disease caused by Wuchereria bancrofti, Brugia malayi or Brugia timori that can clinically manifest itself in the form of lymphedema or elephantiasis.
The Attributable Fraction of the Lymphatic Filariasis Burden to Water Resource Development and Management Report prepared for the WHO commissioned study Burden of water-related vector-borne diseases: An analysis of the fraction attributable to components of water resources development and management. Investigators: Tobias E. Erlanger, Jennifer Keiser, Marcel Tanner, Jürg Utzinger Swiss Tropical Institute, P.O. Box, CH-4002 Basel, Switzerland
Marcia Caldas de Castro, Burton H. Singer Office of Population Research, Princeton University, Princeton, NJ 08544, USA
Robert Bos, Jamie Bartram and Laurence Haller Water, Sanitation and Health (WSH/PHE), World Health Organization, Avenue Appia 20, CH-1211 Geneva 27, Switzerland
Contents: Main Objective
page 3
Approach
page 3
Outcomes
page 4
Conclusion
page 5
Outlook and Perspectives
page 5
Appendix 1
Search strategy and selection criteria for the comprehensive literature review
page 12
Number of hits for lymphatic filariasis combined with selected keywords in different electronic databases
page 13
Appendix 3
Relevant literature to address the main research objective
page 14
Appendix 4
Published review paper American Journal of Tropical Medicine and Hygiene
page 15
Table summarising geographical distribution of the three Filaria species, ecology of their vectors and environmental changes leading to increased vector densities
page 36
Appendix 2
Appendix 5
Appendices 6.1–6.14 Key information form of all retrieved publications
2
page 37
Main Objective In order to strengthen the evidence base in support of decision-making on different intervention options for vector-borne disease prevention and control in the context of water resources development and management, WHO commissioned systematic literature reviews on the association between such development and the burden of four vector-borne diseases. In accordance with this mandate, the main objective of work reported here was to: strengthen and expand the current evidence-base of contextual determinants of lymphatic filariasis (LF) and to assign and quantify attributable fractions of the disease burden to specific components of water resources development and management. This implies the need for (i) the definition and characterization of the contextual determinants of LF; (ii) the compilation of critical LF statistics on a global and regional scale (stratified according to the 14 WHO sub-regions of the world); (iii) a systematic literature review; and (iv) preparation of an analytical report, including questions that remained unanswered and a mapping out of directions for future work. This report summarises the research approach taken and provides an outlook and possible perspectives on how to move forward in view of identified research priorities. The main findings of the systematic review can be found in appendix 4, which contains the text of the article published in the peer-reviewed literature.
Approach For the sake of the current report, LF is defined as a communicable parasitic disease caused by Wuchereria bancrofti, Brugia malayi or Brugia timori that can clinically manifest itself in the form of lymphedema or elephantiasis. Other diseases caused by Filarioidea (e.g. onchocerciasis and dracunculiasis) are not considered here. A systematic literature review was carried out to identify all published studies that examined the effect of water-related environmental changes on the frequency and transmission dynamics of LF. Special consideration was given to publications that (i) presented information on the sequential cause-and-effect relationships between waterrelated environmental change, abundance of vector populations, entomological transmission parameters, microfilaria infection prevalence and rates of clinical 3
manifestations, and/or (ii) compared .the epidemiological conditions in sites where ecosystems had been modified by water resources development with those in ecologically similar settings without such change. The main findings were synthesised and formatted into a review paper (Appendix 4). First, a schematic concept of the contextual determinants of LF was developed. Next, the fraction of the population at risk of LF was estimated in all 76 countries that are currently endemic for this disease. At-risk populations in rural and urban areas of all WHO sub-regions were linked with the most recent burden of disease statistics expressed in disability adjusted life years (DALYs). We employed the recent classification, as presented in the appendices of the WHO World Health Report 2004, which stratifies the world into 14 epidemiological sub-regions. At-risk populations in rural settings were defined by people living in close proximity to irrigated agro-ecosystems in those sub-regions where rural LF transmission occurs; in urban settings they were defined as people lacking access to improved sanitation in those sub-regions where urban LF transmission occurs. In the context of this report, improved sanitation systems include facilities which are designed and maintained in a way that they do not favour the proliferation of LF vectors. The size of the rural population was estimated by multiplying the average population density in rural areas with the total area under irrigation in the LF endemic countries. Statistics on urban dwellers lacking access to improved sanitation were taken from the World Health Report 2004.
Outcomes Appendix 1 summarises the search strategy and selection criteria that we employed to address this objective. Appendix 2 shows the number of hits for different key words derived from a set of electronically-available databases that are widely used for literature reviews. PubMed/Medline was found to be the most comprehensive database, as it comprised most of the literature cited in the other databases that we screened. Our extensive literature search pertaining to the main objective of this project yielded only 14 articles, all of which were published in the peer-reviewed literature (Appendix 3). Two out of these 14 contain descriptive data and were therefore not included in the systematic review article that was published in the peer reviewed literature. The review article (Appendix 4) presents a panel that lists the population at risk in all 76 countries endemic for LF, a figure that summarises the contextual determinants of LF and four tables that show the main findings of vector abundance, transmission and clinical manifestation rates. An estimate of the size of the population at higher risk due to 4
irrigation and inadequate sanitation is also presented in the article. As a general guidance, a table that shows the geographical distribution and ecology of major LF vectors was compiled (Appendix 5). Finally, from all selected publications, including the ones with descriptive data, the key information (KIF) was retrieved and summarised in a standardised format, which is presented in Appendices 6.1–6.14.
Conclusion The objective of this study was to strengthen and expand the current evidencebase of contextual determinants of LF, and to assign and quantify attributable fractions of the disease burden to specific components of water resources development and management and generally of water-related environmental change. It is part of a larger investigation examining the effect of water resources development and management on four vector-borne diseases, the other three being malaria, schistosomiasis and Japanese encephalitis. It was intended to quantify the burden of LF attributable to ecosystem change, with special reference to the changes in local hydrology, through the use of comparative risk assessment (CRA). Utilising counterfactual analysis, studies were required that examine alternative scenarios, thereby describing changes in the exposure to risk factors. Hence, an ideal study would be one that presents data in the initial steady-state (often natural) environment, e.g. prior to the implementation of a water resources development project. It would then describe changes that occurred during project implementation and, finally, it would assess the impact on filarial transmission, prevalence and morbidity several years after implementation. Unfortunately, not a single study was identified which fulfilled these criteria. Most studies either simply quantified prevalence rates or entomological parameters in a specific region (e.g. without differentiating between communities with and without water resource development) or were carried out after the completion of a water resource development project. The fact that the majority of publications are lacking profound environmental, ecological or socio-economic data made it difficult to link outcome measures with water-related or other risk factors. It had been suggested that due to scarcity of detailed analyses, the adoption of indirect methods may be required by calculation of relative risk (RR). Compared with other diseases (e.g. schistosomiasis), the application of indirect methods to calculate RR in the case of LF is very delicate. This is explained by the fact that prevalence rates and morbidity of LF strongly depend on socio-economic, environmental and ecological factors. 5
To date, however, these dynamics have been poorly explored and are far from fully understood. A comparison between areas affected by water resources development projects and areas with similar ecological and epidemiological characteristics but with no water resource projects might be possible but analysis requires great care. Applying study results from a specific site to other more distant areas with similar climatic characteristics is even more arguable, because of small-scale heterogeneities (e.g. prevalence rates vary between villages while climatic parameters are similar). In view of the very few high-quality studies and the lack of a good understanding of basic transmission dynamics of LF in relation to water resources development and management, we conclude that it is currently not possible to calculate the population attributable fraction of risk factors or assign and quantify attributable fractions of disease burdens caused by LF. It was not feasible to derive DALY estimates for the 10 relevant WHO-designated sub-regions of the world.
Outlook and Perspectives Undoubtedly, the potential impact of water resources development and management impact on transmission parameters of LF is considerable, but many critical questions remain unanswered. Obviously, the dearth of LF-based studies pertaining to water resources development and management cannot be caught up with within the next few years. Here we propose a selection of research priorities in the field of LF, without considering research pertaining to clinical aspects, drug development, immunology and molecular parasitology. In our view, some important questions include: (1) What is the impact of water resources development and management projects on the frequency and transmission dynamics of LF in different eco-epidemiological settings? (2) What is the relationship between filaria transmission and infection prevalence or infection intensity and clinical manifestation rates, particularly in regions with altered transmission (vector species succession, transmission intensification) due to water resources projects?
6
(3) What are the specific risk factors, including ecological, epidemiological and socioeconomic, of LF in a certain region? (4) How big is the impact of rapid urbanisation, accompanied with lack of sufficient sanitation facilities, water-storage, urban and peri-urban subsistence agriculture or wastewater mismanagement on the transmission of filariae? In the following part we further explore these research priorities 1-4 and propose concrete strategies to tackle these questions. (1) The fundamental question, whether or not, how and how much water resources development and management projects impact on the transmission dynamics and prevalence of LF remains to be answered. On the basis of our comprehensive literature search and the identification of 14 publications, we were able to carry forward this task. By a meta-analysis of the identified publications, we reviewed and synthesised the outcomes and produced an article that has been published in the peer-reviewed literature. Special attention has been paid to typical agricultural practices currently employed in the high burden areas of LF and the predicted development of water resources (e.g. irrigated rice agriculture). The question how water resource development and management projects impact on LF can, however, only be answered satisfactorily when the epidemiological parameters in affected communities and control communities are monitored prior, during and after their implementation. In other words, environmental and social determinants, transmission indicators and infection and disease prevalences should be kept under rigorous surveillance for an adequate number of years. Conceivably, the LF situation could be compared with that of malaria. This is justified on several grounds: First, in some regions of the world the same vectors transmit both LF and malaria (e.g. in West Africa Anopheles funestus and Anopheles gambiae). Second, both diseases depend on similar risk factors, e.g. low socio-economic status, poor housing conditions, and limited access to health care systems. Third, while the majority of the global LF disease burden is concentrated in Asia – where the main vectors are Culex ssp. – the malaria burden is currently concentrated in sub-Saharan Africa. Importantly, Asia provides suitable environmental conditions for Anopheles species, harbours more people than any other continent and contains the biggest proportion of irrigated agriculture. The correct interpretation of this discrepancy will provide further 7
insight into the complex issue of “parasite-vector-environment” relations, both for LF and malaria. (2) The relation between transmission, the prevalence of infection and clinical manifestations, particularly infection intensity (worm load) and LF morbidity is still not fully understood. In areas where ecological transformations have occurred, e.g. through the development of irrigation systems, this issue gains in importance. Such transformations often lead to the creation of breeding sites or they diminish or alter existing breeding habitats suitable for filaria vectors. As a consequence, the density of vector populations fluctuates and vector species composition changes. Therefore, environmental alterations potentially have an impact on filaria transmission. It has to be assumed that higher filaria transmission gradually increases infection prevalence and the worm burden. Thus, morbidity and clinical manifestation rates are expected to increase. In the case of LF, the connection between transmission, infection and morbidity is complex and often contradictory. Previous studies showed that people with elephantiasis are often amicrofilaraemic while others have a high grade of infection but show no clinical signs. The study design already described in (1) could also be applied for investigating the connection between filaria infection and clinical disease manifestations. Special attention has to be given to the confounding factors resulting from the fact that regions with improved water resources facilities attract people. The influx of people from areas with either low or high LF transmission can bias the outcome of a study. To tackle these questions, we propose to design a research project of the following kind: The investigation should be implemented in an area where a water resources development project is planned. Prior to its implementation, a baseline cross-sectional survey will assess infection prevalence, clinical manifestation rates and transmission parameters of all filaria vectors. Further, demographic parameters such as ethnicity, socio-economic status, and migration patterns of the affected population should be recorded. At the next phase, designed as a cohort-study, the research will assess infection prevalence, clinical manifestation rates and transmission parameters of all filaria vectors during the construction and implementation of the water resources project. After the project’s completion, the area should be further monitored and all parameters reassessed longitudinally for at least another five years. The outcomes of such a study holds promise to examine and quantify how transmission parameters, infection prevalence and clinical manifestation rates are interrelated. Further, it will expand the 8
current evidence base of adverse determinants attributable to water resources development. Future water resource development should include in-depth assessment of potential health impacts, including LF. Indeed, institutionalisation of health impact assessments (HIAs) for development projects quite generally, analogous to environmental impact assessments, would lead to information requirements that could fill many of the data gaps described in this review. In addition, mitigation strategies to alleviate potential negative health impacts – of which LF might be only one component – would also be part of the process of implementing new water projects. Introduction of monitoring and surveillance systems proximal to such water projects would facilitate systematic evaluation of the impact of these ecosystem interventions over time. This, in turn, would greatly improve our understanding of the role of dams and irrigation systems in either promoting or reducing LF transmission. Shedding light on these dynamics is an essential step towards a complete understanding of the disease. It will also help sustain the achievements of the Global Alliance to Eliminate Lymphatic Filariasis (GAELF). This initiative aims to reduce filaria infection prevalence and clinical manifestation rates to nearly zero by wide-spread chemotherapy. Complete elimination is, however, not feasible and vector populations will be unaffected by the GAELF. To prevent resurgence and proliferation of LF in the future, the strategy of GAELF has to be upgraded to include a vector control component. A better understanding of the disease is the basis for a sustainable control strategy that also targets the vector population. (3) For prevention and sustainable control of LF it is crucial to have a clear perception of its risk factors. To date, the vast majority of studies focused on transmission rates, infection prevalence or frequencies of clinical symptoms but did not define risk factors and other determinants of LF. It is of great importance to investigate the major risk factors for LF in the context of specific areas. As already mentioned under item (1), the development of water resources potentially creates several new risks and aggravates common risk factors. Other determinants also alter key parameters of LF. Taking the construction of irrigation systems as an example, we here describe what kind of risk factors and determinants have to be considered and how their magnitude can be estimated:
9
Socio-economic factors People in irrigated areas benefit from higher agricultural yields and can improve their socio-economic status. This translates into potentially better access to health services, increased means to purchase health services and products, and improved nutritional status. Alleviation of poverty will, therefore, have an impact on LF morbidity. Irrigation schemes also attract a work force which may have a different socio-economic status. In the case of LF, studies should always consider the socio-economic status of an affected population and differentiate groups with different levels of vulnerability, as well as the evolution of vulnerability over time. Population density, immigration Areas in which irrigated agriculture is practiced or where man-made reservoirs are created attract people and this results in higher population densities. It significantly alters the demographic structure around water resources development projects. This may lead to the creation of several new risk factors, including those linked to wastewater accumulation, waste mismanagement and poor housing conditions. Furthermore, immigrants may be more susceptible to LF infections if they come from regions with less filaria transmission. In turn, immigrants from regions where LF is highly endemic will introduce filariae and thus change transmission intensity in regions where LF transmission used to be non-existent or low. Studies are needed to elucidate how higher population densities and human movement, in connection with water resources development, affect transmission and morbidity of LF. Artificial breeding-sites and habitat change Irrigation creates or changes breeding sites that are suitable for filaria vectors. New plants and animals or the marginalisation of species can lead to shifts in vector species composition, and can introduce new vector species. As a consequence, vector transmission parameters change and eventually the frequency and intensity of clinical manifestation will also change. To investigate these determinants, transmission parameters, vector species composition, infection prevalence and clinical manifestation rates have to be investigated prior, during and after the construction of irrigation systems. As the transmission can vary from year to year it is crucial to monitor those LF parameters over a period of several years.
10
Exposure Exposure is a factor that directly influences vectorial capacity. If the human–vector contact is altered, this affects prevalence rates as well. The factors described above also influence
human
exposure
to
filaria-transmitting
mosquitoes.
Socio-economic
improvement can result in better housing conditions or an improved capacity to purchase insecticide-treated mosquito nets. Migration of labour force, e.g. farmers, into areas where Culex is active, can be expected to result in an increased exposure to vectors. Vector species composition shifts can promote mosquitoes whose “time of biting activity” and host preference is different. We suggest these factors be considered integrally in future investigations. (4) Currently, the connection between rapid, uncontrolled urbanisation and the proliferation of LF is not well understood. This topic is, however, of considerable public health significance and is expected to further gain in importance, particularly in view of the rapid pace of urbanisation, notably in areas where LF poses high levels of risk (Asia and sub-Saharan Africa). In shanty towns, for example, the building of small-scale irrigation systems, the storage of water for household consumption and the lack of improved sanitation facilities can influence the frequency and transmission dynamics of LF. Due to rapid environmental transformation and population growth, peri-urban settings are considered to be particularly challenging for health research and planning. Our systematic review underscores the need to assess the importance and magnitude of urban LF and we suggest this had best be achieved by the following study design: First, infection prevalence and morbidity of LF can be assessed by means of crosssectional studies carried out in various urban settings, e.g. in shanty towns, areas with subsistence agriculture and in inner cities. Second, breeding-sites of filaria vectors should be defined and the mechanics of their creation described. Third, all important risk factors and determinants should be assessed. Tackling these three issues will lead to a better understanding of the dynamics and the contextual determinants of LF in relation to water resources development and management, infection, morbidity and urbanisation. Findings from these studies will form an important basis for the design and implementation of LF-control strategies, more appropriate planning of water resources development and management projects and the incorporation of effective health safeguards in urban planning and development.
11
Appendix 1: Search strategy and selection criteria for our comprehensive literature review First, a literature search with special emphasis on research findings published over the past 25 years was carried out using the National Library of Medicine’s PubMed database, OVID Technologies (WebSPIRS 5.02), Cambridge Scientific Abstracts Internet Database Service and Thomson ISI (previously known as Institute for Scientific Information). With special consideration of potential bias of research findings during the DDT era, we also included published work between 1945 and 1975. PubMed/Medline contains citations published mostly from 1966 to the present, whereas Thomson ISI database dates back to 1945. The following keywords were employed to search the above-mentioned databases and websites: “lymphatic filariasis” in combination with “malaria”, “epidemics”, “water”, “sanitation”, “water supply”, “water development”, “irrigation”, “dam(s)”, “recreation”, “diversion”, “pool(s)”, “drainage”, “water reservoir(s)”, “water management”, “drinking water”, “downstream”, “upstream”, “sea water”, “environmental management” (“modification”, “manipulation”), “water storage”, “flood control”, “water purification”, “impoundment”, “barrage”, “navigation”, “humidity”, “environment” and “environmental”. Second, this search was complemented with an iterative proceeding in which we consistently reviewed reference lists of all those publications that were of relevance to address our main objective. The bibliographies of all these recovered manuscripts were retrieved again and the searching strategies repeated until no new information was forthcoming. Third, we also performed computer-aided searches of the websites of the following organisations and institutions: World Health Organisation (WHO), Food and Agriculture Organisation of the United Nations (FAO), World Bank, Centers for Disease Control and Prevention (CDC, Atlanta), online catalogues of the University of Basel and Princeton University. The yields of these searches were found to be meagre. Third, dissertation abstracts and unpublished documents (‘grey literature’) were reviewed. Dissertation abstracts were searched in following databases (accessed on 23.12.2004): - www.google.com - ProQuest Digital Dissertations (http://wwwlib.umi.com/dissertations). - Wageningen Dissertation Abstracts (http://www.agralin.nl/wda/). - Index of Theses. A comprehensive listing of abstracts by universities in Great Britain and Ireland (http://www.theses.com/). - COPAC union online catalogue of the members of the Consortium of University Research Libraries (CURL) (http://www.copac.ac.uk/). - Cambridge Scientific Abstracts Internet Database Service: (http://www.lib.ecu.edu/erdbs/csa.html). - M25 Consortium of Academic Libraries (http://www.m25lib.ac.uk/). - The Unicorn Online Catalogue (WEBCAT) of the London School of Hygiene and Tropical Medicine (http://193.63.251.23/uhtbin/cgisirsi). - IRIS Interdisciplinary Online-Databases (www.libiris1.ict.ac.uk). - Library Online Catalogue IDS Basel/Bern (http://aleph.unibas.ch). - University of Chicago, Center for Research Libraries, Foreign Doctoral Dissertations (http://wwwcrl.uchicago.edu/content.asp). - University of Berkeley Digital Library (http://sunsite.berkeley.edu/Libweb/). For this search we employed the same keywords as described above for the peer-reviewed literature search. Through these databases no useful additional data could be found. The peer-review literature and dissertation abstract search made it clear that in this field of research only a small number of studies were done and even fewer published. Since the “grey literature” is mostly not listed in any database it cannot be retrieved remotely by electronic search-engines.
12
Appendix 2. Number of hits for “lymphatic filariasis” combined with selected keywords in different electronic databases (accessed January 15 2005) Search term
Database PubMed OVID Web of Cambridge Scientific Abstracts World MedLine Technologies Science Internet Database Service Cat Environmental Agricola GeoRef Sciences & Pollution Management
Lymphatic filariasis Lymphatic filariasis and epidemics Lymphatic filariasis and water Lymphatic filariasis and sanitation Lymphatic filariasis and water supply Lymphatic filariasis and water development Lymphatic filariasis and irrigation Lymphatic filariasis and dam(s) Lymphatic filariasis and recreation Lymphatic filariasis and diversion Lymphatic filariasis and pool(s) Lymphatic filariasis and drainage Lymphatic filariasis and reservoir(s) Lymphatic filariasis and management Lymphatic filariasis and drinking Lymphatic filariasis and downstream Lymphatic filariasis and upstream Lymphatic filariasis and sea water Lymphatic filariasis and environmental management (modification, manipulation) Lymphatic filariasis and storage Lymphatic filariasis and flood control Lymphatic filariasis and water purification Lymphatic filariasis and impoundment Lymphatic filariasis and barrage Lymphatic filariasis and navigation Lymphatic filariasis and humidity Lymphatic filariasis and environment Lymphatic filariasis and environmental
1584 425
659 1
953 3
8
4
0
77 0
33 13
10 4
15 3
1
0
0
1 0
5
1
1
0
0
0
0
13
4
0
0
0
0
0
6 0 (2) 0
3 0 (1) 0
3 0 (1) 0
0 0 0
0 0 0
0
0 0 0
0 13 (8) 6 9 (14)
0 12 (9) 0 2(2)
0 13 (10) 3 4 (3)
0 0 0 0
0 0 0 0
0 0 0 0
0 0
46
29
26
0
0
0
6
3 0
1 0
2 0
0 0
0 0
0 0
1 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3 0
0 0
2 0
0 0
0 0
0 0
0 0
0
0
0
0
0
0
0
0
0
0
0
0 0
0 0
0 0
0
0
0
0
4 52
3 5
2 11
0 0
0 0
0 0
0 0
29
22
17
0
0
0
0
13
0
0
Appendix 3. Relevant literature to address our main research objective (in inverse chronological order) Smith A. The transmission of bancroftian filariasis on Ukara Island, Tanganyika II. The distribution bancroftian microfilaraemia compared with the distribution hut-haunting mosquitoes and their breeding-places. Bulletin of Entomology Research. 1955;46:437444. Jordan P. Filariasis in the lake province of Tanganyika. East African Journal. 1956 Jun;33(6):237-42. Basu PC. Filariasis in Assam state. Indian Journal of Malariology. 1957;11:293-308. Partono F, Pribadi PW, Soewarta A. Epidemiological and clinical features of Brugia timori in a newly established village. Karakuak, West Flores, Indonesia. American Journal of Tropical Medicine and Hygiene. 1978;27(5):910-5. Samarawickrema WA, Kimura E, Spears GF, Penaia L, Sone F, Paulson GS, Cummings RF. Distribution of vectors, transmission indices and microfilaria rates of subperiodic Wuchereria bancrofti in relation to village ecotypes in Samoa. Transactions of the Royal Society of Tropical Medicine and Hygiene. 1987;81(1):129-35. Rajagopalan PK, Panocker KN, Das PK. Control of malaria and filariasis vectors in south India. Parasitology Today. 1987;3(8):233-40. Raccurt CP, Lowrie RC Jr, Katz SP, Duverseau YT. Epidemiology of Wuchereria bancrofti in Leogane, Haiti. Transactions of the Royal Society of Tropical Medicine and Hygiene. 1988;82(5):721-5. Amerasinghe FP, Ariyasena TG, 1991. Survey of Adult Mosquitos (Diptera, Culicidae) During Irrigation Development in the Mahaweli Project, Sri-Lanka. J Med Entomol 28: 387-393. Hunter JM. Elephantiasis: a disease of development in north east Ghana. Social Science and Medicine. 1992;35(5):627-45; discussion:645-9. Gad AM, Feinsod FM, Soliman BA, Nelson GO, Gibbs PH, Shoukry A. Exposure variables in bancroftian filariasis in the Nile Delta. Journal of the Egyptian Society of Parasitology. 1994;24(2):439-55. Appawu MA, Baffoe-Wilmot A, Afari EA, Nkrumah FK, Petrarca V. Species composition and inversion polymorphism of the Anopheles gambiae complex in some sites of Ghana, west Africa. Acta Trop. 1994 Feb;56(1):15-23. Dzodzomenyo M, Dunyo SK, Ahorlu CK, Coker WZ, Appawu MA, Pedersen EM, Simonsen PE (1999). Bancroftian filariasis in an irrigation project community in southern Ghana. Tropical Medicine and International Health. 4(1):13-8. Appawu MA, Dadzie SK, Baffoe-Wilmot A, Wilson MD. Lymphatic filariasis in Ghana: entomological investigation of transmission dynamics and intensity in communities served by irrigation systems in the Upper East Region of Ghana. Tropical Medicine and International Health.2001;6(7):511-6. Supali T, Wibowo H, Ruckert P, Fischer K, Ismid IS, Purnomo, Djuardi Y, Fischer P. High prevalence of Brugia timori infection in the highland of Alor Island, Indonesia. American Journal of Tropical Medicine and Hygiene. 2002;66(5):560-5. 14
Appendix 4. Review article published in the American Journal of Tropical Medicine and Hygiene 73(3), 2005, pp. 523-533
EFFECT OF WATER RESOURCE DEVELOPMENT AND MANAGEMENT ON LYMPHATIC FILARIASIS, AND ESTIMATES OF POPULATIONS AT RISK TOBIAS E. ERLANGER1, JENNIFER KEISER1, MARCIA CALDAS DE CASTRO2, ROBERT BOS3, BURTON H. SINGER4, MARCEL TANNER1 AND JÜRG UTZINGER1 1
Swiss Tropical Institute, Basel, Switzerland Geography Department, University of South Carolina, Columbia, South Carolina, USA 3 Water, Sanitation and Health, World Health Organization, Geneva, Switzerland 4 Office of Population Research, Princeton University, Princeton, New Jersey, USA 2
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) sub-regions 1 and 2 mosquito densities of the Anopheles gambiae complex and An. 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 People living in tropical and sub-tropical countries have long suffered under the yoke of lymphatic filariasis (LF). This chronic parasitic disease is of great public health and socio-economic significance and is currently endemic in 80 countries/territories of the world.1--3 LF accounts for serious disfiguration and incapacitation of the extremities and the genitals and causes hidden internal damage to lymphatic and renal systems.4--6 Disease, disability, and disfiguration are responsible for a loss of worker productivity, significant treatment costs and social stigma.7,8 At present, the global burden of LF is estimated at 5.78 million disability adjusted life years (DALYs) lost annually.9 Hence, its 15
estimated burden is almost 3.5-fold higher than that of schistosomiasis and approximately one seventh of that of malaria.9 LF is caused by Wuchereria bancrofti, Brugia malayi and B. timori, with > 90% of cases attributable to W. bancrofti.1 Transmission occurs through various mosquito species, primarily Culex (57%), followed by Anopheles (39%), Aedes, Mansonia, and Ochlerotatus. Detailed information on the geographical distribution of the most important LF vectors can be found elsewhere.2 More than 60% of all LF infections are concentrated in Asia and the Pacific region, where Culex is the predominant vector. In Africa, where an estimated 37% of all infections occur, Anopheles is the key vector.2 In 1993, the World Health Organization (WHO) declared LF to be one of six eliminable infectious diseases.10 After several years of preparation and endorsement by the World Health Assembly in 1997, the Global Programme to Eliminate Lymphatic Filariasis (GPELF) was initiated in 1998.11 Large-scale operations were launched in 2000, alongside the forging of a worldwide coalition, the Global Alliance to Eliminate Lymphatic Filariasis, which is a free and non-restrictive partnership forum. WHO serves as its secretariat and is being reinforced by an expert technical advisory group.12--14 GPELF’s goal is to eliminate the disease as a public health problem by 2020. It mainly relies on mass drug administration using albendazole plus either ivermectin or diethylcarbamazine (DEC). At the end of 2003, approximately 70 million people were treated and 36 countries had an active control program in place.14 Sustained political and financial commitment and rigorous monitoring and surveillance are essential elements of the global program, as otherwise LF could reemerge since a small fraction of the population will continue to carry microfilaria. Furthermore, the vector population is unlikely to be significantly affected by GPELF. Employing a mathematical modeling approach, it was shown that vector control programs, in addition to mass drug administration would substantially increase the chances of meeting GPELF’s ambitious target.15 Indeed, some of the most successful control programs in the past demonstrate that an integrated approach, readily adapted to specific eco-epidemiological settings, was a key factor for controlling and even eliminating LF.16--19 In rural areas undergoing ecological transformations, particularly due to the construction of irrigation schemes and dams, new breeding sites suitable for filaria vectors are created.16,20 As a consequence, the transmission dynamics of LF is expected to change. In Africa, where Anopheles transmit malaria and filaria, the estimated surface area of 12 million ha 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 design and lack of maintenance of infrastructures for drainage of sewage and storm-water, waste-water management, water storage, and urban subsistence agriculture can facilitate the proliferation of mosquitoes, including those transmitting filaria. Although the proportion of urban dwellers in the least developed countries was only 27% in 1975, it rose to 40% in 2000 and is predicted to further increase. Nearly 50% of the world’s urban population is concentrated in Asia. Currently, the annual growth rate in Asian cities is 2.7%.24 This implies that in the future, an increasing number of habitats with organically polluted water will be available for Culex vectors. The objectives of the systematic literature review presented in this paper were (i) to assess the current size of the population at risk of LF with particular consideration of water resource development and management, both in rural and urban settings, and (ii) 16
to assess the effect of these ecological transformations on the frequency and transmission dynamics of LF. Our working hypothesis was that environmental changes resulting from water resource development and management adversely affect vector frequencies, filaria transmission, prevalence of infection, and clinical occurrence of LF. These issues are of direct relevance for GPELF and evidence-based policy-making, and for integrated vector management programs and optimal resource allocation for disease control more generally. MATERIALS AND METHODS Contextual determinants and estimation of population at risk in endemic countries. As a first step, we outlined the contextual determinants of LF transmission in a simplified flow chart (Figure 1). For regional estimates of populations at risk of LF, we used the recent classification set forth in the appendices of the annual World Health Report of WHO, which stratifies the world into 14 epidemiological sub-regions.9 For estimation of population fractions at risk of LF due to water resource development and management, we adopted setting-specific definitions. Hence, for rural areas we considered those people at risk of LF who live in close proximity to irrigated agroecosystems, employing data sources from the Food and Agricultural Organization (FAO; http://www.fao.org). We followed a similar approach as in our preceding work with an emphasis on the malaria burden attributable to water resource 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 currently under irrigation in LF-endemic countries/territories. In urban settings the size of the population at risk of LF was defined by the proportion that currently lacks access to improved sanitation. Country-specific percentages of urban dwellers without access to improved sanitation were taken from the World Health Report 2004.9 Justification for this indicator is derived from the following experiences. First, there is evidence that, besides common water-borne diseases, lack of access to clean water and improved sanitation increases the risk of acquiring vector-borne diseases.23,26,27 As will be shown in our review and has been noted before, LF transmission is spurred by rapid urbanization in the absence of accompanying waste management and sanitation facility programs.28-32 Second, a large-scale campaign built around chemotherapy and improved sanitation proved successful to control LF in the Shandong province, People’s Republic of China.33 Third, Durrheim and colleagues recently suggested that chronic parasitic diseases, including LF, could be utilized as viable health indicators for monitoring poverty alleviation, as the root ecological causes of these health conditions depend on poor sanitation, inadequate water supply and lack of vector control measures.27 Search strategies and selection criteria. With the aim of identifying all published studies that examined the effect of water resource development and management on the frequency and transmission dynamics of LF, we carried out a systematic literature review. Particular consideration was given to publications that contained specifications on (i) entomological transmission parameters, abundance of vector populations, microfilaria infection prevalence and rates of clinical manifestations as a result of water resource development, and (ii) studies that compared sites where environmental 17
changes occurred with ecologically similar settings where no water resource developments were implemented. As a first step, we performed computer-aided searches using the National Library of Medicine’s PubMed database, as well as BIOSIS Previews, Cambridge Scientific Abstracts Internet Database Service and ISI Web of Science. We were 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 “water”, “water management”, “reservoir(s)”, “irrigation”, “dam(s)”, “pool(s)”, “sanitation”, “ecological transformation”, and “urbanization”. No restrictions were placed on language of publication. In a next step, the bibliographies of all recovered articles were hand-searched to obtain additional references. In an iterative process, this approach was continued until no new information was forthcoming. Dissertation abstracts and unpublished documents (‘grey literature’) were also reviewed. Dissertation abstracts were searched in online databases, i.e., ProQuest Digital Dissertations, and the Unicorn Online Catalogue (WEBCAT) of the London School of Hygiene and Tropical Medicine. Finally, online databases of international organizations and institutions, namely WHO and FAO of the United Nations, and the World Bank, were scrutinized, adhering to the same search strategy and selection criteria explained above. RESULTS Contextual determinants. The contextual determinants of LF can be subdivided into three broad categories, namely (i) environmental, (ii) biological, and (iii) socioeconomic (Figure 1). They act on different temporal and spatial scales, adding to the complexity of the local LF eco-epidemiology. In the first category, LF transmission is mainly determined by climatic factors and the formation or disappearance of suitable breeding sites for the vector. Breeding sites can be either natural or man-made, and their productivity exhibits strong heterogeneity, even on a small scale, which in turn governs filarial transmission dynamics. In rural settings, the most prominent man-made breeding sites are water bodies created by irrigation systems and dams. Here, the weight of environmental determinants is strongly associated with biological factors, notably vector and parasite species, and various socio-economic factors such as human migration patterns, access to, and performance of, health systems, and individual protective measures. In urban areas, artificial breeding sites are often created by waste-water mismanagement, resulting from poor sanitation systems in private dwellings and industrial units, or the absence of them entirely. Here, biological factors shape the epidemiology of LF after environmental changes have occurred, and socio-economic factors strongly interact with the environmental determinants. The local quality of domestic and industrial waste-water management, access to clean water and improved sanitation, and the construction of roads and buildings depend on the socio-economic status of specific sub-populations.
18
Figure 1. Contextual determinants of lymphatic filariasis
Environmental changes due to waterresource development and management Environmental Factors Climate
Rural
Urban
Agriculture & irrigation
Industrial waste-water (mis)-management
Large hydroelectric dam construction
Domestic water storage
Small dams & barrages for agriculture &
Construction of roads & buildings
domestic use
Local sewerage systems
Water supply & sanitation
Integrated vector management
Intervention Mass-treatment Intervention with filariacides
MassBiological Factors
treatment with filariacides
Poverty alleviation
Socio-economic Factors Poverty alleviation Poverty
Parasite
Mosquito
Human
Population density
Population density
Population density
Species & strain
Species & strain
Immigration & emigration
Survival
Insecticide resistance
Sex, age, ethnicity & immunity Exposure
Knowledge, attitudes & practices
Health systems
Longevity
19
Endemic countries/territories. Table 1 shows estimates of populations at risk of LF for all the countries/territories where the disease is currently endemic. Only politically independent countries were listed (n = 76). Hence, the populations at risk of French Polynesia, New Caledonia, Réunion, and Wallis and Futuna, which belong to France, and American Samoa, which belongs to the United States of America, were assigned to the geographically closest independent states. Timor-Leste, which recently became independent, is also included. However, no estimates for at-risk populations are currently available for the following LF-endemic countries: Cambodia, Cape Verde, Lao People’s Democratic Republic, Republic of Korea, Solomon Islands, and Sao Tome and Principe. In view of relatively small population sizes living in these countries, neglecting at-risk population of LF there, only marginally influences estimates on regional and global scales. Table 1. Estimates of population at risk in all lymphatic filariasis (LF)-endemic countries/territories of the world, stratified into WHO epidemiological sub-regions (population at risk of LF in thousands. Africa WHO sub-region 1a (24 countries) Angola (10,423), Benin (6,736), Burkina Faso (12,963)b, Cameroon (9,338), Cape Verde (n.d.), Chad (6,216), Comoros (768)b, Equatorial Guinea (89), Gabon (896), Gambia (1,235), Ghana (6,200)b, Guinea (8,336), Guinea-Bissau (1,253), Liberia (34), Madagascar including Reunionc (15,841), Mali (11,329), Mauritius (12)d, Niger (10,416), Nigeria (121,901), Sao Tome and Principe (n.d.), Senegal (9,247), Seychelles (81), Sierra Leone (890), Togo (1,182)b WHO sub-region 2a (14 countries) Burundi (1,112), Central African Republic (765), Congo (3,396), Côte d’Ivoire (14,253), Democratic Republic of the Congo (22,481), Ethiopia (3,534), Kenya (10,108), Malawi (11,948), Mozambique (15,336), Rwanda (3,355)e, Uganda (23,399), United Republic of Tanzaniaf (14,421), Zambia (9,980), Zimbabwe (10,816) The Americas WHO sub-region 4 (6 countries) Brazilg (3,569)h, Costa Ricag (83)h, Dominican Republic (1,854)h, Guyana (623)h, Surinameg (< 4)i, Trinidad and Tobagog (< 13)h WHO sub-region 5 (1 country) Haiti (6,078)b Eastern Mediterranean WHO sub-region 7 (3 countries) Egyptf (2,446)b, Sudan (8,302)h, Yemen (100)k South-East Asia WHO sub-region 11 (3 countries) Indonesia (27,046)h [B. malayi: 27,046, B. timori: 3,900]l, Sri Lanka (9,900)b, Thailandm (10,116)k [B. malayi: 7,791]k WHO sub-region 12 (6 countries) Bangladesh (93,984)h, India (494,374)h [B. malayi:190,718]h, Maldives (< 3)n, Myanmar (28,000)b, Nepal (1,359)h, Timor-Leste (778)i [B. timori: 778]i Western-Pacific WHO sub-region 13 (1 country) Brunei Darussalam (40)o WHO sub-region 14 (18 countries) Cambodia (n.d.), China (925,979)h [B. malayi: 63,906]h, Cook Islands including French Polynesiac (248)k, Federated States of Micronesia (109)k, Fiji including Wallis and Futunac (854)k, Kiribati (88)k, Lao People’s Democratic Republic (n.d.), Malaysiag (2,736)h [B. malayi: 2,736]h, Niue (2)k, Papua New Guinea (3,000)p, Philippines (23,800)b [B. malayi: 23,800]b, Republic of Korear (n.d.), Samoaf including American Samoac (248)k, Solomon Islandsr (n.d.), Tonga (104)k, Tuvalu (11)k, Vanuatuf including New Caledoniac (422)k, Viet Nam (12,288)h
20
n.d.: no data currently available) a Except Mauritius percentages of the population at risk from Lindsay & Thomas (2000),59 re-calculated with recent figures from United Nations (2004)60 b Weekly Epidemiological Record (2004)14 c Réunion, French Polynesia, Wallis and Futuna, and New Caledonia belong to France; American Samoa belongs to the United States of America d WHO (2002)61 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 recorded in the United Republic of Tanzania, Egypt, Samoa and Vanuatu3 g In Brazil, Costa Rica, Suriname, Trinidad and Tobago, and Malaysia smaller endemic foci have been eliminated3 h Percentage of people at risk in 1990 taken from Michael et al. (1996),62 re-calculated with recent figures from United Nations (2004)60 i Pan American Health Organization (2002)63 k Weekly Epidemiological Record (2003)64 l Supali et al. (2002)39 m Thailand has recently eliminated filaria transmission3 n People at risk estimated < 1%13 o It has been assumed that Brunei Darussalam has the same percentage of people at risk as Malaysia in 1995 as described by Michael et al. (1996)62 p Kazura & Bockarie (2003)65 r Korea and the Solomon Islands using diverse control strategies have eliminated transmission3
People at risk of LF at global and regional scale. We estimate that approximately half of all people currently living in LF-endemic countries are at risk of the disease, which translates to approximately 2 billion. This is considerably higher than the 1-1.2 billion estimates put forth in the literature.1,2,11 The difference is largely explained by at-risk estimates for China. In urban areas, there are 394.5 million at risk of LF due to lack of access to improved sanitation. This is almost twice the estimated size in rural areas, namely 213 million, which is attributed to living in close proximity to irrigated agriculture. The largest percentages in terms of LF burden, as expressed in DALYs lost (52%), people at risk (29%), size of the population at risk due to proximity to irrigated land (69%), and lack of improved sanitation (33%) are in WHO sub-region 12. This sub-region includes Bangladesh, India, Maldives, Myanmar, Nepal and Timor-Leste (Table 2). Studies identified and qualitative overview. Overall, 12 studies fulfilled the selection criteria of our literature review. These studies were all published in the peerreviewed literature, that is, in specialized entomology, parasitology and/or tropical medicine journals. None of the work retrieved from electronic databases other than PubMed or ISI Web of Science was deemed of sufficient quality to justify study inclusion. Table 3 summarizes the main findings of the selected studies, stratified by rural and urban settings. As a common theme, LF vector composition frequencies shifted in all settings. Water resource developments favored An. gambiae, An. funestus, An. barbirostris, Culex quinquefasciatus, Cu. pipiens pipiens, Cu. antennatus and Aedes polynesiensis, but disfavored An. pharoensis, An. melas, An. subpictus and Ae. samoanus. Transmission parameters were higher in ecosystems altered by water resource projects, and clinical disease manifestation rates often elevated.
21
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 (n.d.: no data currently available) DALYs in 2004 Total population in LF- Population at risk of LF Population in LF-endemic Urban population in LF-endemic (x 103) (from Table 1) countries living in proximity countries without access to caused by LF (103)a endemic countries 3 b improved sanitation (x 103)a (x 10 ) to irrigated areas (x 103) c g 1 3 976 284,551 235,382 574 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
