Mosquito Vector Abundance Immediately Before and After Tropical Storms Alma and Arthur, Northern Belize, 2008

Investigación original / Original research

Mosquito vector abundance immediately before and after Tropical Storms Alma and Arthur, northern Belize, 2008

Meredith G. Morrow,1 Richard N. Johnson,1 Jorge Polanco,2 and David M. Claborn 3

Suggested citation Morrow MG, Johnson RN, Polanco J, Claborn DM. Mosquito vector abundance immediately before and after Tropical Storms Alma and Arthur, northern Belize, 2008. Rev Panam Salud Publica. 2010; 28(1):19–24.

ABSTRACT Objective. To monitor adult mosquito abundance in northern Belize before/after the first tropical storm of the wet season to estimate the time required for development/recovery of potential vector populations; determine which species predominate post-storm; and compare the effectiveness of two types of mosquito traps—octenol-baited Mosquito Magnets® and U.S. Centers for Disease Control and Prevention (CDC) light traps (with/without octenol). Methods. Field experiments were conducted in Orange Walk Town, Belize, 21 May to 3 June 2008. Incidence rate ratios and exact binomial 95% confidence intervals were reported and trap-nights calculated to compare species abundance pre- and post-storm as well as trap- type effectiveness. Results. Twice as many species and three times more Anopheles spp. were trapped pre- storm versus post-storm. However, greater numbers of Aedes taeniorhynchus and Culex (Culex) spp. were trapped post-storm. Mosquito Magnets® were consistently more effective than the CDC traps, obtaining twice as many Anopheles spp. and four times as many culicine species as the octenol-baited version (which collected 14 times more mosquitoes overall and 3.5 times more culicine species than the unbaited version). The unbaited CDC trap did not trap any Anopheles spp. during the study period. Conclusions. Results indicated octenol is an effective attractant for An. crucians in northern Belize; malaria risk in Belize declines immediately post-storm (i.e., mosquito abundance drops); and arboviral risk associated with the rapid increase in culicine mosquitoes post-storm may represent a greater public health threat than malaria (although further research and active disease surveillance is necessary to validate this hypothesis).

Key words Insect vectors; mosquito control; vector control; malaria; tropical storm; Belize.

Tropical storms and their effects on important factors in disease control and useful to disaster response planners who vector mosquito population dynamics are therefore should be considered in public must decide what disease control mea- health preparedness and disaster re- sures should be implemented and when sponse. Data generated by studies of they should begin (2–4). 1 Department of Preventive Medicine and Biomet- mosquito population dynamics in areas This study attempted to record vector rics, Uniformed Services University of the Health that receive significant rainfall during species abundance in northern Belize Sciences (USUHS), Bethesda, Maryland, United storm seasons can be used to estimate immediately before and after Tropical States of America. Send correspondence to: Mere- dith G. Morrow, [email protected] which species will predominate after Storms Alma and Arthur, which occurred 2 Ministry of Health, Belmopan, Belize. tropical weather events and how quickly simultaneously at the end of the 2008 Be- 3 Master of Public Health Program, Missouri State University, Springfield, Missouri, United States of potential vector populations may develop lizean dry season. Tropical Storm Alma America. and recover (1). These estimates can be began on 27 May 2008 as an area of low

Rev Panam Salud Publica 28(1), 2010 19 Original research Morrow et al. • Mosquito abundance pre- and post-tropical storms in Belize pressure that strengthened into a tropical Control and Prevention (CDC) light trap Octenol-baited Mosquito Magnets® depression off the coast of Nicaragua. It (with and without octenol attractant). versus octenol-baited CDC trap. Along was designated as the first tropical storm with the tents used to house the four of the season on 29 May 2008 and reached MATERIALS AND METHODS octenol-baited Mosquito Magnet® traps, peak winds of 104.6 km/h just before an additional tent was set up to house making landfall on the northwestern Mosquito abundances pre- and one CDC light trap, which was baited coast of Nicaragua near León (5, 6). Form- post-tropical storms with the same 1 600-mg octenol cartridge ing at 86.5°W, Alma developed further supplied with each Mosquito Magnet® east than any other Pacific tropical cy- Field experiments were conducted on and run for the same daily sampling pe- clone on record and was also the first approximately 607 ha of cattle pasture riod (1800 to 0600 h). No carbon dioxide tropical storm to make landfall along the bordered by mixed brush, marsh, and was used. To evaluate the effectiveness Pacific coast of Central America since sugarcane habitat in a malarious area in of the two types of traps, Mosquito Mag- 1949 (5, 7). Heavy rainfall across Central northern Belize (Orange Walk Town). The net® trap data from 20–28 May (4 traps × America (including Belize) caused flash first trial was conducted with four of the 8 nights or 32 trap-nights) were com- flooding and landslides in Costa Rica and Mosquito Magnet® traps, which were pared with data from the single baited Nicaragua and left 42 000 people without used to monitor mosquito activity from 21 CDC light trap run from 20–27 May (1 power (6, 7). Damage was estimated at May to 3 June. As per manufacturer in- trap × 7 nights or 7 trap-nights). $33 million USD (7). structions, the traps were sheltered from On 30 May 2008, Alma emerged into rainfall—in this case inside two-person Octenol-baited CDC light trap housed the Gulf of Honduras and fused with a tents utilized by the U.S. military (Na- inside tent versus unbaited CDC light tropical wave off the coast of Belize that tional Stock Number 8340 01 026 6096). traps outside tent. Two unbaited CDC became Tropical Storm Arthur one day The distance from the tent opening to light traps were also run, outside the before the official start of the 2008 At- the Mosquito Magnet® was approxi- tents. Data from these traps were com- lantic hurricane season (5). Tropical mately 1 m. Although the tents had been pared with data from the single baited Storm Arthur was the first Atlantic tropi- treated with insecticides as part of another CDC light trap housed inside the tent to cal storm to form during the month of study, statistical analysis indicated the evaluate the effectiveness of octenol as May since 1981 (5). It made landfall on treatments were not effective (i.e., the trap an attractant. The two unbaited CDC the Yucatan Peninsula on 31 May 2008 catches from the treated tents did not dif- light traps were run for five nights (28 and remained a tropical storm over land fer from those of the control tent [P = May and 31 May–3 June), and the single for nearly 24 hours before weakening to a 0.161]). The four Mosquito Magnet® traps octenol-baited CDC light trap was run tropical depression (5, 8). The rainfall were run nightly, 12 hours per sampling for 10 nights (23–28 May, 31 May, and from Arthur—approximately 25.4 cm in day (1800 to 0600 h), four nights before the 1–3 June), generating a total of 10 trap- Belize over 36 hours—compounded the first storm of the wet season (25–28 May) nights for each (2 traps × 5 nights and 1 effects of flooding caused by Tropical and four nights after (31 May–3 June), trap × 10 nights respectively). Both traps Storm Alma (8). This heavy rainfall yielding 16 trap-nights (number of traps were run for the same daily sampling pe- caused rivers in both southern and north- run [4] × number of nights [4] = 16) for riod (1800 to 0600 h). As in the previous ern Belize to overflow, damaging bridges each of the two comparison time periods trial, no carbon dioxide was used. and highways and forcing evacuations in (pre- and post-storm). All mosquitoes from the daily trap Orange Walk District in the north as well Each of the four Mosquito Magnet® catches were counted and identified to as parts of Corozal, Mexico (8). Flash traps utilized a 3.8-liter tank of 60% species using a dichotomous key (9, 10). flooding reportedly killed nine people in propane and 40% butane to produce heat Voucher specimens were later deposited Belize and affected 100 000 more. Dam- and carbon dioxide (a by-product of in the collections of the U.S. Army’s Wal- age was estimated at $78 million USD (8). the combustion process). The propane/ ter Reed Biosystematics Unit (WRBU) The authors of the current study, who butane mix was chosen because it was at the Smithsonian Institution National were working in the area where the early widely available in Central America and Museum of Natural History (NMNH). tropical storms occurred, designed and consistent with Mosquito Magnet® man- implemented the current research to take ufacturer’s instructions. Each Mosquito Statistical analysis advantage of the opportunity to compare Magnet® was baited with the 1 600-mg the mosquito vector abundance immedi- octenol cartridge included with each The abundances of mosquitoes per ately before and after the first heavy rain- trap and operated according to manufac- trap-night were calculated to determine fall of the wet season. This type of data turer’s instructions. any notable differences pre- and post- may be useful in estimating the effect of The tents were spaced 20 m apart and storm (comparing the Mosquito Magnet® tropical storms on local vector popula- arranged in a line parallel to and approx- trap data before and after the rainfall) and tions and subsequent changes in disease imately 100 m from a local marsh (a mos- to assess the effectiveness of the octenol risks. A secondary goal of the study was quito breeding site). The sides of the tents attractant (comparing the data from the to compare the efficacy of two types of facing the marsh (60°NNE) were left octenol-baited CDC trap inside the tent mosquito traps in northern Belize: the open. The backs of the tents were staked with data from the unbaited CDC traps commercially produced Mosquito Mag- to the ground but left unbuttoned to pre- outside the tent). Incidence rate ratios and net® (Woodstream Corp., Lititz, PA, vent overheating of the trap and avoid exact binomial 95% confidence intervals USA) and the U.S. Centers for Disease the repellent effect of excessive heat. (CIs) were calculated using Stata Statisti-

20 Rev Panam Salud Publica 28(1), 2010 Morrow et al. • Mosquito abundance pre- and post-tropical storms in Belize Original research cal Software version 10.0 (StataCorp LP, creased by half (pre-storm total = 1 137; more likely to capture any mosquitoes College Station, TX, USA) (11). post-storm total = 614). attracted by the presence of the animals). An. crucians (Wiedemann, 1949) was If the data for the abnormally high catch- RESULTS the most abundant species before the ment of An. crucians during that one storm, followed by Coquillettidia nigricans night are removed from the analysis, the Mosquito abundances pre- and (Coquillet, 1904) and Mansonia titillans mosquito abundance for the CDC light post-tropical storms (Walker, 1848). No Ae. taeniorhynchus and trap is reduced to 11.5 mosquitoes per only five Culex spp. specimens were trap-night (Table 2) and the Mosquito There was a significant difference in trapped immediately before the storm. Magnet® trap rate becomes almost five the abundance of mosquitoes trapped by The species trapped most often post- times higher for all species versus the the octenol-baited Mosquito Magnets® storm was Ma. titillans, followed by An. octenol-baited CDC light trap. If the data run immediately before the storm versus crucians. Ae. taeniorhynchus and Culex spp. for the one night of abnormally high An. those run immediately after the storm were trapped more often immediately crucians catchment are included, the for all species except for Anopheles vesti- following the storm. Mosquito Magnet® is almost three times tipennis (Dyar and Knab, 1906), Psoro- more effective at trapping all species of phora ferox (Von Humboldt, 1819), and Octenol-baited Mosquito Magnets® ver- mosquitoes and two times more effective Ps. albipes (Theobald, 1907). The fact that sus octenol-baited CDC trap. There was at trapping Anopheles spp. In either sce- the trap rates for these latter three a significant difference in trap rates be- nario, the Mosquito Magnet® was four species did not differ is most likely due tween the octenol-baited Mosquito Mag- times more effective at trapping culicine to the low numbers of specimens ob- nets® and CDC light trap housed in tents species than the CDC trap. tained (Table 1). While the total number for all species except Ps. ferox and Ps. al- An. crucians was the species trapped of culicine species collected by the traps bipes. The small number of trapped spec- most often in the Mosquito Magnet®, fol- did not change pre- versus post-storm, imens for these latter two species pre- lowed by Ma. titillans and Cq. nigricans. the species composition was altered, vented reliable statistical analysis. While the total number of specimens col- with increases in the number of Aedes The octenol-baited Mosquito Mag- lected per trap-night by the octenol-baited taeniorhynchus (Wiedemann, 1958) and nets® produced an average of 54.7 mos- CDC trap was lower than that for the Culex (Culex) spp. post-storm. All Culex quitoes per trap-night (total trapped = Mosquito Magnets®, the species distribu- spp. were identified down to the sub- 1 750) whereas the octenol-baited CDC tion for both types of traps was the same. genus Culex. As explained by J. Pecor light trap caught an average of 20.4 mos- (NMNH/WRBU, personal communica- quitoes per trap-night (total trapped = Octenol-baited CDC light trap inside tion, 5 July 2007), the Cx. (Cx.) spp. could 143) (Table 2). However, the bulk of the tent versus unbaited CDC light traps not be identified down to any of the 11 trap rate from the CDC light trap was outside tent. The mosquito abundance species found within this subgenus in due to an abnormally high yield of An. per trap-night and associated P-values Belize due to the lack of male specimens. crucians on a single night that was most and 95% CIs for the octenol-baited CDC Conversely, the total number of Anophe- likely due to the presence of livestock in light trap housed inside the tent versus les spp. was three times higher before the the vicinity during that period (the CDC the unbaited CDC light traps outside the storm compared to afterward. The total trap was closer to the livestock than the tents are presented in Table 3. There was number of species caught post-storm de- Mosquito Magnets® and was therefore a significant difference between the trap rates of the baited trap versus the unbaited traps for some species (Anopeheles TABLE 1. Number of mosquitoes collected in four octenol-baited Mosquito Magnet® traps run in- spp., An. crucians, Culicine spp., and Ma. side tents for four nights before the first tropical storm of the wet season (25–28 May) and four titillans) but no difference for others (An. nights after (31 May–3 June), Orange Walk Town, Belize, 2008 albimanus, An. vestitipennis, Cq. nigricans, Culex spp., Ae. taeniorhynchus, Ps. ferox, Pre- Post- Pre-storm Post-storm storm storm (per trap- (per trap- and Ps. albipes), primarily due to the low Species (total) (total) night)a night) IRRb 95% CIc P trap rates of the latter group of species. However, the octenol-baited CDC trap Anopheles spp. 722 223 45.1 13.9 3.2 (2.8, 3.8) <0.001 (run inside the tent) was more effective An. albimanus 128 30 8 1.9 4.2 (2.8, 6.6) <0.001 An. crucians 587 183 36.7 11.4 3.2 (2.7, 3.8) <0.001 overall than the unbaited traps (run out- An. vestitipennis 7 10 0.4 0.6 0.6 (0.2, 2.0) 0.48 side the tent), obtaining 19.3 total mos- Culicine spp. 415 391 25.9 24.4 1.1 (0.9, 1.2) 0.40 quitoes per trap-night (total trapped = Coquillettidia nigricans 264 42 16.5 2.6 6.3 (4.5, 8.9) <0.001 193) versus only 1.4 total mosquitoes per Mansonia titillans 146 204 9.1 12.8 0.7 (0.6, 0.9) 0.02 Culex spp. 5 56 0.3 3.5 0.1 (0.03, 0.2) <0.001 trap-night (total trapped = 14) for the un- Aedes taeniorhynchus 0860 5.4…d (0, 0.04) <0.001 baited traps and thus almost 14 times Psorophora ferox 0 0 0 0 … … 1.00 more mosquitoes overall (and 3.5 times Ps. albipes 0 3 0 0.2 … (0, 2.4) 0.13 more culicine species). An. crucians was Total 1 137 614 71.1 38.4 1.9 (1.7, 2.0) <0.001 the species trapped most often in the a Number of traps run (4) × number of nights (4) = 16 trap-nights per comparison period (pre- and post-storm). baited CDC trap. The unbaited trap b Internal rate of return. c Confidence interval. failed to collect any specimens from this d Insufficient data for calculation. species throughout the study period.

Rev Panam Salud Publica 28(1), 2010 21 Original research Morrow et al. • Mosquito abundance pre- and post-tropical storms in Belize

TABLE 2. Number of mosquitoes collected in four octenol-baited Mosquito Magnet® traps and Hurricane Katrina there was a sharp in- one CDCa light trap run inside tents for eight nights (20–28 May) and seven nights (20–27 May) crease in the number of reported cases of respectively, Orange Walk Town, Belize, 2008 neuroinvasive West Nile virus disease in hurricane-affected regions (13). They also Mosquito Mosquito CDC Magnet® CDC found a >2-fold increase in incidence of Magnet® light traps light trap neuroinvasive West Nile virus disease in traps trap (per trap- (per trap- hurricane-affected areas for 2006 versus Species (total) (total) night)b night)c IRRd 95% CIe P previous years (13). Many of these cases occurred among construction workers Anopheles spp. 945 100 29.5 14.3 2.1 (1.7, 2.6) <0.001 An. albimanus 158 0 4.9 0 …f (9.3, …) <0.001 and other cleanup crew, underscoring An. crucians 770 100 24.1 14.3g 1.7 (1.4, 2.1) <0.001 the need to provide adequate shelter and An. vestitipennis 17 0 0.5 0 … (0.9, …) 0.03 mosquito control for disaster relief vol- Culicine spp. 805 43 25.2 6.1 4.1 (3.0, 5.7) <0.001 unteers and workers as well as residents Coquillettidia nigricans 306 14 9.6 2.0 4.8 (2.8, 8.9) <0.001 Mansonia titillans 350 26 10.9 3.7 2.9 (2.9, 2.0) <0.001 of affected areas in the days and weeks Culex spp. 61 2 1.9 0.3 6.3 (1.8, 56.3) <0.001 that follow the storm (12, 14, 15). During Aedes taeniorhynchus 86 0 2.7 0 … (5.0, …) <0.001 the post-storm cleanup of Hurricane An- Psorophora ferox 0 1 0 0.1 … (0, 8.5) 0.18 drew, repair and cleanup crews were ex- Ps. albipes 3 0 0.1 0 … (0.09, …) 0.55 posed to high densities of mosquitoes, Total 1 750 143 54.7 20.4 2.7 (2.2, 3.2) <0.001 which increased the potential for mos- Excluding abnormally high CDC trap rate recorded for one night of sampling quito-borne disease transmission as well An. crucian 770 26 24.1 4.3 5.6 (3.8, 8.5) <0.001 All species 1 750 69 54.7 11.5 4.8 (3.7, 6.1) <0.001 as bacterial infections from mosquito bites. Like the post-storm species compo- a Centers for Disease Control and Prevention (USA). sition in Belize found in the current b Number of traps run (4) x number of nights (8) = 32 trap-nights. c Number of traps run (1) x number of nights (7) = 7 trap-nights. study, the most common species col- d Internal rate of return. lected post–Hurricane Andrew in both e Confidence interval. Florida and Louisiana were mainly Culi- f Insufficient data for calculation. g Includes abnormally high trap rate recorded for one night (possibly attributable to the presence of livestock near the CDC trap cine spp. (15). It should be noted, how- during that period). ever, that the studies cited above do not reliably link increased mosquito abundance with increased risk of disease be- TABLE 3. Number of mosquitoes collected in one octenol-baited CDCa light trap run inside a tent cause the increased disease incidence for ten nights (23–28 May, 31 May, 1–3 June) and two unbaited CDC light traps run outside tents could have been due to increased human for five nights (28 May and 31 May–3 June), Orange Walk Town, Belize, 2008 exposure to vectors rather than increased vector numbers. Octenol- Octenol- baited Unbaited In the current study, an increase in the baited Unbaited trap traps variety of culicine species was noted as trap traps (per trap- (per trap- soon as four days post-storm. However, Species (total) (total) night)b night)c IRRd 95% CIe P a drastic drop in Anopheles spp. was also noted post-storm, which may have im- Anopheles spp. 143 0 14.3 0 NAf (38.3, …g) <0.001 An. albimanus 1 0 0.1 0 NA (0.03, …) 0.50 plications for malaria prevention during An. crucians 142 0 14.2 0 NA (38.0, …) <0.001 storm seasons (16). In northern Belize, An. vestitipennis 0 0 0 0 NA … 1.00 malaria is mesoendemic and moderately Culicine spp. 50 14 5.0 1.4 3.6 (1.9, 7.0) <0.001 unstable, with seasonal epidemic exacer- Coquillettidia nigricans 18 11 1.8 1.1 1.6 (0.7, 3.8) 0.20 Mansonia titillans 29 3 2.9 0.3 9.6 (3.0, 49.6) <0.001 bations showing a fairly close correlation Culex spp. 2 0 0.2 0 NA (0.2, …) 0.25 with alterations in rainfall (17). As docu- Aedes taeniorhynchus 0 0 0 0 NA … 1.00 mented in this study, habitat damage Psorophora ferox 1 0 0.1 0 NA (0.03, …) 0.50 from tropical weather events can disturb Ps. albipes 0 0 0 0 NA … 1.00 Anopheles spp. populations enough to Total 193 14 19.3 1.4 13.8 (8.0, 25.7) <0.001 temporarily reduce their abundance and a Centers for Disease Control and Prevention (USA). thus their role in malaria transmission b Number of traps run (1) × number of nights (10) = 10 trap-nights. c Number of traps run (2) × number of nights (5) = 10 trap-nights. (18). This finding suggests that public d Internal rate of return. health interventions to prevent malaria e Confidence interval. immediately after a storm may not be as f Not applicable (IRRs were not estimable for species that were not collected in the unbaited traps). g Insufficient data for calculation. important as previously assumed (19), although this hypothesis is somewhat controversial in the literature because DISCUSSION understanding the effects of such events malaria transmission is not always di- on local vector-borne disease epidemiol- rectly correlated with number of mos- Because many mosquito-borne dis- ogy is important for directing appro- quitoes or amount of rainfall (e.g., very eases are found in regions of the world at priate public health responses (12). Cail- low numbers of mosquitoes can also risk for hurricanes and tropical storms, louët et al. (2008) showed that after produce significant disease transmis-

22 Rev Panam Salud Publica 28(1), 2010 Morrow et al. • Mosquito abundance pre- and post-tropical storms in Belize Original research sion) (20, 21)). Once the Anopheles vectors and culicines) than the octenol-baited how other climatic, temporal, and envi- are reestablished in the environment, CDC light trap, and 2) octenol is a reli- ronmental factors facilitate breeding in however, the increased rainfall provides able attractant for An. crucians in north- certain species post-storm (21). After a suitable breeding habitat, thus increas- ern Belize (based on the observation that Tropical Storm Doreen hit California in ing the chance for a possible outbreak, as the unbaited CDC light traps failed to 1977, Gordon et al. hypothesized that was seen after Hurricane Flora swept collect any specimens from this species many breeding cycles are obligatory in across the southern peninsula of Haiti whereas the octenol-baited CDC trap nature and are modified primarily by (17). Conversely, this re-stabilization pe- attracted a high number of them). These temperature rather than precipitation riod can take many weeks to months, findings may prove useful for public (21). Comparing these results to those of most likely depending on the strength of health planners attempting to set up the current study, it should be noted the storm and associated wind speed (17, field surveillance programs after tropical that the rains caused by Tropical Storm 19). For this reason, natural disasters do weather has devastated an area. Doreen were not thought to have cre- not usually cause an immediate increase ated favorable breeding habitats due in vector-borne diseases. In areas that are Limitations to lack of vegetation and larval activ- heavily damaged, however, vector con- ity, whereas the current research site trol may be inappropriately delayed This study had several limitations. For had profound vegetation in its water during the most paramount of times— example, in the trap comparison studies, sources. In any case, it is clear that mos- immediately after the storm—when most there was an uneven number of repli- quito abundance and composition can residents of the affected areas are still liv- cates (due to the unpredictable nature of vary tremendously in a region affected ing in evacuation areas (4, 22). storms). If this type of study were at- by tropical storms. This underscores the The current study found that Culex spp. tempted in the future, comparison of an importance of species abundance re- in Belize are able to increase rapidly fol- equal number of trap-nights and trial search pre- and post-storms to facilitate lowing a storm and should potentially be nights would be preferable. In the pre- and improve public health planning for regarded as a possible vector of interest and post-storm study, the main weak- devastated areas. It also further justifies immediately following heavy rainfall, es- ness was the shortened temporal scope. the need for active disease surveillance pecially at the end of the dry season. If A longer collection time (both pre- and during these time frames to identify any hurricanes strike early in transmission post-storm) would improve the analysis increases in certain species that may season, there could be a late increase in and allow for consideration of the vary- lead to a subsequent increase in disease risk after vector and host populations are ing life cycles of potential vectors found risk or transmission. reestablished. Culex spp. are able to trans- in northern Belize. Although the current mit a number of pathogens to humans results indicating a parallel decrease in Acknowledgments. The authors thank (e.g., West Nile virus and Venezuelan density of mosquitoes during the first 48 Cara Olsen for her help in the statistical equine encephalitis virus) and therefore hours after storms are supported by analyses and the anonymous reviewers may require immediate control measures studies conducted in 2000–2001 on the whose comments helped improve the after heavy rainfall in northern Belize. It effect of tropical storms on adult mos- manuscript. should be noted, however, that although quito abundance in Calcutta, India, it these results indicate the abundance of would be useful to determine how long Financial support. Funding for this certain vector species can increase rapidly this decrease in total species lasts and work was provided by Uniformed Ser- after heavy rainfall, further studies and what factors play a role in rebuilding af- vices University of the Health Sciences, increased surveillance of diseases associ- fected mosquito populations (23). Bethesda, Maryland, under grant num- ated with these vector species are needed In addition, future studies examining ber GS87K3. to ascertain whether or not this increase mosquito populations before and after in species abundance leads to a direct in- several tropical storms and hurricanes Disclaimer. The opinions, interpreta- crease in disease risk. would further clarify how mosquito tions, conclusions, and recommenda- As mentioned above, based on the re- abundance is altered by these weather tions contained herein are those of the search results, the authors of the current events and how it varies on an annual authors and do not reflect official views study concluded that 1) the Mosquito basis. Comparing mosquito populations or policy of the Uniformed Services Uni- Magnet® was more effective at trapping during two or more different storm versity of the Health Sciences or the U.S. all species of mosquitoes (Anopheles spp. years could prove useful in determining Department of Defense.

REFERENCES

1. Juliano SA. Population dynamics. J Am Mosq habitat management strategies for the reduc- NOAA; c2008 [cited 2009 Sept 18]. State of the Control Assoc. 2007;23(2 Suppl):265–75. tion of malaria vectors in northern Belize. J climate—hurricanes & tropical storms; [about 2. Fillinger U, Sombroek H, Majambere S, van Vector Ecol. 2005;30(2):235–43. 2 screens]. Available from: http://www. Loon E, Takken W, Lindsay SW. Identifying 4. Watson JT, Gayer M, Connolly MA. Epi- ncdc.noaa.gov/sotc/?report=tropical- the most productive breeding sites for demics after natural disasters. Emerg Infect cyclones&year=2008&month=5. malaria mosquitoes in The Gambia. Malar J. Dis. 2007;13(1):1–5. 6. International Federation of Red Cross and 2009;8:62, doi:10.1186/1475-2875-8-62. 5. National Oceanic and Atmospheric Adminis- Red Crescent Societies. Costa Rica and 3. Grieco JP, Vogtsberger RC, Achee NL, Vanzie tration, National Climatic Data Center (US) Nicaragua: Tropical Storm Alma. DREF oper- E, Andre RG, Roberts DR, et al. Evaluation of [homepage on the Internet]. Washington: ation no. MDR43003 final report. Geneva:

Rev Panam Salud Publica 28(1), 2010 23 Original research Morrow et al. • Mosquito abundance pre- and post-tropical storms in Belize

IFRC; 2009 [cited 2009 Sept 18]. Available 11. Rothman KJ. Modern epidemiology. Boston: America with emphasis on Anopheles dar- from: http://www.ifrc.org/docs/appeals/ Little Brown; 1986. p. 166. lingi. J Vector Ecol. 2006;31(1):45–57. rpts08/MDR43003dfr.pdf. 12. Diaz JH. Global climate changes, natural dis- 19. Lehman JA, Hinckley AF, Kniss KL, Nasci RS, 7. United Nations Office for the Coordination of asters, and travel health risks. J Travel Med. Smith TL, Campbell GL, et al. Effect of Hurri- Humanitarian Affairs. Weekly note on emer- 2006;13(6):361–72. cane Katrina on arboviral disease transmis- gencies: Latin America & the Caribbean. No. 13. Caillouët KA, Michaels SR, Xiong X, Foppa I, sion. Emerg Infect Dis. 2007;13(8):1273–5. 59, 23 Jun 2008. New York: OCHA; 2008 Wesson DM. Increase in West Nile virus neu- 20. Warrell DA, Gilles HM. Essential malariol- [cited 2009 Sept 18]. Available from: http:// roinvasive disease after Hurricane Katrina. ogy. 4th ed. London: Arnold International; www.reliefweb.int/rw/R-Files2008. nsf/fb9 Emerg Infect Dis. 2008;14(5):804–7. 2002. p. 59. a3459e0c5a152c1257205004f1c3e/e9b2c9a5ca 14. Swygard H, Stafford RE. Effects on health of 21. Gordon EW, Smith TA, Ortiz J, Moisse E. The d313a0c1257461006e7fe0/$FILE/Full. volunteers deployed during a disaster. Am effects of Tropical Storm Doreen on mosquito Report.pdf. Surg. 2009;75(9):747–52; discussion 752-3. populations in Imperial Valley. Proceedings 8. International Federation of Red Cross and 15. Centers for Disease Control and Prevention and Papers of the 46th Annual Conference of Red Crescent Societies. Belize: Tropical Storm (US). Emergency mosquito control associated the California Mosquito and Vector Control Arthur. DREF operation no. MDRBZ001 final with Hurricane Andrew—Florida and Lou- Association; 1978 Jan 29–Feb 1; Yosemite, Cal- report. Geneva: IFRC; 2008 [cited 2009 Sept isiana, 1992. MMWR Morb Mortal Wkly Rep. ifornia, USA. Visalia (CA): CMVCA Press; 18]. Available from: http://www2.reliefweb. 1993;42(13):240–2. 1978. p. 55–7. int/rw/rwb.nsf/db900sid/EDIS-7MKR8Z? 16. Achee NL, Korves CT, Bangs MJ, 22. Nasci RS, Moore CG. Vector-borne disease OpenDocument&query=%20Tropical%20Sto Rejmankova E, Lege M, Curtin D, et al. Plas- surveillance and natural disasters. Emerg rm%20Arthur&emid=TC-2008-000077-CRI. modium vivax polymorphs and Plasmodium Infect Dis. 1998;4(2):333–4. 9. Clark-Gil S, Darsie RF Jr. The mosquitoes of falciparum circumsporozoite proteins in 23. Pramanik M, Raut SK. Impact of kalbaishakhi Guatemala: their identification, distribution Anopheles (Diptera: Culicidae) from Belize, storms on adult mosquito populations in and binomics with keys to adult females and Central America. J Vector Ecol. 2000;25(2): Kolkata (Calcutta), India. Acta Trop. 2003;85 larvae in English and Spanish. Mosq Syst. 203–11. (1):93–5. 1983;15(3):151–284. 17. Mason J and Cavalie P. Malaria epidemic in 10. Pecor JE, Harbach RE, Peyton EL, Roberts Haiti following a hurricane. Am J Trop Med DR, Rejmankova E, Manguin S, et al. Mos- Hyg. 1965;14(4):533–9. quito studies in Belize, Central America: 18. Achee NL, Grieco JP, Rejmankova E, Andre records, taxonomic notes, and a checklist of RG, Vanzie E, Polanco J, et al. Biting patterns species. J Am Mosq Control Assoc. 2002;18(4): and seasonal densities of Anopheles mosqui- Manuscript received on 15 January 2010. Revised version 241–76. toes in the Cayo District, Belize, Central accepted for publication on 9 June 2010.

RESUMEN Objetivo. Comprobar la abundancia de mosquitos adultos en el norte de Belice antes y después de la primera tormenta tropical de la temporada de lluvias, para calcular la cantidad de tiempo que necesitan las poblaciones de vectores potenciales para desarro- Abundancia de mosquitos llarse y recuperarse; determinar cuáles son las especies que predominan después de la vectores inmediatamente tormenta; y comparar la eficacia de dos tipos de trampas de mosquitos: las que utilizan cebos de octenol (Mosquito Magnets®) y las trampas de luz con y sin octenol de los antes y después de las Centros para el Control y la Prevención de Enfermedades de los Estados Unidos (CDC). tormentas tropicales Alma y Métodos. Se llevaron a cabo pruebas de campo en Orange Walk Town, Belice, desde Arthur, norte de Belice, 2008 el 21 de mayo hasta el 3 de junio del 2008. Se registraron las razones de las tasas de incidencia y los intervalos de confianza binomiales exactos de 95% y se evaluaron las trampas nocturnas para comparar la abundancia de las especies y la eficacia del tipo de trampa antes y después de una tormenta. Resultados. Antes de la tormenta, se atrapó el doble de las especies de mosquitos y el triple de Anopheles spp. en comparación con el período que siguió a la tormenta. Sin embargo, el número de Aedes taeniorhynchus y Culex (Culex) spp. atrapados fue mayor después de la tormenta. Las trampas Mosquito Magnets® demostraron uniformemente ser más eficaces que las trampas de los CDC, ya que el número de Anopheles spp. cap- turados fue dos veces mayor que el de la trampa con cebo de octenol y el número de especies de culícidos fue cuatro veces mayor. En general, el número de mosquitos cap- turados por la trampa con octenol fue 14 veces mayor que el de la trampa sin octenol y el número de especies de culícidos fue 3,5 veces mayor. La trampa sin cebo de los CDC no capturó ningún Anopheles spp. durante el período de estudio. Conclusiones. Los resultados demostraron que el octenol es un producto eficaz para atraer a la especie Anopheles crucians en el norte de Belice. En este país, el riesgo de malaria disminuye inmediatamente después de las tormentas (es decir, disminuye la abundancia de mosquitos); y el riesgo de contraer una infección por arbovirus, asociado con el rápido aumento de las especies de culícidos después de las tormentas, podría representar una mayor amenaza para la salud pública que la malaria (aunque se necesitaría más in- vestigación y una vigilancia activa de las enfermedades para validar esta hipótesis).

Palabras clave Insectos vectores; control de mosquitos; control de vectores; malaria; tormenta tropical; Belice.

24 Rev Panam Salud Publica 28(1), 2010