Epidemiology of human Schistosoma haematobium infection around Volta Lake, Ghana, 1973-75

Bulletin of the World Health Organization, 60 (1): 89 - 100 (1982) Epidemiology of human Schistosoma haematobium infection around Volta Lake, Ghana, 1973-75 D. SCOTT, K. SENKER,2 & E. C. ENGLAND3 There was a dramatic rise in the prevalence of urinary schistosomiasis around Volta Lake within ayear ofitsfull impoundment in 1968. Research was undertaken to investigate the epidemiology of the disease in preparation for a controlprogramme. The interplay of threefactors-age, sex, and ethnic affiliation- largely defined the demographic patterns oftheprevalence and the intensity ofinfection. Both ofthese increased inyoung children up to apeak at age 10-14years, and then declined, the intensity ofinfection more rapidly than the prevalence. The prevalence and intensity of infection were both greater in males than females (above ages 15 - 24years and 5 - 9years, respectively), and differences between the two main ethnic groups were related to differences in their lake-related activities. Differ- ences between the patterns ofprevalence and intensity of infection are attributed to the greater sensitivity of the latter measurement in indicating changes in the level of trans- mission. Practical difficulties were encountered in obtaining a precise measurement ofinci- dence, the most important being the considerable degree ofpopulation movement. A field cohort study showed a seasonality of transmission, greatest between January and April, during the period of high level of the lake and in the early part of the draw-down. Research on the intermediate snail host (Bulinus truncatus rohlfsi) and lakesideecology established thefocality of transmission at human water-contact sites serving the shore-line communities and, in conjunction with parasitological surveys, its seasonality: variations in ecology that accompanied the annual rise and fall of the lake led to high levels of transmission when the water level was high and lower levels during the draw-down. The geographical distribution of the infection was also affected by differences in eco- logy, specifically by variations in the distribution and abundance of the aquatic weed Ceratophyllum demersum. A non-seasonal decline in transmission observed in one locality during the period from 1973 to 1975 resulted from a local decrease in the amount of Ceratophyllum. The creation of lakes in endemic zones of schisto- somiasis provides a greatly enhanced potential for transmission of the infection. The construction of a dam across the Volta River at Akosombo, Ghana, 420 km from the sea, began in 1962. Impoundment of the Volta Lake was completed six years later, giving a total surface area of 8500 km2. The original population of the impoundment area, numbering 70 000 persons, was displaced and the people were offered accommodation in 52 new townships built at different distances from the lake. A slightly larger number, mainly fishing families from the lagoons lying along the lower part of the Volta 1 Formerly Project Manager, WHO Schistosomiasis Control Project, Ghana. Present address: P.O. Box M.217, Accra, Ghana. 2 Formerly Epidemiologist, WHO Schistosomiasis Control Project, Ghana. Present address: Via Giuseppe Tartini 2I/I, 52000 Pula, Yugoslavia. 3Technical Officer, WHO Schistosomiasis Control Project, P.O. Box M.190, Accra, Ghana. River and from the delta at its mouth, were attracted to the lake by rich fish harvests in the early years of impoundment. They settled at random along the lake shore creating about 1000 new communities, many being accessible only by boat. Extensive surveys for urinary schistosomiasis among children were undertaken by the Ghanaian Ministry of Health from 1955 over much of the coun- try. While moderate to high prevalences were found around the lower reaches of the Volta River by Onori et al. (1), in summarizing unpublished departmental reports, McCullough & Ali (2) showed that infections with Schistosoma haematobium were, in general, relatively light upstream in the part of the river basin that would be affected by the new lake. The possibility of an increase in transmission of schisto- somiasis following the creation of the lake was foreseen by Macdonald (3), but the high prevalence of infection that actually occurred and the rapidity with 4149 -89- 90 D. SCOTT ET AL. which the lakeside communities were affected was unexpectedly dramatic. As early as 1969, surveys undertaken by Paperna (4) among children at different points along the lake revealed levels of prevalence of 907o, although in 1966 studies in a number of lakeside populations had shown low prevalence (Senker, unpublished data). Paperna's findings were confirmed by Jones,' who demon- strated high levels of transmission, particularly along the dendritic western shores, in the flooded valleys of the Afram, Obossum, Sene, and Pru rivers, results which indicated the need for detailed research into transmission and control of the infection. a JONES, C. R. Assignment report on the health component of the Volta Lake Research Project. Unpublished WHO document AFR/PHA/I 15: AFR/SCHIST/27 (1973). The studies reported in this paper were funded by UNDP and were carried out from 1973 to 1975. MATERIALS AND METHODS The project study area The area selected for study contained 26 com- munities along 80 km of the lake shore, around the Pawmpawm branch and the contiguous part of the southern bank of the Afram branch of the lake(Fig. 1). These communities did not comprise the entire population living along this section of the lake shore, I ~~~~~~~~~~~~~~~~~~II/ 0°15' 00 0005' 0.00 PAWMPAWMNYA 2 ATORTORSI 635' 3 POAKWE DORSE -6-35 st 4 KPONYOKOPE 635 5 FATEM 26 6 KASA24 W AFRAM BRANCH OF LAKE 7 POAKWE PAWMPAWMNYA 2^ 8 DAWA KOFI 9 ANYABONI DZOG9E 10 AKOKOMA KWADIA r i 12 ASIKOKO 6030' 20 ~~~~~~~~~13KUMA KUMA -6-30 20 516tt:14 AMEHIA 6°30 O K M 8 ) t 15 AKRUSU I16 ASAKESO w_-5re ~17 AKOTUI EAST 0* 15' ooio' IS~~~~~~~~1 1 AKOTUI WEST 0° lS oolo of t ~~~~~~~19AKATRI 20 TAMAYESO 30 20 1° 0o 1_ KEY MAP 6025' II. -.L1-y I- ~~ II. P~~~~~~AWAAPAWM -100 lo0 VOLTA ANYABON* LAKE 62' C 10 \ PROJECT '% 6¶*15 21 NYAFUTU 60 AREA 2 OU 23 KUNYASE /AMAFOESU A.CCRA 24 DUKUASE .5 0 KM 25 ODORTOM II0 80KM 26 ODORTOM 3° 20 1° 0° 1° 0005'W 0°00' Fig. 1. Study area, showing the location of the 26 villages investigated. SCHISTOSOMIASIS AROUND VOLTA LAKE since many families were found in small villages, ham- lets, or isolated homesteads situated in areas lying between the study villages. Thus the study area was not a continuous enclosed area. Rather, it consisted of 26 separate small enclaves dispersed along the edge of the lake. A population census of the area was made in 1972 and brought up to date annually. Two main ethnic groups are almost equally represented in 90/o of the population. They are the indigenous Krobo (farmers) and the immigrant Ewe (fishermen). The remainder were of mixed origin. Village populations ranged from 50 to 377, but there was considerable population movement among both the Krobo and the Ewe (5). There was a slight predominance of males (5107); among both sexes, half the population was under 15 years of age. Typically a village is a cluster of one or two-roomed buildings of very simple structure, lying perhaps 50 metres from the high water mark. Paths lead down to access points on the lake, where one or more families carry out all their water-related activities, collection of water, washing clothes, bathing, landing fish catches, cleaning nets, and mooring their canoes. These water-contact sites are particularly attractive to the children. The Krobos, normally, do not fish. The Ewes, in addition to fishing, generally cultivate staple food crops for use in the home. The women often engage in trading in addition to domestic work. The lake season The annual change in lake level amounted to about 3 m and the width of the land exposed at maximum draw-down varied from about 20 m to 120 m. Maxi- mum height is normally achieved by the end of October and a few weeks later the level begins to fall slowly over a period of 8 months, until the end of July. The subsequent annual flooding takes only 3 mnpnths. When the lake level is high the village water contact sites (WCSs) are well defined areas that vary in size and shape, and may cover as little as 50 M2. As the water level becomes lower the WCSs increase in size, and at maximum draw-down the lake's edge is fringed by an open beach. The annual rise and fall of the lake causes a regular cycle of change in the lakeside eco- logy which, in turn, greatly influences the pattern of human/water contact at different seasons. The lake cycle was regular during the period covered by the studies reported here (1973-75), except that an exceptionally high level was recorded in 1974. Parasitological investigations Up to the present, transmission of Schistosoma mansoni has not been associated with the lake; its snail intermediate host has not been found in the lake nor has it been reported elsewhere within 50 km of the study area. Stool examinations were made on a random sample of 252 persons from 5 villages in the study area (46.8%o were male and 60%o were children) using the Kato technique for quantitative diagnosis of intestinal helminth infections. No ova of S. mansoni were detected: hookworm, Ascaris lumbricoides, and Trichuris trichiuria were found with prevalences of 17.1 7o, 407o, and 3.607o, respectively. Degree ofpublic participation in urine surveys. A series of five surveys was carried out in the population of the study area from 1973 to 1975 to provide data on incidence, prevalence, and intensity of the infection. The dates and duration of the surveys, the element of the population investigated and the percentage participation are shown in Table 1. The fifth survey (S5) in 1975 selectively searched for infection among newcomers and those previously negative, in preparation for the introduction of a programme of chemotherapy towards the end of the year. In general, participation was related to the duration of a survey. If children under 2 years of age, who present a special problem as regards urine collection, are Table 1. Details of the five surveys for S. haematobium made in the study area, 1973 - 75 Survey Year Period Duration Element of No. No. Participation no. (months) population registered examined (%) 1 1973 Feb./June 4j Children onlya 1818 1328 73.0 2 1973 July/Sep. 2 Total 3480 2508 72.1 3 1973 Oct./Nov. 2 Children onlya 1835 1254 68.3 4 1974 Mar./Aug. 6 Total 3253 2824 86.8 5 1975 Apr./Nov. 7 (A selective survey - see text) a Born after 31 December 1956. 91 92 D. SCOTT ET AL. excluded the participation rates improve by about 4% in those aged under 15 years and by about 2.50/. in the whole population. The numbers of children examined in surveys 1, 2, 3, and 4 were 1328, 1236, 1254, and 1411, respec- tively. Although these figures are fairly close, it was not the same population on each occasion; in fact only one-third of the children were examined in all 4 surveys; about one-fifth of those registered were examined only once. The main cause of non-participation was absence from the area; with insufficiency of urine (less than 5 ml) in second place. Less than 10o of the population declined to take part in the surveys. Techniques. The parasitological surveys were made by house-to-house visits between 10 h 00 and 14 h 00 for the collection of a specimen of urine from each person present and shown on the census list. The urine was passed into a 370-ml plastic beaker fitted with a snap-on lid. Containers were not left for later use in the case of absentees or persons unable to provide a specimen; in such cases, a second visit was made. Some two hours later the specimen was shaken by hand and a 5-ml sub-sample was withdrawn, using an automatic syringe, and transferred to a container holding an equal quantity of a 10/0 aqueous solution of carbol fuchsin. Subsequently the specimen was passed through a Whatman's No. 1 filter-paper in a Millipore filter (47-mm diameter). The best results were obtained if the urine/stain mixture was held for a few days before filtration. Placing a transparent graticule over the paper facilitated counting of the schistosome ova, using a binocular microscope fitted with x 4 and x 10 objectives. The ova were counted in one quadrant; when the number was below 21, the count was continued over the whole paper; when it was 21-150 a second quadrant was counted and the total multiplied by 2 to give the value for the 5-ml sample of urine. This practice is justified by the demonstration of Dazo & Bilesb that with this technique the schistosome ova are evenly distributed over the filter-paper. In the case of higher counts in the first quadrant, the number was multiplied by four. Tests were made to assess the sensitivity of this tech- nique in detecting low egg counts. From suspensions containing 40 S. haematobium ova in 100 ml of known negative (blank) urine, 20 filter-paper prep- arations were made, each using 5 ml of urine. Similar tests were made using 60 ova in 100 ml. The prob- abilities of recording at least one egg at the expected levels of 2 per 5 ml and 3 per 5 ml were 0.7 and 0.825, respectively, and these values were considered satis- b DAZO, B. C. & BILES, J. E. Investigation on schisto- somiasis in Ghana. Unpublished WHO document, PD/72. 12 (1972). factory for the purposes of the project. Precautions were taken to minimize the risk of contamination of specimens due to the repeated use of laboratory equipment in processing of the urine samples. RESULTS Prevalence of infection and levels ofegg excretion in the urine Data from the two surveys of the total population (S2 and S4) are used to examine the demographic and geographical patterns of prevalence and intensity of infection (egg output). The term "egg output" is used to represent the geo- metric mean of the number of S. haematobium ova counted in the 5 ml random samples of positive urines from the total bladder content. The prevalence rates of urinary schistosomiasis in the total population surveys in 1973 and 1974 were 72.0% and 73.4%, respectively and the geometric mean egg outputs were 54.0 and 47.3 (Table 2). Prevalence rates were high and, since they were based on a single examination of a random 5-ml sample of urine, may well have understated the true figure; the pattern of distribution was strikingly consistent in the two surveys. Demographic distribution. The data on prevalence and intensity of infection (Table 2) are clearly influenced mainly by three factors-age, sex, and ethnic affiliation. In general, prevalence and egg output move together. The most apparent association is with age, both egg output and prevalence rise rapidly in the early years to a peak level in the 10- 14-year age group. While egg output falls sharply in the next decade before assuming a more gentle downward trend, prevalence declines from its peak level only slowly and relatively steadily with advancing years (Table 2). A sex differential, with higher values in males than in females, first becomes apparent in the egg output in the 5 - 9 age group but is not seen in prevalence rates until the age of 15-24 years. With both egg output and prevalence, this feature, once established, is maintained and, with prevalence, it increases with age (Tables 3 and 4). The relationships between ethnic affiliation and prevalence and egg output, respectively, are shown in Tables 3 and 4. There was a tendency for egg output to increase and then decrease more rapidly with age among the Krobo than among the Ewe and the peaks in egg output at age 10- 14 years were slightly higher in the Krobos. SCHISTOSOMIASIS AROUND VOLTA LAKE Similarly, the highest levels of prevalence in the youngest age groups are found among the Krobo. The difference between the two ethnic groups is lost, however, as the prevalence rate approaches 1001o in the 10- 14-year age group. In succeeding age groups, when the prevalence is falling, the rate of decline is- greatest among the Krobo, and after the age of 44 years it is significantly below that of the level found in the Ewe. Geographical distribution. There was wide vari- ation in prevalence of infection among the villages (in S2 ranging between 30.5% and 94.3%: in S4 between 39.3% and 94.5%). Egg output values showed similar variation, ranging from 13.8 to 132.2 in S2 and from 12.2 to 129.9 in S4. In general, the variations in the two parameters were similar in the different villages. There was a marked difference in the level of trans- mission between villages 1 - 14 and 15 - 26, corres- Table 2. Prevalence of S. haematobium and geometric mean of the egg output in 5-mi random samples of urine from those infected, by age No. No. Prevalence Geometric mean examined infected of egg output Age group (%) per random 5-ml urine (years) S2 S4 S2 S4 S2 S4 S2 S4 0 - 4 328 340 166 162 50.6 47.6 32.5 28.1 5 - 9 574 674 499/ 574 85.2 85.1 107.9 87.0 10 - 14 334 397 369 373 92.5 94.0 154.2 135.2 15 - 24 325 371 258 302 79.4 81.4 68.3 54.6 25 - 34 363 384 245 281 67.5 73.2 24.2 22.7 35 - 44 303 349 184 203 60.7 58.2 16.9 14.7 > 45 281 309 155 179 55.2 57.9 12.2 11.2 Total 2508 2824 1807 2074 72.0 73.4. 54.0 47.3 Table 3. Prevalences of S. haematobium by sex, age and ethnic group affiliation Male Female Total Age group Survey (years) Ewe Krobo Ewe Krobo Ewe Krobo 0 - 4 2 52.2 58.5 42.0 50.0 47.2 54.3 4 49.2 58.3 35.7 38.3 43.1 56.5 5 - 9 2 89.7 86.6 78.7 85.7 84.3 86.1 4 81.9 86.7 85.6 87.3 83.7 87.0 10 - 14 2 92.8 91.4 93.7 90.3 93.2 90.8 4 96.6 100.0 92.0 91.2 94.5 95.3 15- 24 2 86.9 87.7 72.0 67.1 79.9 76.6 4 83.5 83.8 79.1 74.1 81.3 78.7 25 - 34 2 83.2 78.0 60.5 46.0 70.8 60.2 4 88.0 84.7 62.2 68.6 73.5 76.0 35 - 44 2 74.2 74.5 43.5 41.5 61.6 56.0 4 69.0 74.5 42.2 42.7 56.8 55.6 > 45 2 69.0 60.6 39.3 31.1 61.6 47.0 4 68.7 60.3 50.0 39.1 64.1 50.7 Total 2 79.5 77.0 65.4 61.7 73.0 69.0 4 77.0 79.0 67.9 68.8 72.7 73.7 93 D. SCOTT ET AL. Table 4. Geometric mean of egg output per random 5-mi urine from those infected, by age and ethnic group Male Female Total Age group (years) Survey Ewe Krobo -Ewe Krobo Ewe Krobo 0 - 4 2 31.6 28.8 26.2 50.7 29.1 37.1 4 24.8 58.5 16.4 23.8 21.3 39.5 5 - 9 2 105.4 150.1 103.2 94.9 104.4 117.6 4 84.9 129.1 63.6 107.8 73.6 117.4 10 - 14 2 150.1 201.5 138.6 139.0 144.8 166.6 4 133.1 200.6 95.1 165.2 114.6 181.9 15 - 24 2 66.8 74.2 59.4 69.8 63.5 72.1 4 60.1 67.8 40.9 50.3 49.9 58.5 25 - 34 2 31.0 36.1 13.8 19.5 21.3 27.8 4 31.5 40.4 16.7 11.6 23.3 22.0 35 - 44 2 20.8 25.5 9.0 5.2 16.3 13.2 4 20.6 24.0 9.9 7.3 16.1 14.0 > 45 2 14.8 14.7 3.5 4.9 11.8 10.5 4 15.2 11.3 8.3 4.6 13.6 8.3 Total 2 53.4 62.2 47.7 50.6 51.0 56.5 4 49.1 63.6 38.4 43.8 44.1 53.1 ponding broadly with the Pawmpawm and the Afram branches of the lake, respectively. The villages in the latter group had significantly higher prevalences and egg output values than those in the Pawmpawm in both S2 and S4 (Table 5). Incidence of infection Two methods are available for investigating the incidence of new infections during a defined period: by comparing data from sequential surveys, or by the selection and repeated examination of a cohort of uninfected persons. In the following sections the problems encountered in obtaining a precise value for incidence by either method in the study area are described. Accordingly the values presented are referred to as crude estimates of the incidence. Datafrom sequential surveys. In theory this is very straightforward, persons becoming infected and excreting eggs in the urine between one survey and the next constitute the numerator, the denominator being the number originally negative. In practice there were problems in implementation and interpretation. Although the surveys were made progressively from east to west, from village No. 1 to No. 26, the system of revisits that was introduced to achieve maximum Table 5. Comparison of prevalence of infection and geometric mean of egg output for those infected, in communities on the Pawmpawm and Afram branches of the lake Geometric mean No. No. Prevalence of egg outputSurvey examined positive (%) per random 5-ml urine Pawmpawm branch 2 1392 910 65.4 41.6 (Villages 1 - 14) 4 1530 988 64.6 33.1 Afram branch 2 1116 897 80.4 70.2 (Villages 15-26) 4 1294 1086 83.9 65.4 Total 2 2508 1807 72.0 54.0 4 2824 2074 73.4 47.3 94 SCHISTOSOMIASIS AROUND VOLTA LAKE participation, and which accounted for about 307o of the population examined, resulted in considerable variation in the interval between one examination and the next. In the case of S2 (July- September 1973) and S4 (March -August 1974) the range of the interval was 8- 12 months. Betweeen S4 and S5 (April- November 1975) it was 9- 13 months. Apart from the variation in the interval between examinations, other factors contributed to the lack of precision. In areas of high transmission the composition of the uninfec- ted group changes as conversions occur unless it is constantly infused with new candidates: and high levels of incidence may be masked where all the persons in incidence-sensitive age groups are already infected at the outset. Population movement was a major constraint in the collection of adequate data. Periodic absence of uninfected persons during the sequence of surveys reduced the numbers on which rates could be based. Long absences occurring between surveys in infection-free areas could give a false impression of the level of local transmission. In Table 6 a comparison is made of the crude esti- mates of incidence between S2 and S4 (1973/74) and between S4 and S5 (1974/75). The data are presented by locality (Pawmpawm and Afram villages and total study area) and by age. The crude incidence estimates were markedly higher on the Afram than on the Pawmpawm. Field cohort study. A cohort of over 100 uninfected persons of all age groups was selected within the study area in 1974 for monthly examination. Possible candi- dates were provided by those whose urine examin- ations had been reported negative on previous surveys. Before admission to the study they were screened for infection by sedimentation of a specimen of urine for one hour and examination of the deposit; if no ova were found, a second preparation was examined. The criterion for acceptance into the study was the absence of ova in three successive daily sedimentations of urine. The absentee rate at the monthly re-examination from the original cohort was very disappointing, and screening surveys had to be repeated over the area. For example, of the 15 candidates admitted in September 1974, only one could be examined in the succeeding 4 months; none of the 10 admitted in November was found at home again. It might, in fact, be expected that in an area where the level of trans- mission is very high the negatives, as a group, would be the least consistent in their local residence, and often unavailable for examination. This proved to be the case. A second major problem was in trying to ensure that the cohort was representative of the whole area. In those villages with lower levels of prevalence, candidates were less difficult to find, but none were found in 5 of the 12 communities on the Afram branch of the lake. Moreover, as conversions occurred in the high transmission villages, the cohort increasingly lost any weighting it had, and replace- ments were difficult or impossible to find to redress the growing imbalance. The results of the study are presented in Table 7, showing the incidence by calendar quarter from October 1974 to September 1975, separately for the Pawmpawm and Afram branches. While the results of these studies are not claimed as a definitive state- ment on the level of incidence in the area they do demonstrate two important features. Firstly, there is a a distinct seasonality of transmission. Taking the time- for maturation of the worm in man as 6 - 8 weeks, it is seen that transmission is negligible from August to October, then it reaches a peak in the next three Table 6. Crude incidence by age group and locality, based on a comparison of data of S2 and S4 (incidence 1973/74) and of S4 and S5 (incidence 1974/75) 5 - 9 years 10 - 14 years Total study population No. in Incidence No. in Incidence No. in Incidence cohorta (%) cohort° (%) cohorta (% S2 - S4 (1973 - 74) Pawmpawm villages 48 25.0 13 38.5 354 27.4 Afram villages 15 60.0 2 100 134 49.3 Total area 63 33.3 15 46.7 488 33.4 S4 - S5(1974-75) Pawmpawm villages 37 32.4 7 14.3 320 16.9 Afram villages 8 50.0 0 - 118 40.7 Total area 45 35.6 7 14.3 438 23.3 a Number of uninfected persons at the first of the two surveys being compared who were re-examined at the second survey. 95 D. SCOTT ET AL. Table 7. Cohort studies of incidence of S. haematobium by calendar quarters, October 1974 - September 1975 Oct. - Dec. Jan. - Mar. Apr. - June July - Sept. Pawmpawm villages No. of persons 28 52 46 54 No. becoming positive 0 3 5 4 Incidence (%) 0 6 11 7 Afram villages No. of persons 13 17 11 1 No. becoming positive 0 7 8 1 Incidence (%) 0 41 73 100 All communities No. of persons 41 69 57 55 No. becoming positive 0 10 13 5 Incidence (%) 0 14 23 9 months, thereafter declining to July. The second characteristic is the contrast in the level of trans- mission between the two branches of the lake, being greater on the Afram than on the Pawmpawm, which accorded with the parasitological findings of preva- lence and intensity of infection. Trends in the level of transmission It is generally considered that of the three epidemi- ological measurements, prevalence rate, egg excretion level in the urine, and incidence of new infections, the latter provides the most sensitive index of the level of transmission and the prevalence rate the least. In a hypothetical case, where transmission ceased abruptly in an endemic area, incidence would fall immediately to zero, whereas the prevalence would decline only as egg excretion finally ceased in infected persons. Egg output would occupy an intermediate position. In the study area, both prevalence and egg output were influenced by a considerable degree of popu- lation mobility. As prevalence is the coarser measure- ment, in a closed community it would be expected that it would retain the same relationship to egg output where both are subjected to the same constraints on precision. In this study, the opportunity to examine trends is restricted to the comparison of data from only two successive surveys. In the case of prevalence and level of egg output in the urine these were S2 and S4, in 1973 and 1974. Trends in crude incidence are obtained by comparing the rate found from S2 to S4 with that occurring between S4 and S5. The findings are presented in Table 8. In drawing conclusions from the parasitological data, the high level of population mobility during the period must be taken into account and, in particular, the marked reduction of the population observed in 1973 and 1974. In examining the incidence data, in addition to the difficulties noted above, a monthly programme of snail control initiated in all 26 villages in May 1975 may also have affected trends by its influence on the S5 survey results. In consequence, the estimates of crude incidence derived from S4 - S5 must be considered as minimal. Interpretation of the findings of the parasitological surveys, in terms of trends in transmission, is greatly assisted by examination of the data from snail and ecological studies made at the same time. For this reason, transmission in the Afram villages will be considered later. However, in the case of the Pawmpawm branch the parasitological findings alone are sufficient to indicate that there was a decline in transmission: the reduction in the level of geometric mean of the egg count in the urine was statistically significant at the 5% level of probability, and in spite of the severe limitations which must be placed on interpretation of the crude rates of incidence the magnitude of the reduction recorded cannot be ignored. Table 8 demonstrates that when transmission appears to be declining, the evidence of the change is most marked in the incidence figures and least in prevalence (Pawmpawm branch) and that a reduction in the more sensitive factors measured, incidence and egg output, may be found when the prevalence is stable or even increasing. Related concurrent studies During the same period (1973 -75), research was undertaken in the study area on the snail intermediate host (B. truncatus rohlfsi) and on the ecological 96 SCHISTOSOMIASIS AROUND VOLTA LAKE Table 8. Changes in the prevalence of S. haematobium infection and in the geometric mean of the egg output found between 1973 and 1974 and changes in the crude incidence of infection between 1973/74 and 1974/75 Pawmpawm villages Afram villages (Numbers 1 - 14) (Numbers 15 - 26) Age group 1973 1974 Change 1973 1974 Change (years) (S2) (S4) (%) (S2) (S4) (%) Prevalence (%) 10 -14 90.4 90.2 -0.2 94.6 97.5 +3.1 All ages 65.4 64.6 -1.2 80.4 83.9 +4.4 Geometric mean of the egg output per random 5-ml urine 10-14 131.2 113.7 -13.3 179.9 157.7 -12.3 All ages 41.6 33.1 -20.4 70.2 65.4 -6.8 Crude incidence (%) 1973/74 1974/75 Change 1973/74 1974/75 Change (S2lS4) (S4/S5) (%) (S2IS4) (S4/S5) (%) 10-14 38.5 14.3 -62.9 60.0a 50.0a -16.7a No. in cohort 13 7 15 8 All ages 27.4 16.9 - 38.3 49.3 40.7 -17.4 No. in cohort 354 320 134 118 a On the Afram branch the figures relate to the age group 5 - 9 years, since all children aged 10 - 14 years who were examined in both surveys (S4 and S5) were already infected at the time of S4. conditions associated with transmission of the infection. In the final twelve months, investigations were also made into aspects of human behaviour relating to transmission. These studies, many of which have been published, emphasized the focality of S. haematobium transmission along the lake shore (6, 7) and the influence of the annual rise and fall of the lake on the seasonal pattern of the snail population and transmission dynamics (8), and confirmed the finding of Odei (9) concerning the close association between the distributions of B. truncatus rohlfsi and the aquatic weed Ceratophyllum demersum (10). Snail ecology. Snail sampling data are shown in Table 9, where the numbers of B.truncatus rohlfsi collected, the numbers infected with mature cercariae, and the infection rates are presented for the period January - May 1974 and compared with the findings for the same period in the following year, these months covering the greater part of the season of high transmission. In both periods there was a higher level of trans- mission on the Afram than on the Pawmpawm, and this is consistent with the ecological observations and with the findings on the parasitology surveys. There was a decline from 1974 to 1975 in the number of infected snails collected on the Pawmpawm and in the Table 9. Numbers of snails collected, numbers infected with mature cercariae, and infection rates in 1974 and 1975 January - May January - May 1974 1975 10 Pawmpawm villages No. of snails collected 581 460 No. infected 52 26 Infection rate (%) 9.0 5.7 6 Afram villages No. of snails collected 1010 1554 No. infected 155 122 Infection rate (%) 15.3 7.8 infection rate. This evidence supports the conclusion based on the parasitological data that the level of transmission was falling on this branch of the lake, and is in accord with the reductions in abundance and distribution of Ceratophyllum at the water contact sites on the Pawmpawm found by Klumpp & Chu (10), and their conclusions relating to the importance of this aquatic weed to snail populations. On the Afram, where Ceratophyllum was 97 D. SCOTT ET AL. becoming more plentiful (10), the total number of B.truncatus rohlfsi taken by sampling in selected water contact sites greatly increased in 1975, as would be expected. The number of infected snails, the more reliable index of transmission, however, declined. In view of the abnormally high levels of the lake in November 1974 it is impossible to make a definitive statement on the trend of transmission on the Afram based on evidence that indicated a manifestly worsening ecological situation, but which was equivocal in relation to the intermediate snail host. Examination of these data together with those of the parasitological surveys (Table 8), which showed a reduction in the level of egg excretion in the urine between 1973 and 1974, and a fall in incidence between 1974 and 1975, suggests that transmission had previously been declining, but provides no clear- cut evidence on the trend in the early part of 1975. Studies ofhuman water contact. The results of the observations made at water contact sites in the second half of 1974 in one of the project villages, and repeated seasonally during the next twelve months have been reported by Dalton & Pole (11). There was no evidence of seasonal variation in water contact. The main finding was a relationship between the degree of water contact (in terms of frequency and duration, and the nature of the associated activity), and the level of urinary egg excretion expressed by the egg count. Age was less important than the degree of contact. These studies showed that contact increases in early childhood to reach a maximum level between the ages of 5 and 9 and thereafter declines progressively with age. The mean duration of contact was greater in males than in females, this feature being evident even in the youngest. A comparison of these findings with those of the contemporary parasitology surveys shows that changes in the pattern of water contact experienced by the two sexes from childhood through adult life is subsequently closely reflected in the level of the intensity of infection and, with less refinement, in prevalence. Water contact data showing the variations associ- ated with the ethnic groups as well as those of age and sex have also been recorded (Dalton, unpublished data, 1978). Observations made on four consecutive days at six villages on 93 Ewe and 28 Krobo children under 15 years of age and 84 Ewe and 19 Krobo adults, indicate that Krobo children are in contact with water longer than Ewe children, but that among adults the reverse is the case. These findings help to explain the higher prevalence of infection in the Krobo than in the Ewe children, the subsequent lack of difference between the two groups up to the age of about 35, and finally, the later appearance of higher' levels of prevalence in the Ewes, being most evident in those aged over 44 years (Tables 3 and 4). The combined results of the parasitological and water contact studies thus confirm the conclusion that the demographic pattern of S. haematobium infection in the study area is basically determined by variations in the level of water contact experienced by the different age, sex, and ethnic groupings of the population. EPIDEMIOLOGICAL ASSESSMENT Snail and ecological surveys undertaken in the study area have established the focality of trans- mission of S. haematobium at the village water con- tact sites (7, 8). They have also demonstrated a seasonality of transmission closely associated with the annual rise and fall of the lake, with a peak level occurring soon after the lake reaches its maximum height, and this has now been corroborated by field studies of incidence. Knowledge of these fundamental epidemiological features is of great importance in examining the feasibility of a control programme and in its planning and execution. Parasitological findings and the results of the snail surveys were also in complete accord in describing the contrast between the Afram and Pawmpawm related villages and in the non-seasonal trend in transmission which became apparent in the widening difference between the two localities that occurred from 1973 to 1974. Lakeside ecology thus plays the dominant role in determining the characteristics of focality and season- ality of transmission. At high water level, conditions are not only most favourable for the snail but also impose a close association between the definitive and intermediate hosts, facilitating transmission between them. The geographical variations in transmission, and the non-seasonal trend were also determined by variations in ecology, specifically in the distribution and abundance of Ceratophyllum demersum, the aquatic weed most closely associated with B. trun- catus in the lake. As a result of repeated exposure at the village water contact sites, few children in the study area remained uninfected by the age of 14 years. The levels of trans- mission to which the different elements of the popu- lation are exposed, and the changes that occur in the course of time, are shown in variations of the crude estimates of incidence, of intensity of infection, and prevalence. The demographic patterns of both inten- sity of infection and prevalence are similar in their rapid rise to a peak level in the age group 10- 14, in demonstrating a subsequent decline, in portraying higher levels in males than females and in showing a clear contrast between the ethnic groups. That vari- ations of egg output found at successive surveys do not run strictly in parallel with those of prevalence is 98 SCHISTOSOMIASIS AROUND VOLTA LAKE attributed to the greater sensitivity of the former in indicating changing levels of transmission over a period of time. For example, whilst the prevalence in the Afram-related villages rose slightly between 1973 and 1974 (S2 - S4) a significant fall in the value of the egg output was recorded simultaneously (Table 8). Evidence of trends from the crude estimates of inci- dence, from the sequence of monthly snail and eco- logical studies also suggest that transmission was declining at that time on this part of the lake, but that the rate of fall, in combination with its duration, was insufficient to be reflected in the prevalence. Data from water contact studies indicate that vari- ations in the level of transmission found by age, sex and ethnic groupings are related to their different degrees of exposure to infection at the contact sites and that it is unnecessary to postulate underlying immunological factors. ACKNOWLEDGEMENTS We wish to express our gratitude to Dr E. G. Beausoleil, Director of Medical Services, Ministry of Health, Ghana, for permission to publish this paper and to the Technical Officers of the Epidemiology Unit on secondment from the Ministry of Health for the high level of competence of their work in the laboratory and in the field and for their very full support. We wish also to thank the staff of the Biology and Sociology Units for their cooperation, and in particular Dr K. Y. Chu, Mr R. K. Klumpp, and Dr P. Dalton for helpful discussions on and for use of their unpublished data. We are particularly grateful to Dr A. Davis, Director, Parasitic Diseases Programme, WHO, Geneva, for his continued advice and support. RtSUMPE tPIDtMIOLOGIE DE L'INFECTION HUMAINE A SCHISTOSOMA HAEMATOBIUM AUTOUR DU LAC VOLTA AU GHANA, 1973 -1975 Dans l'ann6e qui a suivi la pleine mise en eau du Lac Volta (Ghana) ont 6te constatees des pr6valences elevees de schistosomiase urinaire en des lieux ou elles etaient aupara- vant faibles. Les conclusions 6pidemiologiques issues de 2 ans et demi d'investigations effectu&es prealablement au lancement d'un programme de lutte se fondent sur les resul- tats de 5 enquetes par examen des urines realis6es parmi les 3500 habitants de 26 villages s'echelonnant sur 80 km au bord du lac, de meme que sur les resultats d'investigations paralleles faites sur le mollusque h6te interm6diare (Bulinus truncatus rohlfsi) et sur le comportement humain. La popu- lation se composait principalement de deux groupes ethni- ques, les paysans indigenes Krobo et les pecheurs Ewe qui avaient migr6 ves le lac en provenance du cours inferieur de la Volta. Un facteur d'importance epidemiologique majeure reside dans l'elevation et la baisse annuelles du niveau du lac et dans les variations saisonnieres qui se produisaient en conse- quence dans l'ecologie des rives du lac. En periodes de hautes eaux, l'acces au lac 6tait limite a des secteurs relative- ment faibles et bien circonscrits en raison de la densite de la vegetation; en periodes de basses eaux, I'acces etait plus facile grAce aux vastes etendues de plages de sable sans vegetation. Au cours des enquetes parasitologiques, on a preleve des 6chantillons d'urine entre 10 h 00 et 14 h 00 et l'examen microscopique du residu colore sur filtre Millipore fournis- sait des donn6es sur la prevalence et sur le niveau de l'excre- tion d'aeufs dans l'urine (intensite de l'infection) chez les sujets infectes. Les typologies demographiques de la prevalence et de l'intensite de l'infection etaient largement determinees par l'interaction entre trois facteurs: age, sexe et appartenance ethnique. Les deux parametres augmentaient chez les jeunes enfants pour atteindre un pic entre 10 et 14 ans d'age puis declinaient, le phenomene etant plus rapide pour l'intensite de i'infection que pour la prevalence. La pr6valence et l'intensite de l'infection etaient l'une et l'autre plus fortes chez les hommes que chez les femmes (au-dessus des ages 15-24 ans et 5-9 ans respectivement) et les differences entre les deux principaux groupes ethniques etaient attri- buables a la dissemblance de leurs activites en rapport avec le lac. Les differences entre les evolutions de la prevalence et de l'intensite de l'infection sont explicables par le fait que la deuxieme mesure est un indicateur plus sensible des change- ments intervenus dans le niveau de la transmission. L'obtention d'une mesure precise de l'incidence s'est heurtee a des difficultes pratiques, dont la principale etait le degre considerable de mouvement de population. Une 6tude de cohorte menee sur le terrain a montre que la transmission revetait un caractere saisonnier, atteignant son maximum entre janvier et avril, pendant la periode de hautes eaux et dans la premiere partie de la periode d'abaissement des eaux. Des etudes sur l'h8te intermediaire et l'6cologie des rives du lac ont montr6 que la transmission etait focalisee aux points de contact avec l'eau des communautes riveraines, ce qui etaie et explique les indications parasitologiques quant au caractere saisonnier de la transmission. En periodes de hautes eaux, les conditions &cologiques sont les meilleurs pour les mollusques et favorisent une etroite proximite entre h8tes definitifs et h8tes intermediaires en un petit nombre de points de contact restreints, tres frequentes et fortement pollues, des villageois avec l'eau. De telles conditions ne 99 100 D. SCOTT ET AL. peuvent qu'accroi^tre la transmission du parasite. Les resultats des enquetes parasitologiques et des 6chantil- lonnages de mollusques ont fait apparaitre un degr6 plus eleve de transmission dans le secteur Afram du lac que dans le secteur Pawmpawm. Cette disparit6 g6ographique etait egalement lie a 1'cologie bordiere, plus specialement a la -repartition et a I'abondance de Ceratophyllum demersum, plante aquatique la plus 6troitement associ6e a B. truncatus sur le lac. Une baisse non saisonniere de la transmission observ&e dans le secteur Pawmpawm pendant une periode de deux ans resultait d'une diminution locale de I'abon- dance de Ceratophyllum. REFERENCES 1. ONORI, E. ET AL. Schistosomiasis in the Volta Region of Ghana. Annals of tropical medicine and parasitology, 57: 59-70 (1963). 2. MCCULLOUGH, F. S. & ALI, Y. M. The distribution and prevalence of Schistosoma haematobium and Schisto- soma mansoni in Ghana. Ghana medicaljournal, 4(3): 83-84 (1965). 3. MACDONALD, G. Medical implications of the Volta River Project. Transactions .of the Royal Society of Tropical Medicine & Hygiene, 49(1): pp.'8-19 (1965). 4. PAPERNA, I. Study of an outbreak of schistosomiasis in the newly formed Volta Lake in Ghana. Zeitschrift fur Tropenmedizin und Parasitologie, 21: 411-424 (1969). 5. SENKER, K. WHO workshop report. In: Proceedings of the International Conference on Kainji Lake and River BasinsDevelopment inAfrica. Ibadan, 11-17December 1977. Kainji Lake Research Institute, 1979, vol. 2, p.601. 6. CHU, K. Y. & VANDERBERG, J. A. Techniques for esti- mating densities of Bulinus truncatus rohlfsi and its horizontal distribution in Volta Lake, Ghana. Bulletin of the World Health Organization, 54: 441-416 (1976). 7. CHU, K. Y. & KLUMPP, R. K. Focal transmission of Schistosoma haematobium in Lake Volta, Ghana, In: Proceedings of the International Conference on Schistosomiasis, Cairo, October 1975, 1978, pp. 85-88. 8. KLUMPP, R. K. & CHU, K. Y. Ecological studies of Bulinus rohlfsi, the intermediate host of Schistosoma haematobium in the Volta Lake. Bulletin of the World Health Organization, 55: 715-730 (1977). 9. ODEI, M. Observations on some weeds of malacological importance in Volta Lake. Bulletin de l'InstitutFranpais d'Afrique noire, Ser. A, 35: 57-66 (1973). 10. KLUMPP, R. K. & CHU, K. Y. Importance of the aquatic weed Ceratophyllum to transmission of Schistosoma haematobium in the Volta Lake, Ghana. Bulletin ofthe World Health Organization, 58: 791-798 (1980). 11. DALTON, P. R. & POLE, D. Water contact patterns in relation to Schistosoma haematobium infection. Bulletin ofthe World Health Organization, 56: 417-426 (1978).