine, Kazibwe, Tukahebwa, Webster, Fenwick and Stothard2006). Furthermore, it is imperative that the diagnostic approaches need to be modified, and additional tools developed, evaluated and applied, as further progress is made towards control and even potentially, in certain circumstances/localities, the elimination of this important NTD (Bergquist et al. Reference Bergquist, Johansen and Utzinger2009). Even the lightest remaining infection intensities should be easily detectable in the field and any associated morbidity closely monitored. The key issue remains, of course, that the health of the individuals treated continues to improve and morbidity is reduced as a consequence of these large-scale control programmes.
ACKNOWLEDGEMENTS
The Schistosomiasis Control Initiative (SCI) research presented here and/or authors were funded by grants from the Bill and Melinda Gates Foundation (JPW, AK, PHLL, JRS, AF), the Royal Society (JPW), the Medical Research Council (PHLL) and a CONTRAST project grant, a multidisciplinary alliance to optimize schistosomiasis control and transmission surveillance in sub-Saharan Africa, sponsored by the European Union (contract no. 032203) (JPW, PHLL). We are very grateful to Jürg Utzinger for detailed comments on the manuscript.
