446 Jounulr, oF THEArupnlcnu Mosqurro CoNrnol AssocrlrroN Vor,.6, No. 3

FACTORSAFFECTING THE HUMAN.FEEDINGBEHAVIOR OF ANOPHELINE MOSQUITOESIN EGYPTIAN OASES

MOHAMED A. KENAWY.I JOHN C. BEIER.z3CHARLES M. ASIAGO'eNo SHERIF EL SAID'

ABSTRACT. Blood meals were tested by a direct enzyme-linkedimmunosorbent assay(ELISA) for 424 Anophel,essergentii and for 63 An. multicolor collected in Siwa, Farafra and Bahariya oases in the of Eg5pt. Both specieswere highly zoophilic. Human blood-feedingby An. sergentii was lesscommon in Bahariya (2.3Vo)and Farafra (1.3%)than in Siwa (I5.37o).A likely explanationis that large domestic animals are held at night inside houses in Bahariya and in Farafra whereas in Siwa, animals are usually housedoutdoors in sheds.These patterns of An. sergentii human-feedingbehavior may contribute to the persistenceof low-level Plnsmodiurn uiuor transmission in Siwa in contrast to negligible or no transmission in Bahariya and Farafra.

INTRODUCTION sistenceof P. uiuax in Siwa but not in Bahariya and Farafra is interesting becauseresidents in Zoophilic feeding behavior by anopheline ma- theseecologically similar oasesemploy different Iaria vectors representsan important regulatory methods for holding domestic animals such as mechanism in malaria transmission. In , cows,donkeys, goats and sheep.In Bahariya and (Theo- the malaria vectors Anophelessergentii Farafra, Iarge domesticanimals are usually kept bald) and An. pharoensis Theobald, and a sus- inside housesat night whereasin Siwa, animals pectedvector, An. rnulticolor Cambouliu, feed to are kept away from housesin sheds. a large extent on domestic mammals. This has This study examines the possibility that tra- (Kenawy been observedin Governorate ditional animal holding practicesmay affect the (Beier et al. 1987),in Governorate et al. human-feeding behavior of anophelines in the (Hurlbut 1987), in the Delta and Weitz oasesof Siwa, Bahariya and Farafra. From these (Barber 1956),and in the Western Desert oases oases,blood meals were identified for An. ser- and Rice 1937,Kenawy et al. 1986b).However, gentii and,An. multicolar collected from inside factors affecting the degreeof zoophilic feeding houses,from animal shedsand from mixed sites have not been investigated.The geographicvar- (containing animal rooms inside houses). iability in anopheline feeding patterns and so- ciocultural differencesthroughout Eglpt suggest that animal holding practices may be a key MATERIALS AND METHODS factor affecting malaria transmission. Study areas: The study was conducted in 3 Malaria in the oasesof the Western Desert of oasesin the Western Desert of Egypt (Fig. 1). Egypt (Siwa, Bahariya, Farafra, Dakhla and SiwaOasis (29" 07' -29' 21'N,25' 16'-26" 07'E) is Kharga) is currently being evaluated to deter- 82 km long from east to west and lies an average mine appropriate measuresfor control. Histori- of 18 m below sea level. Siwa has over 11,000 these oaseswere endemic for Plasrnodiurn cally, residentsdistributed in 5 districts and about 546 uiuax,P. and P. malarine (Halawani falciparum ha are under cultivation. Bahariya (27'45'- and Shawarby 1957).Today, P. uiuax alone per- 28'21'N, 28'30'-29"15'E) lies 360km southeast sists at rates usually <5Voit the villages of Siwa of Cairo, includes 9 oases, and has >30'000 Oasis, and in El Gara, a small oasis near Siwa residents. Farafra oasis (26'15'-27'30'N, (Kenawy et al. 1986a).In contrast, Bahariya 26"45'-29"00'E) Iies 170 km southwest of Ba- to malaria-free. Hassan and Farafra appear be hariya with 3,000 residents in 5 small villages (1983) persons et al. showed that 2,285 from (ouicensus,March 1988)and 250 ha are under Bahariya were slide negative, and 9,066 blood cultivation. The water supply in the 3 oasesis samples from Bahariya and 734 from smear drawn mainly from springs and wells. Due to were negative during surveysfrom 1979 Farafra water infiItration from wells and improper (Kenawy, unpublished data). The per- to 1989 drainage, Iarval development areas for mosqui- toes are common in seepageareas and in irri- l University Research and Training gation canals.Numerous salt lakes exist in Siwa Center on Vectors of Diseases,Ain ShamsUniversity, oasis. The main crops are dates, olives, barley, ,, Arab Republic of Egypt. wheat and various fruits and vegetablesgrown 2 and the The Kenya Medical Research Institute in gardens around springs. Clover and alfalfa Nairobi, Kenya. U.S. Army Medical ResearchUnit, grown in Farafra and rice is cultivated in 3 address:Department of Immunology and are Present parts Housesare constructed Infectious Diseases,The Johns Hopkins University' some of Bahariya. ceilings supported School of Hygiene and Public Health, 615 N. Wolfe of mu-dwith thick walls and Street,Baltimore, MD 21205. by date-palm logs. This type of construction SEPTEMBER1990 AuopnnuNp Br,ooo-Fnpuuc rn EcyprreNOe,sns

Fig. 1. Map showing the location of Siwa, Farafra and Bahariya oasesin the Western Desert of Egypt. struction provides insulation from the desert which animal rooms were inside the houses. heat. The main difference in housing between Mosquitoes were placed in small paper cartons the 3 oases is that animal sheds in Siwa are and transported to the laboratory in Cairo for usually separatedfrom houses,whereas in Far- processing. Mosquitoes were identified, placed afra and Bahariya, most of the houses include in 1.5-ml cappedplastic vials (up to 2l/vial) by one or more animal rooms where animals are speciesand site of collection, and kept desic- held during the night. cated at room temperature until tested. For the Mosquito collectinnand handling: Mosquitoes Siwa collection in 1986,samples were frozen at were collected in 4 Siwa villages (Meshindit, -70"C and then desiccated1 week before test- Bahi El Din, Aghurmi and Abu Shurufl during ing. April and May 1986 and in July 1982, in 2 _ Preparatinn of m.osquitoesfor ELISA testin4: Farafra villages (Qasr Farafra and El Sheikh Each blood-fed mosquito was cut transversefu Marzouk) during November 1g8? and March with a scalpelblade at thejuncture ofthe thorai 1988,and in 4 Bahariya villages(Mandisha, El and abdomen.Abdomens were preparedindivid- Qubala, and El Qasr) during November ually for testing by trituration in 0.2-ml glass and December1987. microtissuegrinders to which b0-pl0.01 M phos- Indoor resting sites were surveyed by space- phate bufferedsaline (PBS), pH7.4 was added. spraying of pyrethroid insecticide (0.2% niopy- Of this, a 25-p.l aliquot was then diluted with buthrin in kerosene)with a hand sprayer. A l- PBS (1:50)and frozenat -20"C. m2_white sheet supportedby 2 sticki was spread Blood rneal identification: Mosquito blood and movedaround the walls and under furniture meals were identified by a direct enzyme linked by one collector while another sprayed the in- immunosorbent assay (ELISA) (Beier et al. secticide. (22 Collection sites included S0 houses 1988) at the U.S. Army Medical ResearchUnit, in Siwa and 8 in Farafra), 46 animal shedsigt Kenya Medical Research Institute, Nairobi. in Siwa and 15 in Farafra) and bg mixed sites Kenya. Each mosquito sample was tested (3 for in Siwa, 26 in Farafra and B0 in Bahariya) in blood mealsof t hosts: human, cow, goat, horse, 448 JouRNer,oF THE ArurRrceN Mosourro CoNrRol AssocrlrroN Vol.6, No.3 dog, cat, rabbit, rat and chicken. Negative sam- RESULTS ples were retested for all hosts. Positive and negative controls and methods for determining A total of 544An. sergentii and 71 An. multi- positivity were describedpreviously (Beier et al. colorfemales were collectedin Siwa, Farafra and 1988). Bahariya oases.Anophelcs sergentii represented Host census:Humans and domestic animals 98.2%of the total anophelinescollected in Siwa were counted in Siwa whereas in Farafra and (4 : 164),70.7% in Farafra(n:232') and 100% Bahariya oasescensus data were basedon infor- in Bahariya(n:219). mation from local councils, veterinary stations Blood meals ftom 424 An. sergentii were and malaria stations. A censusin May 1986was tested by ELISA (Table 1). Positive reactions conductedin all houses(n:62) and their ad- were obtained fot 80.4%, 66.7% and 76.3% of jacent animal sheds in Bahi EI Din, one of 4 the samplestested in Siwa, Farafra and Baha- villages sampled in Siwa Oasis. Potential hosts riya, respectively.Annpheles sergentii collected included:humans (450), cows (81), sheepand inside houses in Siwa fed mainly on humans goats (719), donkeys (80), rabbits (142), geese (42.9%),and in animal sheds,most of the feed- (50),ducks (43) and chickens(377). In Farafra, ingswere on donkeys(35.lVo) and cows(31.1%). the estimated numbers of hosts were 3,000 hu- In Farafra, only LSVo(1/78) ofthe An. sergentii mans,665 cowsand buffaloes,L,740 sheep and contained human blood. This speciesfed pre- goats,311 donkeysand horsesand 370 camels. dominantly on donkeys (70%) inside houses, In Bahariya, the estimated numbers of hosts in and on cows in animal sheds (70.8%) and in 3 areas of mosquito collections were 13,500hu- mixed sites (59.1%).In Bahariya, 55.0% and mans,2,650 cows and buffaloes,5,550 goats and 35.7%of the 129 blood meals were identified as sheep, 2,500 donkeys and 600 camels. Other donkey and cow, respectively; only 2.3% con- common animals included cats, dogs, rabbits, tained human blood. Overall, the proportion of ducks, geese,chickens, wild birds, rodents, bats An. sergentiifeeding on humans in Siwa (L5.3Vo, and reptiles but these were not counted. n : 111)was significantly higher than in Farafra Forage rotios.' Forage ratios were calculated (x' : 9.3, df : 1, P < 0.005)or in Bahariya (x' for An. sergentii and,An. multicolar in Siwa and : 14.'1.,df : 1, P < 0.005). Farafra oases to provide a standard index of Blood meals from 63 An. multicolnr were host selection.This ratio was calculated for hu- tested by ELISA (Table 2). In Siwa, 3 fed mos- mans and domesticanimals as the percentageof quitoes contained human, goat and horse blood. positive blood meals divlded by the percentage In Farafra, a low percentageof human feedings of these hosts. Ratios f 1 indicate preferences, (2.9%\ was observed and most of the feedings ratios ( 1 indicate avoidance, and ratios ap- were on cows. proaching 1 indicate no preferenceor avoidance Mixed blood meals were detected for 6% (L9/ (Hesset al. 1968). 318)of An. sergentiiand for 10.8%(4/37) ot An.

Table 1. Host blood meal sourcesfor Anophelessergentii from inside houses,animal shedsor mixed sites containing animal rooms inside housesin Siwa, Farafra and Bahariya oases,Eg;pt.

% bloodmeal hosts*

Sites/ Total no. No. (%) Goat/ Donkey/ Mixed habitats tested identified Human Cow sheep horse mealsi+ Siwa House 28 (82.4) 42.9 3.6 r7.9 17.9 L7.9 Shed 95 74 (37.9) 4.1 31.1 2r.6 35.1 8.1 Mixed 9 I (100.0) 22.2 22.2 22.2 22.2 11.1 TotaI 138 111 (80.4) 15.3 23.4 20.7 29.7 10.8a Farafra House t2 10 (83.3) 0.0 30.0 0.0 70.0 0.0 Shed 28 24 (85.7) 0.0 70.8 16.7 8.3 4.2 Mixed 77 44 (57.r) 2.3 59.1 4.5 29.5 4.5 Total lr7 78 (66.7) 1.3 59.0 7.7 28.2 3.8b Bahariya Mixed 169 r29 (76.3) 2.3 35.7 3.9 55.0 3.1c * No reactions to rat, cat, dog and chicken antisera, +* Mixed meals included: (a) cow * donkey (n : l), goat * donkey (z : 3), donkey + rabbit (n = 1), human * donkey (n = 6) and human + rabbit (n : 1), (b) cow * donkey (n: 1), human * cow (n: 1), and human * rabbit (n : 1), and (c) cow-t donkey (a : 2) and human + cow (n : 2). SEPTEMBER1990 ANopnslrNo BLooD-FEEDTNGrN EcvprreN Onsps 449

Table 2. Host blood meal sources fot Arnplwlcs multicolar fuom inside houses, animal sheds or mixed sites containing animal rooms inside houses in Siwa and in Farafra oases, Egypt.

Sites/ Total no. No. (%) Goat/ Donkey/ Mixed habitats tested identified Human Cow sheep horse meals** Siwa House 1 I (100.0) 0.0 0.0 0.0 100.0 0.0 Shed 2 2 (100.0) 50.0 0.0 50.0 0.0 0.0 Mixed 0 TotaI 3 (100.0) oo.D 0.0 tr3 0.0 Farafra House 5 I (33.3) 0.0 0.0 0.0 100.0 0.0 Shed 56 JJ (58.e) 3.0 57.6 18.2 9.1 12.L Mixed 1 0 (0.0) Total 60 34 (56.7) 2.9 55.9 r7.6 11.8 11.8 * No reactions to rat, cat, dog and chicken antisera. ** Mixed mealsincluded cow * donkey(n:2), cow * rabbit (n : 1) and cow * donkey+ goat (n : 1).

multicolor. These were mostly feedingson more afra and Bahariya, respectively, 2 areas where than one type of large mammal. InB An. sergen- animals are usually kept inside housesat night. tii and in one An. rnulticolar, rabbit blood was Anophclcs sergentii forage ratios for humans detected in conjunction with blood from large were 11 times lower than the highest nonhuman mammals, including human. A triple feeding on forage ratio (equines)in Siwa, and 174 and 134 cow, donkey and goat was detected for one An. times lower than forage ratios for equines in multicolor in Farafra. Farafra and Bahariya, respectively. Thus, the Forageratios were calculated for An. sergentii inherent zoophilic feeding behavior of An. ser- and,An. multicol,orfrom Siwa, Farafra and Ba- gentii appearsto be amplified when large do- hariya (Table 3). Around 93.5% (332/gS5)of the mestic animals are housed inside houses at feedings were on humans or on large domestic night. mammals (excluding mixed meals); thus, forage Both An. sergentii and.An. rnulticolor may feed ratios were calculatedonly for humans, bovines, on more than one type of host per gonotrophic ovines and equines. Highest forage ratios were cycle as evident by mixed blood meals in G.0% observed for equines and bovines. Anopheles (19/318)of the An. sergentiiand in 10.870(4/ sergentii forage ratios for humans were low in 37) of An. multicolar blood meals. Host-seekins (0.45) Siwa but were 15 times lower in Farafra behavior after blood meal interruption may bi (0.03) and 11 times lower in Bahariya (0.04). related to animal holding practices.Mixed blood Low forageratios for humans were also observed meals for An. sergentii collected in houses, for An. multicolor in Farafra (0.06). sheds,or in mixed structuresin Siwa (l0.8Vo) weremore commonthan in Farafra (3.8%)or in DISCUSSION Bahariya (3.L%).The low percentageof mixed meals in Farafra and Bahariya, where animals Arwph,elessergentii and.An. multirolnr from B are kept inside houses,may simply indicate that Western Desert oases fed primarily on large An. sergentiihas a high degreeoffeeding success domestichosts. Such zoophilic feeding behavior on Iarge mammals. has been reported previously for these 2 species The percentageof negative ELISA reactions in Siwa (Barber and Rice 1937, Kenawv et al. was relatively high (19.6 to 33.3% fot An. ser- 1986b) and in areas outside Egypt iBtrce- gentii and 43.3%for multi,col.or).Using the same Chwatt and Gockel 1960). Over 95% ofthe An. assays,Beier et al. (1988) identified 94.6% of sergentii in Bahariya and Farafra, andAn. mul- blood meals from mosquitoescollected in west- ticolor in Farafra, fed on large domesticanimals. ern Kenya. The non-reacting samplescould in- This is the first evidence that anophelines in dicate feedings on hosts which were not tested Bahariya and in Farafra feed to a greater extent but this is unlikely; precipitin tests identified on large domestic animals than in Siwa. 700% of 497 blood meals from Siwa using anti- Patterns of human-feeding by An. sergentii sera to human and domestic hosts (Kenawv et may be affectedstrongly by animal holding prac- al. 1986b). A more likely explanation is ihat tices. In Siwa, where animals are usually kept handlingprocedures(e.g., drying and storagefor away from houses, 153% of the feedings were up to 2 years) degradedhost IgG in blood meals. on humans.This was 11.8and 6.7 times higher The differences in host-feeding patterns for than the proportion of human feedings in Far- An. sergentii in the 3 Western Desert oasesmav 450 Jounuu oF THEAupRrclr.t Moseurro CoNrRor,AssocrlrroN VoL.6,No.3

have important implications for the low-level maintenance of P. uiuax transmission. The ex- a Sror$ u oNic6 tremely low levels of human-feeding(<3.0%) by k ctd;cjro .6 the main vector speciesin Farafra and in Ba- a .c 90 hariya are consistent with recent parasite sur- s veys which indicate a complete interruption of q transmissionin theseoases (Hassan et al. 1983: N\Om\O !g ca ro Kenawy et al., unpublished data). The higher B proportions of human-feedings in Siwa (up to o 42.97oinside houses)appears sufficient to main- tain low rates of P. uiuor transmission (Kenawy o @Oo)co roic\o et al. 1986a). Ultimately, vector infectivity in o roi6ii vq the oasisenvironment dependsupon susceptible

ro(oroLo vectors feeding on infective human hosts during d z X m6troN o the first or the secondgonotrophic cycle. Based upon limited sampling, the probability of hu- s?9i man-feeding by An. sergentii appears to be at az (O6l@O) Ieast5 times more likely in Siwa than in Farafra I qoqqn O.+ o 6l or in Bahariya. In the absenceof other obvious ht social or ecologicaldifferences among the oases, d6 practices a major 6@ animal holding appear to be ;o o- qeqq factor affecting vector-feedingpatterns. NrOrd z> roii

qo 6H k^ 6Y ACKNOWLEDGMENTS ro 6l qqq6'lc6 0) iE onon grateful Ministry of ur_ We are to the Egyptian ;;: o Health, to A. Merdan, Director of the Ain Shams 90 Center, to M. Mugambi, former Director of the

oa cooF-N Kenya Medical ResearchInstitute (KEMRI), to lio 'ioib.-d { !F 6N Davy K. Koech, former Director of Biomedical SciencesResearch Center (KEMRI), and to 90 C. R. Roberts, Commander,U.S. Army Medical ResearchUnit, Nairobi, Kenya, for facilitating .i 9? ro(0s{ pro- NiO6 this research.Expert field assistancewas roic6v t> vA vided by Zakaria Morsy, Mohamed Abdel Rah- '!ri o@n'* man and Ibrahim Botros from the Ain Shams RO z c6i Center, Mohamed El Sagher from Farafra ma- tr> laria unit, and staff of Siwa and Bahariya ma- a laria units, Ministry of Health, Egypt. Technical :ot Lo$@ro nqo?o? assistancewas provided by the following from st the Walter Reed Project of KEMRI: John Ka- 6a q) manz4 Juma Makasa, Josephat Mwangi, Mi- o P o a o chael Ouko, Lucy Wanjiru and Rose Ongwang. j cr)$rr This study was supported by the Regional I rjc.io6; "Epidemiology .oi iNNN o Project entitled: and Control of k Arthropod-Borne Diseases in Egypt-Nol AI

X 22667"between the Researchand Training Cen- * o a ter on Vectors of Diseases.Ain Shams Univer- cardio 5; 60 c9.+ (9 sity, Abbassia, Cairo, Egypt, and the National a Institute of Allergy and Infectious Diseases OHOrO d utcoi@ (NIAID), National Institutes of Health (NIH)' , $r o by the U.S. Army u Bethesda, Maryland, and Medical Research Unit, Nairobi, Kenya. The Eo o views of the authors do not purport to reflect EtrOg o E.'.8'E m the position of the U.S. Department of the o PO>6 Army, the Department of Defense,or the gov- F ernments of Egypt or Kenya. Citation of trade names in this report does not constitute an SEPTEMBER1990 ANopHELINEBlooo-Fnnoruc nq Ecvptrlx Oasns

official endorsement or approval of the use of Eg)"t. J. Egypt. Med. Assoc.40:753-792. such items. Hassan,2., M. Gebril, M. Kenawy, F. Mahmoud, B. El Sawaf,S. El Said and L. EI-Okbi. 1983.Parasi- REFERENCES CITED tological study of malaria in Egypt. J. Egypt. Public Health Assoc.(Spec. Issue 3 & 4) 58:242-272. Barber, M. A. and J. B. Rice. 1937.A surveyof malaria Hess, A. D., R. O. Hayes and C. H. Tempelis. 1968. in Egypt. Am. J. Trop. Med. 17:413-436. The use of forage ratio technique in mosquito host Beier,J. C., J. H. Zimmerman,M. A. Kenawy, S. El preferencestudies. Mosq. News 28:386-389. Said and M. M. Abbassy.1987. Host-feeding pat- Hurlbut, H. S. and B. Weitz. 1956.Some observations terns of the mosquito community (Diptera: Culici- on the bionomics of the common mosquitoesof the dae) in two Faiyum Governoratevillages, Eg)"t. J. .Am. J. Trop. Med. Hyg. 5:901-908. Med. Entomol. 24:28-34. Kenawy, M. A., J. C. Beier and S. El Said. 1986a.First Beier, J. C., P. V. Perkins, R. A. Wirtz, J. Koros, D. record of malaria and associatedArnph.eles in El Diggs, T. P. Gargan and D. K. Koech. 1988.Blood Gara Oasis, Egypt. J. Am. Mosq. Control Assoc. meal identification by direct enzyme-linked immu- 2:101-103. nosorbentassay (ELISA), tested,onAnopheles (Dip- Kenawy,M. A., J. H. Zimmerman,J. C. Beier, S. EI tera: Culicidae) in Kenya. J. Med. Entomol. 25:g- Said and M. M. Abbassy.1986b. Host-feeding pat- 16. terns of Anophelessergentii and,An.multirolor (Dip- Bruce-Chwatt, L. J. and C. W. Gockel. 1960.A studv tera: Culicidae) in Siwa and El Gara oases,Egypt. of blood-feedingpatterns of Annphelesmosquitoes J. Med. Entomol. 23:576-577. through precipitin tests. Results of collaborative Kenawy,M. A., J. C. Beier, J. H. Zimmerman,S. EI work for the period of 1955-59and their application Said and M. M. Abbassy. 1987. Host-feeding pat- to malaria eradicationprogrammes. Bull. W.H.O terns of the mosquito community (Diptera: Culici- 22:685-720. dae) in ,Egypt. J. Med. Ento- Halawani, A. and A. A. Shawarby. 1957. Malaria in mol. 24:35-39.