RESEARCH
Effectiveness of residual spraying of peridomestic ecotopes with deltamethrin and permethrin on Triatoma infestans in rural western Argentina: a district-wide randomized trial
Efficacité sur Triatoma infestans de la pulvérisation des écotopes péridomestiques par la deltaméthrine et la perméthrine rémanentes dans un district rural de l'ouest de l'Argentine : essai randomisé de district
Eficacia contra Triatoma infestans del rociamiento de acción residual de ecotopos peridomésticos con deltametrina y permetrina en zonas rurales del oeste de la Argentina: ensayo aleatorizado en un distrito
Ricardo E. GürtlerI, 1; Delmi M. CanaleII; Cynthia SpillmannII; Raúl StarioloII; Oscar D. SalomónIII; Sonia BlancoII; Elsa L. SeguraIV
IAssociate Professor, Laboratorio de Ecología General, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, Ciudad Universitaria, C1428EHA Buenos Aires, Argentina (email: gurtler@biolo.bg.fcen.uba.ar)
IIResearcher, Coordinación Nacional de Control de Vectores, Córdoba, Argentina
IIIDirector, Centro Nacional de Diagnóstico e Investigación en Enfermedades Endemo-epidémicas, Administración Nacional de Laboratorios e Institutos de Salud "Dr Carlos G. Malbrán", Buenos Aires, Argentina
IVSenior Researcher, Instituto Nacional de Parasitología "Dr Mario Fatala Chaben" "Dr Carlos G. Malbrán", Buenos Aires, Argentina
ABSTRACT
OBJECTIVE: To compare the effectiveness of a single residual spraying of pyrethroids on the occurrence and abundance of Triatoma infestans in peridomestic ecotopes in rural La Rioja.
METHODS: A total of 667 (32.8%) peridomestic sites positive for T. infestans in May 1999 were randomly assigned to treatment within each village, sprayed in December 1999, and reinspected in December 2000. Treatments included 2.5% suspension concentrate (SC) deltamethrin in water at 25 mg active ingredient (a.i.)/m2 applied with: (a) manual compression sprayers (standard treatment) or (b) power sprayers; (c) 1.5% emulsifiable concentrate (EC) deltamethrin at 25 mg a.i./m2; and (d) 10% EC cis-permethrin at 170 mg a.i./m2. EC pyrethroids were diluted in soybean oil and applied with power sprayers. All habitations were sprayed with the standard treatment.
FINDINGS: The prevalence of T. infestans 1-year post-spraying was significantly lower in sites treated with SC deltamethrin applied with manual (24%) or power sprayers (31%) than in sites treated with EC deltamethrin (40%) or EC permethrin (53%). The relative odds of infestation and catch of T. infestans 1-year post-spraying significantly increased with the use of EC pyrethroids, the abundance of bugs per site before spraying, total surface, and host numbers. All insecticides had poor residual effects on wooden posts.
CONCLUSION: Most of the infestations probably originated from triatomines that survived exposure to insecticides at each site. Despite the standard treatment proving to be the most effective, the current tactics and procedures fail to eliminate peridomestic populations of T. infestans in semiarid rural areas and need to be revised.
Keywords: Pyrethrins/administration and dosage; Triatoma/drug effects; Insect control/methods; Randomized controlled trials; Argentina (source: MeSH, NLM).
RÉSUMÉ
OBJECTIF: Comparer l'efficacité d'une pulvérisation rémanente unique de pyréthrinoïdes sur la présence et l'abondance de Triatoma infestans dans les écotopes péridomestiques du district rural de La Rioja.
MÉTHODES: Un total de 667 sites péridomestiques (soit 32,8 %) où la présence de T. infestans avait été décelée en mai 1999 ont été tirés au sort dans chacun des villages pour être traités; la pulvérisation a eu lieu en décembre 1999, et ils ont été de nouveau inspectés en décembre 2000. Le traitement était le suivant : suspension concentrée à 2,5 % de deltaméthrine dans l'eau, épandue à la dose de 25 mg de substance active /m2, appliquée au moyen : a) d'un pulvérisateur à pression préalable actionné manuellement (traitement classique) ou b) d'un pulvérisateur à moteur ; c) concentré émulsifiable à 1,5 % de deltaméthrine, épandu à la dose de 25 mg de substance active /m2 ; d) concentré émulsifiable à 10 % de cis-perméthrine, épandu à la dose de 170 mg de substance active /m2. Les concentrés émulsifiables de pyréthrinoïdes ont été dilués dans l'huile de soja et épandus au moyen de pulvérisateurs à moteur. Dans toutes les habitations l'épandage a été fait selon la méthode classique.
RÉSULTATS: Un an après la pulvérisation, la fréquence de T. infestans était nettement plus faible dans les sites traités par la deltaméthrine en suspension concentrée appliquée manuellement (24 %) ou au moyen de pulvérisateurs à moteurs (31 %) que dans les sites traités par le concentré émulsifiable de deltaméthrine (40 %) ou de perméthrine (53 %). Un an après la pulvérisation, la probabilité relative d'infestation ou de capture de T. infestans était nettement augmentée par les facteurs suivants : utilisation d'une concentré émulsifiable de pyréthrinoïde, abondance des triatomes sur le site avant pulvérisation, surface totale et nombre d'hôtes. Tous les insecticides avaient peu d'effet rémanent sur les poteaux en bois.
CONCLUSION: La plupart des infestations sont probablement dues à des triatomes qui ont survécu à l'exposition aux insecticides dans chacun des sites. Même si le traitement classique se révèle être le plus efficace, la stratégie et les méthodes actuelles ne parviennent pas à éliminer les populations péridomestiques de T. infestans en milieu rural semi-aride ; elles devront être révisées.
Mots clés: Pyréthrine/administration et posologie; Triatoma/action des produits chimiques; Lutte contre insecte/méthodes; Essai clinique randomisé; Argentine (source: MeSH, INSERM).
RESUMEN
OBJETIVO: Comparar la efectividad de un único rociamiento de acción residual de piretroides contra la presencia y abundancia de Triatoma infestans en ecotopos peridomésticos en La Rioja rural.
MÉTODOS: Un total de 667 (32,8%) sitios peridomésticos positivos para T. infestans en mayo de 1999 fueron asignados aleatoriamente a tratamiento dentro de cada aldea, rociados en diciembre de 1999, y reinspeccionados en diciembre de 2000. Las opciones de tratamiento fueron una suspensión concentrada (SC) de deltametrina al 2,5% en agua a razón de 25 mg ai/m2 aplicada mediante: (a) pulverizadores manuales de compresión (tratamiento ordinario) o (b) pulverizadores eléctricos; (c) un concentrado emulsionable (CE) al 1,5% de deltametrina a 25 mg ai/m2; y (d) un CE al 10% de cis-permetrina a 170 mg ai/m2. Los CE de piretroides se diluyeron en aceite de soja y se aplicaron con pulverizadores eléctricos. Todas las viviendas se rociaron aplicando el tratamiento ordinario.
RESULTADOS: La prevalencia de T. infestans al cabo de un año del rociamiento fue significativamente menor en los sitios tratados con SC de deltametrina aplicada con pulverizadores manuales (24%) o eléctricos (31%) que en los sitios tratados con CE de deltametrina (40%) o CE de permetrina (53%). Las posibilidades relativas de infestación y captura de T. infestans al cabo de un año del rociamiento aumentaban sensiblemente con el uso de CE de piretroides, la abundancia de insectos por sitio antes del rociamiento, la superficie total y el número de huéspedes. Todos los insecticidas tuvieron un escaso efecto residual en los postes de madera.
CONCLUSIÓN: La mayoría de las infestaciones se debieron probablemente a triatominos que sobrevivieron a la exposición a los insecticidas en cada sitio. A pesar de que el tratamiento ordinario demostró ser el más eficaz, los procedimientos y tácticas actuales no logran eliminar las poblaciones peridomésticas de T. infestans en las zonas rurales semiáridas, y deberían por tanto ser revisados.
Palabras clave: Piretrinas/administración y dosificación; Triatoma/efectos de drogas; Control de insectos/métodos; Ensayos controlados aleatorios; Argentina (fuente: DeCS, BIREME).
Introduction
Triatoma infestans, the main vector of Chagas disease, has been the target of the regional elimination programme "Southern Cone Initiative" since 1991 (1). Based mostly on the residual application of pyrethroid insecticides, this programme has been much more successful in Brazil, Chile, and Uruguay, than in the semiarid Chaco extending over Argentina, Bolivia, and Paraguay. Resurgence of peridomestic T. infestans populations after spraying with pyrethroids has frequently been reported by official control programmes in Argentina (S. Blanco, unpublished data, 2000), but the reasons remain unclear. Pyrethroids have proven much more effective in human habitations (2, 3) than in peridomestic ecotopes housing domestic animals and various species of triatomine bugs (410).
The formulation of the insecticide affects its residual activity (2). For example, emulsifiable concentrate (EC) cyfluthrin had a greater impact on sand flies that rested on sylvatic vegetation, probably because the oil carrier allowed a greater adsorption of the insecticide to the substrate (11). In addition, we hypothesized that newer formulations of EC pyrethroids applied with back-pack power sprayers might have greater effect than wettable powder (WP) or suspension concentrate (SC) formulations applied with manual sprayers, because power sprayers might increase the penetration of insecticides into deep refuges and provide a greater initial impact (3). EC or WP deltamethrin suppressed peridomestic infestations by T. infestans for 500 days (4), though not always (8). Our objectives were to compare the effectiveness on peridomestic populations of T. infestans of new emulsifiable formulations of deltamethrin and cis-permethrin applied with power sprayers relative to that of SC deltamethrin applied with manual sprayers (the standard treatment) or motor sprayers in a district-wide infested area. We also sought to identify factors that modify the local effectiveness of pyrethroids.
Methods and materials
Study area
The study was conducted in rural areas around Olta (30.4ºS, 66.1ºW), Departments General Belgrano and Chamical, Province La Rioja, Argentina (Fig. 1). The area is on a semiarid plain with xerophytic vegetation, and has been described elsewhere (12). The houses had last been sprayed with beta-cyfluthrin approximately 56 years before this trial. Most houses were of adobe bricks and thatched roofs, and nearly all had numerous typical peridomestic structures (range, 110 per house), for domestic animals (10). The area was selected because it had shown recurrent reinfestation after insecticide spraying and high peridomestic infestation rates; most houses were relatively easy to access, and local authorities were cooperative. The study area was extended to adjacent areas of Chamical to reach about 100 houses per treatment.
Study design
The trial was divided into a baseline survey, intervention, and assessment. At baseline, three teams numbered all 369 houses with a metal plaque (Fig. 1), and surveyed 362 houses between 22 April and 20 May 1999. Each team was composed of two highly experienced bug collectors from the National Vector Control Programme (NVCP) staff, one bug collector from the local control programme, and one supervisor. Supervisors explained the objectives and phases of the trial to every household, and canvassed each family head to record the number of residents, type and number of domestic animals, and use of insecticides (type and frequency of use). The number and type of domestic and peridomestic structures were sketched on a map, building materials were noted, and distances from each structure to human habitations were paced out. Two highly experienced NVCP bug collectors concurrently assessed the intensity of peridomestic infestation by T. infestans and other triatomine species using timed manual collections with an irritant spray (0.2% tetramethrin, Icona, Buenos Aires) for 30 min per house (10). The local bug collector searched for bugs in domestic areas. All bugs were examined in the field laboratory to determine numbers by species and stage (10).
The interventions aimed at full coverage (blanket spraying) and were carried out between 29 November and 23 December 1999, when the mean ambient temperature was 23.3 ºC (average minimum, 8.3 ºC; average maximum, 38.4 ºC). All houses were listed alphabetically by village and randomly assigned to a treatment within every village. Peridomestic sites were sprayed with 2.5% SC deltamethrin (K-Othrina, Agrevo, at 25 mg active ingredient (a.i.)/m2) in water applied with (a) 5-litre manual compression sprayers with Teejet 8002 fan nozzles, considered the standard treatment, or (b) back-pack power sprayers of 15-litre capacity (Guarany, Brazil); (c) 1.5% EC deltamethrin (Cislin, Agrevo, at 25 mg a.i./m2), and (d) 10% EC cis-permethrin (Chemotecnica, Buenos Aires, at 170 mg a.i./m2), both applied with the back-pack power sprayers. All domestic sites were sprayed with the standard treatment (4). EC-based pyrethroids were diluted in soybean oil to exclude the possibility that local alkaline waters might reduce their insecticidal activity. To prevent bias, EC insecticides were transferred from the original containers labelled by the manufacturer into 5-litre clean containers identified by a two-letter and two-number code, following criteria used in randomized trials (13). Two people not involved in the trial labelled the containers and stored the codes in duplicate.
All domestic and peridomestic sites were sprayed by four fixed teams, the members of which had no background information regarding the intensity of infestation in each site. Each team was composed of a highly experienced NVCP sprayman who only treated peridomestic sites, two local spraymen who only treated domestic habitats, and one supervisor. A total of 352 houses were sprayed, including 22 houses that could only be accessed in March 2000 and were sprayed with the standard treatment. A total of 17 houses, including three with infested peridomestic sites, remained unsprayed due to householders' absence or inaccessible dirt roads.
Before starting spraying operations, the discharge volume per minute of every power sprayer fitted with each type of plastic nozzle tip was measured three times using water and soybean oil. To keep the codes closed and to simplify spraying operations, we mixed 400 ml of EC insecticides with 6.6 litres of soybean oil. Nozzle tips No. 1 were used during the first week, but they frequently clogged and discharged small amounts of aerosol. We replaced them with nozzle tips No. 3 in the second week, but the rate of consumption of oil and the drop size were large. Therefore, nozzle tips No. 2 were used from the third week on. To keep the target dose, the spray coverage speed was readjusted according to the nozzle tip-specific discharge volume per minute. The applied insecticide dose in a daily sample of each team's sprays was estimated by measuring the area of each structure, the time used to spray it, and the volume of insecticide applied per site.
At 7 or 11 days post-spraying, sets of three replicates of 510 fifth instar nymphs totalling 163 T. infestans were exposed to cement blocks from four storerooms and wooden posts from four goat corrals under the shade. To fit to irregular surfaces, the insects were held in cones (10 cm in height and 14 cm in diameter) made of X-ray developed film, which were attached to the substrate using nails and adhesive tape. After a 24-h exposure, the insects were transferred to clean jars containing folded filter paper and kept at 2530 ºC and 5070% relative humidity. Mortality and knockdown effects were evaluated at 0, 1, 2, 3, 7, 14, 30, and 75 days post-exposure. The bugs were fed once on chicken at 30 days post-exposure. A second set of bioassays had to be entirely discarded because many of the insects reached the field in bad condition.
The assessment of intervention effects only included houses with peridomestic sites positive for T. infestans before spraying, and was carried out by four teams between 1 and 23 December 2000. Each team was composed of one supervisor and the two bug collectors who conducted the baseline inspections; the latter had no information regarding the insecticide used in each house. Bug collectors searched for triatomines in each peridomestic site positive before spraying using 0.2% tetramethrin for 15 min, and in a sample of the previously negative sites most likely to be infested, for a maximum of 40 min per house. Evaluations comprised 325 houses; these included 799 peridomestic sites randomly treated with insecticides in December 1999 and 61 sites sprayed in March 2000. Many sites positive at baseline could not be reinspected or were excluded because of lack of access to the house, disappearance, or physical modification. The collected bugs were processed as in the baseline survey. Infestation was taken to mean the capture of at least one live or moribund T. infestans of any stage (except eggs) in a given site. Colonization meant the finding of at least one live or moribund T. infestans nymph, regardless of the presence of other stages or triatomine species. Abundance referred to the number of live or moribund T. infestans bugs captured per peridomestic site over a defined search period.
Data analysis
The unit of analysis was the peridomestic site infested by T. infestans at baseline, randomly assigned to an insecticide treatment and sprayed in December 1999, and reinspected 1-year post-spraying. Six such sites with missing data were excluded from regression analyses, which were based on 513 sites. For analytical purposes, the houses were tentatively grouped into 35 clusters according to a minimum distance between adjacent clusters (2 km), and the degree of connectivity among houses within a given cluster (Fig. 1). The outcome variables were: (a) the proportion of sites infested by T. infestans 1-year post-spraying, analysed by logistic-binomial random effects regression for distinguishable data (14); and (b) the log-total number of T. infestans collected per site 1-year post-spraying, analysed by multiple linear regression. The explanatory variables were insecticide treatment (four levels, with SC deltamethrin applied with manual sprayers as the reference level); log-number of T. infestans in site i before spraying; spray team (four levels); opportunity (or week) of treatment (three levels); type of peridomestic ecotope (two levels, representing roofed structures, such as storerooms or sheds or kitchens, versus unroofed structures, such as corrals and piles); and surface and reported number of domestic animal hosts in each site, transformed to log (x + 1). Statistical significance was judged at the 5% level. Interactions between significant predictors were added stepwise one by one and kept in the model if they proved significant. The survival of bugs used in bioassays was analysed using Gehan's generalized Wilcoxon test.
Results
A total of 5251 T. infestans bugs were collected from 667 (38.2%) of 1748 peridomestic sites inspected at baseline (mean, 2.1 infested sites and 1.8 colonies per house). At a household level, the prevalence of peridomestic T. infestans decreased from 90.1% at baseline to 70.5% 1-year post-spraying. Sites treated with SC deltamethrin and manual sprayers showed the lowest prevalence of T. infestans 1-year post-spraying (24%), followed by those treated with SC deltamethrin (31%), EC deltamethrin (40%), and EC permethrin (53%) applied with power sprayers (Fig. 2 a)). The percentage of colonization by T. infestans showed a similar pattern as infestation but with slightly lower rates.
Infestation rates 1-year post-spraying increased markedly with the abundance of T. infestans per site before spraying (Fig. 2 b)d)). In peridomestic sites with 0 or 14 bugs per site at baseline, both SC deltamethrin-based treatments presented lower infestation rates (56% and 2023%, respectively) than EC-based treatments (1518% and 3344%, respectively) (Fig. 2 b) and Fig. 2 c)). In sites with five or more bugs before spraying, SC deltamethrin achieved a lower infestation rate when applied with manual rather than power sprayers (24% versus 39%), and both were more effective than EC-based treatments (4965%) (Fig. 2 d)).
At baseline, the prevalence of infestation by T. infestans peaked in goat or sheep corrals (70%), followed by storerooms or sheds (51%), chicken coops (46%), and kitchens (42%) (Table 1). One year post-spraying, goat or sheep corrals showed the largest mean infestation rate (57%), followed by chicken coops (33%), storerooms or sheds (28%), and pig pens (25%). Goat or sheep corrals represented 44% of all the infested, sprayed, and reinspected sites, but they contributed to 68% of the infestations detected.
Using multiple logistic regression, the relative odds of infestation 1-year post-spraying differed in a highly significant way among treatments but not between those based on SC deltamethrin (Table 2). The successive addition of the log-number of T. infestans per site before spraying, log-total surface, and log-number of domestic hosts improved the fit of the model in a highly significant way, whereas the spray team, treatment opportunity, and type of ecotope exerted no significant effects on the likelihood of infestation. Addition of the random parameter measuring household clustering effects improved the fit significantly (P = 0.008). Only the interaction between treatment and bug abundance before spraying was statistically significant. Logistic regression results were qualitatively similar when all 799 peridomestic sites were included in the analysis.
The relative reductions of T. infestans abundance 1-year post-spraying achieved by SC deltamethrin applied with manual (76.1%) or power sprayers (76.4%) were significantly greater than those achieved by EC deltamethrin (62.6%) or EC permethrin (48.7%) (Table 3). Median T. infestans catches per infested site 1-year post-spraying increased from 2 bugs for SC deltamethrin (first-third quartiles, Q1Q3, 16 bugs) to 5 (18) and 6 (311) bugs for EC deltamethrin and EC permethrin, respectively. Multiple linear regression of the log-number of T. infestans collected per site 1-year post-spraying identified the same set of significant predictors as logistic regression (data not shown) and, additionally, a highly significant interaction between surface and host numbers. The total bug catch increased significantly with both total surface and local number of hosts in unroofed ecotopes, but not in storerooms, sheds, or kitchens (Fig. 3).
The insecticidal activity of pyrethroids differed significantly on cement blocks (c2 = 3.63; df = 3; P <0.0001) but not on wooden posts (c2 = 2.99; df = 3; P >0.3) (Fig. 4). SC deltamethrin killed all the insects exposed to sprayed cement blocks, but none of the nymphs exposed to wooden posts. Bug survival at 75-days post-exposure to EC pyrethroids was very high both on wooden posts (7380%) and cement blocks (6373%). The surviving bugs were able to feed and most reached the adult stage.
Discussion
The standard treatment proved to be the most effective, but it failed to eliminate T. infestans in peridomestic sites just 1-year post-spraying. The ranking of treatment effectiveness was robust to changes in nozzle tips during the trial, spray team composition, treatment opportunity, types of regression analysis, and outcome variables. The estimates of the relative reduction in bug abundance actually underestimated the true impact of all insecticides because baseline collections were conducted when T. infestans populations and temperatures were decreasing (mid-autumn), whereas the post-spraying evaluation was carried out when bug populations were increasing and their flushing-out response to the irritant spray was enhanced (early summer). Considering effectiveness, cost, need of fuel, and low acceptability by spraymen, power sprayers did not provide any net benefit relative to manual sprayers in the control of T. infestans in peridomestic ecotopes. Randomization of treatments within clusters of houses and blind assessment procedures are distinctive features of this large-scale trial.
Residual foci
Several pieces of evidence clearly suggest that most of the infestations originated from triatomine bugs that survived exposure to the insecticides at each site (i.e. residual foci). First, the short-lasting residual effects of pyrethroids in peridomestic sites was well below the time taken for T. infestans eggs to hatch (range, 821 days) at local ambient temperatures (15). Marked exposure to sunlight and high temperature induce photolysis of pyrethroid molecules and thus reduce insecticidal activity (8). Additionally, power sprayers lifted dust that eventually would deposit on the sprayed surface and mask the insecticide. Second, a large fraction of all foci were high density and included late-instar nymphs. These instars, which may survive 1 year under fasting conditions and take 36 months to develop from eggs depending on local temperatures and host availability (15), have lower susceptibility to pyrethroids than younger instars (16). Third, the spatial distribution of foci 1-year-post-spraying was apparently equally widespread over the whole district. Fourth, the effectiveness of peridomestic treatments was modified by the abundance of T. infestans per site before spraying, as observed in another semiarid rural area (17, 18). The greater the abundance of bugs during their egg-laying season, the greater the likelihood that some of the late-instar nymphs or eggs inside crevices or hollow logs would emerge after the insecticides' residual effects waned and initiate a new generation. Clearly, the effectiveness of pyrethroids on triatomine bug populations is bug-density dependent.
Peridomestic foci of T. infestans detected just 13 months after applying SC pyrethroids with manual sprayers (58, 18) were most probably residual foci, but molecular methods are needed to provide conclusive evidence regarding the source of reinfestants. Other sources of reinfesting bugs were much less likely in our study. The likelihood of invasion by T. infestans by flight or passive transport from adjacent untreated areas was extremely unlikely given the large spraying coverage (95%). In the past, sylvatic colonies of T. infestans in semiarid rural Argentina were very rarely found despite intensive searches (15). The significant effects of household clustering on the likelihood of infestation suggests that a fraction of all foci detected 1-year after spraying might have originated from T. infestans dispersing by flight from nearby sites treated with the least effective treatments, or from other unidentified cluster-specific attributes. The small percentage of infested sites found among those that had been negative before spraying may be explained by the limited sensitivity of the capture method in peridomestic sites (19), or by new colonizations that occurred between the baseline survey in May and interventions in December. Because T. infestans collected from 14 villages in our study area were fully susceptible to pyrethroids (20), resistance cannot explain the low degree of effectiveness of the tested pyrethroids in peridomestic sites.
Comparison with other studies
At a household level, peridomestic infestation rates 1-year post-spraying were much greater than those recorded using the same insecticides, doses, and qualified staff in other semiarid rural areas (47, 18). As these trials were conducted under cooler temperatures (in MayOctober) than ours, the observed variation in effectiveness may be related to the inverse temperature-dependence of the insecticidal activity of pyrethroids (8). This important factor has seldom been considered and is especially relevant for peridomestic ecotopes (21). It is noteworthy that the ranking of post-treatment infestation of peridomestic ecotopes matched their lack of capacity to damp temperature fluctuations, which increased from goat corrals to storerooms and human habitations (21). The joint effects of temperature, exposure to sunlight and dust, and type of substrate in field operations will very likely predominate over small variations in effectiveness among insecticides determined in laboratory settings.
SC pyrethroids showed highly variable (112 months) residual activity against T. infestans on indoor porous or mud walls protected from sunlight (4, 5, 22, 23), but we failed to find any data for peridomestic structures. Although our bioassays were limited, the average residual effects of the insecticides were consistent with their ranking of effectiveness. Moreover, the very poor residual effects of all pyrethroids on exposed wooden posts were consistent with the high rates of infestation observed before and after the blanket spraying, especially in such corrals. Typically made of wooden posts, hollow logs, and piled branches of thorny scrub in semiarid regions, goat or sheep corrals provide abundant refuges for the bugs and host a sizable number of animals year-round. Both refuge and host availability significantly increased the population abundance and growth rate of T. infestans in experimental huts under natural climatic conditions (24, 25). Consistent with this evidence, our trial is the first to show that the joint effects of type of material, total surface, and host numbers modify the likelihood of local infestation and abundance of T. infestans in peridomestic sites after spraying with pyrethroids. Clearly, the study goat or sheep corrals are key sites for maintaining residual populations of T. infestans after insecticide spraying.
Conclusion
The difficulty of eliminating T. infestans from heavily treated areas in the semiarid Chaco may be explained by the joint effects of high-density infestation, numerous peridomestic structures with materials that provide refuges for the bugs, high temperature, exposure to sunlight and dust, and probably inappropriate insecticide application dosages, frequency, and timing. The spray coverage and surveillance of sparsely populated, impoverished rural areas with frequent migration is complex and also contributes to persistent infestations. Blanket or selective insecticide sprays are frequently conducted during the hot season, when triatomine bugs increase in numbers and become more apparent, but unfortunately under these conditions the effectiveness of pyrethroid sprays is strongly reduced. The remedial nature of such actions is enhanced in community-based control programmes against T. infestans because rural villagers do not perceive peridomestic infestations as a direct nuisance or hazard to their animals or themselves. Early and persistent peridomestic infestation after spraying with the standard treatment against T. infestans (46, 8), Triatoma brasiliensis, and Triatoma pseudomaculata in Brazil (9, 26), and Triatoma pallidipennis in Mexico suggests that this may be a generalized problem in arid or semiarid areas, although reinfestations by several species other than T. infestans may also be driven by invasion from sylvatic foci. The evidence herein provided demonstrates that the current tactics and procedures fail to eliminate peridomestic populations of T. infestans in semiarid rural areas and need to be revised. Until more cost-effective and environmentally friendly tools become available, triatomine control programmes may improve the use of available or related SC pyrethroids using specific tactics tailored to the peridomestic environment. This would include determination of optimal dosages, frequency, and timing of insecticide spraying; reinforced surveillance of key peridomestic sites using simple sensing devices (12); and development of environmental management methods aimed at reducing bug abundance in key structures housing domestic animals (2). 
Conflicts of interest: none declared.
Acknowledgements
We thank Aventis (ex Agrevo), Chemotecnica, Rentokil, Fundación Mundo Sano, R. Chuit, R. González Llanos, R. Di Lello, G. Vazquez-Prokopec, G. Wallace, and the ECLAT network for support. F. Petrocco, F. Kessler (Plagfree), and H. Fernández Alza kindly assisted field operations, and Kathe Rogerson provided editorial assistance. We also thank the NVCP and La Rioja staff who conducted bug assessments and insecticide spraying. REG, ODS, and ELS are members of the Researcher Career from CONICET, Argentina.
References
1. Schofield CJ, Dias JCP. The Southern Cone Initiative against Chagas disease. Advances in Parasitology 1999;42:1-27.
2. WHO. Vector control. Methods for use by individuals and communities. Geneva: World Health Organization; 1998. p. 326.
3. Schofield CJ. Field testing and evaluation of insecticides for indoor residual spraying against domestic vectors of Chagas disease. Geneva: World Health Organization; 2001. WHO document WHO/CDS/WHOPES/GCDPP/2001.1.
4. Gualtieri JM, Ríos CH, Cichero JA, Váez R, Carcavallo RU. Ensayo de campo con Decametrina en su formulación líquido emulsionable y floable en el control del Triatoma infestans en la Provincia de Córdoba. [Field trial of Decamethrin in emulsifiable and flowable liquid formulation in the control of Triatoma infestans in Cordoba Province.] Chagas (Córdoba) 1984;1:17-20. In Spanish.
5. Guillén G, Díaz R, Jemio A, Cassab JA, Teixeira Pinto C, Schofield CJ. Chagas disease vector control in Tupiza, Southern Bolivia. Memorias do Instituto Oswaldo Cruz 1997;92:1-8.
6. Cecere MC, Gürtler RE, Canale D, Chuit R, Cohen JE. The role of the peridomiciliary area in the elimination of Triatoma infestans from rural Argentine communities. Pan American Journal of Public Health 1997;1:273-9.
7. Zerba EN, Wallace G, Picollo MI, Casabe N, Licastro S, Wood E, et al. Evaluation of beta-cypermethrin for the control of Triatoma infestans. Revista Panamericana de Salud Pública 1997;1:133-7.
8. Roussel Uclaf. Deltamethrin. Avignon: l'Imprimerie Aubanel Press; 1982. p. 268-71.
9. Oliveira Filho AM. Recent advances in the use of slow-release insecticide formulations against triatomines. Revista de la Sociedade Brasileira de Medicina Tropical 1995;28 Suppl 3:74-8.
10. Canale DM, Cecere MC, Chuit R, Gürtler RE. Peridomestic distribution of Triatoma garciabesi and Triatoma guasayana in north-west Argentina. Medical and Veterinary Entomology 2000;14:383-90.
11. Perich MJ, Hoch AL, Rizzo N, Rowton ED. Insecticide barrier spraying for the control of sand fly vectors of cutaneous leishmaniasis in rural Guatemala. American Journal of Tropical Medicine and Hygiene 1995;52:485-8.
12. Vazquez-Prokopec GM, Ceballos LA, Salomón OD, Gürtler RE. Field trials of an improved cost-effective device for detecting peridomestic populations of Triatoma infestans (Hemiptera: Reduviidae) in rural Argentina. Memorias do Instituto Oswaldo Cruz 2002;97:971-7.
13. Smith PG, Morrow RH. Methods for field trials of interventions against tropical diseases: a toolbox. Oxford: Oxford Medical Publications; 1991.
14. Egret. Epidemiological graphics, estimation and testing package. Seattle: Statistics and Epidemiology Research Corporation; 1993.
15. Canale DM, Carcavallo RU. Triatoma infestans (Klug). In: Carcavallo RU, Rabinovich JE, Tonn RJ, editors. Factores biológicos y ecológicos en la Enfermedad de Chagas. Buenos Aires: Servicio Nacional de Chagas; 1985. p. 237-50.
16. Zerba EN, de Licastro SA, Wood EJ, Picollo de Villar MI. Insecticides: mechanism of Action. In: Brenner RR, Stoka AM, editors. Chagas' disease vectors. Boca Raton: CRC Press; 1987. p.101-23.
17. Gürtler RE, Petersen RM, Cecere MC, Schweigmann NJ, Chuit R, Gualtieri JM, et al. Chagas disease in north west Argentina: risk of domestic reinfestation by Triatoma infestans after a single community wide application of deltamethrin. Transactions of the Royal Society of Tropical Medicine and Hygiene 1994;88:27-30.
18. Cecere MC, Gürtler RE, Canale DM, Chuit R, Cohen JE. Effects of partial housing improvement and insecticide spraying on the reinfestation dynamics of Triatoma infestans in rural northwestern Argentina. Acta Tropica 2002;84:101-16.
19. Gürtler RE, Vazquez Prokopec GM, Ceballos LA, Lund Petersen C, Salomón OD. Comparison between two artificial shelter units and timed manual collections for detecting peridomestic Triatoma infestans (Hemiptera: Reduviidae) in rural northwestern Argentina. Journal of Medical Entomology 2001;38:429-36.
20. Picollo MI. Avances en el monitoreo de resistencia en Triatominos y necesidades futuras. [Advances in monitoring resistance in Triatominae and future needs.] In: Red Latinoamericana de Control de Triatominos. Monitoreo de la resistencia a insecticidas en Triatominos en América Latina. Buenos Aires: Fundación Mundo Sano; 2001. p.13-20. In Spanish.
21. Vazquez-Prokopec GM, Ceballos LA, Cecere MC, Gürtler RE. Seasonal variations of microclimatic conditions in domestic and peridomestic habitats of Triatoma infestans (Hemiptera: Reduviidae) in rural northwest Argentina. Acta Tropica 2002;84:229-38.
22. Diotaiuti L, Pinto CT. Biological susceptibility of Triatoma sordida and Triatoma infestans to deltamethrin and lambdacyhalothrin under field conditions. Revista de la Sociedade Brasileira de Medicina Tropical 1991;24:151-5.
23. Ferro EA, Arias AR, Ferreira ME, Simancas LC, Rios LS, Rosner JM. Residual effect of lambdacyhalothrin on Triatoma infestans. Memorias Instituto Oswaldo Cruz 1995;90:415-19.
24. Cecere MC, Canale DM, Gürtler RE. Effects of refuge availability on the population dynamics of Triatoma infestans in central Argentina. Journal of Applied Ecology 2003;40:742-56.
25. Gorla DE, Schofield CJ. Population dynamics of Triatoma infestans under natural climatic conditions in the Argentine Chaco. Medical and Veterinary Entomology 1989;3:179-94.
26. Diotaiuti L, Faria Filho O, Carneiro FCF, Dias JCP, Pires HHR, Schofield CJ. Aspectos operacionais do controle do Triatoma brasiliensis. Cadernos de Saúde Pública 2000;16:7-14.
Submitted: 04 April 02
Final revised version received: 04 March 03
Accepted: 03 July 03
1 Correspondence should be sent to this author.