The effect of different treatment regimens on the epidemiology of seasonally transmitted Schistosoma haematobium infections in four villages in the Senegal River Basin, Senegal

TRANSACTIONSOFTHEROYALSOCIETYOFTROPICALMEDICINEANDHYGIENE(1999)93,142-150

The effect of different treatment regimens on the epidemiology of seasonally transmitted Schistosoma haematobium infections in four villages in the Senegal River Basin, Senegal 1Department of Parasitology, Faculty of D. J. Shawl+, J. Vercruyssel*, M. Picquet’*‘t, B. Sambouz and A. Ly’ Veterinary Medicine, University of Gent, Salisbulylaan 133, B-9820 Merelbeke, Belgium; 2Programme ESI’OIR, R@m Midicale de St Louis, BP 394, St Louis, Senegal Abstract This paper describes the present epidemiological situation of Schistosoma haematobium in 4 villages in the middle vallev of the Senegal River Basin, in terms of level and intensity of infection, seasonality of transmission; and intermediate hosts, and the effect of different treatment schedules with praziquantel on the overall infection levels and re-infection rates. The longitudinal study involving 7 surveys was carried out between June 1995 and March 1997 in Diatar, Guia, Donaye and Niandane. The prevalence and intensity of infection remained low throughout the survey (<55% and cl2 eggs/l0 mL urine), and there were no svstematic differences in the mevalence or intensitv of infection between men and women. Before treatment, infections were highly aggregated in individuals and were concentrated in children (aged < 15 years) with 85% of the potential contamination; no individual aged >24 years produced >50 eggs/l0 mL urine. Using WHO guidelines mass treatment was given to all Diatar and Guia villagers in December 1995, whereas in Donaye and Niandane only individuals positive for eggs were treated. Six weeks post-treatment cure rates in all villages were >80%, with marked declines in levels of infection (<20% and c4.5 eggs/l0 mL). By March 1997 infection levels in Donaye and Niandane had returned to pre-treatment levels, whereas in the 2 mass-treated villages (Diatar and Guia) infection levels were still markedly reduced compared to pre-treatment levels. Rates of conversion were very low between all surveys; however, there was an apparent high level of reversion (>20%), due to the alternation of individuals apparently positive and negative between surveys. Water and infected snails were present from June to March. Therefore, owing to the high aggregation of infections in children, the low overall infection levels and the transmission period, it is suggested that in this area the best treatment schedule would be selective treatment of school-aged children in March/April, probably on an annual basis. Keywords:

Schistosomahaemarobium,

chemotherapy, treatment regimens, transmission, epidemiology, Senegal

Introduction Urinary schistosome infections are present throughout sub-Saharan Africa (TORIIAN et al., 1993), with manv of these infections endemic (e.g., BRAD&% & McC~LLOUGH, 1973; CI-IANDIWANA et al., 1988). However, the construction of large-scale water development schemes has led to increases in infection levels and/or new epidemics being observed (e.g., BETTERTON et al., 1988). During the 1980s 2 dams were built in the Senegal Rive; Basin (SRB) in northern Senegal (PICQUET et Gl., 1996). The subsequent dramatic outbreak and rapid increase of Schistosoma mansoni infections in the lower valley of the SRB have been well documented (TALLA et al., 1990; STELMA et al., 1993, 1995; PICQUET et al., 1996, 1998). However, there has been little consideration of urinarv schistosomiasis in the SRB. and this work has either looked at mixed S. mansonifi haematobium infections in the delta region (BELOT etal., 1993; VERDI? et al., 1994), or provided an overview of the whole of the SRB (PICQUET etal., ~~~~;SOUTHGATE, 1997). There has been no detailed study of the epidemiology of S. haematobium infections, esueciallv in the middle vallev of the SRB (where S. ma&o& infections are absent). T&s has prevented the development of treatment programmes (how frequently to treat and to what proportion, etc.) in the area. These programmes can be constructed only after the efficacy of treatment and re-infection pressures have been established (WHO, 1993). Before dam completion S. haematobium levels in the middle valley were low (<5%) (VERCRUYSSE et al., 1985). The source of infection was considered to be temporary laterite pools and rice fields that filled during the rainy season (July-November) but dried out during the dry season. These water-bodies were known to ipresent address: Centre de Recherches sur 1’8pidemiologie des Desastres, Universite Catholique de Louvain, Ecole de Santk Publique, 30.94 Clos Chapelle-aux-Champs, 1200 Brussels, Belgium. *Author for correspondence; phone +32 9 264 7400, fax +32 9 264 7496, e-mail [email protected]

support Bulinus senegalensis, the intermediate host for S. haematobium in the area (BROWN. 1994). This intermittent water supply results in maiked seasonality in transmission (ELIAS et al., 1994). However, the standard measures of infection (prevalence and intensity of infection) do not distinguish between past and recent transmission (CHANDI~ANA et al., 1987). For this, longitudinal studies of incidence and reversion are required. Such studies are rare (FAROOQ & HAIRSTON, 1966; SCOTT etal., ~~~~;CHANDI~ANA etal., 1987),owingto population movement, and the lack of longitudinal &keys (CHANDI~AN~ et al., 1987). The obiectives of the oresent work were 3-fold. First, to determine the preseni epidemiological situation of S: haematobium in the middle valley of the SRB, in terms of level and intensity of infection, transmission period and intermediate hosts. Secondly, to find the effect the seasonality of transmission has on incidence and apparent loss of infection. Finally, to ascertain the effect of different treatment strategies with praziquantel on the overall infection levels and re-infection rates. For this work, 4 villages were selected for a longitudinal study between June 1995 and March 1997. Materials and Methods Study site and population A large Eurouean (STD3 /INCO DC) Programme on the research aid control of’schistosomiaiis i; the SRB, instigated in 1994, established that in the Departement de Podor only S. haematobium infections were present (PICQUET et al., 1996). The region is a typical sahelian area with annual rainfall <300 mm, concentrated between May and September/October. The population (Toucoleurs) in this area are sedentary, and use irrigation mainly for rice culture. The extent of the irrigation has expanded greatly since the completion of the dams. Men are mainlv involved in cultivation and, while women do help, they are generally concerned with household duties. Children help with the cultivation, but also play in and around water-bodies. In June 1995, 4 villages (Diatar, Donaye, Guia and

TREATMENT

AND EPIDEMIOLOGY

OF S. HAEMATOBIUM

SENEGAL

IN

Niandane) within 25 km of Podor were chosen for a longitudinal study of the epidemiology of S. haematobium. A (good) primary health care centre was available, and the villages were not part of any other control programme. Common nearby water-contact sites were identified in each village, and 1 (large) rice field, 1 (large) temporary laterite pool and 1 major irrigation canal per village were then selected. Prior to the start of the study, transmission at these sites was known to be highly seasonal (assumed June-November). Experimental design A population census was carried out in the 4 villages immediately before the start of the study (Table). The objectives were explained in detail to villagers and individuals were invited to narticinate in the studv. The number of people per village who were registered was calculated using WHO guidelines (WHO, 1993). Households were then selected at random, and individuals in these households were invited to participate until the number of people registered per village had reached the requirednumber. The narticinationthroughout the studv is also shown in the Table. Approximately;qual numbers of males (47%) and females (53%) were reaistered. with the age-groups 4-9, lo- 14,‘15- 19,20-25,30-39 and 240yearsrepresentedby29%, 17%, lo%, 15%,9%,and 20% of the individuals, respectively.

143

Parasitological techniques On each sampling date all registered villagers were invited to provide a single urine sample. Ten mL of the urine were filtered with Nytrel@ filters, which were stained with iodine (WHO, 1983), and samples were examined the following day. Mean egg counts were calculated as the geometric mean of infected individuals, and expressed as eggs/ 10 mL. Malacological surveys From January 1996 until March 1997 monthly malacological surveys were carried out in the 4 villages. The 3 transmission sites (rice field, irrigation canal and temporary laterite pool) selected per village were visited; the presence or absence of water was noted; and for 15 min per site the snail abundance was monitored by 2 experienced snail samnlers. Snail abundance is aiven as the total number of snails collected per site-during the 15 min. All collected snails were identified and screened for patent schistosome infections by exposure to light for 1 h. Schistosome cercariae were distinguished from other trematodes, but no attempt was made to discriminate S. haematobium from S. bovis and S. curassoni. Snail samples were later sent to the Experimental Taxonomy Unit, Natural Historv Museum, for confirmation of identification. At the same time water temperature and pH were recorded using a Hanna water tester.

Table. Summary of participation of individuals in the 4 villages Senegal River Basin from June 1995 to March 1997 Diatar

Guia

Total

1310 354 166 114 74

1983 708 421 316 213

1390 335 170 146 147

2178 536 313 281 292

6861 1933 1070 857 726

100 115 97 103 85

315 335 274 244 209

148 141 134 125 104

267 260 234 226 202

830 851 739 698 600

45 28

113 65

69 40

143 85

370 218

Donaye Population Registered Tune 1995 November 199 5 December 1995 (praziquantel) January 1996 April 1996 July 1996 November 1996 March 1997 Gave 7 samples Aged < 15 years and gave 7 samples

in the

Niandane

All registered individuals were invited to provide a single urine sample for examination on each of the 7 sampling dates over the 2 l-month period: late June 1995 (before the arrival of water in the transmission sites); early November 1995 (end of the transmission period); late January 1996 (6 weeks post-treatment); early April 1996 (to allow all me-patent infections at treatment to have developed to adults); late July 1996; early November 1996; and mid-March 1997 (end of the survey). In mid-December 1995 a single treatment of praziquantel (PZQ) at 40 mg/kg was offered to the villagers. In 2 of the villages (Diatar and Guia) the nrevalence of infection in chil&en‘(aged < 15 years)‘was 550% and the prevalence of heavy infections (>50 eggs/l 0 mL) was >5%. Therefore, following WHO recommendations on treatment according to levels of infection (WHO, 1993) mass treatment was given. In Donaye and Niandane where levels were below 50% and 5%, respectively, only those individuals who were positive received PZQ. Prior to treatment it was explained to the villagers who was to be treated, and why. At the end of the study, levels in all 4 villages were <40% and <3%, resnectivelv, and thereforeonly children in the 4 villages were treated.

Data manipulation and statistical analysis Compliance of registered individuals was never 100% (Table), and therefore to minimize potential bias from particular individuals attending for particular samples, only individuals who gave samples on all 7 dates were considered for the analysis (Table). However, this meant that there were insufficient samples to plot representative age-prevalence and -intensity curves, particularly following treatment. Therefore, given the closeness of the villages to each other, data from different villages were combined according to treatment regimen: Diatar and Guia (which had mass treatment) and Donaye and Niandane (where only positive individuals were treated). No statistical difference in prevalence or intensity of infection between villages within groups was found, Differences between groups in the prevalence of infection were analysed by Pearson x2 statistics, the intensity of infection by general linear models with negative binomial errors on positive egg counts only, and the shape of age-prevalence and -intensity curves by 2-sample Kolgmorov-Smimov tests (D statistic). JORDAN’s method (1985) was used to calculate the corrected index of potential contamination. Differences

D. J. SHAW ETAL.

144

The time between surveys ranged from 77 to 133 days, and therefore a relative measure of reversion and conversion was estimated. This allows a better description of the relative changes in infection pressure during different parts of the year, particularly useful in areas of distinct seasonal transmission. Usually annual reversion and conversion rates are calculated (e.g., CHANDNVANA et al., 1987); therefore, a modified equation was used:

in the degree of aggregation were tested using a recently developed method of analysis of dispersion (see SHAW et al. (1998) for a description). In brief, maximum likelihood estimates of K (the inverse measure of aggregation estimated in the negative binomial distribution) were obtained, along with a common k (BLISS & OWEN, 1958). The differences in the deviances associated with these maximum likelihood estimates are approximately x2 distributed. Standard definitions of reversion and conversion are used in the present paper: reversion (and cure rate when treatment has occurred) is the proportion of children aged Cl5 years becoming negative between 2 surveys; conversion is the proportion of children aged < 15 years becoming positive between 2 surveys. Transmission ratio is defined as the number of children aged <15 years becoming positive between 2 samples/number of children aged < 15 years becoming negative between the same 2 samples. As with the age-prevalence and intensity analysis, data from Diatar were combined with those for Guia, and data for Donaye were combined with those for Niandane. Statistical differences between groups in the reversion and conversion rates were analysed using Pearson x2 statistics. In all cases a Pvalue CO.05 was taken to indicate significance.

R,

JUT 95

1 1

Nov 95

1 1

1 -

C,

=

1 -

neg(77/Y);

Results Prevalence and intensity of infection prior to treatment Figure 1 shows the change in prevalence and intensity of infection in the 4 villages from June 1995 to March I EZd 623 0

Jan 96

&77/Y);

where R,, C, and T, are the relative reversion, conversion and transmission rates; inf the proportion remaining positive between 2 samples; neg the proportion apparently remaining negative; y the number of days between the 2 samples, 77 days the minimum sampling period; N the number positive in the first of the 2 samples; and d the number negative. R,, C, and T, are used for the remainder of this paper.

Treated

1

=

Diatar (mass) Guia (mass) Donaye (pos) Niandane (pcs)

1 1

Apr 96

1 [

Jul 96

1 I

Nov 96

1 I

Mar 97

1

1 1

Apr 96

1 I

Jul 96

1 I

Nov 96

I

Mar 97

1

Treated

1

Jun 95

I

1

Nov 95

1 I

Jan 96

1

Fig. 1. Overall (a) prevalence and (b) intensity (geometric mean ofpositive egg counts) of S. haematobium infection in the 4 villages in the middle valley of the Senegal River Basin, from June 1995 to March 1997. Also included on the figures is a line denoting the single treatment with praziquantel in December 1995, and a key indicating whether this treatment was given to the entire population (mass: Diatar & Guia) or just those individuals with positive egg counts (pos: Donaye & Niandane).

TREATMENT

AND EPIDEMIOLOGY

OF S. HAEMATOBIUM

1997. In June 1995 the prevalence of infection ranged from 33% to 54% (Fig. la), but there was no significant difference between villages (x2 = 5.25, P = 0.155). While mean intensities were low (< 10.8 eggs/ 10 mL; Fig. lb), there were significant differences between villages (x2 = 9.207, P = 0.027). Five months later (November 1995) the prevalence had declined in 3 villages, but not significantly (x2 < 1.19, P > 0.275), and there was still no difference between villages (x2 = 3.33, P = 0.344). The mean intensity of infection in Diatar and Guia increased to 12 and 11.4 eggs/ 10 mL during this period, but not significantly (~2 <3.327, P > 0.067). No increase was observed in Donaye and Niandane. Diatar and Guia also had a higher proportion of heavily infected individuals (>50 eggs/ 10 mL: 7.3% and 7.7%, respectively), compared to Donaye and Niandane (4.4% and 1+3%, respectively). The degree of aggregation at this time in all 4 villages was relatively high (k = 0.071-0.095), with no significant difference between villages (x2 = 1.723, P = 0.632). In November 1995 combined data from Diatar plus Guia produced a typical age-prevalence curve (Fig. 2a), with in this case the peak (60%) in the lo-14-years agegroup, followed by a decline in older age-groups. The prevalence of heavy infections was age-related, with no individual aged >23 years producing more than 50 (a) Diatar+Guia

145

IN SENEGAL

eggs/l0 mL in any village. A similar age-prevalence pattern was observed for Donaye plus Niandane (Fig. 2b; D = 0.6, P = 0.329). There was no difference between the age-intensity curves of Diatar plus Guia and Donaye plus Niandane (D = 0.4, P = 0.819, Figs 2c and 2d), with the highest intensities in those aged 4-9 years (17.1 and 8.7 eggs/ 10 mL, respectively), and a decline to 3.2 eggs/ 10 mL in the oldest age-group. The prevalence of infection in males (182 in total) and females ( 188 in total) did not differ within any village in either June 1995 or November 1995 (x2 C3.2, P BO.07): e.g., Diatar 57.1% of males and 52.6% of females infected in June 1995. However, there were differences in the intensities of infection between the sexes. In June 1995 males in all 4 villages had higher mean intensities (9.31-12.83 eggs/l0 mL) than females (3.61-7.20 eggs/l0 mL), with the differences significant in Guia and Niandane (x2 >6.460, P always >O.Oll). Five months later, the intensity of infection in males in Donaye, Guia, and Niandane (7.07-16.47 eggs/ 10 mL) was still significantly greater (x2 B5.56, Palways CO.019) than in females (2.05-6.72 eggs/l0 mL). In contrast, there was a non-significant difference in intensitv of infection between sexes in Diatar IF = 16.22 M = 9.34 eggs/l0 mL; x2 = 0.004, eggs/i 0 mL; P = 0.947). (b) Donaye+Niandane A - A Pre-treatment l ..... l 15 months post +ve’s treatment

(c) Diatar+Guia 1 A-A Pretreatment l ..... 0 15 months post -mass treatment

(d) Donaye+Niandane A -A Pre-treatment l -.... . 15 months post +ve’s treatment

Fig. 2. S. haemarobium(a) age-prevalence and (c) age-intensity curves of individuals in Diatar & Guia (combined) prior to treatment (November 1995), and 15 months post-treatment (March 1997), where mass treatment of all villagers in December 1995 occurred. The curves (b) and (d) as for (a) and (c) but for Donaye & Niandane (combined), where treatment in December 1995 was of only individuals positive for eggs in 10 mL urine.

D. J. SHAW ET/X..

146

The corrected index of potential contamination revealed that approximately 85% (86.3% Diatar + Guia, 83.4% for Donaye + Niandane) of the contamination of water is from children aged ~15 years, compared to <14% of individuals aged >25 years (3% and 13.5%, respectively). Prevalence

and intensity

of infection following

treatment

In November 1995, 53.8% and 50.1% of children in Diatar and Guia, respectively, had positive egg counts and 7.3% and 7.7% of all individuals were producing >50 eggs/ 10 mL urine. Following WHO guidelines (see Materials and Methods), in December 1995 a single treatment with PZQ (40 mg/kg) was offered to all villagers in Diatar and Guia. In Donaye and Niandane the prevalence of infection in children was 35.7% and 44.6%, respectively, and the proportion of heavy infections 4.4% and 1.8%. Therefore, in Donaye and Niandane treatment was offered only to individuals producing eggs. Six weeks after treatment (January 1996), cure rates of 80-93% were obtained in the 4 villages, with a significant decline in the urevalence of infection (Fig. 1a; P always 11.36, P2d years had a positive egg count, compared to > 15% pretreatment. In addition, the age-related mean intensity of infection was still markedly reduced (Fig. 2~). In contrast, in Donaye + Niandane the prevalence of infection in all age-groups in March 1997 (37% in children, 23% in adults) was back to pre-treatment levels. There were also no significant differences between November 1995 and March 1997 in the shape of the age-prevalence (D = 0.4, P = 0.819) and -intensity curves (Fig. 2d, D = 0.6, P = 0.329), or the corrected index of potential contamination of children aged <15 years (76%, t = 0.011, P = 0.991). However, there had been an increase in the proportion of those aged <25 years who were infected (from 35% to 50%), despite treatment. There was no significant difference in the prevalence of infection between males and females in any village in any ofthe 5 surveys carried out after January 1996 @ always <2.06. and Palwavs >0.151). Furthermore. which sex had the highest prevalence of infection in a particular village changed between surveys. There were consistent differences in the intensity of infection observed in males and females between villages, but the differences were not related to treatment schedule. IV

Reversion,

conversion,

and transmission

rates

Figure 3 shows the relative reversion, conversion and transmission rates for the 2 sets of data. Reversion remained >20% throughout the survey, even during the transmission periods. This result can be explained by

the fact that only a third of children provided positive egg counts in consecutive surveys. The remainder alternated between negative and positive egg counts. Before treatment (June to November 1995), there was no significant difference in the rate of either conversion or reversion between Diatar + Guia and Donaye + Niandane (x2 = 0.007, P = 0.934 and x2 = 1.482, P = 0.224, respectively), and the transmission ratio for both sets of villages was < 1. Six weeks post-treatment, conversion in Donaye + Niandane (33.3%) was significantly higher than Diatar + Guia (9.8%. yx2= 8.206, P = 0.004), and remained higher for the ‘remainder- of the surveys. Transmission ratios for Diatar + Guia remained low throughout the study (< 1.6), with net transmission occurring only in April and November 1996 (Fig. 3a). In contrast, ratios >3 were observed in Donaye + Niandane at this time (Fig. 3b). Malacological

surveys

Transmission of S. haematobium was assumed to be restricted to the period July-November, i.e., the rainy season (VERCRIJYSSE et al., 1985). However, water was still present in the temporary pools and irrigation canals way beyond November in all 4 villages JJanuary and February in Diatar and Guia, April in Donaye and Niandane (Fig. 4)]. In Diatar and Guia no water was present in any of the transmission sites between March and the beginning of July 1996. In Donaye water was present from June 1996 until March 1997, and for Niandane from August 1996 to March 1997. The temperature of the water observed in the 4 villages never fell below 16.l”C (Februarv). and never exceeded 38.5”C (July/August), &l-i an a&rage temperature of 28.8”C (+ 0.363). The pH of the water was always greater than 6 and never exceeded 10.3 (mean 8.24 f 0.10). In Diatar and Guia B. senepalensis were found predominantly between July and September (Fig. 4a), with rice fields havine the hiahest numbers in Tulv and August 1996, irrigationcanals-in August and September 1996 and temporary pools in September and October 1996. Maximum infection levels (12%) were observed in July 1996. After September very few B. senegalensis we& found (<35). \ , Onlv low numbers of uninfected B. truncatus were present in these 2 villages (maximum 1 l), and all were found in the irrigation canals. The uattern for Donave and Niandane was suite different. B. ienegalensis levels-started to rise in June, reached a peak of 247 snails in September/October and were still present in relativelv large numbers (>60) until February 1997. Infection rates with cercariae increased slowly f&m 1% to 20%. The peak in the rice fields was between July and September 1996, whereas in the irrigation canals it was September to November 1996, and in the temporary ~001s between Sentember 1996 to Februarv 1997. Manv kore B. truncatui were found (maximum-261), mainly between December and February, with infection rates >30% in February 1997. Discussion Epidemiolopikal

situation

before treatment

-Seven years after the completion of the water development in the SRI3 the nrevalence of infection in the 4 infected villages has risen from <5 % (CHAINE & WK, 1983) to between 33% and 54%. The 4 villages were chosen because the 1994 transverse survey had shown that there was an S. haematobium focus-in this area (PICQUET et al., 1996). The vast majority of villages in the middle valley at this time were either not infected, or had verv low levels of infection (< 10%). Such focalitv is corn&n with S. haematobium infections (e.g., KOUG et al., 1981). In other words, the 4 villages probably experience the most intense S. haematobium transmission in this area. Intensity levels (< 12 eggs/ 10 mL) are comparable to those in other low endemicity foci INigeria (BETTERTON et al., 1988), Somalia (KOURA et al., 1981) and Zimbabwe (CHANDIWANA et al., 1988)].

TREATMENTANDEPIDEMIOLOGYOF

INSENEGAL

SHAEMATOBZUM

147

(a) Diatar + Guia Reversion (%)

z7

-

J”A.95

8_

NO:.95

Jail96

ApA

J"l196

NOd.96

Ma:.97

Jan.96

Apr.96

Jul.96

Nov.96

Mar.97

Conversion (%)

8

Transmission

%

ratio

3 z Jun.95

Nov.95

(b) Donaye + Niandane

8_

Reversion (%)

5:

Jun.95

Nov.95

Jan.96

Apr.96

Jul.96

Nov.96

Mar.97

Jan.96

Apr.96

Jul.96

Nov.96

Mar.97

Jan.96

Apr.96

Jul..96

Nov.96

Ma;.97

Conversion (%)

Jun.95

R

Transmission

Nov.95 ratio

x x Jun.95

Noi.

Fig. 3. Comparison of the change in status of S. haematobium infection in children (aged < 15 years) from June 1995 to March 1997 in (a) Diatar & Guia (combined), where all villagers were treated in December 1995 (T); and (b) Donaye & Niandane (combined), where in December 1995 only individuals with positive egg counts were treated (T). Nevertheless, the degree of aggregation in the villages was high (k 50 eggs/ 10 mL in 2 villages prior to treatment, thereby contributing >80% of the potential contamination. DE VLAS et al. (1992) have demonstrated that aggregation in egg counts reflects aggregation in worm burdens, which has consequences in terms of morbidity (CHANDIWANA et al., 199 1): ultrasonography from these villages has revealed severe urinary tract lesions in some children (DELEGUE et al., 1998). The concentration ofinfections in children was despite the fact that the majority of the adult men are involved in intensive irrigation agricultural practices, with adult women helping as well as perfoming household duties. Children do help both sets of adults, but also play in and around water-bodies. Unfortunately, no detailed watercontact data are available from the present study. There was some evidence of greater exposure of males compared to females (particularly prior to treatment), but the differences were not consistent across all villages and surveys. Another possible reason for the differences in infection levels between children and adults could be acquired immunity in older individuals WOOLHOUSE et al.. 1991). However, prior to the dam construction the prevalence of infection was very low (CHAINE & MALEK,

1983), and therefore effective transmission can have occurred for only a maximum of 6 years. This means that children aged 7-15 years have had the same duration of exposure as adults. There is some evidence of a quicker acquisition of immunity to Schistosoma infections with inCreaiitIg infection intensities (WOOLHOUSE et al., 1991; MUTAPI et al., 1997). but such hiah levels of infection are not present in this study. Effect of d#erent

treatment regimens

One ofthe most striking results are the clear epidemiological differences observed due to the different treatment schedules given to the 4 villages. The decision on whether to mass treat or to treat only egg-positive individuals was made according to standard guidelines (WI-IO, 1993). Post-treatment, cure rates in the 4 villages were good (>80%) and typical for field situations (Fig. la). However, by 15 months post-treatment (March 1997) the prevalence of infection had returned to pre-treatment levels in the 2 villages where only positive individuals were treated, whereas in the 2 mass-treated villages it was still low. In addition, there was no difference between the age-intensitv curves vreand 15 months post-treatment -in Donaye-+ Niandane (Fig. 2d), whereas in Diatar + Guia age-related intensities of infection were still markedly reduced (Fig. 2~).

D. J. SHAW ETAL.

148

(a) Diatar + Guia

TP

150

1

B. senegalensis 0

150 B. truncatus 0

1

-

J’F’~‘A’~‘J’J’A’S’O’NID

1996

J’F’M’

1997

(b) Donaye + Niandane RF TP IC

8. senegalensis

250 B. truncatus 0I J’F’M’A’M’J’J’A’S’O’N’D

1996

J’F’M’

1997

Fig. 4. Monthly ecological and malacological changes in water and snail infection levels from January 1996 to March 1997 in (a) Diatar & Guia (combined) and (b) Donaye At Niandane (combined); with presence (solid bars) or absence (white bars) of water in rice fields (RF), irrigation canals (IC), and/or temporary laterite pools (TP); and the total number of B. senegalensisand B. truncatus (shaded bars) found at the different sites per month (see text for details of malacological survey); and the number of infected B. senegalensis and B. tmncatw (white bars) found.

These results suggest that selective treatment in this and similar epidemiological areas will not be successful in reducing disease in the long term. WOOLHOUSE et al. (1997) have shown that if treatment does not include most of the individuals who contaminate transmission sites then treatment will be ineffectual. In the present area this group are the children. At low intensities of infection there will be a relatively high proportion of infected people who produce a negative egg count at the time of treatment4ue to day-to-day variation in egg et al., 1997) and the single examinaoutput (VAN EON tion-and therefore who are not treated. However, these individuals are still producing eggs, thereby contaminating the environment as well as sustaining further pathological damage. Therefore, making decisions on control in this area (and other areas of low infection intensity) based purely on the WHO guidelines on whether X0% of children are excreting eggs and 15% of individuals are excreting X0 eggs/l0 mL is not the best method of evaluating control strategies. Mass treatment would seem to be the most appropriate strategy for such areas. However, the cost of mass treatment must be considered in relation to the benefits.

A more optimal treatment schedule would be selective chemotherapy of individuals most at risk of infecet al., tion/heavy levels of infection (WOOLHOUSE 1997), i.e., the school-aged children. This scheme has been applied in many other endemic situations, e.g., Cameroon (BAUSCH & CLJNE, 1993, Kenya (KING et al., 1992), Mauritania (ETARD etal., 1990), andhas been shown to be highly effective in reducing prevalence and intensities of infection. Community-based control programmes also require cost-effective decisions based on the estimated force of re-infection in order to plan treatment intervals. The aim is to treat as effectively as possible to reduce morbidity, et al., but to treat as infrequently as possible (ETARD 1995). CHAN et al. (1995) have approached this problem from a theoretical framework for S. munsoni infections with some success. In the present paper we adopt a more epidemiological approach. The interval between treatments can be estimated from examination ofre-infection rates and the degree of aggregation. Higher re-infection rates mean that worm levels in individuals post-treatment will reach morbidity-inducing levels sooner, thereby requiring treatment more frequently, and the lower

TREATMENTANDEPIDEMIOLOGYOF

S.HAEMATOBIUMIN

the degree of aggregation the greater proportion of the population that will be affected. Conversion rates were low throughout the present study, indicating that re-infection rates are low. The degree of aggregation was high (k < 0.1). As ANDERSON & MEDLEY (1985) have shown, these 2 estimates are related-re-infection rates should be at their most rapid when the degree of aggregation is low. Therefore, a longer interval between treatments can be recommended for the Podor area, compared to more intense foci. In addition, 15 months post-treatment the proportion of individuals excreting >50 eggs/l0 mL was still <3%. This suggests that annual treatment would be a reasonable approach in this area, as has been used in other endemic situations (KING et al.. 1992: ETARD et al.. 1995). The mala‘cological results ’ suggest that March/April would be the best time for this treatment-at the end of one transmission period and before the next. Even the treatment given during December in the present study was sufficient to remove the vast maioritv of visible lesions in the bladder (DELEGUE et al.,’ 1998). Therefore treatment in March/April is likely to have an even greater effect in reducing burdens below the pathological level. Such a treatment schedule is being evaluated in another longitudinal survey in 2 more villages in the Podor area (D. De Clercq, personal communication). reversion and transmission One important measure of transmission is the rate of conversion (incidence) (FAROOQ & HAIRSTON, 1966; SHIPP. 1973; CHANDI~ANA et al., 1987). In the present study -we were particularly interested -in whether the seasonal pattern of transmission could be detected in terms of the proportion of children (<15 years) who become positive for egg counts (i.e., net transmission) between surveys. To do this we used modified formulae such that relative estimates of reversion and conversion between surveys were calculated. While an increase in conversion was observed during the transmission period, levels were very low (Fig. 3). The apparent rate of conversion was also augmented by individuals who alternated between positive andnegative egg counts in surveys. This alternation was most apparent in the relatively high rate of reversions observed: >20% for all surveys. S. haemawbium worms have been estimated to live for any period between 4 and 12 years BRADLEY & MCCULLOUGH, 1973; WILKINS et al., 1984), which cannot explain the rate of reversions observed in the present study. These high rates occurred because the average level of infection was low, and therefore with the w&l-documented day-to-day variation in egg counts (e.g., VAN E'ITEN et al., 1997), individuals who were in fact infected with S. haemawbium worms (i.e., had produced eggs in a previous survey) were producing too few eggs at the time of sampling to be detected in 10 mL of urine. This problem was also observed bv GOLL et al. (1984) in The Gambia. This suggests that if accurate data on rates of reversion and conversion are required in such low endemicity areas then 10 mL of urine should be filtered on 2 consecutive days, thereby increasing the sensitivity of the technique. Another possibility would be the use of statistical models to estimate the true reversion and conversion rates, along with appropriate confidence limits. NAGELKERKE et al. (1990) have proposed such a model for malaria and Giardiu infections, but as DE VLAS et al. (1993) have pointed out the model depends crucially on assumptions based on detectability, and some of these may not hold for Schistosoma infections, indicating that further theoretical work is required.

SENEGAL

149

B. senegulensis in this area was limited to non-permanent water-bodies (i.e., temporary rain-fed pools), which underwent considerable diurnal temperature fluctuations. Snails were found between August and November, with water present from June/July until November. Cercarial infection levels were very low (0.4%). B. truncatus (B. geurnez) was never found in these pools. Laboratory work had also established non-compatibility with the Senegalese S. haematobium. In 1983, 5 years before the completion of the second dam upstream at Manantali, B. senegalensis was found in low numbers in the temporary rain-fed laterite pools and in the limited rice fields present at the time, but only B. truncatus (infected with S. bovis) was found in the relatively few irrigation canals (VERCRUYSSE et al., 1985). In the present study, which started 7 years after the completion of the dam at Manantali, the effects in terms of presence of water and snail abundance have been profound. Water is now present for far longer periods of time (June/July - at least February/April, depending on the village and vear), and as a conseauence snails are present For longer periods (Fig. 4). B: senegalensis was found in all 3 transmission sites-rice fields. urination canals and the laterite pools-depending on me &me of year. This means that the transmission period has increased from the 5/6 months assumed before the present study (VERCRUYSSE etal., ~~~~;PICQIJET et&, 1996)to at least 9 months of the year.

Rates of conversion,

Malacologv

surveys and transmission

period

In 1978 CHAINE & hbLEK (1983) found that, asin other endemic areas [Nigeria (B‘E-ITE~ToN et al., i988), The Gambia (GOLL & WILKINS, 1984)], the presence of

Acknowledgements

The authors gratefully acknowledge the participation of all the villagers and C. Hann, C. Thiam, M. Sy, S. Ndiongue, P. Diouf and F. Lv of the Grandes Endemies de Podor for assistancein

the field. Special thanks go to the European Delegation in Dakar for logistic support. We also thank Drs D. De Clercq, J. De Bont and V. Southgate for useful discussions in the preparation of this manuscript, and Dr P. Hagan for critical reading of an early manuscript. This work was supported by the VLIR-Flemish Inter-University Council (Belgium), the Commission of the European Communities Research Programme (ERB35 14PL950268) and is associated with the ESPOIR programme for research and control of schistosomiasis in the Senegal River Basin.

Anderson, R. M. & Medley, G. F. (1985). Community control of helminth infections of man by mass and selective chemotherapy. Parasitology, 90, 629-660. Bausch, D. & Cline, B. L. (1995). The imuact of control measures on urinary schistosomiasis in primary school children in northern Cameroon: a uniaue ouoortunirv for controlledobservations. AmericanJourn~lof l%picalM~dicineand Hygiene, 53,577-580. Belot, J., Geerts, S. & Diouf, M. (1993). Observations on the population dynamics of snail hosts for schistosomes in the delta of the Senegal river basin. .7oumal of Molhcan Studies, 59,7-13. Betterton, C., Ndifon, G. T., Bassey, S. E., Tan, R. M. & Oyeyi, T. (1988). Schistosomiasis in Kano State, Nigeria I. Human infections near dam sites and the distribution and habitat preferences of potential snail intermediate hosts. Annals of Tropical Medicine and Parasitology, 82, 56 l-570. Bliss, C. I. & Owen, A. R. (1958). Negative binomial distributions with a common k. Biomenika, 45, 37-58. Bradley, D. J. &McCullough, F. S. (1973). Egg output stability and the epidemiology of Schistosoma haematobium Part II. An analysis of the epidemiology of endemic S. haematobium. Transactions of the Royal Society of Tropical Medicine and Hygiene,

61,491-500.

Brown, D. S. (1994). Freshwater Snails ofAfrica and theirMedical Importance, 2nd edition. London: Taylor & Francis. Chaine, J. P. & Malek, E. A. (1983). Urinary schistosomiasis in the Sahelian region of the Senegal river basin. Tropical and Geographical Medicine, 35,249-256. Chan, M.-S., Guyatt, H. L., Bundy, D. A. P., Booth, M., Fulford, A. J. C. &Medley, G. F. (1995). The development of an age-structured model for schistosomiasis transmission dynamics and control and its validation for Schistosoma munsoni. Epidemiology and Infection, 115, 325-344. Chandiwana, S. K., Makaza, D. & Taputaira, A. (1987). Variations in incidence of schistosomiasis in the highveld

D. J. SHAW ETAL.

150

region of Zimbabwe. Tropical Medicine and Parasitology, 39, 313-319. Chandiwana, S. K., Taylor, I’. & Clarke, V. de V. (1988). Prevalence and intensity of schistosomiasis in two rural areas in Zimbabwe and their relationship to village location and snail infection rates. Annals qf Tropical Medicine and Parasitology, 82, 163- 1’73. Chandiwana. S. K., Woolhouse, M. E. 1. &Bradlev, M. (1991). Factors affecting-the intensity of reinfection withSch&osoka haematobium following treatment with praziquantel. Parasitology, 102,73-83. Delegue, I’., Picquet, M., Shaw, D. J., Vercruysse, J., Sambou, B. & Ly, A. (1998). Morbidity induced by Schistosoma haematobium infections, as assessed by ultrasound before and after treatment with praziquantel, in a recently expanded focus (Senegal river basin). Annals of Tropical Medicine and Parasitology, 92, 775-783. de Vlas, S. J., Gryseels, B., van Oortmarssen, G. J., Polderman, A. M. & Habbema, J. D. F. (1992). A model for variations in single and repeated egg counts in Schistosoma mansoni infections. Parasitology, 104,45 l-460. de Vlas, S. J., Nagelkerke, N. J. D., Habbema, J. D. F. & van Oortmarssen, G. J. (1993). Statistical methods for estimating prevalence and incidence of parasitic diseases. Statistical Methods in Medical Research, 2, 3-21. Elias, E., Daffalla, A. A., Lassen, J. M., Madsen, H. & Christensen, N. 0. (1994). Schistosoma haematobium infection patterns in the Rahad Irrigation Scheme, Sudan. Acza Tropica, 58, 115-125. Etard, J. F., Bore], E. & Segala, C. (1990). Schistosoma haematobium infection in Mauritania: two years of follow-up after a targeted chemotherapy-a life-table approach of the risk of reinfection. Parasitology, 100, 399-406. Etard, J. F., Audibert, M. & Dabo, A. (1995). Age-acquired resistance and predisposition to reinfection with Schistosoma haematobium after treatment with praziquantel in Mali. American Journal of Tropical Medicine and Hygiene, 52, 549-558. Farooq, M. & Hairston, N. G. (1966). The epidemiology of Schistosoma haematobium and S. mansoni infections in the Egypt-49 Project Area. Bulletin of the World Health Organization, 35, 331-338. Gall, P. H. & Wilkins, H. A. (1984). Field studies on Bulinus senegalensis and transmission of Schistosoma haematobium infection in a Gambian community. Tropenmedizin und Parasitologie, 35, 29-36.

Gall, I’. H., Wilkins, H. A. & Marshall, T. F. de C. (1984). Dynamics of Schistosoma haematobium infection in a Gambian community. II. The effect on transmission of the control of Bulinus senegalensis by the use of niclosamide. Transactions of the Royal

Society

of Tropical

Medicine

and Hygiene,

18,

222-226. Jordan, P. (1985). Schistosomiasis-the St Lucia Project. Cambridge: Cambridge University Press. Jordan, I’., Webbe, G. & Sturrock, R. F. (1993). Human Schistosomiasis. Wallingford: CAB International. King, C. H., Much&i, E. M. & Ouma, J. H. (1992). Agetargeted chemotherapy for control of urinary schistosomiasis in endemic populations. Memorias do Institute Oswald0 Crux, 87, SIV, 203-210. Koura, M., Upatham, E. S., Awad, A. H. & Ahmed, M. D. (198 1). Prevalence of Schistosoma haematobium in the Koryole and Merca districts of the Somali Democratic Republic. Annals of Tropical Medicine and Parasitology, 75, 53-6 1. Mutapi, F., Ndhlovu? I’. D., Hagan, I’. & Woolhouse, M. E. J. (1997). A comparison of humoral responses to Schistosoma in areas with low and high levels of infection. haematobium Parasite Immunology, 19,255-263. Nagelkerke, N. J. D., Chunge, R. N. & Kinoti, S. N. (1990). Estimation of parasitic infection dynamics when detectability is imperfect. Statistics in Medicine, 9, 12 1 1- 12 19. Picquet, M., Emould, J. C., Vercruysse, J., Southgate, V. R., Mbaye, A., Sambou, B., Niang, M. & Rollinson, D. (1996). The epidemiology of human schistosomiasis in the Senegal

river

basin.

Medicine

Transactions of the Royal and Hygiene, 90,340-346.

Society

of Tropical

Picquet, M., Vercruysse, J., Shaw, D. J., Diop, M. & Ly, A. (1998). Efficacy of praziquantel against Schistosoma mansoni in northern Senegal. Transactions of the Royal Society of Tropical Medicine and Hygiene, 92, 90-93. Scott, D., Senker, K. &England, E. C. (1982). Epidemiology of human Schistosoma haematobium infection around Volta Lake, Ghana, 1973-5. Bulletin of the World Health Oranization, 60, 89%iO0.

Shaw. D. 1.. Grenfell. B. T. & Dobson, A. I’. (1998). Patterns of macroparasite aggregation in wildlife host populations. Parasitology, 117, 597-610. Shiff, C. J. (1973). The value of incidence for the assessment of schistosomiasis control. A study in southern Rhodesia. B&etin of the World Health Organization, 48, 409-4 14. Southgate, V. R. (1997). Schistosomiasis in the Senegal River Basin: before and after the construction of the dams at Diama, Senegal and Manantali, Mali and future prospects. Journal of Helminthology, 71, 125-132. Stelma, F. F., Talla, I., Polman, K., Niang, M., Sturrock, R. F., Deelder, A. M. & Gryseels, B. (1993). Epidemiology of Schistosoma mansoni infection in a recently exposed community in northern Senegal. American Joumalof Tropical Medicine andHygiene, 49, 701-706. Stelma, F. F., Talla, I., Sow, S., Kongs, A., Niang, M., Polman, K., Deelder, A. M. & Gryseels, B. (1995). Efficacy and side effects of praziquantel in an epidemic focus of Schistosoma mansoni. American Journal of Tropical Medicine and Hygiene, 53,167-170. Talla, I., Kongs, A., Verle, I’., Belot, J., Sarr, S. & Coll, A. M. (1990). Outbreak of intestinal schistosomiasis in the Senegal river basin. Annales de la Soci& Belge de Midecine Tropicale, 70, 1733180. van Etten, L., Kremsner, I’. G., Krijger, F. W. & Deelder, A. M. (1997). Day-to-day variation of egg output and schistosome circulating antigens in urine of Schistosoma haematobiuminfected school children from Gabon and follow-up after chemotherapy. American Journal of Tropical Medicine and Hygiene,

57, 337-341.

Vercruysse, J., Southgate, V. R. & Rollinson, D. (1985). The epidemiology of human and animal schistosomiasis in the Senegal river basin. Acta Tropica, 42, 249-259. Verli, I’., Stelma, F., Desreumaw, I’., Dieng, A., Diaw, O., Kongs, A., Niang, M., Sow, S., Talla, I., Sturrock, R. F., Gryseels, B. & Capron, A. (1994). Preliminary study of urinary schistosomiasis in a village in the delta of the Senegal river basin, Senegal. Transactions of the Royal Society of Tropical Medicine and Hygiene, 88,40 l-405. WHO (1983). Urznefiltration technique for Schistosoma haemaGeneva: World Health tobium. WHO document I’DI’/834. Organization. WHO (1993). The Control of Schistosomiasis. Second Report of the WHO&pert Committee. Geneva: World Health Organization, Technical Report Series, no. 830. Wilkins, H. A., Goll, I’. H., Marshall, T. F. de C. &Moore, P. J. (1984). Dynamics of Schistosoma haematobium infection in a Gambian community. III. Acquisition and loss of infection. Transactions

of the Royal

Society

of Tropical

Medicine

and

Hygime, 78,227-232. Woolhouse, M. E. J., Taylor, I’., Matanhire, D. & Chandiwana, S. K. (1991). Acquired immunity and the epidemiology of Schistosoma

haematobium.

Nature,

351,757-759.

Woolhouse, M. E. J., Dye, C., Etard, J.-F., Smith, T., Charlwood, J. D., Gamett, G. I’., Hagan, I’., Hii, J. K. L., Ndhlovu, P. D., Quinell, R. J., Watts, C. H., Chandiwana, S. K. & Anderson, R. M. (1997). Heterogeneities in the transmission of infectious agents: implications for the design of control programs. Proceedings of the NationalAcademy of Sciences of the USA,

94,338-342.

Received 15 June 1998; revised 2 December 1998; accepted 9 December 1998

for publication