Acta Tropica xxx (2005) xxx–xxx

3 Effect of irrigated rice agriculture on ,

4 including challenges and opportunities for integrated

5 vector management

a,∗ b a c 6 Jennifer Keiser , Michael F. Maltese , Tobias E. Erlanger , Robert Bos , a d a 7 Marcel Tanner , Burton H. Singer ,Jurg¨ Utzinger

a 8 Swiss Tropical Institute, P.O. Box, CH-4002 Basel, Switzerland b 9 St. Antony’s College, Oxford University, Oxford OX2 6JF, UK c 10 Water, Sanitation and Health, World Health Organization, Avenue Appia 20, CH-1211 Geneva 27, Switzerland d 11 Office of Population Research, Princeton University, Princeton, NJ 08544, USA Received 29 October 2004; received in revised form 31 March 2005; accepted 5 April 2005

12

13 Abstract

14 Japanese encephalitis (JE) is a disease caused by an that is spread by marsh birds, amplified by pigs, and mainly 15 transmitted by the bite of infected tritaeniorhynchus mosquitoes. The estimated annual incidence and mortality rates are 16 30,000–50,000 and 10,000, respectively, and the estimated global burden to JE in 2002 was 709,000 disability-adjusted life years 17 lost. Here, we discuss the contextual determinants of JE, and systematically examine studies assessing the relationship between 18 irrigated rice agriculture and clinical parameters of JE. Estimates of the sizes of the rural population and population in irrigated 19 areas are presented, and trends of the rural population, the rice-irrigated area, and the rice production are analyzed from 1963 to 20 2003. We find that approximately 1.9 billion people currently live in rural JE-prone areas of the world. Among them 220 million 21 people live in proximity to rice-irrigation schemes. In 2003, the total rice harvested area of all JE-endemic countries (excluding 2 22 the Russian Federation and Australia) was 1,345,000 km . This is an increase of 22% over the past 40 years. Meanwhile, the total 23 rice production in these countries has risen from 226 millions of tonnes to 529 millions of tonnes (+134%). Finally, we evaluate 24 the effect of different vector control interventions in rice fields, including environmental measures (i.e. alternate wet and dry 25 irrigation (AWDI)), and biological control approaches (i.e. bacteria, nematodes, invertebrate predators, larvivorous fish, fungi 26 and other natural products). We conclude that in JE-endemic rural settings, where vaccination rates are often low, an integrated 27 vector management approach with AWDI and the use of larvivorous fish as its main components can reduce vector populations, 28 and hence has the potential to reduce the transmission level and the burden of JE. 29 © 2005 Published by Elsevier B.V.

30 Keywords: Japanese encephalitis; Geographical distribution; Global burden; Vector control; Rice agriculture; Irrigation; Integrated vector 31 management; Environmental control; Biological control

32 ∗ Corresponding author. Tel.: +41 61 284 8218; fax: +41 61 284 8105. E-mail address: [email protected] (J. Keiser).

1 0001-706X/$ – see front matter © 2005 Published by Elsevier B.V. 2 doi:10.1016/j.actatropica.2005.04.012

ACTROP 1666 1–18

UNCORRECTED PROOF 2 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

1 1. Introduction 50 years (Solomon et al., 2000). At present, the geo- 48 graphical distribution of JE ranges from Japan, mar- 49 2 Japanese encephalitis (JE) is a -borne viral itime Siberia and the Republic of Korea in the North, 50 3 disease. The JE virus is a member of the family Fla- to most parts of China and the Philippines in the East, 51 4 viviridae. Clinically apparent infection takes place in Papua New Guinea in the South, and and to 52 5 one out of 200–300 infected patients (Pugachev et al., the West (Broom et al., 2003). Recent outbreaks of JE 53 6 2003). The disease is characterized by a wide range of have been reported Southward in Australia, and west- 54 7 presentations, as both the symptoms and the clinical ward in Pakistan (Solomon et al., 2000). 55 8 course can differ broadly among patients. They range Currently, approximately 90% of the world’s rice 56 9 from mild flu-like symptoms to considerable neuro- is produced in Asia (Consultative Group on Interna- 57 10 logic symptoms, such as rigors, convulsions, polio-like tional Agricultural Research et al., 1998). In most of 58 11 flaccid paralysis, seizures or encephalomyelitis. Severe the countries, where JE outbreaks have been reported, 59 12 clinical cases are likely to have life-long neurological rice is not only a staple food, but rice growing also is 60 13 sequelae. Mostly children and young adults are affected a major economic activity and key source of employ- 61 14 (Solomon et al., 2000; Tsai, 2000; Ding et al., 2003; ment and income generation (Consultative Group on 62 15 Halstead and Jacobson, 2003). The annual incidence International Agricultural Research et al., 1998). Ef- 63 16 and mortality estimates for JE are 30,000–50,000 and forts to further enhance the high annual rice production 64 17 10,000, respectively (Solomon, 2004). However, there in these areas are essential to maintain food security. 65 18 is considered to be severe under-reporting of JE and It has been estimated that in the next 25 years the de- 66 19 one study estimated the annual incidence at 175,000 per mand for rice will rise by 65% in the Philippines, 51% 67 20 year (Tsai, 2000). JE outbreaks occur in cycles that may in , 45% in Viet Nam and 38% in 68 21 be linked to climatic patterns and the immune status of (http://www.biotech-info.net/rice expert.html). Thai- 69 22 the populations. The great majority of cases and death land, for example, is already in the planning stages 70 23 occur in World Health Organization (WHO) regions of of designing new rice-irrigation schemes for year- 71 24 South-East Asia and the Western Pacific. In 2002, the round irrigation (Consultative Group on International 72 25 estimated global burden of JE was 709,000 disability- Agricultural Research et al., 1998). Hence, there is con- 73 26 adjusted life years (DALYs) lost (WHO, 2004). At siderable concern in public health circles, as the in- 74 27 present there are no established antiviral treatments tensification of rice production systems as well as the 75 28 against JE. Interferon alpha was the most promising extension of the flooded surface area, particularly in 76 29 drug in small open-label trials, but it failed to affect the semi-arid areas, contributes greatly to increased fre- 77 30 outcome in children with JE (Solomon et al., 2003). quencies and intensities of JE outbreaks. 78 31 There has been a changing pattern in the epidemiol- The objectives of this paper were (i) to evaluate the 79 32 ogy of JE. On the one hand, primarily due to extensive effect of irrigated rice agriculture on the burden of JE; 80 33 vaccination campaigns, JE has been almost eliminated and (ii) to review different vector control interventions 81 34 in many economically advanced countries of East Asia and discuss challenges and opportunities for integrated 82 35 and South-East Asia (i.e. Japan, Republic of Korea and vector management (IVM). The remainder of this paper 83 36 Taiwan) and the burden of JE has been substantially re- is structured as follows. First, we put forward contex- 84 37 duced in many other endemic countries (Halstead and tual determinants of JE, highlighting the important role 85 38 Jacobson, 2003). On the other hand, intensified trans- of rice agroecosystems on JE vector populations. Sec- 86 39 mission has been observed in other parts of South-East ond, we review studies assessing the relationship be- 87 40 Asia and the Western Pacific, most likely due to an ex- tween irrigated rice growing and clinical parameters of 88 41 pansion of irrigated agriculture and pig husbandry, as JE. Third, we present estimates of the population living 89 42 well as changing climatic factors. Waterresource devel- in proximity to irrigation and rice-irrigation schemes in 90 43 opment and management, in particular flooded rice pro- JE-prone areas, stratified by relevant WHO sub-regions 91 44 duction systems, are considered among the chief causes of the world (WHO, 2004). Fourth, we quantify the 92 45 for several JE outbreaks (Amerasinghe and Ariyasena, changes of the rice-irrigated area, rice production, and 93 46 1991; Akiba et al., 2001). Conversely, the occurrence rural population sizes over the past 40 years, in coun- 94 47 of the disease has changed considerably over the past tries, where JE is currently endemic. Finally, we re- 95

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 3

96 viewed studies that employed different environmental Although JE vectors are able to breed in ground wa- 120 97 and biological control interventions to reduce larval JE ter habitats, sunlit pools, roadside ditches, tidal marshes 121 98 vector populations, which have been implemented in of low salinity, or man-made containers, one of their 122 99 irrigated rice production systems. major preferred larval habitats are rice fields (Mogi, 123 1984; Sucharit et al., 1989). The ecology of Culex 124 spp. in rice fields has been studied and reviewed in 125

100 2. Contextual determinants great detail (Suzuki, 1967; Lacey and Lacey, 1990). 126 Since very high densities of JE vector were 127

101 Fig. 1 depicts the contextual determinants of JE. found consistently in rice fields, it was concluded 128 102 The most important epidemiological features that gov- that the impact of these man-made breeding sites is 129 103 ern the transmission of JE include (i) environmental much more important than that of natural breeding 130 104 factors, i.e. agricultural practice, altitude, climate and places. For example, a significant increase in the abun- 131 105 the presence of pigs and marsh birds; (ii) human im- dance of Cx. tritaeniorhynchus and increased human- 132 106 munization rates and vector control measures; and (iii) vector contacts have been noted following completion 133 107 socioeconomic parameters. of the large rice-irrigation scheme in the Mahaweli 134 108 The principal vector of JE is Culex tritae- project, (Amerasinghe and Ariyasena, 1991; 135 109 niorhynchus. Female specimens are infective 9–10 Amerasinghe, 1995). 136 110 days after having taken the viraemic blood meal, JE vector abundance is closely related to agro- 137 111 having undergone three gonotropic cycles (Gajanana climatic features (Peiris et al., 1993; Phukan et al., 138 112 et al., 1997). Other culicine mosquitoes that can trans- 2004), most notably temperature and monthly rain- 139 113 mit JE include Cx. bitaeniorhyncus, Cx. epidesmus, Cx. fall (Suroso, 1989; Solomon et al., 2000; Bi et al., 140 114 fuscocephala, Cx. gelidus, Cx. pseudovishnui, Cx. si- 2003). In addition, potential JE vectors were rarely 141 115 tiens, Cx. vishnui and Cx. whitmorei (Sehgal and Dutta, found at altitudes above 1200 m (Peiris et al., 1993). 142 116 2003). In Australia, Cx. annulirostris was found to be However, the most important causative factors of JE 143 117 the major JE vector species (Hanna et al., 1996). In is the management of paddy water, and the peak peri- 144 118 a recent study in Kerala, South India, JE was isolated ods of mosquito abundance are associated with cycles 145 119 from indiana (Arunachalam et al., 2004). in local agricultural practices. In , the highest 146

Fig. 1. Contextual determinants of Japanese encephalitis.

ACTROP 1666 1–18

UNCORRECTED PROOF 4 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

147 numbers of larvae and pupae of JE vectors were col- of different environmental and biological control in- 195 148 lected when the rice fields were ploughed with the water terventions in rice fields. Though JE vectors tend to 196 149 in the fields (also termed puddling). The vector popula- bite and rest outdoors, self-protection behaviour, such 197 150 tion decreased after transplanting when the fields were as sleeping under insecticide-treated nets (ITNs), us- 198 151 flooded, and stayed low until harvesting (Somboon ing repellants, and insect-proofing homes and 199 152 et al., 1989). In , small plots in the rice fields, work places, might also assist in reducing JE trans- 200 153 which are common before planting and contain veg- mission (WHO, 1997). For example, a population- 201 154 etation, were found to be conductive factors to facili- based case-control study in China, which evaluated 202 155 tate enhanced breeding of JE vectors; up to 40 pupae the protective effect of ITNs against JE, showed that 203 2 156 were collected per m (Heathcote, 1970). Furthermore, the risk of infection among children below 10 years 204 157 the height of the rice plants, water temperature, dis- was greatly reduced (Luo et al., 1994). On the other 205 158 solved oxygen, ammonia nitrogen and nitrate nitrogen hand, all 187 serologically confirmed JE cases in re- 206 159 strongly influence the abundance of immatures (Sunish cent outbreaks in Northeast India reported that they 207 160 and Reuben, 2001). The practice of paddy cultivation, had slept under a bed-net (Phukan et al., 2004). Fur- 208 161 proximity of houses to water bodies and suitable cli- thermore, application of deet-permethrin soap (“Mos- 209 162 matic factors were the most important environmental bar”) let to an 89–100% reduction in man-vector con- 210 163 factors associated with several recent JE outbreaks in tact, including vectors transmitting JE (Mani et al., 211 164 Northeast India (Phukan et al., 2004). 1991b). 212 165 The presence of pigs and marsh birds is crucial in The third broad category of determinants that gov- 213 166 the etiology of JE, as the virus is carried by birds and ern JE transmission is people’s socioeconomic status. 214 167 amplified by pigs (Broom et al., 2003). The latter are In Central China, more JE cases were observed among 215 168 the most important natural host for transmission of JE children living in poor quality houses, and whose par- 216 169 to humans. Pigs have high and prolonged viraemias, ents had lower income. However, the sample size of the 217 170 are often common in endemic countries, and are gen- study was small, which might explain why the associa- 218 171 erally reared in open and unroofed pigpens, which tions were not statistically significant (Luo et al., 1995). 219 172 are located near houses (Mishra et al., 1984; Solomon Religion, exposure to domestic , and household 220 173 et al., 2000). In 2004, the world’s pig population was crowding were also described as risk factors associated 221 174 estimated at 951 million, of which 60% are found in with JE (Halstead and Jacobson, 2003). 222 175 Asia (www.faostat.fao.org). In this part of the world 176 pig farming has increased considerably over the past 177 10 years, from 490 million pigs in 1994 to 574 million 3. Rice irrigation and JE incidence 223 178 in 2004 (+17.1%) (www.faostat.fao.org). In Asia pigs 179 stocks increased. Humans, goats, and horses are As summarized in the previous section, the man- 224 180 considered dead-end hosts (Reuben et al., 1992). For agement of paddy water strongly influences the trans- 225 181 example, in the Thanjavur district, India, an area with mission of JE. Hence, we were motivated to conduct a 226 182 extensive rice agriculture, a very low JE incidence has systematic literature review to identify published work 227 183 been reported, which has been explained by a high cat- with an emphasis on the relationship between rice ir- 228 184 tle to pig ratio (400:1) (Vijayarani and Gajanana, 2000). rigation and JE. We searched Biosis previews, Ovid 229 185 The important role of birds was demonstrated in Indian Technologies, Medline and the Web of Science ap- 230 186 villages with or without herons in close proximity; in plying the following keywords: “Japanese encephali- 231 187 rice-growing villages without herons, seroconversion tis” and “water”, or “rice”, or “irrigation”, or “rice ir- 232 188 rates in children aged 0–5 and 6–15 years were 0% and rigation and agriculture”, or “paddy”, or “field”. Pa- 233 189 5%, respectively. In ecologically-similar villages with pers published in English, French or German were in- 234 190 herons, the corresponding rates were 50% and 56%, cluded. We also considered manuscripts written in Chi- 235 191 respectively (Mani et al., 1991a). nese, Japanese or Korean, if an abstract of these pa- 236 192 The second category of contextual determinants pers was accessible in English on the above-mentioned 237 193 comprises vaccination and transmission interruption databases. The key findings of our literature review are 238 194 strategies. Later in this paper, we review the effects summarized here. 239

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 5

240 Four studies in India analyzed the presence of rice risk-table.htm). The countries were grouped into differ- 285 241 irrigation and vector abundance in relation to the inci- ent epidemiological sub-regions of the world, accord- 286 242 dence of JE. In the Gorakhpur district in Uttar Pradesh, ing to recent classifications of WHO, which is based 287 243 and the Mandya district of Karnataka, areas extensively primarily on child and adult mortality rates (WHO, 288 244 developed for irrigated rice agriculture, the occurrence 2004). The JE-endemic countries are located in seven 289 245 of JE was closely associated with high vector densi- of the 14 WHO sub-regions (Table 1). 290 246 ties, breeding in the fields or the canal system. The Though sporadic JE cases have been reported 291 247 highest numbers of JE cases were observed shortly af- from peri-urban areas, disease occurrence is largely 292 248 ter the mosquito densities peaked (Mishra et al., 1984; restricted to rural settings (Self et al., 1973; Solomon 293 249 Kanojia et al., 2003). In addition, in the Mandya dis- et al., 2000). Hence, in Table 1 we show data on 294 250 trict, a high incidence of JE was found in extensively ir- the rural population for the individual JE-endemic 295 251 rigated areas, while few cases occurred in villages with countries only (United Nations, 2004). For those 296 252 less irrigation or no irrigation systems (Geevarghese countries, where JE is endemic only in certain parts 297 253 et al., 1994). In Assam, 78.6% of the JE cases occurred of the rural areas (e.g. in the Russian Federation JE 298 254 in families practicing rice cultivation (Phukan et al., outbreaks have been observed only in far Eastern 299 255 2004). maritime areas South of Chabarovsk (CDC, 2004; 300 256 Based on an epidemiological, serological and clin- http://www.cdc.gov/ncidod/dvbid/jencephalitis/risk- 301 257 ical study of 54 JE patients in the Northern Philip- table.htm), we determined the rural endemic popula- 302 258 pines, 41 cases (76%) were associated with irrigated tion of these areas consulting Encarta Encyclopedia 303 259 rice fields and only one patient (2%) could not be linked (Microsoft Corporation, 2004). 304 260 to rice irrigation (Barzaga, 1989). In Taiwan, JE was We find that approximately 1.9 billion people cur- 305 261 monitored over 4 years and the highest numbers of rently live in rural JE-prone areas of the world, the 306 262 cases were observed in the counties with the highest majority of them in China (766 million) and India (646 307 263 number of rice paddies (e.g. in Ilan county in 1969 a million) (Tables 1 and 2). This figure is in line with a 308 264 morbidity rate of 5.5 per 100,000 inhabitants). Fewer recent “population at risk estimate”, quoting a popu- 309 265 cases occurred in areas where dry farming had been lation of 2 billion, 700 million of these children under 310 266 adopted (e.g. in Yunlin county in 1969 a morbidity rate the age of 15 years living in JE-endemic areas (Tsai, 311 267 of 0.64 per 100,000 inhabitants) and no cases were re- 2000). 312 268 ported in non-rice cultivated provinces (Okuno et al., As discussed above surface irrigation is an impor- 313 269 1975). tant risk factor in the epidemiology of JE. Conse- 314 270 In 1985–1986 in the Mahaweli System H, Sri Lanka, quently, in Table 2 we present the size of the total ir- 315 271 an epidemic of JE occurred, resulting in more than rigated and the rice-irrigated land in JE-endemic ar- 316 272 400 cases and 76 deaths. In 1987–1988 a second out- eas, together with data on populations at risk from 317 273 break took place with >760 cases and 138 deaths. these areas and estimated DALYs, stratified by the 318 274 The promotion of smallholder pig husbandry was sus- relevant WHO sub-regions. The numbers on the ir- 319 275 pected to be responsible for these outbreaks. The high- rigated and the rice-irrigated areas in JE-prone ar- 320 276 est number of cases occurred in areas with irriga- eas were calculated by multiplying the total national 321 277 tion and pig husbandry, while no cases were reported irrigated and rice-irrigated area (http://www.fao.org) 322 278 from non-irrigated areas with few pigs (Amerasinghe, by the “endemic fraction” (e.g. 17.7% of Pakistan’s 323 279 2003). population is estimated to live in the endemic area, 324 hence we assume that 17.7% of both the total area 325 and the rice-irrigated areas are located in JE-prone 326 280 4. Population at risk areas). The sizes of the populations in JE-endemic 327 irrigated areas and JE-endemic rice areas were esti- 328 281 In order to estimate the current population at risk of mated by multiplying the average national rural pop- 329 282 JE, we first compiled a list of all countries where JE ulation densities (United Nations, 2004) by the to- 330 283 epidemics or sporadic cases have been reported (CDC, tal endemic area under irrigation and rice paddies, 331 284 2004; http://www.cdc.gov/ncidod/dvbid/jencephalitis/ respectively. 332

ACTROP 1666 1–18

UNCORRECTED PROOF 6 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

Table 1 Rural population/population in irrigated areas in JE-endemic countries stratified by relevant WHO sub-regions of the world (all numbers in thousand; n.d.: no data currently available) The Americas WHO sub-region 3 United States of America: Guam (150/n.d.) and Saipan (40/n.d.)

Eastern Mediterranean WHO sub-region 7 Pakistana (16,750/4,001)

Europe WHO sub-region 10 Russian Federationb (614/n.d.)

South-East Asia WHO sub-region 11 Indonesia (119,589/3033), Sri Lanka (15,057/1461), Thailand (42,796/4114)

WHO sub-region 12 Bangladeshc (111,165/28,958), Bhutanc (2,065/17), Democratic People’s Republic of Korea (8,818/1,066), Indiad (646,100/100,873), (34,927/809), Nepale (12,500/329), Timor-Leste (719/51)

Western-Pacific WHO sub-region 13 Australiaf (15/n.d.), Brunei Darussalam (85/<1), Japang (25,113/2376), Singapore (<100/n.d.)

WHO sub-region 14 (11,514/172), Chinah (766,757/24,796), Lao People’s Democratic Republic (4489/30), Malaysia (8814/99), Papua New Guinea (4958/n.d.), Philipines (31,182/1612), Republic of Korea (9,395/1081), Viet Nam (60,441/5460)

a Areas around Karachi and in the Lower Indus Valley, province of Sind (CDC, 2004, http://www.cdc.gov/ncidod/dvbid/jencephalitis/risk- table.htm); population estimates taken from (United Nations, 2004) and (Microsoft Corporation, 2004). b Far Eastern maritime areas South of Chabarovsk (administrative district of Primorskij Kraij) (CDC, 2004, http://www.cdc.gov/ncidod/ dvbid/jencephalitis/risk-table.htm); population estimates taken from (United Nations, 2004) and (Microsoft Corporation, 2004). c Bangladesh and Bhutan are potential JE-endemic countries, but due to lack of data the situation has to be clarified (CDC, 2004, http://www.cdc.gov/ncidod/dvbid/jencephalitis/risk-table.htm). d Reported cases from all states except Arunachal, Dadra, Daman, Diu, Gujarat, Himachal, Jammu, Kashmir, Lakshadweep, Meghalaya, Nagar Haveli, Orissa, Punjab, Rajasthan and Sikkim (CDC, 2004, http://www.cdc.gov/ncidod/dvbid/jencephalitis/risk-table.htm); population estimates taken from (United Nations, 2004) and (Microsoft Corporation, 2004). e Hyperendemic in Southern Terai lowlands (BBIN, 2004). f Outbreaks on the Islands of Torres Strait and on mainland Australia at Cape York Peninsula (Hanna et al., 1996); at-risk to JE assumed for all inhabitants at Cape York Peninsula; population estimates from (United Nations, 2004) and (Microsoft Corporation, 2004). g Rare-sporadic cases on all islands except Hokkaido; population estimates taken from (United Nations, 2004) and (Microsoft Corporation, 2004). h Cases in all provinces except Xizang (Tibet), Xinjiang, Qinghai (CDC, 2004, http://www.cdc.gov/ncidod/dvbid/jencephalitis/risk-table.htm); population estimates taken from (United Nations, 2004) and (Microsoft Corporation, 2004).

333 According to our estimates 1,025,000–1,080,000 ing to our estimates 132–167 million people live in 340 2 334 km land is irrigated in JE-prone areas. We find that JE-endemic irrigated areas in this WHO sub-region 341 335 currently 180–220 million people are living in prox- (Table 2). 342 336 imity to irrigation or rice-irrigation schemes in the The estimated global burden of JE in 2002 was 343 337 JE-endemic regions, and thus are at risk of con- 709,000 DALYs.WHO sub-regions 12 and 14 currently 344 338 tracting the disease (Table 2). Irrigated agriculture is bear 84% of this burden (597,000 DALYs).The remain- 345 339 most pronounced in WHO sub-region 12. Accord- ing 16% are thought to occur in WHO sub-regions 7

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 7 h h g ) 3 10 × 1,517 (1.0%) 32,749 (2.1%) 220,859 (6.1%) Population in JE-endemic rice areas ( United Nations, h,i i densities ( h g ) 3 10 × Population in JE-endemic irrigated areas ( . f ) 3 10 × Rural endemic population ( 190 (100%) n.d. n.d. e ) 3 10 × Microsoft Corporation (2004) 5 151,9963,623,815 25,313 (16.6%) 1,934,154 (53.3%) 2,376 (1.6%) 180,338 (4.9%) Total population in 2003 ( ) and h h , Rome. d ) 2 en.asp 12,971 (0.16%) 308,844 (2.7%)1,080,999 (2.4%) 1,560,475 897,550 (57.5%) 33,249 (2.1%) Rice paddies in JE-endemic areas in 2003 (km h,i i h d ) 2 Irrigated land in JE-endemic areas in 2002 (km c ) 2 http://www.fao.org/waicent/portal/statistics Total JE-endemic areas (km http://www.cdc.gov/ncidod/dvbid/jencephalitis/risk-table.htm b . ) 2 Total area (km . a . DALYs lost due to JE in 2002 ) by the total area under irrigation/of rice paddies in JE-endemic areas. Omitting and Australia. Omitting Papua New Guinea. The size of the endemic irrigated population and endemic population in rice areas was estimated by multiplying the average national rural population Microsoft Corporation (2004) Data for the whole country obtainedUnited from Nations the (2004) Food and Agriculture Organization 2004 and multiplication by the endemic fraction. Food and Agriculture Organization 2004, Estimated with aid of CDC, 2004 ( WHO (2004) i f c e a g b h d 7 83,000 796,095 140,914 (17.7%) 31,507 (4.0%) 3912 (0.5%) 153,578 16,750 (10.9%) 4,001 (2.6%) 497 (0.3%) 3 n.d. 663 663 (100%) n.d. n.d. 190 Total 709,000 44,457,730 13,425,469 (30.1%) 1,025,639 (2.3%) 10111213 n.d. 29,000 277,000 n.d. 17,075,400 2,483,300 4,437,432 165,760 (1.0%) 2,483,300 (100%) 3,401,672 (76.7%) 8,125,410 104,100 (4.2%) n.d. 506,195 (11.4%) 330,670 (4.1%) 233,888 521,384 (9.4%) (11.7%) 20,317 (0.25%) 1,312,548 301,782 n.d. 816,294 (62.2%) 177,442 (58.8%) 132,103 (10.1%) 8,609 (2.9%) 143,246 167,647 (12.8%) 18,448 614 (0.4%) n.d. n.d. 14 320,000 11,539,430 6,902,490 (59.8%) 363,520 (3.2%) WHO sub-region n.d.: no data. Table 2 Burden of JE and population at risk in endemic areas and living in close proximity to irrigation, stratified by relevant WHO sub-regions 2004

ACTROP 1666 1–18

UNCORRECTED PROOF 8 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

346 and 11, whereas no information is currently available WHO sub-region 13. And finally, in WHO sub-region 391 347 for WHO sub-regions 3, 10 and 13. 14, a substantial expansion of the rice-irrigated area has 392 occurred between 1963 and 1973, but has decreased 393 thereafter. 394 348 5. Trend of rice agriculture in endemic WHO The total rice production has risen from 226 mil- 395 349 sub-regions lions of tonnes in 1963 to 529 millions of tonnes in 396 2003 (+134%). The most significant increases of rice 397 350 We compiled data for the relevant WHO sub-regions production occurred in WHO sub-regions 7, 11, 12 and 398 351 on the rice harvested area and rice production between 14, as shown in Fig. 2. On the other hand, in WHO sub- 399 352 1963 and 2003, which is available at http://faostat. region 13 (mainly Japan) a reduction from 16.6 to 9.7 400 353 fao . org / faostat / form?collection = Production .Crops. millions of tonnes (−43%) has taken place. 401 354 Primary&Domain=Production&servlet=1&hasbulk=0 We also depict in Fig. 2 the change of the rural pop- 402 355 &version=ext&language=EN. No distinction could ulation of the five most relevant WHO sub-regions, 403 356 be made between rain-fed rice and irrigated rice and which overall increased from 1325 million inhabitants 404 357 no information was available on the intensity of the in 1963 to 2197 million inhabitants in 2003 (+66%). In 405 358 irrigated rice production (single, double or triple the past decade a decrease in the rural population has 406 359 cropping). been observed in WHO sub-regions 11 and 14. 407 360 No data was obtainable for the JE-endemic coun- 361 tries of WHO sub-regions 3 (Guam and Saipan) and 362 Singapore. We excluded data on the rice harvested area 6. Intervention strategies in rice fields 408 363 and rice production in the Russian Federation, because 364 the current size of the JE-endemic area there are small Our aim was to systematically review the litera- 409 365 compared to the rest of the JE-endemic countries. We ture to identify published work on biological control 410 366 furthermore excluded Australia as outbreaks have so strategies against Cx. tritaeniorhynchus larvae in rice 411 367 far only be reported Torres Strait Islands and on main- fields. We did not include studies on the application of 412 368 land Australia at Cape York Peninsula (Hanna et al., synthetic larvicides and insecticides against Cx. tritae- 413 369 1996). niorhynchus in rice fields, as the use of chemical control 414 370 As mentioned above rice agriculture accounted for was found to be of no operational value due to the short 415 2 371 up to 1,080,000 km in all JE-prone areas in 2003. In activity of these products, high costs and resistance de- 416 372 the same year the total rice harvested area of these velopment (Wada, 1988). We searched the same elec- 417 373 countries, including the non-JE-endemic regions, was tronic databases as mentioned in Section 3 and used the 418 2 374 1,345,000 million km . This is an increase of 22% over following keywords: “Japanese encephalitis” or “Culex 419 375 the past 40 years. In some of these countries the rice- tritaeniorhynchus” in combination with “alternate wet 420 376 irrigated area has almost doubled over the period in- and dry irrigation”, or “Azolla”, or “Bacillus”, or “bio- 421 377 vestigated. For example, in Pakistan the irrigated area logical control”, or “bacteria”, or “control”, or “fungi”, 422 2 2 378 increased from 12,861 km in 1963 to 22,100 km in or “intermittent irrigation”, or “invertebrates”, or “lar- 423 379 2003 (+72%). Marked declines in the rice-irrigated vivorous fish”, or “natural products”, or “nematodes”, 424 380 area have been observed in other countries. In Japan, or “predator”, or “water management”. 425 381 for example, the rice-irrigated area has been halved 2 2 382 (32,760 km in 1963 to 16,650 km in 2003). In Fig. 2 6.1. Environmental control strategies 426 383 we present the changes of the rice-irrigated area over 384 the past 4 decades for 5 of the 7 relevant WHO sub- 6.1.1. Alternate wet and dryirrigation (AWDI) 427 385 regions, where JE has been reported to be a significant Traditionally, rice fields are flooded, which provides 428 386 public health problem. In WHO sub-regions 11 and 12, ideal breeding places for several mosquito species, in- 429 387 a substantial increase in the rice-irrigated area is appar- cluding those that transmit JE. The management of ir- 430 388 ent. In WHO sub-region 7 (Pakistan) the rice area grew rigation water that leads to the alternate wetting and 431 389 until 1993, but remained stable over the past 10 years. drying of fields, including the canals and ditches, can 432 390 Meanwhile, the rice-irrigated area has decreased in be active or passive. An important feature of this tech- 433

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 9

Fig. 2. Changes of rice growing area, rice production and rural population at risk of JE in 5 WHO sub-regions between 1963 and 2003.

ACTROP 1666 1–18

UNCORRECTED PROOF 10 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

434 nique is that the soil can dry out, which in turn curtails 435 the life cycle development of the mosquito from lar- 436 vae and pupae to adult. In order to achieve a significant 437 reduction of mosquito larvae, AWDI (also termed inter- 438 mittent irrigation) has to be applied during the entire 439 cropping season and should cover all rice fields that 440 are connected by irrigation canals over a large area. 441 This method is particularly feasible in places where 442 control of the water supply and drainage is possible, 443 hence where land and climatic conditions are suitable Rice yield Water consumption 444 (Mogi, 1988). Growing water shortages in many areas 445 create an incentive to better control irrigation water 446 management. AWDI is one such strategy. The poten- 447 tial of AWDI is, however, limited in areas where there 448 is a threat of insufficient resources to re-flood the fields 449 and where farmers perceive a risk of reduced yields by 450 letting their fields dry out. C. tritaeniorhynchus adults Not known Not known Not known 451 The effects of AWDI on the abundance of Anophe- 452 les, which are the vector species of , rice yield, 453 water consumption and methane emission have been 454 reviewed recently (van der Hoek et al., 2001; Keiser 455 et al., 2002). In brief, this method had been introduced 456 in Asia about 300 years ago, primarily to obtain higher 457 rice yields. The first study of AWDI as a potential tool 458 to reduce mosquito vectors was conducted in Bulgaria C. tritaeniorhynchus immatures 459 on mosquitoes of the An. maculipennis complex in the 14.3–48.2% 460 1920s (Keiser et al., 2002). AWDI was compulsory in 461 Portugal in the 1930s, when malaria was still a problem larvae in rice fields 462 there. 463 We found four studies analyzing the effect of AWDI 464 on densities of JE vectors and the key findings are sum- 465 marized in Table 3. Overall, Cx. tritaeniorhynchus im- 466 matures were reduced by 14–91% in rice fields apply- 467 ing AWDI. One study investigated the effect of AWDI

468 on the Cx. tritaeniorhynchus adult population, which Culex tritaeniorhynchus 469 decreased by 55–70%. The crop yield was examined

470 in two trials and increases between 4 and 13% were ) Irrigation interval: 3–5 days Decrease: 75–88% Not known Increase: 4% Not known 471 observed in AWDI rice fields. The effect this method ) Midseason drying Decrease of fourth instars: 472 has on the incidence of JE remains to be investigated. ) Irrigation interval: 5 days Decrease: 81–91% Decrease: 55–70% Increase: 13% Decrease: 50%

473 6.2. Biological control strategies ) Irrigation interval: Several days Decrease Not known Not known Not known

474 6.2.1. Bacteria Rajendran et al., 1995 Takagi et al., 1995 Mogi, 1993 475 The larvicidal activity of the spore forming bac- Lu-Baulin, 1988 476 teria Bacillus sphaericus and B. thuringiensis israe- 477 lensis have been discovered in 1965 and 1976, re- 478 spectively (Mittal, 2003). Subsequently, these biocides Tsu City, Japan ( Kinryu, Japan ( Study site; referenceHenan, China ( Tamil Nadu, India ( Specificity of AWDI Outcome 479 have been evaluated in various formulations against Table 3 Alternate wet and dry irrigation (AWDI) against

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 11

Table 4 Application of Bacillus spp. against Culex tritaeniorhynchus larvae in rice fields Study site; period (if known); Intervention strategy Outcome reference India, 1983–1984 (Kramer, 1984) Application of Bacillus sphaericus 91–99% reduction with doses of 0.5–1.5 kg/ha Culex and Bacillus thuringiensis H-14 in tritaeniorhynchus and subpictus, however rice fields activity did not subsist beyond a few days India, 1982 (Balaraman et al., 1983) Application of Bacillus thuringiensis 100% reduction with doses of 27 × 105 spores/ml, but first H-14 in rice fields and second instars reappeared 3 days after application India (Sundararaj and Reuben, 1991) Application of microgel droplet 44–79% reduction of early instar and 82–100% reduction of formulation of Bacillus sphaericus in late instar of Culex tritaeniorhynchus,Cx. vishnui and Cx. rice fields pseudovishnui for at least 5 weeks applying a dose of 4.3 kg/ha Korea (Rhee et al., 1983) Application of Bacillus thuringiensis 95–98% reduction of Culex tritaeniorhynchus. The residual israelensis H-14 in rice field effects lasted only 24 h at doses of 0.8–1.2 l/ha of Bacillus thuringiensis israelensis H-14, SAN 402 I suspension concentrate containing 600 IU/mg Rourkela city, India, June 1993 to Application of Bacillus sphaericus Application of B. sphaericus, when sprayed at 1 g/m2 October 1994 (Yadav et al., 1997) (strain B-101, serotype H5a, 5b) in significantly reduced larval and pupal counts (P < 0.0001) in rice field rice fields. Duration of effect could not be determined as fields received periodically Bacillus sphaericus treated wastewater

480 mosquito vectors worldwide (Lacey and Lacey, 1990). B. sphaericus 1593 M resulted in a 44–79% reduc- 507 481 B. thuringiensis produces four key insecticidal pro- tion of early instar and 82–100% reduction of late 508 482 teins, while B. sphaericus produces a single binary instar larvae (C. fuscanus, C. pseudovish- 509 483 toxin (Federici et al., 2003). These protein toxins bind nui and C. tritaeniorhynchus) for at least 5 weeks 510 484 to cells of the gastric caecum and posterior midgut of (Sundararaj and Reuben, 1991). Similarly, up to 511 485 the mosquitoes causing intoxication, which eventually 95–98% of C. tritaeniorhynchus larvae were reduced 512 486 leads to death (Mittal, 2003). The advantages of these in three other field sites evaluating B. sphaericus or 513 487 bacterial insecticides in terms of efficacy, specificity B. thuringiensis formulations (Balaraman et al., 1983; 514 488 and environmental safety are well documented (Zahiri Rhee et al., 1983; Kramer, 1984). However, the larvi- 515 489 et al., 2004). A major drawback of these products, cidal activity did not persist in these rice fields beyond 516 490 however, is their high cost of production, the labor- a couple of days; in the Republic of Korea the residual 517 491 intensive delivery, as well as first reports of resistance effect of B. thuringiensis H-14 was found to last only 518 492 (Sundararaj and Reuben, 1991; Federici et al., 2003). 24 h (Rhee et al., 1983). 519 493 The effect of the individual biolarvicides depends on

494 water temperature, pH, aquatic vegetation, the formula- 6.2.2. Nematodes 520 495 tion applied, type of habitat, and the target mosquitoes Romanomermis culicivorax is probably the best- 521 496 (Mittal, 2003). B. sphaericus, for example, is known to studied nematode parasite of mosquitoes. The prepar- 522 497 exhibit a high activity against Culex mosquitoes, while asitic nemas are applied to the fields, where they locate 523 498 certain Aedes species are not affected (Mittal, 2003; a host, penetrate the cuticle and develop within the 524 499 Zahiri et al., 2004). Hence, no general conclusion mosquito larvae (Lacey and Lacey, 1990). Although 525 500 can be drawn about whether B. sphaericus and B. this method has not been widely studied, we identified 526 501 thuringiensis israelensis formulations are suitable for three publications assessing the effect of nematodes on 527 502 JE vector control in rice fields. The results on the JE vectors, which are summarized in Table 5.InTai- 528 2 503 application of these bacterial formulations in rice wan, application of 7000 nematodes per m rice field 529 504 fields against Cx. tritaeniorhynchus are summarized yielded a reduction of 11–18% C. tritaeniorhynchus 530 505 in Table 4. In Tamil Nadu, India, the application (Lacey and Lacey, 1990). In two studies in China the 531 506 of 4.3 kg/ha of a microgel droplet formulation of nematode R. yunanensis has been distributed in rice 532

ACTROP 1666 1–18

UNCORRECTED PROOF 12 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

Table 5 Application of nematodes against Culex tritaeniorhynchus larvae in rice fields Study site; period (if known); reference Intervention strategy Outcome China, 1986–1995 (Peng et al., 1998) Predation efficacy of Romamomermis 2000–4000 nematodes/m2 yielded parasitism rates of yunanensis in rice fields 52.2–96.7% for Culex tritaeniorhynchus China (Song and Peng, 1996) Predation efficacy of Romamomermis 1000–3000 nematodes/m2 resulted in parasitism rates of yunanensis and a second 60.8–95.5% of Culex tritaeniorhynchus and Romamomermis species in rice fields Cx.quinquefasciatus and Anopheles spp. Taiwan, 1974 (Lacey and Lacey, 1990; Predation efficacy of Romamomermis Larvae reduction of 11–18% of Culex tritaeniorhynchus Mitchell et al., 1974) culicivorax in rice fields

2 533 fields (2000–4000 nematodes/m ), which resulted in duction of exotic predators such as Gambusia might 568 534 parasitism rates of 52.2–96.7% and 60.8–95.5%, re- displace the native fish populations to reduce their nat- 569 535 spectively, of C. tritaeniorhynchus larvae (Song and ural value, as being observed in Japanese rice fields 570 536 Peng, 1996; Peng et al., 1998). (Wada, 1988). Therefore, the compatibility of the cho- 571 sen fish with local fauna and flora is of high importance 572 (Lacey and Lacey, 1990). 573 537 6.2.3. Invertebrate predators 538 Invertebrate predators, i.e. Coleoptera, Hemiptera 539 or Odonata, though less common than the use of fish, 6.2.5. Fungi 574 540 are also known to substantially reduce mosquito lar- Fungi that have been studied extensively for their 575 541 val populations in rice fields (Lacey and Lacey, 1990). potential as biological mosquito control agents include 576 542 However, they are highly sensitive to temperature, pres- Coelomomyces spp. and Lagenidium giganteum. The 577 543 ence of vertebrates, growth of rice and chemical pollu- former have been investigated how they impact the 578 544 tants (Lacey and Lacey, 1990). In India, the presence of development of JE vectors in China. Field observa- 579 545 notonectids was negatively associated with larval abun- tions there showed a strong effect of the fungus Coelo- 580 546 dance of Cx. pseudovishnui, Cx. tritaeniorhynchus and momyces indica on Cx.tritaeniorhynchus, as infected 581 547 Cx. vishnui (Sunish and Reuben, 2002). larvae were unable to develop into adults (Liu and Hsu, 582 1982). However, fungi have not been applied for bio- 583

584 548 6.2.4. Larvivorous fish logical control of JE vectors on a large scale so far, as 585 549 The use of larvivorous fish to reduce mosquitoes has practical problems, for example their production, have 586 550 a history of over 100 years (Lacey and Lacey, 1990). yet to be solved (Lacey and Lacey, 1990). 551 The mosquitofish, Gambusia affinis, is the most widely 552 used predator. Other fish species include Tilapia spp., 6.2.6. Other natural products 587 553 Poecilia reticulata or Cyprinidae (Lacey and Lacey, Natural products might also have high potential for 588 554 1990). A detailed summary of the most common preda- reducing the proliferation of culicine mosquitoes in 589 555 cious fish is given in (Lacey and Lacey, 1990). rice fields. Our literature search on the control of JE 590 556 After stocking rice fields with 1–10 natural preda- vectors in rice fields yielded in-depth investigation of 591 2 557 tor fishs per m , larval populations of Cx. tritae- only two natural products. In Tamil Nadu, India ap- 592 558 niorhynchus were reduced by 55.2–87.8% (Table 6). plication of the floating water fern Azolla microphylla 593 559 Larvivorous fish cannot be applied in rice fields, where greatly reduced immature mosquito populations. How- 594 560 irregular irrigation is practiced. It should also be noted ever, the infestation of the rice field with Azolla was 595 561 that predator populations are strongly influenced by difficult to achieve, 80% coverage by Azolla was ac- 596 562 temperature, rice growth, vegetation, use of pesticides complished only 13–14 days after rice transplantation, 597 563 or chemical pollutants (Lacey and Lacey, 1990). In limiting its wider use as a biological mosquito control 598 564 addition, recent research has shown that ovipositing agent (Rajendran and Reuben, 1991). 599 565 mosquitoes may move to other breeding sites in re- Neem cake powder, made from freshly harvested or 600 566 sponse to the stocking of rice fields with predatory fish stored neem kernels, from the neem tree (Azadirachta 601 567 (Angelon and Petranka, 2002). Furthermore, the intro- indica) rich in the active principle azadirachtin. At a 602

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 13

Table 6 Application of fish predators against Culex tritaeniorhynchus larvae in rice fields Study site; period (if known); reference Intervention strategy Outcome Bulkyo, Bosong-gun, Chollanamdo, Presence of Misgurnus mizolepis in Coefficients of correlation between Misgurnus Korea 1995–1996 (Lee, 1998) rice fields mizolepis and abundance of mosquito larvae showed negative correlations in Anopheles sinensis (−0.66) and Culex tritaeniorhynchus (−0.47) Suwon, 1992–1993 (Kim et al., 1994) Moroco oxycephalus and Misgurnus 4 fish/m2 gave a reduction of 55.2% of immatures of anguillicaudatus in rice fields Anopheles sinensis, Culex pipiens pallens and Culex tritaeniorhynchus compared to control field after 8 weeks South Delhi, May–June 1980 (Mathur Presence of Gambusia affinis in rice 10 fish/m2 gave a mean reduction of 63.9% of et al., 1981) fields immatures of Culex tritaeniorhynchus in 4 weeks China (Liao et al., 1991) Predation efficacy of 78 days after treatment with 1.9 fish/m2 a reduction of Ctenopharyngdon idella 80% Culex tritaeniorhynchus larvae was observed India, June to October 1991 (Prasad Predation efficacy of Gambusia 87.8% mosquito (6 Anopheline and 4 Culicine) larval et al., 1993) affinis in rice fields control in well submerged rice fields, no effect of Gambusia affinis in rice fields applying AWDI Quanzhou county, China, 1988–1989 Predation efficacy of Cyprinius Population size of larvae and adult Culex (Wu et al., 1991) carpio,Ctenopharyngdon idella and tritaeniorhynchus decreased Tilopia spp. South Korea, 1979 and 1981 (Yu et al., Predation efficacy of Aphyocypris Fish stocking of 1.5/m2 in Wondang-Ni rice paddy 1982) chinensis resulted in mosquito larval reduction of 98.9% in the third week after fish introduction against Anopheles sinensis, Culex pipiens, Culex crientalis and Culex tritaeniorhynchus Jindo Island, Chollanam-do Province, Predation efficacy of Aphyocypris Aphyocypris chinensis (1.5 fish/m2) achieved a South Korea, 1989 (Yu and Kim, chinensis in combination with reduction of 60.8-77.0% against both, Anopheles 1993) Bacillus thuringiensis (H-14) sinensis and Culex tritaeniorhynchus in the first 2 month. In the third month Bacillus thuringiensis (H-14) treatment was made, which yielded a satisfactory degree of control maintained above 93.1% for 2 weeks

603 dosage of 500 kg/ha it yielded an 81% reduction of late panying ecological transformations are manifold, and 619 604 instar culicine larvae and a 43% reduction in the pupal inherently difficult to assess and quantify. Negative ef- 620 605 production (Rao et al., 1992). Stable product neem cake fects include increased frequencies and transmission 621 606 coated urea and the combination of neem cake coated dynamics of water-based (e.g. schistosomiasis) and 622 607 urea and water management practices (i.e. AWDI) led water-related vector-borne diseases, including malaria 623 608 to a 70–95% reduction of culicine immatures. The ben- (Keiser et al., in press), lymphatic filariasis (Erlanger 624 609 efits were also found to be cost-effective as neem treat- et al., in press) and JE. On the other hand, such water 625 610 ment gave higher values for mean number of produc- projects are key for hydroelectric power production and 626 611 tive tillers, plant height and mean number of grains per food security; hence they can stimulate social and eco- 627 612 panicle (Rao et al., 1995). nomic development. In this review, we have focused 628 on irrigation schemes, with particular emphasis on rice 629 growing. Currently, there is a paucity on high quality 630

613 7. Discussion and conclusion data on the effects of large and small dams on JE. 631 Over the last four decades rice production has ex- 632

614 Water resource development and management, in panded considerably in most countries where JE is cur- 633 615 particular the construction and operation of small and rently endemic. This growth will most likely continue 634 616 large dams and irrigation schemes, occurred at an enor- for food security reasons. We find that in 2003 approx- 635 617 mous pace over the past 50 years (Gujja and Perrin, imately 180–220 million people live near irrigation 636 618 1999). The effects of these water projects and accom- schemes in JE-prone countries, hence have a poten- 637

ACTROP 1666 1–18

UNCORRECTED PROOF 14 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

638 tial risk of contracting JE. These figures are unavoid- side effects occur in 10–30% and allergic reactions oc- 686 639 ably subject to a level of uncertainty. The possibility cur in 0.6% of vaccinated adults (Jones, 2004). Fur- 687 640 that we have overvalued the population at risk cannot thermore, clinically significant neurological adverse 688 641 be ruled out, which is justified on four grounds. First, events, such as a meningoencephalitis, have been re- 689 642 pigs, which are the amplifying hosts, are not present ported. In addition, these vaccines are expensive and 690 643 in all rural irrigated sites. Second, the disease affects production can hardly keep up with demand (Jones, 691 644 mainly children under the age of 15 years. This segment 2004). In China a live attenuated vaccine (SA14-14- 692 645 of the rural population in Asia accounts for approx- 2) was developed in 1988 and the efficacy of a single 693 646 imately one-third of the total rural Asian population dose was estimated to be 99.3% (Bista et al., 2001). 694 647 (Tsai, 2000). Third, it is presently not possible to pre- However, its use in other affected countries has been 695 648 cisely map the population of irrigated areas in partially restricted by regulatory concerns over manufacturing 696 649 JE-endemic countries. Forth, no data are available on and control (Solomon et al., 2003). 697 650 vaccination coverage in rural irrigated areas. However, There is ongoing research to develop improved JE 698 651 it is also conceivable that we might have underesti- vaccines. These efforts received a boost in December 699 652 mated the actual population at risk of JE due to irriga- 2003, when the Bill and Melinda Gates Foundation 700 653 tion, as we have applied a rather conservative proxy, donated US$ 27 million to the Children’s Vaccine Pro- 701 654 namely the national rural population density. Irrigated gram at the Program for Appropriate Technology in 702 655 areas, however, are often characterized by a large pop- Health for this purpose (UNO, 2003). For example, a 703 656 ulation density as they attract a great number of people newly developed live attenuated vaccine (ChimeriVax- 704 657 (Keiser et al., in press). In addition, due to the predicted JE) that uses yellow fever 17D as a live vector for the 705 658 expansion of irrigation schemes, the rural population in envelope genes of SA14-14-2 virus has been tested in 706 659 JE-endemic areas living in proximity to irrigated rice phase 2 clinical trials and appears to be well tolerated 707 660 agro ecosystems will further rise in the years to come. (Monath et al., 2003). It is currently undergoing a 2- 708 661 Unfortunately and in contrast to malaria (Keiser year clinical study in Australia, to assess the duration 709 662 et al., in press), very few studies are presently available of immunity and to gain further knowledge on safety 710 663 that investigated the effect of an introduction or an ex- and immunogenicity (Jones, 2004). 711 664 pansion of rice-irrigation schemes on the incidence of It is important to recognize that strategies other than 712 665 JE. The few studies reviewed here clearly link JE inci- vaccination may play an important role for prevention 713 666 dence with rice agriculture. There are important gaps and control of JE, particular in rural areas, where vac- 714 667 in our understanding of the effects of different types of cination coverage are sometimes low or where there 715 668 water-related projects on the frequency and transmis- is no history of immunization against JE at all, as re- 716 669 sion dynamics of JE. In our view longitudinal studies cently documented for Northeast India (Phukan et al., 717 670 that track both a change in vector abundance and JE 2004). Vaccination of the far rural population is strate- 718 671 infectivity rates in humans, for example at large water gically difficult and costly, in particular as three doses 719 672 projects in the framework of health impact assessment are required to achieve adequate neutralizing antibody 720 673 studies, covering the entire course of a water resource levels. As demonstrated in China self-protection be- 721 674 development project, are warranted. haviour, such as sleeping under insecticide-treated nets 722 675 Clearly, vaccination is the current strategy of choice (ITNs) can significantly reduce the risk of infection 723 676 for prevention and control of JE outbreaks in Asia, (Luo et al., 1994). On the other hand, all JE patients of 724 677 and extensive government-supported vaccination cam- a recent outbreak in Northeast India reported that they 725 678 paigns will remain the mainstay of control. Formalin- had slept under a bed-net (Phukan et al., 2004). Con- 726 679 inactivated vaccines, based on wild type Nakayama sequently, irrigation schemes should be implemented 727 680 or Beijing-1 strains grown in adult mouse brain (e.g. and maintained in a way that adverse health effects to 728 681 JE-Vax), are licensed for immunization against JE by rice growers and residents of the area are minimized 729 682 several Asian countries. However, in order to achieve and social and economic development improved. 730 683 100% seroconversion, 2–3 primary doses and one We evaluated and discussed several environmental 731 684 booster dose after a year are required, with subsequent and biological vector control measures in rice fields: 732 685 boosts every 3–4 years (Pugachev et al., 2003). Minor Bacillus sphaericus and B. thuringiensis israelensis 733

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 15

734 were found to greatly reduce JE mosquito larvae. How- Amerasinghe, F.P., 1995. Mosquito vector ecology: a case study from 777 735 ever, it is currently not economical to apply these bac- Sri Lanka and some thoughts on research issues. In: Tropical 778 Diseases, Society and Environment, Proceedings from a WHO- 779 736 terial toxins in rice fields, as they are costly, labor- TDR/SAREC Research Seminar, SAREC Documentation: Con- 780 737 intensive and often only have a short duration of activ- ference Reports 1995, pp. 135–156. 781 738 ity. Few studies are available on invertebrates, fungi, Amerasinghe, F.P., 2003. Irrigation and mosquito-borne diseases. J. 782 739 nematodes or the neem cake powder in rice fields, Parasitol. 89, 14–22. 783 740 which allows assessing and quantification of their po- Amerasinghe, F.P., Ariyasena, T.G., 1991. Survey of adult 784 mosquitoes (Diptera: Culicidae) during irrigation development 785 741 tential as JE mosquito control agents. Reviewing the in the Mahaweli Project Sri Lanka. J. Med. Entomol. 28, 387– 786 742 literature has shown that in settings where the irriga- 393. 787 743 tion water in the rice fields can be managed, AWDI Angelon, K.A., Petranka, J.W., 2002. Chemicals of predatory 788 744 has considerable potential to reduce JE vector densi- mosquitofish (Gambusia affinis) influence selection of oviposi- 789 745 ties. A similar conclusion can be made for the use of tion site by Culex mosquitoes. J. Chem. Ecol. 28, 797–806. 790 Arunachalam, N., Samuel, P.P., Hiriyan, J., Thenmozhi, V., Gajanana, 791 746 larvivorous fish. An integrated vector management ap- A., 2004. Japanese encephalitis in Kerala, South India: can Man- 792 747 proach with AWDI and the use of larvivorous fish as sonia (Diptera: Culicidae) play a supplemental role in transmis- 793 748 its main components can reduce JE vector populations, sion? J. Med. Entomol. 41, 456–461. 794 749 and hence has the potential to reduce the transmission Balaraman, K., Balasubramanian, M., Jambulingam, P., 1983. Field 795 750 level and the burden of JE. It should be emphasized trial of Bacillus thuringiensis H-14 (VCRC B-17) against Culex 796 and Anopheles larvae. Ind. J. Med. Res. 77, 38–43. 797 751 that these intervention strategies must be tailored care- Barzaga, N.G., 1972–1985. A review of Japanese encephalitis cases 798 752 fully to a specific setting, which renders it difficult to in the Philippines Southeast Asian. J. Trop. Med. Public Health 799 753 generalize reported experiences and results. 20, 587–592. 800 754 Consequently, there is a pressing need to implement Bi, P., Tong, S., Donald, K., Parton, K.A., Ni, J., 2003. Climate 801 755 and monitor the performance of well-designed stud- variability and transmission of Japanese encephalitis in Eastern 802 China. Vector Borne Zoonotic Dis. 3, 111–115. 803 756 ies to further strengthen our understanding of the con- Bista, M.B., Banerjee, M.K., Shin, S.H., Tandan, J.B., Kim, M.H., 804 757 textual determinants of environmental and biological Sohn, Y.M., Ohrr, H.C., Tang, J.L., Halstead, S.B., 2001. Efficacy 805 758 control methods on vector abundance and clinical out- of single-dose SA 14-14-2 vaccine against Japanese encephalitis: 806 759 comes of JE in different ecological, epidemiological a case control study. Lancet 358, 791–795. 807 760 and socio-cultural settings. Broom, A.K., Smith, D.W., Hall, R.A., Johansen, C.A., Macken- 808 zie, J.S., 2003. Arbovirus infections. In: Cook, G., Zumla, A. 809 (Eds.), Manson’s Tropical Diseases, 21st ed. Saunders, London, 810 pp. 725–764. 811 761 Acknowledgements Consultative Group on International Agricultural Research, Techni- 812 cal Advisory Committee, and CGIAR Secretariat, 1998. Report 813 762 This work was part of a project entitled “Burden of the Fifth External Programme and Management Review of 814 763 of water-related vector-borne diseases: an analysis of International Rice Research Institute (IRRI), FAO, Rome, Italy. 815 Ding, D., Kilgore, P.E., Clemens, J.D., Wei, L., Zhi-Yi, X., 2003. 816 764 the fraction attributable to components of water re- Cost-effectiveness of routine immunization to control Japanese 817 765 sources development and management”, which par- encephalitis in Shanghai. Chin. Bull. World Health Organ. 81, 818 766 tially funded by the World Health Organization. J. 334–342. 819 767 Keiser (Project no. PMPDB-10622) and J. Utzinger Erlanger, T.E., Keiser, J., Castro, M., Bos, R., Singer, B.H., Tanner, 820 768 (Project no. PPOOB-102883) are grateful to the Swiss M., Utzinger, J., in press. Effect of water resource development 821 and management on lymphatic filariasis, and estimates of popu- 822 769 National Science Foundation for financial support. lations at risk. Am. J. Trop. Med. Hyg. 823 770 We thank Dr. B. Kay for carefully reviewing this Federici, B.A., Park, H.W., Bideshi, D.K., Wirth, M.C., Johnson, J.J., 824 771 manuscript. 2003. Recombinant bacteria for mosquito control. J. Exp. Biol. 825 206, 3877–3885. 826 Gajanana, A., Rajendran, R., Samuel, P.P., Thenmozhi, V., Tsai, T.F., 827

772 References Kimura-Kuroda, J., Reuben, R., 1997. Japanese encephalitis in 828 South Arcot district, Tamil Nadu India: a three-year longitudi- 829 nal study of vector abundance and infection frequency. J. Med. 830 773 Akiba, T., Osaka, K., Tang, S., Nakayama, M., Yamamoto, A., Ku- Entomol. 34, 651–659. 831 774 rane, I., Okabe, N., Umenai, T., 1997. Analysis of Japanese en- Geevarghese, G., Mishra, A.C., Jacob, P.G.,Bhat, H.R., 1994. Studies 832 775 cephalitis epidemic in Western Nepal in. Epidemiol. Infect. 126, on the mosquito vectors of Japanese encephalitis virus in Mandya 776 81–88.

ACTROP 1666 1–18

UNCORRECTED PROOF 16 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

833 District, Karnataka India. Southeast Asian J. Trop. Med. Public with pyrethroid insecticide and vaccination against Japanese en- 890 834 Health 25, 378–382. cephalitis. Trans. R. Soc. Trop. Med. Hyg. 88, 632–634. 891 835 Gujja, B., Perrin, M., 1999. A place for dams in the 21st Century. Mani, T.R., Rao, C.V., Rajendran, R., Devaputra, M., Prasanna, Y., 892 836 Discussion Paper, World Wildlife Fund, p. 108. Hanumaiah, Gajanana, A., Reuben, R., 1991a. Surveillance for 893 837 Halstead, S.B., Jacobson, J., 2003. Japanese encephalitis. Adv. Virus Japanese encephalitis in villages near Madurai, Tamil Nadu, In- 894 838 Res. 61, 103–138. dia. Trans. R. Soc. Trop. Med. Hyg. 85, 287–291. 895 839 Hanna, J.N., Ritchie, S.A., Phillips, D.A., Shield, J., Bailey, M.C., Mani, T.R., Reuben, R., Akiyama, J., 1991b. Field efficacy of “Mos- 896 840 Mackenzie, J.S., Poidinger, M., McCall, B.J., Mills, P.J., 1995. bar” mosquito repellent soap against vectors of bancroftian filar- 897 841 An outbreak of Japanese encephalitis in the Torres Strait Aus- iasis and Japanese encephalitis in Southern India. J. Am. Mosq. 898 842 tralia. Med. J. Aust. 165, 256–260. Control Assoc. 7, 565–568. 899 843 Heathcote, O.H., 1970. Japanese encephalitis in Sarawak: studies on Mathur, K.K., Rahman, S.J., Wattal, B.L., 1981. Integration of lar- 900 844 juvenile mosquito populations. Trans. R. Soc. Trop. Med. Hyg. vivorous fish and temephos for the control of Culex tritae- 901 845 64, 483–488. niorhynchus breeding. J. Commun. Dis. 13, 58–63. 902 846 Jones, T., 2004. A chimeric live attenuated vaccine against Japanese Microsoft Corporation, 2004. Encarta Encyclopedia, Redmond, 903 847 encephalitis. Expert Rev. Vaccines 3, 243–248. USA. 904 848 Kanojia, P.C.,Shetty, P.S.,Geevarghese, G., 2003. A long-term study Mishra, A.C., Jacob, P.G., Ramanujam, S., Bhat, H.R., Pavri, K.M., 905 849 on vector abundance & seasonal prevalence in relation to the 1983. Mosquito vectors of Japanese encephalitis epidemic in 906 850 occurrence of Japanese encephalitis in Gorakhpur district Uttar Mandya district (India). Indian J. Med. Res. 80, 377–389. 907 851 Pradesh. Indian J. Med. Res. 117, 104–110. Mitchell, C.J., Chen, P.S., Chapman, H.C., 1974. Exploratory tri- 908 852 Keiser, J., Castro, M., Maltese, M., Bos, R., Tanner, M., Singer, B.H., als utilizing a mermithid nematode as a control agent for Culex 909 853 Utzinger, J., in press. The effect of irrigation and large dams on in Taiwan. Taiwan Yi Xue Hui Za Zhi 73, 241–254. 910 854 the burden of malaria on global and regional scale. Am. J. Trop. Mittal, P.K., 2003. Biolarvicides in vector control: challenges and 911 855 Med. Hyg. prospects. J. Vector Borne Dis. 40, 20–32. 912 856 Keiser, J., Utzinger, J., Singer, B.H., 2002. The potential of inter- Mogi, M., 1984. Mosquito problems and their solution in relation to 913 857 mittent irrigation for increasing rice yields, lowering water con- paddy rice production. Protect. Ecol. 7, 219–240. 914 858 sumption, reducing methane emissions, and controlling malaria Mogi, M., 1988. Water management in rice cultivation and its rela- 915 859 in African rice fields. J. Am. Mosq. Control Assoc. 18, 329–340. tion to mosquito production in Japan. In: Vector Borne Disease 916 860 Kim, H.C., Kim, M.S., Yu, H.S., 1994. Biological control of vector Control in Humans Through Rice Agroecosystem Management. 917 861 mosquitoes by the use of fish predators Moroco oxycephalus and International Rice Research Institute in collaboration with the 918 862 Misgurnus anguillicaudatus in the laboratory and semi-field rice WHO/FAO/UNEP Panel of Experts, pp. 101–109. 919 863 paddy. Korean J. Entomol. 24, 269–284. Mogi, M., 1993. Effect of intermittent irrigation on mosquitoes 920 864 Kramer, V.,1984. Evaluation of Bacillus sphaericus & B. thuringien- (Diptera: Culicidae) and larvivorous predators in rice fields. J. 921 865 sis H-14 for mosquito control in rice fields. Indian J. Med. Res. Med. Entomol. 30, 309–319. 922 866 80, 642–648. Monath, T.P., Guirakhoo, F., Nichols, R., Yoksan, S., Schrader, R., 923 867 Lacey, L.A., Lacey, C.M., 1990. The medical importance of rice- Murphy, C., Blum, P., Woodward, S., McCarthy, K., Mathis, D., 924 868 land mosquitoes and their control using alternatives to chemical Johnson, C., Bedford, P., 2003. Chimeric live, attenuated vaccine 925 869 insecticides. J. Am. Mosq. Control Assoc. 2 (Suppl.), 1–93. against Japanese encephalitis (ChimeriVax-JE): phase 2 clini- 926 870 Lee, D.K., 1998. Effect of two rice culture methods on the seasonal cal trials for safety and immunogenicity, effect of vaccine dose 927 871 occurrence of mosquito larvae and other aquatic animals in rice and schedule, and memory response to challenge with inacti- 928 872 fields of Southwestern Korea. J. Vector Ecol. 23, 161–170. vated Japanese encephalitis antigen. J. Infect. Dis. 188, 1213– 929 873 Liao, S., Xu, B.Z., Chen, C.Y.,Liang, Z.P.,Zhang, H.F., 1991. Breed- 1230. 930 874 ing grass carp against mosquitoes in rice field. Zhongguo Ji Sheng Okuno, T., Tseng, P.T., Hsu, S.T., Huang, C.T., Kuo, C.C., 1975. 931 875 Chong Xue Yu Ji Sheng Chong Bing Za Zhi 9, 219–222. Japanese encephalitis surveillance in China (Province of Taiwan) 932 876 Liu, S.L., Hsu, Y.C.,1982. Effect of the fungus Coelomomyces indica during 1968–1971. I. Geographical and seasonal features of case 933 877 on the viability of Culex tritaeniorhynchus larvae. Kun Chong outbreaks. Jpn. J. Med. Sci. Biol. 28, 235–253. 934 878 Xue Bao 25, 409–412. Peiris, J.S., Amerasinghe, F.P., Arunagiri, C.K., Perera, L.P., 935 879 Lu-Baulin, 1988. Environmental management for the control of Karunaratne, S.H., Ratnayake, C.B., Kulatilaka, T.A., 936 880 ricefield-breeding mosquitoes in China. In: VectorBorne Disease Abeysinghe, M.R., 1993. Japanese encephalitis in Sri Lanka: 937 881 Control in Humans Through Rice Agroecosystem Management. comparison of vector and virus ecology in different agro-climatic 938 882 International Rice Research Institute in collaboration with the areas. Trans. R. Soc. Trop. Med. Hyg. 87, 541–548. 939 883 WHO/FAO/UNEP Panel of Experts, pp. 111–121. Peng, Y., Song, J., Tian, G., Xue, Q., Ge, F., Yang, J., Shi, Q., 940 884 Luo, D., Ying, H., Yao, R., Song, J., Wang, Z., 1995. Socio-economic 1998. Field evaluations of Romamomermis yunanensis (Nema- 941 885 status and micro-environmental factors in relation to the risk of toda: ) for control of Culicinae mosquitoes in China. 942 886 Japanese encephalitis: a case-control study. Southeast Asian J. Fundam. Appl. Nematol. 21, 227–232. 943 887 Trop. Med. Public Health 26, 276–279. Phukan, A.C., Borah, P.K., Mahanta, J., 2004. Japanese encephalitis 944 888 Luo, D., Zhang, K., Song, J., Yao, R., Huo, H., Liu, B., Li, Y., in Assam Northeast India. Southeast Asian. J. Trop. Med. Public 945 889 Wang, Z., 1994. The protective effect of bed nets impregnated Health 35, 618–622. 946

ACTROP 1666 1–18

UNCORRECTED PROOF J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx 17

947 Prasad, H., Prasad, R.N., Haq, S., 1993. Control of mosquito breeding Sundararaj, R., Reuben, R., 1991. Evaluation of a microgel droplet 1004 948 through Gambusia affinis in rice fields. Indian J. Malariol. 30, formulation of Bacillus sphaericus 1593 M (Biocide-S) for con- 1005 949 57–65. trol of mosquito larvae in rice fields in Southern India. J. Am. 1006 950 Pugachev, K.V.,Guirakhoo, F., Trent, D.W., Monath, T.P., 2003. Tra- Mosq. Control Assoc. 7, 556–559. 1007 951 ditional and novel approaches to flavivirus vaccines. Int. J. Para- Sunish, I.P., Reuben, R., 2001. Factors influencing the abundance 1008 952 sitol. 33, 567–582. of Japanese encephalitis vectors in ricefields in India-I. Abiotic 1009 953 Rajendran, R., Reuben, R., 1991. Evaluation of the water fern Azolla Med. Vet. Entomol. 15, 381–392. 1010 954 microphylla for mosquito population management in the rice- Sunish, I.P., Reuben, R., 2002. Factors influencing the abundance 1011 955 land agro-ecosystem of South India. Med. Vet. Entomol. 5, of Japanese encephalitis vectors in ricefields in India-II Biotic. 1012 956 299–310. Med. Vet. Entomol. 16, 1–9. 1013 957 Rajendran, R., Reuben, R., Purushothaman, S., Veerapatran, R., Suroso, T., 1989. Studies on Japanese encephalitis vectors in In- 1014 958 1995. Prospects and problems of intermittent irrigation for con- donesia. Southeast Asian J. Trop. Med. Public Health 20, 627– 1015 959 trol of vector breeding in rice fields in Southern India. Ann. Trop. 628. 1016 960 Med. Parasitol. 89, 541–549. Suzuki, T., 1967. Bibliography on Culex tritaeniorhynchus in Japan 1017 961 Rao, D.R., Reuben, R., Nagasampagi, B.A., 1995. Development of 1945–1966. Jpn. Environ. Sanit Cent. 1018 962 combined use of neem (Azadirachta indica) and water manage- Takagi, M., Sugiyama, A., Maruyama, K., 1995. Effect of rice cultur- 1019 963 ment for the control of culicine mosquitoes in rice fields. Med. ing practices on seasonal occurrence of Culex tritaeniorhynchus 1020 964 Vet. Entomol. 9, 25–33. (Diptera: Culicidae) immatures in three different types of rice- 1021 965 Rao, D.R., Reuben, R., Venugopal, M.S., Nagasampagi, B.A., growing areas in Central Japan. J. Med. Entomol. 32, 112– 1022 966 Schmutterer, H., 1992. Evaluation of neem, Azadirachta indica, 118. 1023 967 with and without water management, for the control of culicine Tsai, T.F., 2000. New initiatives for the control of Japanese encephali- 1024 968 mosquito larvae in rice-fields. Med. Vet. Entomol. 6, 318–324. tis by vaccination: minutes of a WHO/CVI meeting, Bangkok, 1025 969 Reuben, R., Thenmozhi, V., Samuel, P.P., Gajanana, A., Mani, T.R., Thailand, 13–15 October 1998. Vaccine 18 (Suppl. 2), 1–25. 1026 970 1992. Mosquito blood feeding patterns as a factor in the epidemi- United Nations, 2004. World Urbanization Prospects: The 2003 Re- 1027 971 ology of Japanese encephalitis in Southern India. Am. J. Trop. visions, Population Division Department of Economics and So- 1028 972 Med. Hyg. 46, 654–663. cial Affair of the United Nations, New York. 1029 973 Rhee, H.I., Shim, J.C., Kim, C.L., Lee, W.J., 1983. Small scale UNO, 2003. Gates foundation grant to help fight Japanese Encephali- 1030 974 field trial with Bacillus thuringiensis israelensis H-14 for control tis. United Nations Wire Service, Washington, http://www. 1031 975 of the vector mosquito (Culex tritaeniorhynchus) larvae in rice unwire.org/UNWire/20031209/449 11099.asp. 1032 976 fields. Korean J. Entomol. 13, 39–46. van der Hoek, W., Sakthivadivel, R., Silver, J.B., Konradsen, F., 2001. 1033 977 Sehgal, A., Dutta, A.K., 2003. Changing perspectives in Japanese Alternate wet/dry irrigation in rice cultivation: a practical way 1034 978 encephalitis in India. Trop. Doct. 33, 131–134. to save water and control malaria and Japanese encephalitis? 1035 979 Self, L.S., Shin, H.K., Kim, K.H., Lee, K.W., Chow, C.Y., Hong, International Water Management Institute; Colombo, Sri Lanka. 1036 980 H.K., 1973. Ecological studies on Culex tritaeniorhynchus as a Research Report 47. 1037 981 vector of Japanese encephalitis. Bull. World Health Organ. 49, Vijayarani, H., Gajanana, A., 2000. Low rate of Japanese encephalitis 1038 982 41–47. infection in rural children in Thanjavur district (Tamil Nadu), an 1039 983 Solomon, T., 2004. Flavivirus encephalitis. N. Engl. J. Med. 351, area with extensive paddy cultivation. Indian J. Med. Res. 111, 1040 984 370–378. 212–214. 1041 985 Solomon, T., Dung, N.M., Kneen, R., Gainsborough, M., Vaughn, Wada, Y., 1988. Strategies for control of Japanese encephalitis in rice 1042 986 D.W., Khanh, V.T., 2000. Japanese encephalitis. J. Neurol. Neu- production systems in developing countries. In: Vector Borne 1043 987 rosurg. Psychiatry 68, 405–415. Disease Control in Humans Through Rice Agroecosystem Man- 1044 988 Solomon, T., Dung, N.M., Wills, B., Kneen, R., Gainsborough, M., agement. International Rice Research Institute in collaboration 1045 989 Diet, T.V., Thuy, T.T., Loan, H.T., Khanh, V.C., Vaughn, D.W., with the WHO/FAO/UNEP Panel of Experts, pp. 153–160. 1046 990 White, N.J., Farrar, J.J., 2003. Interferon alfa-2a in Japanese en- WHO, 1997. Vector Control—Methods for Use by Individuals 1047 991 cephalitis: a randomised double-blind placebo-controlled trial. and Communities. World Health Organization, WHO, Geneva, 1048 992 Lancet 361, 821–826. http://www.who.int/docstore/water sanitation health/vectcontrol 1049 993 Somboon, P., Choochote, W., Khamboonruang, C., Keha, P., Suwan- /ch08.htm. 1050 994 phanit, P., Sukontasan, K., Chaivong, P., 1989. Studies on the WHO, 2004. The World Health Report 2004, The World Health Re- 1051 995 Japanese encephalitis vectors in Amphoe Muang, Chiang Mai port 2004, Geneva, World Health Organisation. 1052 996 Northern Thailand. Southeast Asian J. Trop. Med. Public Health Wu, N., Liao, G.H., Li, D.F., Luo, Y.L., Zhong, G.M., 1991. The 1053 997 20, 9–17. advantages of mosquito biocontrol by stocking edible fish in 1054 998 Song, J., Peng, Y., 1996. Field trials of combined use of two species of rice paddies. Southeast Asian J. Trop. Med. Public Health 22, 1055 999 mermithid nematodes to control Anopheles and Culex breeding 436–442. 1056 1000 in China. Indian J. Malariol. 33, 161–165. Yadav, R.S., Sharma, V.P., Upadhyay, A.K., 1997. Field trial of Bacil- 1057 1001 Sucharit, S., Surathin, K., Shrestha, S.R., 1989. Vectors of Japanese lus sphaericus strain B-101 (serotype H5a, 5b) against filariasis 1058 1002 encephalitis virus (JEV): species complexes of the vectors. and Japanese encephalitis vectors in India. J. Am. Mosq. Control 1059 1003 Southeast Asian J. Trop. Med. Public Health 20, 611–621. Assoc. 13, 158–163. 1060

ACTROP 1666 1–18

UNCORRECTED PROOF 18 J. Keiser et al. / Acta Tropica xxx (2005) xxx–xxx

1061 Yu, H.S., Kim, H.C., 1993. Integrated control of encephalitis vector breeding habitats of rice paddies and stream seepage in South 1067 1062 (Culex tritaeniorhynchus) with native fishes (Aplocheilus latipes Korea. Korean J. Entomol. 12, 61–68. 1068 1063 and Aphyocypris chinensis) and Bacillus thuringiensis (H-14) Zahiri, N.S., Federici, B.A., Mulla, M.S., 2004. Laboratory and 1069 1064 in marshes in Jindo Island of Korea. Korean J. Entomol. 23, simulated field evaluation of a new recombinant of Bacillus 1070 1065 221–230. thuringiensis ssp. israelensis and Bacillus sphaericus against 1071 1066 Yu, H.S., Lee, D.K., Lee, W.J., 1982. Mosquito control by the re- Culex mosquito larvae (Diptera: Culicidae). J. Med. Entomol. 1072 lease of fish predator Aphyocypris chinensis in natural mosquito 41, 423–429. 1073

ACTROP 1666 1–18

UNCORRECTED PROOF