Effect of Water Resource Development and Management on Lymphatic Filariasis, and Estimates of Populations at Risk

Home , Elephantiasis, Mansonia annulifera

EFFECT OF WATER RESOURCE DEVELOPMENT AND MANAGEMENT ON LYMPHATIC FILARIASIS, AND ESTIMATES OF POPULATIONS AT RISK

TOBIAS E. ERLANGER, JENNIFER KEISER, MARCIA CALDAS DE CASTRO, ROBERT BOS, BURTON H. SINGER, MARCEL TANNER, AND JÜRG UTZINGER* Swiss Tropical Institute, Basel, Switzerland; Department of Geography, University of South Carolina, Columbia, South Carolina; Water, Sanitation and Health, World Health Organization, Geneva, Switzerland; Office of Population Research, Princeton University, Princeton, New Jersey

Abstract. Lymphatic filariasis (LF) is a debilitating disease overwhelmingly caused by Wuchereria bancrofti, which is transmitted by various mosquito species. Here, we present a systematic literature review with the following objectives: (i) to establish global and regional estimates of populations at risk of LF with particular consideration of water resource development projects, and (ii) to assess the effects of water resource development and management on the frequency and transmission dynamics of the disease. We estimate that globally, 2 billion people are at risk of LF. Among them, there are 394.5 million urban dwellers without access to improved sanitation and 213 million rural dwellers living in close proximity to irrigation. Environmental changes due to water resource development and management consistently led to a shift in vector species composition and generally to a strong proliferation of vector populations. For example, in World Health Organization (WHO) subregions 1 and 2, mosquito densities of the Anopheles gambiae complex and Anopheles funestus were up to 25-fold higher in irrigated areas when compared with irrigation-free sites. Although the infection prevalence of LF often increased after the implementation of a water project, there was no clear association with clinical symptoms. Concluding, there is a need to assess and quantify changes of LF transmission parameters and clinical manifestations over the entire course of water resource developments. Where resources allow, integrated vector management should complement mass drug administration, and broad-based monitoring and surveillance of the disease should become an integral part of large-scale waste management and sanitation programs, whose basic rationale lies in a systemic approach to city, district, and regional level health services and disease prevention.

INTRODUCTION scale operations were launched in 2000, alongside the forging of a worldwide coalition, the Global Alliance to Eliminate People living in tropical and subtropical countries have Lymphatic Filariasis (GAELF), which is a free and nonre- long suffered under the yoke of lymphatic filariasis (LF). This strictive partnership forum. WHO serves as its secretariat and chronic parasitic disease is of great public health and socio- is being reinforced by an expert technical advisory group.12–14 economic significance and is currently endemic in 80 coun- GPELF’s goal is to eliminate the disease as a public health tries/territories of the world.1–3 LF accounts for serious dis- problem by 2020. It mainly relies on mass drug administration figuration and incapacitation of the extremities and the geni- using albendazole plus either ivermectin or diethylcarba- tals and causes hidden internal damage to lymphatic and renal mazine (DEC). At the end of 2003, approximately 70 million systems.4–6 Disease, disability, and disfiguration are respon- people were treated and 36 countries had an active control sible for a loss of worker productivity, significant treatment program in place.14 costs, and social stigma.7,8 At present, the global burden of LF Sustained political and financial commitment and rigorous is estimated at 5.78 million disability adjusted life years monitoring and surveillance are essential elements of the (DALYs) lost annually.9 Hence, its estimated burden is al- global program, as otherwise LF could reemerge because a most 3.5-fold higher than that of schistosomiasis and approxi- small fraction of the population will continue to carry microfi- mately one seventh of that of malaria.9 LF is caused by laria. Furthermore, the vector population is unlikely to be Wuchereria bancrofti, Brugia malayi, and Brugia timori, with significantly affected by GPELF. Employing a mathematical > 90% of cases attributable to W. bancrofti.1 Transmission modeling approach, it was shown that vector control pro- occurs through various mosquito species, primarily Culex grams, in addition to mass drug administration, would sub- (57%), followed by Anopheles (39%), Aedes, Mansonia, and stantially increase the chances of meeting GPELF’s ambitious Ochlerotatus. Detailed information on the geographical dis- target.15 Indeed, some of the most successful control pro- tribution of the most important LF vectors can be found else- grams in the past demonstrate that an integrated approach, where.2 More than 60% of all LF infections are concentrated readily adapted to specific eco-epidemiologic settings, was a in Asia and the Pacific region, where Culex is the predomi- key factor for controlling and even eliminating LF.16–19 nant vector. In Africa, where an estimated 37% of all infec- In rural areas undergoing ecological transformations, par- tions occur, Anopheles is the key vector.2 ticularly due to the construction of irrigation schemes and In 1993, the World Health Organization (WHO) declared dams, new breeding sites suitable for filaria vectors are cre- LF to be one of six eliminable infectious diseases.10 After ated.16,20 As a consequence, the transmission dynamics of LF several years of preparation and endorsement by the World is expected to change. In Africa, where Anopheles transmit Health Assembly in 1997, the Global Program to Eliminate malaria and filaria, the estimated surface area of 12 million ha Lymphatic Filariasis (GPELF) was initiated in 1998.11 Large- under irrigation in 1990 is estimated to increase by one third until 2020.21 Rapid and uncoordinated urbanization often leads to new habitats for filaria vectors.22,23 Especially poor * Address correspondence to Jürg Utzinger, Department of Public design and lack of maintenance of infrastructures for drainage Health and Epidemiology, Swiss Tropical Institute, CH-4002 Basel, of sewage and storm water, waste-water management, water Switzerland. E-mail: [email protected] storage, and urban subsistence agriculture can facilitate the 523 524 ERLANGER AND OTHERS proliferation of mosquitoes, including those transmitting fi- that chronic parasitic diseases, including LF, could be used as laria. Although the proportion of urban dwellers in the least viable health indicators for monitoring poverty alleviation, as developed countries was only 27% in 1975, it rose to 40% in the root ecological causes of these health conditions depend 2000 and is predicted to further increase. Nearly 50% of the on poor sanitation, inadequate water supply and lack of vec- world’s urban population is concentrated in Asia. Currently, tor control measures.27 the annual growth rate in Asian cities is 2.7%.24 This implies Search strategies and selection criteria. With the aim of that in the future, an increasing number of habitats with or- identifying all published studies that examined the effect of ganically polluted water will be available for Culex vectors. water resource development and management on the fre- The objectives of the systematic literature review presented quency and transmission dynamics of LF, we carried out a in this paper were (i) to assess the current size of the popu- systematic literature review. Particular consideration was lation at risk of LF with particular consideration of water given to publications that contained specifications on (i) en- resource development and management, both in rural and tomological transmission parameters, abundance of vector urban settings, and (ii) to assess the effect of these ecological populations, microfilaria infection prevalence and rates of transformations on the frequency and transmission dynamics clinical manifestations as a result of water resource develop- of LF. Our working hypothesis was that environmental ment, and (ii) studies that compared sites where environmen- changes resulting from water resource development and man- tal changes occurred with ecologically similar settings where agement adversely affect vector frequencies, filaria transmis- no water resource developments were implemented. sion, prevalence of infection, and clinical occurrence of LF. As a first step, we performed computer-aided searches us- These issues are of direct relevance for GPELF and evidence- ing the National Library of Medicine’s PubMed database, as based policy-making, and for integrated vector management well as BIOSIS Previews, Cambridge Scientific Abstracts In- programs and optimal resource allocation for disease control ternet Database Service, and ISI Web of Science. We were more generally. interested in citations published as far back as 1945. The following keywords (medical subject headings and technical terms) were used: “lymphatic filariasis” in combination with MATERIALS AND METHODS “water,”“water management,”“reservoir(s),”“irrigation,” “dam(s),”“pool(s),”“sanitation,”“ecological transformation,” and “urbanization.” No restrictions were placed on lan- Contextual determinants and estimation of population at guage of publication. risk in endemic countries. As a first step, we outlined the contextual determinants of LF transmission in a simplified In a next step, the bibliographies of all recovered articles flow chart. For regional estimates of populations at risk of LF, were hand-searched to obtain additional references. In an we used the recent classification set forth in the appendices of iterative process, this approach was continued until no new the annual World Health Report of WHO, which stratifies the information was forthcoming. world into 14 epidemiologic subregions.9 For estimation of Dissertation abstracts and unpublished documents (“gray population fractions at risk of LF due to water resource de- literature”) were also reviewed. Dissertation abstracts were velopment and management, we adopted setting-specific defi- searched in online databases, that is, ProQuest Digital Disser- nitions. Hence, for rural areas we considered those people at tations and the Unicorn Online Catalogue (WEBCAT) of the risk of LF who live in close proximity to irrigated agro- London School of Hygiene and Tropical Medicine. ecosystems, employing data sources from the Food and Ag- Finally, online databases of international organizations and ricultural Organization (FAO; http://www.fao.org). We fol- institutions, namely WHO and FAO of the United Nations, lowed a similar approach as in our preceding work with an and the World Bank, were scrutinized, adhering to the same emphasis on the malaria burden attributable to water research strategy and selection criteria explained above. source development and management.25 In fact, the size of the rural irrigation population was estimated by multiplying the average population density in rural areas by the total area RESULTS currently under irrigation in LF-endemic countries/territories. In urban settings, the size of the population at risk of LF Contextual determinants. The contextual determinants of was defined by the proportion that currently lacks access to LF can be subdivided into three broad categories, namely (i) improved sanitation. Country-specific percentages of urban environmental, (ii) biological, and (iii) socioeconomic (Figure dwellers without access to improved sanitation were taken 1). They act on different temporal and spatial scales, adding from the World Health Report 2004.9 Justification for this to the complexity of the local LF eco-epidemiology. indicator is derived from the following experiences. First, In the first category, LF transmission is mainly determined there is evidence that, besides common water-borne diseases, by climatic factors and the formation or disappearance of lack of access to clean water and improved sanitation in- suitable breeding sites for the vector. Breeding sites can be creases the risk of acquiring vector-borne diseases.23,26,27 As either natural or man-made, and their productivity exhibits will be shown in our review and has been noted before, LF strong heterogeneity, even on a small scale, which in turn transmission is spurred by rapid urbanization in the absence governs filarial transmission dynamics. of accompanying waste management and sanitation facility In rural settings, the most prominent man-made breeding programs.28–32 Second, a large-scale campaign built around sites are water bodies created by irrigation systems and dams. chemotherapy and improved sanitation proved successful to Here, the weight of environmental determinants is strongly control LF in the Shandong province, People’s Republic of associated with biologic factors, notably vector and parasite China.33 Third, Durrheim and colleagues recently suggested species, and various socioeconomic factors such as human EFFECT OF WATER PROJECTS ON LYMPHATIC FILARIASIS 525

FIGURE 1. Contextual determinants of lymphatic filariasis. migration patterns, access to, and performance of, health sys- systems in private dwellings and industrial units, or the ab- tems, and individual protective measures. sence of them entirely. Here, biological factors shape the epi- In urban areas, artificial breeding sites are often created by demiology of LF after environmental changes have occurred, waste-water mismanagement, resulting from poor sanitation and socioeconomic factors strongly interact with the environ- 526 ERLANGER AND OTHERS mental determinants. The local quality of domestic and industrial wastewater management, access to clean water and im- TABLE 1 proved sanitation, and the construction of roads and buildings Estimates of population at risk in all lymphatic filariasis (LF)- depend on the socioeconomic status of specific subpopulations. endemic countries/territories of the world, stratified into WHO epidemiological subregions (population at risk of LF in thousands) Endemic countries/territories. Table 1 shows estimates of populations at risk of LF for all the countries/territories Africa where the disease is currently endemic. Only politically inde- WHO subregion 1a (24 countries) Hence, the popula- Angola (10,423), Benin (6,736), Burkina Faso (12,963),b .(76 ס pendent countries were listed (N Cameroon (9,338), Cape Verde (n.d.), Chad (6,216), Comoros tions at risk of French Polynesia, New Caledonia, Réunion, b and Wallis and Futuna, which belong to France, and Ameri- (768), Equatorial Guinea (89), Gabon (896), Gambia (1,235), Ghana (6,200),b Guinea (8,336), Guinea-Bissau (1,253), can Samoa, which belongs to the United States, were assigned Liberia (34), Madagascar including Reunionc (15,841), Mali to the geographically closest independent states. Timor-Leste, (11,329), Mauritius (12),d Niger (10,416), Nigeria (121,901), which recently became independent, is also included. How- Sao Tome and Principe (n.d.), Senegal (9,247), Seychelles b ever, no estimates for at-risk populations are currently avail- (81), Sierra Leone (890), Togo (1,182) WHO subregion 2a (14 countries) able for the following LF-endemic countries: Cambodia, Cape Burundi (1,112), Central African Republic (765), Congo Verde, Lao People’s Democratic Republic, Republic of Ko- (3,396), Coˆ te d’Ivoire (14,253), Democratic Republic of the rea, Solomon Islands, and Sao Tome and Principe. In view of Congo (22,481), Ethiopia (3,534), Kenya (10,108), Malawi (11,948), Mozambique (15,336), Rwanda (3,355),e Uganda relatively small population sizes living in these countries, ne- f glecting at-risk population of LF there, only marginally influ- (23,399), United Republic of Tanzania (14,421), Zambia (9,980), Zimbabwe (10,816) ences estimates on regional and global scales. The Americas People at risk of LF at a global and regional scale. We WHO subregion 4 (6 countries) estimate that approximately half of all people currently living Brazilg (3,569),h Costa Ricag (83),h Dominican Republic h h g i in LF-endemic countries are at risk of the disease, which (1,854), Guyana (623), Suriname (< 4), Trinidad and Tobagog (< 13)h translates to approximately 2 billion. This is considerably WHO subregion 5 (1 country) higher than the 1–1.2 billion estimates put forth in the litera- Haiti (6,078)b ture.1,2,11 The difference is largely explained by at-risk esti- Eastern Mediterranean WHO subregion 7 (3 countries) mates for China. In urban areas, there are 394.5 million at risk f b h k of LF due to lack of access to improved sanitation. This is Egypt (2,446), Sudan (8,302), Yemen (100) Southeast Asia almost twice the estimated size in rural areas, namely 213 WHO subregion 11 (3 countries) million, which is attributed to living in close proximity to Indonesia (27,046)h [B. malayi: 27,046, B. timori: 3,900],l Sri irrigated agriculture. The largest percentages in terms of LF Lanka (9,900),b Thailandm (10,116)k [B. malayi: 7,791]k burden, as expressed in DALYs lost (52%), people at risk WHO subregion 12 (6 countries) Bangladesh (93,984),h India (494,374)h [B. malayi: 190,718],h (29%), size of the population at risk due to proximity to Maldives (< 3),n Myanmar (28,000),b Nepal (1,359),h irrigated land (69%), and lack of improved sanitation (33%) Timor-Leste (778)i [B. timori: 778]i are in WHO subregion 12. This subregion includes Bangla- Western Pacific WHO subregion 13 (1 country) desh, India, Maldives, Myanmar, Nepal, and Timor-Leste o (Table 2). Brunei Darussalam (40) WHO subregion 14 (18 countries) Studies identified and qualitative overview. Overall, 12 Cambodia (n.d.), China (925,979)h [B. malayi: 63,906],h Cook studies fulfilled the selection criteria of our literature review. Islands including French Polynesiac (248),k Federated States These studies were all published in the peer-reviewed litera- of Micronesia (109),k Fiji including Wallis and Futunac (854),k k ture, that is, in specialized entomology, parasitology, and/or Kiribati (88), Lao People’s Democratic Republic (n.d.), Malaysiag (2,736)h [B. malayi: 2,736],h Niue (2),k Papua New tropical medicine journals. None of the work retrieved from Guinea (3,000),p Philippines (23,800)b [B. malayi: 23,800],b electronic databases other than PubMed or ISI Web of Sci- Republic of Korear (n.d.), Samoa including American Samoac ence was deemed of sufficient quality to justify study inclu- (248),k Solomon Islandsr (n.d.), Tonga (104),k Tuvalu (11),k Vanuatuf including New Caledoniac (422),k Viet Nam sion. h Table 3 summarizes the main findings of the selected stud- (12,288) n.d., no data currently available. ies, stratified by rural and urban settings. As a common a Except Mauritius, percentages of the population at risk from Lindsay and Thomas,59 theme, LF vector composition frequencies shifted in all set- recalculated with recent figures from United Nations.60 b Weekly Epidemiological Record.14 tings. Water resource developments favored An. gambiae, c Re´union, French Polynesia, Wallis and Futuna, and New Caledonia belong to France; American Samoa belongs to the United States. An. funestus, An. barbirostris, Culex quinquefasciatus, Cu. d WHO.61 pipiens pipiens, Cu. antennatus, and Aedes polynesiensis, but e For Rwanda the same “at-risk” percentage as for Burundi was taken. f A significant reduction in prevalence and intensity of microfilaria has recently been disfavored An. pharoensis, An. melas, An. subpictus, and Ae. recorded in the United Republic of Tanzania, Egypt, Samoa, and Vanuatu.3 g In Brazil, Costa Rica, Suriname, Trinidad and Tobago, and Malaysia, smaller endemic samoanus. Transmission parameters were higher in ecosys- foci have been eliminated.3 h Percentage of people at risk in 1990 taken from Michael and others,62 recalculated with tems altered by water resource projects and clinical disease recent figures from United Nations.60 manifestation rates often elevated. i Pan American Health Organization.63 k Weekly Epidemiological Record.64 Vector densities. In total, seven studies investigated either l Supali and others.39 m Thailand has recently eliminated filaria transmission.3 the shift of LF vector composition frequencies or the change n People at risk estimated < 1%.13 in vector abundance, as shown in Table 4. In two study sites o It has been assumed that Brunei Darussalam has the same percentage of people at risk as Malaysia in 1995 as described by Michael and others.62 in Ghana and one in the United Republic of Tanzania, com- p Kazura and Bockarie.65 r Korea and the Solomon Islands using diverse control strategies have eliminated trans- position frequencies of An. gambiae increased in irrigated mission.3 sites compared with An. funestus.34–36 In turn, the relative dominance of An. gambiae was found to be smaller in irri- EFFECT OF WATER PROJECTS ON LYMPHATIC FILARIASIS 527

TABLE 2 Current global and regional estimates of lymphatic filariasis (LF), including studies identified in our systematic literature review, disability adjusted life years (DALYs), total population, population at risk, population living in proximity to irrigated areas, and urban population without access to improved sanitation

Total population in Population in LF-endemic Urban population in LF-endemic WHO Studies DALYs in 2002 LF-endemic countries Population at risk of LF countries living in proximity countries without access to subregiona identified caused by LF (× 103)a (× 103)b (× 103) (from Table 1) to irrigated areas (× 103) improved sanitation (× 103)a 1 3 976 284,551 235,382c 574g 38,445k 2 2 1,035 312,344 144,903 305 25,956 4 0 9 193,892 6,147 306 25,570l 5 1 1 8,326 6,078 < 1 1,561 7 1 122 125,551 10,847 1,646 2,265 9 0 1 n.d. n.d. n.d. n.d. 11 1 242 302,781 47,062d 8,262 31,212 12 3 2,977 1,287,945 618,496d 147,894h 131,157 13 0 0 358 40 < 1 n.d. 14 1 411 1,565,246 970,589d,e,f 54,034i 176,791m Total 12 5,777 4,079,995 2,039,548 213,021 394,511 n.d.: no data currently available. a Source: World Health Report.9 b Source: United Nations Urbanization Prospects—The 2003 Revisions.60 c Without Cap Verde, and Sao Tome and Principe. d In all countries both endemic for W. bancrofti and B. malayi or B. timori, “population at risk” from the predominant filaria species was taken. e Without Cambodia, Lao People’s Democratic Republic, Republic of Korea, and Solomon Islands. f China has considerably reduced LF transmission, therefore those figures are likely to be significantly smaller. g Without Equatorial Guinea and Seychelles. h Without Maldives and Timor-Leste. i Without Cook Islands, Federated States of Micronesia, Kiribati, Niue, Papua New Guinea, Samoa, Solomon Islands, Tonga, Tuvalu, and Vanuatu. k Without Liberia, Sao Tome and Principe, and Seychelles. l Without Trinidad and Tobago. m Without Federated States of Micronesia, Malaysia, Tonga, and Tuvalu. gated areas in the Upper East region of Ghana and in the mission potential (0.5 versus 13.8), and the percentage of in- United Republic of Tanzania.37,38 fective An. gambiae (0.3% versus 2.5–3.3%) were notably In absolute numbers (i.e., mosquito counts), changes mani- higher in irrigated villages compared with control villages.38 fested themselves more prominently. In all settings where This study also found a higher percentage of infective An. water resource developments were implemented, 1.7–24.6 funestus (0% versus 1.3%) and a higher worm load per infec- times more An. gambiae were caught when compared with tive vector (1.0 versus 1.8) when compared with the nonirri- control sites. Similar numbers were found for An. funestus. gated villages. A different study that assessed the prevalence Another common LF vector in Africa, namely An. melas, of infective filaria in vectors in irrigated villages in southern could not maintain itself in irrigated areas. Hence, this species Ghana recorded even higher fractions of infective An. gam- disappeared. Most likely, it was replaced by the strongly pro- biae (8%) and An. funestus (2%).36 In Sri Lanka, the geo- liferating An. gambiae s.s. population.35 In Indonesia, An. metric mean of female Cu. quinquefasciatus per man-hour subpictus was exclusively found in areas without irrigation was 1.6 times higher after the implementation of a large irri- and An. barbirostris, a typical rice-field breeder, proliferated gation system.40 in villages with irrigated paddies.39 An integrated, community-based bancroftian filariasis and In urban areas on Upolu Island (Samoa), domestic water- malaria control program was carried out in the first half of the storage and waste accumulation provided suitable breeding 1980s in urban Pondicherry, India, which aimed at transmis- sites for Ae. polynesiensis, which in turn became the predomi- sion reduction by simultaneous implementation of biologic, nant vector in those areas. On the other hand, Ae. samoanus chemical and physical vector control measures.29 Source re- seemed to favor less populated areas where the relative abun- duction by means of environmental management was given dance of Ae. polynesiensis was small.30 High numbers of high priority. It consisted of draining water-bodies, deweed- Culex vectors were found in urban areas dominated by waste- ing, and sealing of tanks and cisterns. Regarding biological water mismanagement and domestic water storage.29,31,32 control, larvivorous fish were released in permanent water Transmission parameters. Table 5 summarizes the five bodies. Larvicides and oil were used as chemical methods, studies that assessed the impact of water resource develop- and physical control measures included application of poly- ment and management on transmission parameters. Three styrene expanded beads in wells. Within five years, the annual studies were carried out in irrigation schemes,36,38,40 one biting rate for W. bancrofti-transmitting Cu. quinquefasciatus study evaluated the impact of water mismanagement in the decreased from 26,203 to 3,617, the numer of infective bites face of urbanization,30 and one study was undertaken after a per person per year decreased from 225 to 22, and the annual water management control program had been launched.29 transmission potential decreased from 450 to 77. On the other Overall, it was found that irrigation, waste-water mismanage- hand, the worm load increased during the program from 2.0 ment, water storage, or waste accumulation generally lead to to 3.5. increased biting rates, higher transmission potentials, and a The effect of urbanization on transmission parameters of higher proportion of vectors infective or infected with mi- LF has been documented in Samoa. In areas affected by eco- crofilaria. system transformation, the biting density per man per hour In east Ghana, the annual biting rate (188 versus 299), the (26 versus 8), the fraction of infected (2.2% versus 1.7%) and annual infective biting rate (0.5 versus 7.7), the annual trans- infective (0.4% versus 0.3%) Ae. polynesiensis were greater 528 ERLANGER AND OTHERS

TABLE 3 Overview of studies meeting our inclusion criteria that assessed the effect of water resource development and management on changes of lymphatic filariasis (LF), including vector composition, vector abundance, transmission parameters, filaria infection prevalence, and clinical manifestation rates, as stratified by rural and urban settings in different WHO subregions of the world

Water resource Shift in Human WHO Country, year of development and vector Vector Transmission infection Clinical Setting subregion study (reference) management Vector species (filaria species) composition abundance parameters prevalence manifestation Rural 1 Ghana, 2000 Irrigated An. gambiae (W. bancrofti) ↑ ↑ – – (Appawu agriculture An. funestus (W. bancrofti) ↑ ↑ – – – and others38) Cu. quinquefasciatus (none) ↑ – – – An. pharoensis (none) ↑ ––– An. nili, An. rufipens, ↑↑ ––– Ae. aegypti (none) Rural 1 Ghana, 1995 Irrigated An. gambiae s.l. ↑ ↑ – – – (Dzodzomenyo agriculture (W. bancrofti) and others36) An. funestus (W. bancrofti) ↓ ↓ – – – Cu. quinquefasciatus (none) ↓ ↓ – – – ––– ס ↓ An. pharoensis (W. bancrofti) Rural 1 Ghana, 1993 Rice irrigation An. gambiae s.s. ↓ ↑ – – – (Appawu (W. bancrofti) and others35) An. melas (W. bancrofti) ↑ ↓ – – – Rural 2 United Republic Rice irrigation An. gambiae (W. bancrofti) ↑ ↑ – ↑ – of Tanzania, An. funestus (W. bancrofti) ↓ ↑ – – – 1956 (Jordan34) Rural 2 United Republic Rice irrigation An. gambiae (W. bancrofti) ↑ ↑ – ↑ – of Tanzania, An. funestus (W. bancrofti) ↓ ↑ – – – 1951–1953 (Smith37) Rural 11 Indonesia, 2001 Rice irrigation An. subpictus (W. bancrofti) ↓ ↓ – ↓ ↓a (Supali and An. barbirostris (B. timori) ↑ ↑ – ↑ ↑b others39) Rural 12 Sri Lanka, Rice irrigation Cu. quinquefasciatus ↑↑ ––– 1986–1987 (W. bancrofti) (Amerasinghe and others40) Rural 12 India, 1957 Irrigation, Cu. quinquefasciatus – ↑ – ↑↑ (Basu41) sullage, (W. bancrofti, B. malayi) storm-water drains Urban 5 Haiti, 1981 Water storage, Cu. quinquefasciatus – ↑ – ↑ – (Raccurt waste-water (W. bancrofti, B. malayi) and others31) management Urban 7 Egypt, 1986 Waste-water Cu. pipiens pipiens, – ↑ – ↑ – (Gad and pools Cu. antennatus others32) (W. bancrofti) Urban 12 India, 1987 Waste-water Cu. quinquefasciatus – ↑↑c –– (Rajagopalan canals, pits, (W. bancrofti) and others29) reservoirs Urban 14 Samoa, 1978–1979 Man-made Ae. polynesiensis ↑ ↑ ↑ – – (Samarawickrema breeding sites, (W. bancrofti) and others30) water storage Ae. samoanus ↓ ↓ ↓ – – (W. bancrofti) .no change :ס ;increase in sites where water-related change occurred; ↓: decrease in sites where water-related change occurred :↑ a Genital lymphedema. b Elephantiasis. c Except “number of infective larvae per mosquito,” which was decreasing.

than in areas without ecosystem transformation. On the other In 2002, Supali and colleagues39 found that in Indonesian hand, biting density per man per hour (67 versus 33) and the villages with irrigated rice agriculture, An. barbirostris was percentage of infected (0.5% versus 0.2%) and infective responsible for B. timori transmission. The infection preva- (0.2% versus 0.04%) Ae. samoanus were found to be lence of B. timori among villagers was 6%, while W. bancrofti smaller.30 infections were not found. As many as 7% of all people were Filarial prevalence and clinical manifestation rates. Infec- diagnosed with leg elephantiasis, which was associated with tion prevalence and clinical manifestations were assessed in brugian filariasis. In irrigation-free villages, the main vector seven and two studies, respectively. Table 6 points out that was An. subpictus and human filarial infection prevalence was water resource developments had a strong effect on microfi- 12%, but both An. barbirostris vectors and B. timori filaria, laria infection prevalence. In six settings, prevalence rates were absent. Clinical symptoms appeared as genital lymph- were between 0.5% and 19% higher (median: 7%) compared edema in 5% of all people. with control areas. The most dramatic impact of a water resource development EFFECT OF WATER PROJECTS ON LYMPHATIC FILARIASIS 529

TABLE 4 Absolute and relative change in abundance of different filaria vectors in areas where water resource development and management (WRDM) occurred compared to similar control-sites without WRDM

Absolute and relative Control site WRDM occurred change in abundance Country, year of study (reference) Type of change Vector species No. % No. % No. Factor Ghana, 2000 (Appawu Irrigated agriculture An. gambiae s.l. 756 87.7 1,256/1,831 81.9/73.1 +500/+1,075 1.7/2.4 and others38) (site 1/site 2) An. funestus 48 5.6 254/471 16.5/18.8 +206/+423 5.3/9.8 Cu. quinquefasciatusa 51 5.9 0/128 0/5.1 −51/+77 dis./2.5 An. pharoensisa 2 0.2 0/27 0/1.1 −2/+25 dis./13.5 An. nili,a An. rufipens,a 5 0.6 24/47 1.6/1.9 +19/+42 4.8/9.4 and Ae. aegypti Ghana, 1995 Irrigated agriculture An. gambiae s.l. 15 12 141 77 +126 9.4 (Dzodzomenyo An. funestus 101 82 40 22 −61 0.4 and others36) Cu. quinquefasciatusa 5 4 0 0 −5 dis. An. pharoensis 32 3 1 0 1 Ghana, 1993 (Appawu Rice irrigation An. gambiae s.s. 27/17 96/94 50 100 +23/+33 1.9/2.9 and others35) (site 1/site 2) An. melasa 1/1 4/6 0 0 −1/−1 dis. Sri Lanka, 1986–1987 Rice irrigation Cu. quinquefasciatus 209 48.3 467 79.8 +258 2.2 (Amerasinghe Cu. pseudovishnuia 224 51.7 118 20.2 −106 0.5 and others40) Samoa, 1978–1979 Man-made breeding Ae. polynesiensis – ↓ – ↑ –– Samarawickrema sites, water storage Ae. samoanus – ↑ – ↓ –– and others30) United Republic of Rice irrigation An. gambiae 29 96.7 714 99.6 +685 24.6 Tanzania, 1956 An. funestus 1 3.3 3 0.4 +2 3 (Jordan34) United Republic of Rice irrigation An. gambiae 2,057 99.9 3,959 99.7 +1,902 1.9 Tanzania, 1951–1953 An. funestus 2 0.1 29 0.3 +27 14.5 (Smith37) dis.: disappearance of vector after WRDM; ↑: increase; ↓: decrease. a Not filaria transmitting. on LF was found in villages of the United Republic of Tan- systematically reviewed the literature and estimated the cur- zania a half-century ago. Microfilaria prevalence in two vil- rent magnitude of urban malaria in Africa45 and examined lages with irrigated rice plantations were 11% and 19% the effect of irrigation and large dams on the burden of ma- higher compared with two nearby villages where no irrigation laria on a global and regional scale.25 Here, we extended our systems had been constructed.34 preceding work from malaria to LF, with an emphasis on the In a north Indian area served by irrigation, infection preva- effect of water resource development and management, and lence for W. bancrofti was found to be 0.5% and disease estimates of at-risk populations. manifestation 1.5% higher compared with a similar setting It is important to note that estimates of populations at risk without irrigation. Close by, in another irrigated plot, but of LF, as presented in Table 1, differ considerably according inhabited by people of a different ethnic origin, microfilaria to the source of publication. Also, some countries/territories prevalence was 9% greater. Disease manifestations, on the were highly successful in lowering filaria transmission over 41 other hand, were almost at the same level (−0.5%). the past 10–20 years (e.g., China), and therefore care is Very high W. bancrofti infection prevalence in the popula- needed in the interpretation of at-risk population. Our esti- tion of Leogane, Haiti (39% and 44%), could be attributed to mate of 2 billion might thus be a significant overestimation.1–3 waste-water discharge by factories located in the city. Infec- The term “at-risk” raises problems with its definition, because tion prevalence in control districts without waste-water pools in most countries where transmission has been interrupted, 31 were much lower (27%). High prevalence (17%) in a town the population is still likely to face the risk of reemerging LF in the Egyptian Nile delta was due to sewage ponds of public epidemics as parasites and vector species continue to be 32 facilities (prevalence of control site: 12%). On Samoa, in present and environmental conditions are suitable for trans- contrast, in areas affected by human settlements, the preva- mission. lence of W. bancrofti infections was 1.1% smaller than in Our population estimates in LF-endemic countries regard- 30 control areas. ing proximity to irrigated areas (i.e., 213 million) are rather conservative. Irrigated areas often attract people, and thus DISCUSSION the population density is usually disproportionately high. However, depending on the vector species and the practice of Previous studies have shown that the establishment, opera- irrigation, the risk profile of LF could also be lower when tion and poor maintenance of water resource development compared with nonirrigated control areas. For transmission projects, and the process of rapid and uncoordinated urban- of bancroftian filariasis outside of Africa, it is less the practice ization, have a history of facilitating a change in the frequency of irrigated agriculture per se, but rather the presence of pol- and transmission dynamics of vector-borne diseases.16,20,22,23 luted peridomestic man-made breeding sites that are suitable However, detailed analyses on the contextual determinants are habitats for LF vectors (mostly Culex). sparse.42–44 In recent attempts to fill some of these gaps, we Care should also be exhibited in the interpretation of our 530 ERLANGER AND OTHERS

TABLE 5 Transmission parameters of different filaria vectors in areas where water resource development and management (WRDM) occurred compared to control areas without WRDM

Country, year of study (reference) Type of change Transmission parameters of different filaria vectors Control site WRDM occurred Relative change Ghana, 2000 (Appawu Irrigated agriculture Annual biting rate of An. gambiae and 188 299 1.6 and others38) (site 1/site 2) An. funestus Annual infective biting rate of An. gambiae 0.5 7.7 15.4 and An. funestus Worm load of An. gambiae and An. funestus 1.0 1.8 1.8 Annual transmission potential of An. gambiae 0.5 13.8 27.6 and An. funestus Infective An. gambiae 0.3% 3.3%/2.5% 11/8.3 Infective An. funestus 0% 0%/1.3% n.a. Ghana, 1995 Irrigated agriculture Infective An. gambiae – 8% – (Dzodzomenyo Infective An. funestus – 2% – and others36) Infected An. gambiae – 27% – Infected An. funestus – 16% – Sri Lanka, 1986–1987 Rice irrigation Geometric mean female Cu. quinquefasciatus 4.6 7.4 1.6 (Amerasinghe and per man-hour others40) India, 1979–1985 Vector control Annual biting rate of Cu. quinquefasciatus 26,203 3,617 0.1 (Rajagopalan program Annual infective biting rate of 225 22 0.1 and others29) Cu. quinquefasciatus Worm load of Cu. quinquefasciatus 2.0 3.5 1.8 Annual transmission potential of 450 77 5.8 Cu. quinquefasciatus Samoa, 1978–1979 Man-made breeding Biting density per man-hour of 8263.3 (Samarawickrema sites, water storage Ae. polynesiensis and others30) Infected Ae. polynesiensis 1.7% 2.2% 1.3 Infective Ae. polynesiensis 0.3% 0.4% 1.3 Biting density per man-hour of Ae. samoanus 67 33 0.5 Infected Ae. samoanus 0.5% 0.2% 0.4 Infective Ae. samoanus 0.2% 0.04% 0.2 n.a.: not applicable. at-risk population estimates in urban settings. We used access causal webs of the various levels of disease causality, with to improved sanitation as the underlying risk factor to derive outcomes shaped by a combination of distal, proximal, and our estimates. However, the current definition of access to physiologic/pathophysiological causes.46 In fact, settings with improved sanitation is primarily constructed by an aggrega- access to improved sanitation, as defined by WHO, on the tion of different social and infrastructure determinants rather “least improved end” can include highly productive mosquito than setting-specific eco-epidemiologic features. Arguably, breeding sites, while mosquito breeding is unlikely to occur in this is an oversimplification, as it fails to capture the complex settings on the “most improved end.” Hence, the nature of

TABLE 6 Filaria prevalence and frequencies of clinical manifestations in areas where water resource development and management (WRDM) occurred compared to similar areas without WRDM

Country, year of study Filaria vector or WRDM Change in (reference) Type of WRDM clinical symptoms Control site occurred absolute terms Indonesia, 2001 (Supali and others39) Rice irrigation W. bancrofti 12% 0% Absence B. timori 0% 6% +6% Genital lymphedema 5% 0% Absence Elephantiasis 0% 7% +7% Egypt, 1986 (Gad and others32) Areas around large cesspit/ W. bancrofti 12% 17%/7% +5%/−5% small cesspit Haiti, 1981 (Raccurt and others31) Waste-water area/area with W. bancrofti 27% 39%/44% +12%/+17% water-storage Samoa, 1978–1979 (Samarawickrema Man-made breeding sites, W. bancrofti 5.3% 4.2% −1.1% and others30) water storage India, 1957 (Basu41) Rice irrigation, sullage, and Mixed infection of 5%/2% 5.5%/12% +0.5%/+9% storm-water drains in 2 B. malayi and sites (site 1/site 2) W. bancrofti (ratio 74:26) Genital lymphedema and 3.5%/3% 5%/2.5% +1.5%/−0.5% elephantiasis United Republic of Tanzania, 1956 Rice irrigation W. bancrofti 7% 26% +19% (Jordan34) United Republic of Tanzania, Rice irrigation W. bancrofti 12% 23% +11% 1951–1953 (Smith37) EFFECT OF WATER PROJECTS ON LYMPHATIC FILARIASIS 531 water-resource development and management in urban areas bancroftian filariasis transmitting An. subpictus vectors were exhibits strong spatiotemporal heterogeneity, often at very replaced by timorian filariasis transmitting An. barbirostris, small scales. In addition, the fine-grained detail about waste- resulting in a shift from genital lymphedema to elephantia- water management that would be essential for a precise ap- sis.39 In Egypt and Senegal, a similar phenomenon was ob- praisal of potential vector breeding sites is not available on a served for schistosomiasis. The construction of large dams led scale that would sharply reduce uncertainties in the present to a shift from Schistosoma haematobium to Schistosoma report. Nevertheless, the estimates in Table 2 do provide a mansoni, most likely because of a shift in intermediate host good approximate indication of the magnitude of the prob- snails. This was paralleled by a change of clinical manifesta- lem. Unfortunately, LF is too far down on virtually all disease tion.52,53 priority lists to get serious attention and serve as a basis for Our review only identified two studies that investigated establishing the financial resources and political will for wa- clinical manifestation rates in connection with water projects. ter-related improvements in urban areas. It is conceivable Thus, it is difficult to set forth conclusions about whether that endemic countries could get major LF reductions as a water resource development projects positively or adversely by-product of multifaceted water campaigns that aim to im- affect clinical manifestations due to LF. It is delicate to use prove overall health in a systemic manner. results on filaria infection prevalence and transmission pa- The 12 studies we identified through our systematic review rameters as proxies, since microfilaremia and clinical symp- can be grouped into two broad categories, namely (i) those toms are not implicitly associated. People with clinical mani- that looked at ecosystems influenced by irrigated rural agri- festations are often amicrofilaremic, while others who are free culture and (ii) those that investigated urban environments of symptoms have microfilariae in their blood.54,55 Currently, affected by poor design and lack of maintenance of infrastruc- there is no clear evidence of acquired or innate immunity to tures for drainage of sewage and storm water. Despite the filaria infection. Thus, it is uncertain if lower infection rates different nature of these studies, entomological parameters and clinical manifestation among the local residents could be, revealed a quite consistent shift in species composition fre- at least partially, explained by acquired immunity or innate quencies, and a proliferation of the overall vector population. immunity genes that govern susceptibility to infection and High abundances were recorded for An. funestus, and espe- lymphatic pathology.56,57 cially for An. gambiae, in irrigated agro-ecosystems, particu- Another important finding of our systematic literature re- larly in West Africa. Members of the An. gambiae complex view is that urbanization, especially in connection with waste- are the most anthropophilic filaria vectors.47 In Africa, the water mismanagement and water-storage, resulted in signifi- fraction of irrigated arable land is still small (8.5%) but is cant shifts in LF transmission parameters, as demonstrated in expected to increase significantly in the decades to come.48 Haiti, India, and Samoa. Reverse shifts in the abundance of Consequently, it is conceivable that implementation of irriga- Ae. samoanus and Ae. polynesiensis, two vectors with varying tion systems in this region increases transmission of W. ban- infectivity rates, indicated that rapid and uncontrolled urban- crofti.49 Achieving the GPELF’s ambitious goal could be of a ization impacts differently on various vector species. De- particular challenge in Africa, where the burden of LF could creased transmission parameters of Ae. samoanus in city cen- actually increase. ters show that urbanization can also marginalize a vector that Regarding the observation of higher counts of vector spe- fails to adapt to the new condition. cies following water resource developments, these do not au- We have estimated that > 70% of urban dwellers in LF- tomatically translate into a higher LF burden. Due to the endemic areas are currently located in Asia. Cu. quinque- complicated nature of LF pathology and the highly complex fasciatus, the most important LF vector in this region, prefers transmission dynamics, it is possible that after the implemen- polluted waters for breeding. The rapid pace at which urban- tation of an irrigation system in a highly endemic area, the LF ization continues to build inroads in Asian (and African) burden could level off after a few years.15,43 The entomologi- countries, often in the face of declining economies, is paral- cal studies carried out in Sri Lanka during the development of leled by unprecedented pollutions of open waters and sewage the Mahaweli irrigation project in the 1980s revealed that systems beyond organic matters. In fact, industrial pollutants several mosquito species proliferated over the course of proj- and heavy metals transform these water bodies into hostile ect implementation. High densities of Cu. quinquefasciatus, environments for the living biota, including LF vectors. which is the main LF vector in Sri Lanka, were documented, Therefore, the issue of uncontrolled urbanization and poor however, filaria transmission could not be confirmed.40,50 waste-water management as a consequence, gains further im- It is widely acknowledged that vector species shifts de- portance here. pended on a myriad of factors, i.e., seasonality, temperature, In urban settings, integrated vector management compris- plant succession, irrigation practices, total area under irriga- ing environmental management (e.g., draining) and biological tion, water-depth, and water quality.51 In the studies analyzed (e.g., introduction of larvivorous fish), chemical (e.g., appli- here, these aspects were not retrievable from the published cation of larvicides), and physical (e.g., use of bed nets) con- work. Thus, temporal variations cannot be excluded, render- trol measures can have a significant impact on LF transmis- ing study comparison difficult. Future studies should quantify sion. A prominent example is the community-based inte- species composition frequencies and vector populations not grated control program in Pondicherry, India.29 Despite a only between different eco-epidemiologic settings, but also somewhat higher worm load 5 years after the control program during different seasons and according to different irrigation was launched, transmission parameters dropped significantly. practices within the same setting. The reason for the increase of the worm load might be due to Once a vector species is replaced by another that transmits smaller mosquito populations feeding more exclusively on hu- a different filaria species, clinical manifestation rates are mans.58 Another example of how an integrated control ap- likely to shift. This was observed in rural Indonesia, where proach with strong emphasis on environmental management 532 ERLANGER AND OTHERS impacts on LF was described by Chernin.28 In Charleston, tion: a public health success and development opportunity. South Carolina, southern United States, bancroftian filariasis, Filaria J 2: 13. which was introduced by African slaves, disappeared after the 4. Langhammer J, Birk HW, Zahner H, 1997. Renal disease in lymphatic filariasis: evidence for tubular and glomerular disorders municipal sanitation system had been improved. These mea- at various stages of the infection. Trop Med Int Health 2: 875– sures were initially intended to fight typhoid fever and related 884. infectious diseases. However, they indirectly reduced polluted 5. Ottesen EA, Duke BO, Karam M, Behbehani K, 1997. Strategies domestic waters and therefore reduced the available breed- and tools for the control/elimination of lymphatic filariasis. Bull World Health Organ 75: 491–503. ing-sites for filaria transmitting Cu. quinquefasciatus. 6. Dreyer G, Figueredo-Silva J, Neafie RC, Addiss DG, 1998. Lym- To further strengthen and expand the current evidence- phatic filariasis. Nelson AM, Horsburgh CR, eds. Pathology of base of the contextual determinants of LF, additional inves- Emerging Infections. Volume 2. Washington DC: American tigations are warranted. It would be of particular interest to Society for Microbiology, 317–342. document qualitatively and quantitatively both transmission 7. Dreyer G, Noroes J, Addiss D, 1997. The silent burden of sexual disability associated with lymphatic filariasis. Acta Trop 63: and disease parameters, coupled with overall changes in key 57–60. demographic, health, and socioeconomic parameters over the 8. Ramaiah KD, Das PK, Michael E, Guyatt H, 2000. The economic course of major water resource development projects, such as burden of lymphatic filariasis in India. Parasitol Today 16: 251– irrigation schemes and large dams. Moreover, it is essential to 253. investigate the role of urban LF, particularly in the light of 9. WHO, 2004. The World Health Report 2004 – Changing History. Geneva: World Health Organization. rapid and uncontrolled urbanization. These investigations are 10. International Task Force for Disease Eradication, 1993. Recom- likely to be carried out only if they are incorporated as part of mendations of the International Task Force for Disease Eradi- comprehensive waste management and sanitation programs, cation. MMWR Recomm Rep 42: 1–38. driven by the need to establish and finance systemic health 11. Molyneux DH, Bradley M, Hoerauf A, Kyelem D, Taylor MJ, 2003. Mass drug treatment for lymphatic filariasis and on- systems at the city, district, and regional levels. We conclude chocerciasis. Trends Parasitol 19: 516–522. that integrated vector management, taking into account en- 12. Ottesen EA, 2000. The global programme to eliminate lymphatic vironmental, biological and socioeconomic determinants, filariasis. Trop Med Int Health 5: 591–594. should receive more pointed consideration, as it is a promis- 13. WHO, 2000. Preparing and Implementing a National Plan to ing approach to complement mass drug administration pro- Eliminate Lymphatic Filariasis: A Guideline for Programme Managers. A guideline for programme managers. Geneva: grams that form the backbone of the GPELF. Without an World Health Organization (WHO/CDS/CPE/CEE/2000.16). integrated control approach, the ambitious goal to eliminate 14. WHO, 2004. Report on the mid-term assessment of microfila- LF as a public health problem by 2020 might remain elusive. raemia reduction in sentinel sites of 13 countries of the Global Programme to Eliminate Lymphatic Filariasis. Wkly Epide- Received September 19, 2004. Accepted for publication January 22, miol Rec 79: 358–365. 2005. 15. Michael E, Malecela-Lazaro MN, Simonsen PE, Pedersen EM, Barker G, Kumar A, Kazura JW, 2004. Mathematical model- Acknowledgments: The authors thank Dr. Felix P. Amerashinge, ling and the control of lymphatic filariasis. Lancet Infect Dis 4: Prof. David H. Molyneux, Dr. Will Parks, Dr. Erling Pedersen, and 223–234. Dr. Christopher A. Scott for valuable comments on the manuscript. 16. Harb M, Faris R, Gad AM, Hafez ON, Ramzy R, Buck AA, 1993. We also thank Jacqueline V. Druery and her team from Stokes Li- The resurgence of lymphatic filariasis in the Nile delta. Bull brary at Princeton University for help in obtaining a large body of World Health Organ 71: 49–54. relevant literature. 17. Manga L, 2002. Vector-control synergies, between ‘Roll Back Financial support: This investigation received financial support from Malaria’ and the Global Programme to Eliminate Lymphatic the Water, Sanitation and Health unit and the Protection of the Hu- Filariasis, in the African region. Ann Trop Med Parasitol 96 man Environment (WSH/PHE) at the World Health Organization (Suppl. 2): S129–S132. (WHO ref. Reg. file: E5/445/15). The research of J. Keiser and J. 18. Prasittisuk C, 2002. Vector-control synergies, between ‘Roll Back Utzinger is supported by the Swiss National Science Foundation Malaria’ and the Global Programme to Eliminate Lymphatic (Projects PMPDB–106221 and PPOOB–102883, respectively). M. C. Filariasis, in south-east Asia. Ann Trop Med Parasitol 96 Castro is grateful to the Office of Population Research and the Cen- (Suppl. 2): S133–S137. tre for Health and Wellbeing at Princeton University. 19. Burkot T, Bockarie M, 2004. Vectors. Am J Trop Med Hyg 71 (Suppl.): 24–26. Authors’ addresses: Tobias E. Erlanger, Jennifer Keiser, Marcel Tan- 20. Hunter JM, 1992. Elephantiasis: a disease of development in ner, and Jürg Utzinger, Swiss Tropical Institute, P.O. Box, CH–4002 north east Ghana. Soc Sci Med 35: 627–645. Basel, Switzerland. Marcia Caldas de Castro, Department of Geog- 21. Rosengrant MW, Perez ND, 1997. Water Resource Development raphy, University of South Carolina, 125 Callcott Hall, Columbia, SC in Africa: A Review and Synthesis of Issues, Potentials and 29208. Robert Bos, Department of Protection of the Human Envi- Strategies for the Future. International Food Policy Research ronment, World Health Organization; 20 Avenue Appia, CH–1211 Institute, EPTD Discussion Paper No. 28. Geneva 27, Switzerland. Burton H. Singer, Office of Population Re- 22. Mott KE, Desjeux P, Moncayo A, Ranque P, de Raadt P, 1990. search, Princeton University, 245 Wallace Hall, Princeton, NJ 08544. Parasitic diseases and urban development. Bull WHO 68: 691– Reprint requests: Jürg Utzinger, Department of Public Health and 698. Epidemiology, Swiss Tropical Institute, CH–4002 Basel, Switzerland, 23. Knudsen AB, Slooff R, 1992. Vector-borne disease problems in Telephone: +41 61 284-8129, Fax: +41 61 284-8105, E-mail: rapid urbanization: new approaches to vector control. Bull [email protected]. WHO 70: 1–6. 24. United Nations, 2004. Nations Human Settlements Programme. REFERENCES Washington, DC: United Nations. 25. Keiser J, Castro MC, Maltese MF, Bos R, Tanner M, Singer BH, 1. WHO, 2001. Lymphatic filariasis. Wkly Epidemiol Rec 76: 149– Utzinger J, 2005. Effect of irrigation and large dams on the 154. burden of malaria on a global and regional scale. AmJTrop 2. Zagaria N, Savioli L, 2002. Elimination of lymphatic filariasis: a Med Hyg 72: 392–406. public-health challenge. Ann Trop Med Parasitol 96 (Suppl. 2): 26. Fontes G, Rocha EM, Brito AC, Antunes CM, 1998. Lymphatic S3–S13. filariasis in Brazilian urban area (Maceio´ , Alagoas). Mem Inst 3. Molyneux D, 2003. Lymphatic filariasis (elephantiasis) elimina- Oswaldo Cruz 93: 705–710. EFFECT OF WATER PROJECTS ON LYMPHATIC FILARIASIS 533

27. Durrheim DN, Wynd S, Liese B, Gyapong JO, 2004. Lymphatic 47. Costantini C, Sagnon N, della Torre A, Coluzzi M, 1999. Mos- filariasis endemicity—an indicator of poverty? Trop Med Int quito behavioural aspects of vector-human interactions in the Health 9: 843–845. Anopheles gambiae complex. Parassitologia 41: 209–217. 28. Chernin E, 1987. The disappearance of bancroftian filariasis from 48. Keiser J, Utzinger J, Singer BH, 2002. The potential of intermit- Charleston, South Carolina. AmJTropMedHyg37:111–114. tent irrigation for increasing rice yields, lowering water con- 29. Rajagopalan PK, Panicker KN, Das PK, 1987. Control of malaria sumption, reducing methane emissions, and controlling ma- and filariasis vectors in South India. Parasitol Today 3: 233– laria in African rice fields. J Am Mosq Control Assoc 18: 329– 241. 340. 30. Samarawickrema WA, Kimura E, Spears GF, Penaia L, Sone F, 49. Surtees G, 1970. Effects of irrigation on mosquito populations Paulson GS, Cummings RF, 1987. Distribution of vectors, and mosquito-borne diseases in man, with particular reference transmission indices and microfilaria rates of subperiodic to ricefield extension. Int J Environ Stud 1: 35–42. Wuchereria bancrofti in relation to village ecotypes in Samoa. 50. Amerasinghe FP, Munasingha NB, 1988. A predevelopment mos- Trans R Soc Trop Med Hyg 81: 129–135. quito survey in the Mahaweli Development Project area, Sri 31. Raccurt CP, Lowrie RC Jr, Katz SP, Duverseau YT, 1988. Epi- Lanka: adults. J Med Entomol 25: 276–285. Wuchereria bancrofti Trans R demiology of in Leogane, Haiti. 51. Service MW, 1984. Problems of vector-borne disease and irriga- Soc Trop Med Hyg 82: 721–725. tion projects. Ins Sci Appl 5: 227–231. 32. Gad AM, Feinsod FM, Soliman BA, Nelson GO, Gibbs PH, Shoukry A, 1994. Exposure variables in bancroftian filariasis in 52. Abdel-Wahab MF, Strickland GT, El-Sahly A, El-Kady N, the Nile Delta. J Egypt Soc Parasitol 24: 439–455. Zakaria S, Ahmed L, 1979. Changing pattern of schistosomiasis in Egypt 1935-79. Lancet 314: 242–244. 33. Cao WC, Van der Ploeg CPB, Ren ZX, Habbema JDF, 1997. Success against lymphatic filariasis. World Health Forum 18: 53. Southgate VR, 1997. Schistosomiasis in the Senegal River Basin: 17–20. before and after the construction of the dams at Diama, Sene- 34. Jordan P, 1956. Filariasis in the Lake Province of Tanganyika. gal and Manantali, Mali and future prospects. J Helminthol 71: East Afr Med J 33: 237–242. 125–132. 35. Appawu MA, Baffoe-Wilmot A, Afari EA, Nkrumah FK, Pe- 54. Kar SK, Mania J, Kar PK, 1993. Humoral immune response dur- trarca V, 1994. Species composition and inversion polymor- ing filarial fever in Bancroftian filariasis. Trans R Soc Trop phism of the Anopheles gambiae complex in some sites of Med Hyg 87: 230–233. Ghana, West Africa. Acta Trop 56: 15–23. 55. Ravindran B, 2003. Aping Jane Goodall: insights into human 36. Dzodzomenyo M, Dunyo SK, Ahorlu CK, Coker WZ, Appawu lymphatic filariasis. Trends Parasitol 19: 105–109. MA, Pedersen EM, Simonsen PE, 1999. Bancroftian filariasis 56. Hise AG, Hazlett FE, Bockarie MJ, Zimmerman PA, Tisch DJ, in an irrigation project community in southern Ghana. Trop Kazura JW, 2003. Polymorphisms of innate immunity genes Med Int Health 4: 13–18. and susceptibility to lymphatic filariasis. Genes Immun 4: 524– 37. Smith A, 1955. The transmission of bancroftian filariasis on 527. Ukara Island, Tanganyika II. The distribution bancroftian mi- 57. Stolk WA, Ramaiah KD, van Oortmarssen GJ, Das PK, crofilaraemia compared with the distribution hut-haunting Habbema JDF, de Vlas SJ, 2004. Meta-analysis of age- mosquitoes and their breeding-places. Bull Entomol Res 46: prevalence patterns in lymphatic filariasis: no decline in mi- 437–444. crofilaraemia prevalence in older age groups as predicted by 38. Appawu MA, Dadzie SK, Baffoe-Wilmot A, Wilson MD, 2001. models with acquired immunity. Parasitology 129: 605–612. Lymphatic filariasis in Ghana: entomological investigation of 58. Samuel PP, Arunachalam N, Hiriyan J, Thenmozhi V, Gajanana transmission dynamics and intensity in communities served by A, Satyanarayana K, 2004. Host-feeding pattern of Culex quin- irrigation systems in the Upper East Region of Ghana. Trop quefasciatus Say and Mansonia annulifera (Theobald) Med Int Health 6: 511–516. (Diptera: Culicidae), the major vectors of filariasis in a rural 39. Supali T, Wibowo H, Rückert P, Fischer K, Ismid IS, Purnomo, areaofsouthIndia.J Med Entomol 41: 442–446. Djuardi Y, Fischer P, 2002. High prevalence of Brugia timori 59. Lindsay SW, Thomas CJ, 2000. Mapping and estimating the infection in the highland of Alor Island, Indonesia. Am J Trop population at risk from lymphatic filariasis in Africa. Trans R Med Hyg 66: 560–565. Soc Trop Med Hyg 94: 37–45. 40. Amerasinghe FP, Ariyasena TG, 1991. Survey of adult mosqui- 60. United Nations, 2004. World Urbanization Prospects. The 2003 toes (Diptera: Culicidae) during irrigation development in the Revisions. New York: Department of Economic and Social J Med Entomol 28: Mahaweli Project, Sri Lanka. 387–393. Affairs; Population Division of the United Nations (ESA/P/ Indian J Malariol 11: 41. Basu PC, 1957. Filariasis in Assam State. WP.190). 293–308. Defining the Roles of Vector Control and Xeno- 42. Patz JA, Graczyk TK, Geller N, Vittor AY, 2000. Effects of 61. WHO, 2002. monitoring in the Global Programme to Eliminate Lymphatic environmental change on emerging parasitic diseases. Int J Filariasis. Report of the Informal Consultation. Parasitol 30: 1395–1405. Geneva: World 43. Amerasinghe FP, 2003. Irrigation and mosquito-borne diseases. J Health Organization (WHO/CDS/CPE/PVC/2002.3). Parasitol 89 (Suppl.): S14–S22. 62. Michael E, Bundy DAP, Grenfell BT, 1996. Re-assessing the 44. Molyneux DH, 2003. Common themes in changing vector-borne global prevalence and distribution of lymphatic filariasis. Para- disease scenarios. Trans R Soc Trop Med Hyg 97: 129–132. sitology 112: 409–428. 45. Keiser J, Utzinger J, Castro MC, Smith TA, Tanner M, Singer 63. WHO, 2002. Lymphatic Filariasis Elimination in the Americas. BH, 2004. Urbanization in sub-Saharan Africa and implication Report of the Regional Program-Manager’s Meeting. Port-au- for malaria control. Am J Trop Med Hyg 71 (2 Suppl.): 118– Prince, Haiti: Pan American Health Organization. 127. 64. WHO, 2003. Lymphatic filariasis. Wkly Epidemiol Rec 78: 171– 46. Ezzati M, Utzinger J, Cairncross S, Cohen AJ, Singer BH, 2005. 179. Environmental risks in the developing world: exposure indica- 65. Kazura JW, Bockarie MJ, 2003. Lymphatic filariasis in Papua tors for evaluating interventions, programmes, and policies. J New Guinea: interdisciplinary research on a national health Epidemiol Community Health 59: 15–22. problem. Trends Parasitol 19: 260–263.