ABSTRACT.
Mexico has shown an increase in dengue incidence rates. There are factors related to the location that determine housing infestation by Aedes. This study aimed to determine factors associated with housing infestation by immature forms of Aedes spp. in the dengue endemic localities of Axochiapan and Tepalcingo, Mexico, from 2014 to 2016. A cohort study was carried out. Surveys and inspections of front- and backyards were conducted every 6 months, looking for immature forms of Aedes spp. A house condition scoring scale was developed using three variables (house maintenance, tidiness of the front- and backyards, and shading of the front- and backyards). Multiple and multilevel regression logistic analysis were conducted considering the housing infestation as the outcome and the household characteristics observed 6 months before the outcome as factors; this was adjusted by time (seasonal and cyclical variations of the vector). The infestation oscillated between 5.8% of the houses in the second semester of 2015 and 29.3% in the second semester of 2016. The factors directly associated with housing infestation by Aedes were the house condition score (adjusted odds ratio [aOR]: 1.64; 95% CI: 1.40–1.91) and the previous record of housing infestation (aOR: 2.99; 95% CI: 2.00–4.48). Moreover, the breeding-site elimination done by house residents reduced the housing infestation odds by 81% (95% CI: 25–95%). These factors were independent of the seasonal and cyclical variations of the vector. In conclusion, our findings could help to focalize antivectorial interventions in dengue-endemic regions with similar demographic and socioeconomic characteristics.
INTRODUCTION
Dengue virus (DENV) is transmitted by Aedes genus mosquitoes, and the disease caused by this virus is rapidly expanding worldwide, having a high socioeconomic impact and becoming a global public health problem.1 According to the surveillance system, Mexico has had an increase in dengue incidence rates from a median of 18,037 annual confirmed cases in the 1980 s to 127,506 confirmed cases between 2009 and 2018. In addition, the cases appear to be more severe.2
The dengue expansion and incidence rate increase are the result, at least in part, of the anthropophilic behavior of the vector. For example, the spread of Aedes aegypti is related to the increase in urbanization and how water is supplied in urban settlements.3–5 Moreover, there are factors related to the location that determine housing infestation, such as shade, exposure to sunlight, vegetation, temperature, and humidity. In addition, some community practices are considered protective factors, such as mosquito screens in doors and windows and participation in public health education activities.6 However, these characteristics, as well as their association with housing infestation by Aedes, vary among localities, most likely due to microclimate factors.
These considerations and the need to execute control policies have encouraged the development of research to mitigate dengue endemicity. One research topic has been to identify the infestation dynamics in endemic localities to achieve effective vector control. Therefore, this study aimed to determine the factors associated with housing infestation by immature forms of Aedes spp. in two dengue-endemic Mexican localities over 2.5 years. Our analysis included the development of a scoring scale based on house conditions.
MATERIALS AND METHODS
Study area and sample size.
This work was a secondary analysis of the data obtained from a prospective cohort study done in the Axochiapan and Tepalcingo localities in the state of Morelos, Mexico. The cohort was assembled in the project “Peridomestic Infection as a Determinant of DENV Transmission between 2011 and 2012” (stage 1),7 and it was followed up by the subsequent project “Seroprevalence, Neutralizing Titers, and Dengue Incidence Rates on an Endemic Population from 2014 to 2016” (stage 2). Both studies were designed and performed by our research team at the Instituto Nacional de Salud Pública (INSP) of Mexico. All houses that participated in stage 17 and surveyed in at least two consecutive times during stage 2 were included (Figure 1).
Figure 1.
Study design and follow-up.
The cohort assembled in 2011 included an exposed group consisting of clusters of up to five houses (one in which a confirmed dengue case occurred and four neighboring houses within a 50-m radius), and an unexposed group consisting of clusters of up to five houses within a 50-m radius in areas where no confirmed dengue cases were reported within a 100-m radius in the two months before the sampling day. Per each exposed cluster, an unexposed cluster was enrolled in the same location during the following 3 weeks.7
Axochiapan and Tepalcingo are dengue endemic localities. Both localities meet the Official Mexican Standard NOM-032-SSA2-2010, which requires state health services to implement permanent larval control (temephos) and dejunking, as well as focalized fumigation in peridomestic areas where the epidemiological surveillance system identifies suspected dengue cases. Also, areas with high dengue incidence are nebulized.8
According to the Instituto Nacional de Estadística y Geografía, Secretaria de Desarrollo Social, and Consejo Nacional de Evaluación de la Política de Desarrollo Social, as of 2015, Tepalcingo is an urban locality with 12,895 inhabitants; 9.3% of the population older than 14 years is illiterate; 71.1% of dwellings are connected to the water public distribution network, and 76.9% have a sewer system; the extreme poverty is 18.1%; and the Gini coefficient is 0.438. Axochiapan is an urban locality with 18,659 inhabitants; 10.2% of the population older than 14 years is illiterate; 54.6% of dwellings are connected to the water public distribution network, and 78.3% have a sewer system; the extreme poverty is 30.8%; and the Gini coefficient is 0.431.
Data collection.
An adult resident from each house was surveyed every 6 months for up to five surveys. The surveys included entomological and sociodemographic factors, housing characteristics, and an inspection of the front- and backyards and water containers looking for Aedes larvae and pupae. A house was considered infested when Aedes larvae or pupae were detected during inspection. All surveys were done by the same staff specialized in entomology. Maintenance of the house and the backyard, and shaded areas in front- and backyards were also evaluated based on the description made by Tun-Lin et al.9 Houses considered not well maintained were those having a poor structure, poorly organized, dirty, and with unpainted or cracked walls, improvised sections, or broken windows/doors. Untidy front- and backyards were those in disarray and covered with trash or junk and overgrown grass or vegetation. Shaded front- and backyards were those having a total shaded area exceeding 50% of the total area.
Data analysis.
Each evaluation involved calculating the Breteau index (BI; both at the aggregate and locality level) using the formula BI = ((positive recipients for Aedes larvae or pupae × 100)/Number of houses), and the House index (HI), which was calculated as the percentage of houses that were infested. The Mann–Whitney U test was used to compare the BI and χ2 to compare HI.
A panel database was prepared to determine associated factors with housing infestation (outcome). House characteristics observed 6 months before the outcome evaluation were considered independent variables (Figure 1). We represented the relationships between determinants of infestation in a directed acyclic graph (DAG) using DAGitty software.10 In this way, we defined three domains directly related to the infestation, including climatic factors, measures to eliminate breeding sites, and housing conditions (Figure 2). These domains were evaluated with various variables. The selection of the variables and their functional form were performed using statistical procedures.
Figure 2.
Recreated directed acyclic graph. Black = outcome; Green = exposure; Blue = mediator variables; Gray = unobserved variables.
Regarding housing conditions, a scale was developed, including the maintenance of the house, front- and backyard tidiness, and the front- and backyard shaded areas.9 The scale calculation considered the coefficients obtained during the univariable analysis, which were multiplied by two and rounded to the nearest integer. Then, the following values were assigned: house condition good = 0, moderate = 2, not properly maintained = 3; front- and backyard tidiness good = 0, moderate = 1; untidy = 4; shade in front- and backyards limited or no shade = 0, some shade (25–50%) or shaded (> 50%) = 2. A single score per house between 0 and 9 was obtained by adding the values of these three variables.
To consider the climatic factors, we constructed a categorical variable that represented the cyclical variation of BI and HI (Figure 3). This variable had three categories, 1) high infestation semesters included the 2014 survey and the second survey in 2016, 2) moderate infestation semesters included the first surveys in 2015 and 2016, and 3) low infestation semester that was the second survey in 2015. Also, to consider the seasonal variations of the vector, models were adjusted by the exposition month and the squared of this month.
Figure 3.
Breteau and House indices during follow-up. (A) The Breteau index (BI). (B) The House index (HI).
To identify the variables within each domain, a univariable multilevel logistic regression analysis was done considering the house as the grouping level, and variables with a P < 0.20 in the univariable analysis were tested in multiple multilevel models. We performed a backward selection to obtain a model including those variables showing a significant association (P < 0.05) and those variables that changed the OR of associated variables by more than 10%.
Because each house was evaluated up to five times, autocorrelation was a potential problem in this study. Therefore, we evaluated the correlation between the residuals of consecutive surveys and between the residuals and the order of surveys. The residuals were calculated as the observed outcome minus (exp[logit]/(1 + exp[logit])). There was neither correlation between the consecutive residuals nor between residual and visit order (Pearson’s coefficients: −0.0002, P = 0.9960; and 0.0231, P = 0.4486, respectively). The statistical analyses were carried out using the STATA 16 software (StataCorp, College Station, TX).
Interpretation.
We considered the dichotomous variable of infestation as the outcome, which was measured repeatedly in the cohort houses. We chose both the housing condition (measured with the scale) and the activities to eliminate the vector as exposures of interest. To estimate the total effect of these exposures, an adjustment was made for the variables representing the climatic conditions, as suggested by the DAG.
The previous infestation variable, which was measured simultaneously with the exposures of interest, was considered a mediator. This assumption was based on the fact that exposures occurred before the survey, and the infestation was measured directly during the survey. Therefore, the exposure can precede and affect the previous infestation, and the last one can be a direct determinant of the outcome (i.e., infestation measured 6 months later). Consequently, a second multiple regression model was built, including the previous infestation. In this model, measures of the association of housing conditions and actions against the vector were considered as the direct effects.
RESULTS
During the study period, 316 houses were evaluated in the first survey and 273 houses during the fifth survey. The multilevel model considered 313 houses which had been followed up at least two consecutive times during the study period (stage 2 of the cohort study). During the first survey, the median of residents per house was five (interquartile range [IQR]: three to six; minimum: one, maximum: 24). Additionally, the average number of rooms per house was four (standard deviation: two).
Regarding housing conditions, the majority of houses had drains connecting to the public sewage network in all surveys. However, some houses had septic tanks, and more than 50% of toilets were still emptied by manually pouring water. During inspections, most houses had no mosquito screens on access doors/windows. Also, mosquito screens were broken in more than 20% of houses that did have screens. Moreover, there were pots in backyards of most houses. In general, the majority of houses were properly maintained, their front- and backyards were tidy, and their shaded areas comprised 25% to 50% of the total surface area of front- and backyards (Table 1).
Table 1.
Characteristics of houses and antivector interventions per survey
| Characteristic/survey | n (%) | ||||
|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | |
| Drainage | N = 317 | N = 308 | N = 298 | N = 278 | N = 273 |
| Public network | 291 (91.8) | 280 (90.9) | 276 (92.6) | 259 (93.2) | 254 (93.0) |
| Septic tank | 23 (7.3) | 25 (8.1) | 19 (6.4) | 17 (6.1) | 16 (5.9) |
| Ravine | 3 (0.9) | 3 (0.97) | 3 (1.0) | 2 (0.7) | 3 (1.0) |
| Toilet | N = 317 | N = 308 | N = 297 | N = 277 | N = 272 |
| Direct water discharge | 128 (40.3) | 121 (39.2) | 118 (39.7) | 112 (40.3) | 111 (40.8) |
| Manual water discharge | 187 (58.9) | 184 (59.7) | 177 (59.6) | 163 (58.8) | 159 (58.5) |
| No water | 2 (0.63) | 3 (0.97) | 2 (0.67) | 2 (0.72) | 2 (0.73) |
| Mosquito screen on access doors | N = 315 | N = 305 | N = 296 | N = 276 | N = 271 |
| No | 225 (71.4) | 213 (69.8) | 216 (73.0) | 201 (72.8) | 197 (72.7) |
| All | 47 (14.9) | 46 (15.1) | 39 (13.2) | 38 (13.8) | 36 (13.2) |
| Some | 43 (13.6) | 46 (15.1) | 41 (13.8) | 37 (13.4) | 38 (14.0) |
| Condition of mosquito screens on doors* | N = 84 | N = 35 | N = 75 | N = 75 | N = 74 |
| Broken | 24 (28.6) | 7 (20) | 21 (28) | 23 (30.7) | 23 (31.1) |
| Mosquito screen on access windows | N = 316 | N = 306 | N = 296 | N = 276 | N = 271 |
| No | 187 (59.2) | 177 (57.8) | 180 (60.8) | 169 (61.2) | 164 (60.5) |
| All | 68 (21.5) | 67 (21.9) | 59 (19.9) | 56 (20.3) | 53 (19.6) |
| Some | 61 (19.3) | 62 (20.3) | 57 (19.3) | 51 (18.4) | 54 (19.9) |
| Condition of mosquito screens on windows* | N = 127 | N = 62 | N = 116 | N = 106 | N = 106 |
| Broken | 32 (25.2) | 13 (21) | 32 (27.6) | 29 (27.4) | 29 (27.4) |
| Pots in front yards | N = 308 | N = 299 | N = 294 | N = 278 | N = 273 |
| Yes | 26 (8.4) | 38 (12.7) | 31 (10.5) | 31 (11.2) | 29 (10.6) |
| Pots in backyards | N = 308 | N = 302 | N = 295 | N = 278 | N = 273 |
| Yes | 295 (95.8) | 253 (83.8) | 281 (95.3) | 269 (96.8) | 263 (96.3) |
| House maintenance | N = 314 | N = 308 | N = 296 | N = 277 | N = 273 |
| Good | 218 (69.4) | 252 (81.8) | 208 (70.3) | 200 (72.2) | 193 (70.7) |
| Moderately good | 86 (27.4) | 48 (15.6) | 76 (25.7) | 66 (23.8) | 69 (25.3) |
| Not properly maintained | 10 (3.2) | 8 (2.6) | 12 (4.1) | 11 (4.0) | 11 (4.0) |
| Front and backyards tidiness | N = 310 | N = 307 | N = 296 | N = 278 | N = 273 |
| Tidy | 165 (53.2) | 213 (69.4) | 149 (50.3) | 143 (51.4) | 134 (49.1) |
| Moderately tidy | 131 (42.2) | 86 (28.0) | 125 (42.2) | 109 (39.2) | 115 (42.1) |
| Untidy | 14 (4.5) | 8 (2.6) | 22 (7.4) | 26 (9.3) | 24 (8.8) |
| Shade percentage in front and backyards | N = 303 | N = 229 | N = 294 | N = 277 | N = 273 |
| Shade > 50% | 17 (5.6) | 36 (15.7) | 14 (4.8) | 10 (3.6) | 9 (3.3) |
| Some shade (25–50%) | 186 (61.4) | 185 (80.8) | 182 (61.9) | 174 (62.8) | 169 (61.9) |
| Limited or no shade (< 25%) | 100 (33) | 8 (3.5) | 98 (33.3) | 93 (33.6) | 95 (34.8) |
| Antivector measures | N = 318 | N = 309 | N = 297 | N = 278 | N = 272 |
| Performed by residents | 117 (36.9) | 113 (36.5) | 183 (61.6) | 152 (54.6) | 131 (48.2) |
| House fumigation | 73 (62.4) | 35 (31.0) | 82 (44.8) | 46 (30.2) | 39 (29.8) |
| Eliminating breeding sites | 12 (10.2) | 18 (15.9) | 20 (10.9) | 11 (7.2) | 27 (20.6) |
| Other antivector measures | 55 (47) | 72 (63.7) | 135 (73.8) | 123 (80.9) | 114 (87) |
| Performed by public authorities | 93 (44.3) | 117 (38.4) | 189 (65.4) | 174 (69.6) | 102 (39.9) |
| Temephos application | 83 (89.2) | 112 (95.7) | 187 (98.9) | 148 (85.1) | 99 (97.1) |
| Fumigation | 19 (20.4) | 35 (29.9) | 82 (43.4) | 46 (26.4) | 39 (38.2) |
| Dejunking | 3 (3.2) | 0 | 2 (1.0) | 18 (10.3) | 5 (4.9) |
Denominators correspond to houses that had mosquito screens on windows or doors and that were inspected during each survey.
After evaluating vector-control activities done by residents, fumigation was one of the most frequent activities (30–60%) during surveys. Moreover, residents used other vector-control activities (larvivorous fish, smoke as repellent, mosquito nets, and repellent lotions) during the last three surveys (47–87%). Regarding vector-control activities done by public authorities at the municipal and state levels, temephos applications were the most commonly used (85.1–98.9%), followed by fumigation (Table 1). The highest BI was observed in the first survey, followed by the fifth, showing similar results in both localities (P = 0.59) (Figure 3A). The HI showed the highest infestation during the last survey and the lowest infestation during the third survey. The second, third, and fourth surveys showed significantly lower infestation frequencies compared with the first (Figure 3B).
In all surveys, the three main infested types of containers were cement tanks, barrels, and buckets. During the first survey, the most common type of infested container was cement tanks (27.7%; approximately 1,000 L), followed by barrels (25%; 100–200 L), and buckets (15%; containers of 18–20 L). The median of infested containers per household was 3.5 (IQR: 2–13). The same tendency was seen in the second survey for these containers, although the median was lower (1.5; IQR: 1–4). Similar results were observed during the third survey, where cement tanks accounted for 33.3%, and buckets and barrels for 16.6% of infested containers (median: 1 infested container/household; IQR: 0–2.5). In the fourth and fifth surveys, barrels were the most frequent infested containers (48% and 55%, respectively). Infested tires were only observed in the first and second surveys just reaching a frequency of 3%. Furthermore, infested pots were present during the first three surveys. Infested animal waterers, although at low percentages, were present all surveys. In addition, other containers, ranging from 4% to 11%, contributed to infestation indices.
The univariable analysis showed that the manual discharge of water in toilets compared with direct water discharge was associated with a higher risk of infestation. Regarding drainage availability, having a septic tank showed a trend of greater risk of infestation, but it was not significant. Also, eliminating breeding sites by residents was a protective factor, whereas having implemented intervention measures against the adult vector and applied barrier measures or other types of antivector interventions by residents were associated with a higher risk of infestation. The latter measures included the use of smoke repellent, mosquito nets, homemade insecticides, and fish in water tanks (Table 2). Other activities carried out by residents, such as fumigation, as well as activities carried out by public authorities in houses and neighborhoods were not related to the risk of infestation.
Table 2.
Factors associated with housing infestation 6 months later
| House factor | No. obs/no. houses | Univariable OR (95% CI) | P | Total effect aOR* (95% CI) | P | Direct effect aOR* (95% CI) | P |
|---|---|---|---|---|---|---|---|
| House condition scale†‡ | 1,086/314 | 1.84 (1.56–2.18) | < 0.001 | 1.80 (1.50–2.16) | < 0.001 | 1.64 (1.40–1.91) | < 0.001 |
| Eliminating breeding sites in house | 1,117/315 | 0.32 (0.11–0.96) | 0.042 | 0.19 (0.05–0.75) | 0.018 | 0.25 (0.07–0.89) | 0.032 |
| Infestation history | 1,114/314 | 3.30 (2.31–4.73) | < 0.001 | – | – | 2.99 (2.00–4.48) | < 0.001 |
| Drainage | |||||||
| Public network | 1,117/314 | Ref. | – | – | – | – | – |
| Septic tank | 1.84 (0.93–3.62) | NS | – | – | – | – | |
| Ravine or crevice | 2.34 (0.43–12.7) | NS | – | – | – | – | |
| Toilet | |||||||
| Direct water discharge | 1,118/315 | Ref. | – | – | – | – | – |
| Manual water discharge | 1.53 (1.03–2.28) | 0.033 | – | – | – | – | |
| No water | 2.48 (0.32–18.75) | 0.378 | – | – | – | – | |
| Any intervention over the past month | 1,117/315 | 1.50 (1.05–2.13) | 0.023 | – | – | – | – |
| Other vector control measure | 1,117/315 | 1.51 (1.05–2.17) | 0.025 | – | – | – | – |
| Barrier measures | 1,118/315 | 2.30 (1.38–3.69) | 0.001 | – | – | – | – |
aOR = adjusted odds ratio; Obs = observations; OR = odds ratio.
Adjusted model included a total of 313 houses with at least two evaluations done within 6 months. The model was adjusted by cyclical and seasonal variation of the vector.
The scale was developed using the following values. House condition: good = 0; moderate = 2; not properly maintained = 3. Front- and backyard tidiness: good = 0; moderate = 1; untidy = 4. Shade in front- and backyards: limited or no shade = 0; some shade (25–50%) = 2; shaded (> 50%) = 2.
Scores between 0 and 4 were considered as a 0, score of 5 as 1, score of 6 as 2, and so on. Thus, for a score > 4 points, the odds of infestation increased by 80% per each point in the house condition score, which was considered the total effect. The direct effect was 64% odds increase per each point > 4. Bold data represents significant association.
Concerning inspections, the moderately appropriate (odds ratio [OR]: 2.5; 95% CI: 1.64–3.7) or inappropriate (OR: 4.7; 95% CI: 2.0–11.2) house maintenance, and moderately tidy (OR: 1.86; 95% CI: 1.28–2.7) or an untidy (OR: 9.4; 95% CI: 5.02–17.6) front- and backyards were factors associated with housing infestation. In both cases, a dose–response relationship was identified according to the change in category. Also, the presence of shaded areas in front- and backyards showed an increasing trend in relation to Aedes housing infestation. By grouping these three variables and assigning a house condition score, an increase in the infestation frequency was observed when the score was ≥ 5 points (Figure 4).
Figure 4.
Aedes spp. infestation about house condition score.
The multiple analysis showed that the estimated total effect for the house condition score was an 80% increase in the house infestation odds 6 months later per each incremental point in the score above the 4-point score. On the other hand, the estimated total effect for the breeding-site elimination by residents was an 81% decrease in the house infestation odds. In addition, the history of infestation (6 months before) was strongly associated with housing infestation 6 months later (OR: 2.99; 95% CI: 2.00–4.48), whereas the estimated direct effect for the house condition score was a 64% increase of the house infestation odds per each incremental point in the score, and the estimated direct effect for the elimination of breeding sites was a 75% decrease in the house infestation odds. These associations were independent of cyclical and seasonal variations of the vector (Table 2).
DISCUSSION
Aedes indices evaluated in the Axochiapan and Tepalcingo localities showed cyclical and seasonal variations because parabolic trend and monthly changes during the study period were observed. Moreover, the month and the semester of the survey were related to housing infestation. This phenomenon can be attributed to variations in temperature, exposure to sunlight, and relative humidity, which have been correlated to seasonal and cyclical changes in vector abundance.11–13 In addition, among the Aedes-infested containers, the most common were buckets, barrels, and cement tanks; this is similar to previous studies carried out in Cuautla (Mexico),14 Mérida (Mexico),15 and Matamoros (Mexico), and Brownsville (TX).16 Although pots were present in nearly all surveyed backyards, they were rarely infested, which differs from a study conducted in the United States, where pots were the most infested containers.17 Tires were rarely found in houses in both localities, more likely because inhabitants recognize them as potential breeding containers, so they are thrown away during local dejunking campaigns.
Regarding factors associated with housing infestation by immature forms of Aedes, our findings suggest that having an improperly maintained house, an untidy backyard, and a shaded area in the front- and backyards exceeding 25% of its total area (9 points in scale) increased the infestation by 8 times compared with houses with 0 points in the house condition scoring scale. These findings are similar to previous study in Queensland, Australia, where yard condition and shade were related to premises with Aedes aegypti immature forms because untidy premises were 2.5 times more likely to be infested than tidy premises and shaded premises were 2.6 times more likely to be infested than nonshaded premises.9 Also, improper maintenance of houses (adjusted OR: 1.64; 95% CI: 1.15–2.32) and poor hygiene have been associated with housing infestation in Cuba.18,19 Likewise, in shaded areas, Aedes infestation is up to 2.3 times compared with sunny areas under standard conditions.20 Similar findings were reported in Cuba, southern Mexico, and Colombia, where housing conditions, backyard maintenance, and shaded areas were associated with higher vector density.18,21,22 On the other hand, in Kampong Cham, Cambodia, the premise condition index was negatively related to the number of immature forms but was positively related to the number of adult females per house.23 Also, some cross-sectional studies have not found this relationship, possibly due to local and sociocultural factors.5
In addition, infested houses were found to have a greater probability of remaining infested or becoming infested again over the following 6 months, which emphasizes the importance of continuing with the entomological monitoring by public authorities focusing on houses with infestation history. Also, the identification of types of infested containers is important because it allows applying more control measures to them.24 Moreover, the analysis showed the elimination of breeding sites as a protection factor against housing infestation in the localities. This may be due to the residents recognizing the potential of disposable containers to become breeding sites, so they dispose of them.
House fumigations and the availability of water from the public distribution network were not related to housing infestation. These differ from other studies that found fumigation as a protective factor against Aedes infestation,25 and the no availability of water distribution networks as a risk factor.26 This discrepancy may be the result of an intermittent water distribution service in Axochiapan and Tepalcingo, so residents must store water for daily use. The use of organophosphorus larvicides (such as temephos, which was one of the most widely used methods by public authorities) and adulticides were also not related to housing infestation in these localities. This is consistent with findings in Nicaragua, where temephos was not related to a reduction in entomological indices.27 Moreover, the use of temephos was found to increase the risk of DENV infection in Mexico and Nicaragua.28 Consequently, it is necessary to study the impact of these chemicals on vector density as well as the possible mosquito resistance in these localities because resistance to temephos was up to 58% of Aedes larvae in Tapachula, México. Also, resistance to malathion, permethrin, and deltamethrin was between 50% and 90% of Aedes mosquitoes.29 In Tapachula as well as in Axochiapan and Tepalcingo, these chemicals have been frequently used.
The use of other vector control measures by residents, such as using smoke and mosquito coil repellents, was found to be an infestation risk factor as described by Morales-Pérez, who showed that the domestic use of insecticides increases the risk of DENV infection (up to 1.68 times) compared with houses where this measure was not applied.30 Similarly, there is no solid evidence to recommend peridomestic fumigation as the only effective control intervention.31
Regarding the BI and HI, they were high despite the various antivector activities carried out in Axochiapan and Tepalcingo. These indices are similar to those observed in the state of Guerrero, where BI ranged from 25 to 27, and HI ranged from 13% to 14%,30 and in Colombia were the HI ranged from 5% to 18%.5 Knowing the minimum value of Aedes indices to prevent dengue cases continues to be a challenge. For instance, a study in Brazil described that although BI was 4, dengue cases still occurred.32 Likewise, although the housing infestation in this study showed a decrease during the second semester in 2015, the number of dengue cases kept an endemoepidemic trend according to data reported by the surveillance system (64 confirmed dengue cases in 2014, 27 in 2015, and 46 in 2016). Nevertheless, it should be noted that the HI was previously related to DENV infection in Axochiapan and Tepalcingo ([aPR = adjusted prevalence rate]: 1.1; 95% CI: 1.03–1.17),33 similar to results reported in two longitudinal cohorts from Iquitos, Peru.34
Some limitations of this study are related to the nonevaluation of certain variables such as weather-related factors and other indices like the container, pupae, and adult indices. Weather-related variables were not specifically evaluated because each locality only has one weather station. However, the model was adjusted by cyclical and seasonal variation of the vector using the semester and month of each survey. Other indices, such as quantitative indices of pupae that are highly correlated to the numbers of the adults and adult indices,35 were not evaluated because their measurement is more complex and are beyond the scope of this study. Another limitation is that nonresidential places were not sampled because householders were interviewed and inspections inside the houses were conducted. Thus, the results of this study cannot be extrapolated to inhabited premises or other kind of places like vacant lots and construction worksites.
CONCLUSION
The house condition score was associated with a greater probability of housing infestation 6 months later, independently from seasonal and cyclic variations of the vector. Also, activities by the residents to eliminate vector breeding sites were associated with a lower probability for housing infestation. On the other hand, the history of infestation was an independent predictor of the same condition 6 months later. When we analyzed the infestation history as a mediator of the exposures of interest, the direct effect of house condition and antivectorial activities still stood strong, suggesting an independent durable impact.
Our findings bring up new research ideas for these localities, such as evaluating possible causes of the limited impact of temephos and fumigation on reducing vector infestation and developing alternative options such as educational campaigns to eliminate potential vector breeding sites, and to properly maintain front- and backyards. We consider that this risk indicator identified could be used to focalize local antivectorial interventions in dengue endemic areas with similar demographic and socioeconomic characteristics in Latin America, Asia, and Africa.
Financial Disclosure
This work was supported in part by Instituto Nacional de Salud Pública de México (CI-986/CI-494) and Sanofi Pasteur (DNG22-EXT) [J. R.-C.] and Universidad de Boyacá [A. L. M.-D.]. F. A. D.-Q. was granted a fellowship for research productivity from the Brazilian National Council for Scientific and Technological Development–CNPq, process/contract identification: 312656/2019-0. The American Society of Tropical Medicine and Hygiene (ASTMH) assisted with publication expenses.
REFERENCES
- 1. Gubler DJ, 2011. Dengue, urbanization and globalization: the unholy trinity of the 21st century. Trop Med Health 39: S3–S11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Pan American Health Organization and World Health Organization , 2022. Health Information Platform for the Americas. Available at: https://www3.paho.org/data/index.php/en/mnu-topics/indicadores-dengue-en/dengue-nacional-en/252-dengue-pais-ano-en.html?start=2. Accessed August 7, 2022.
- 3. Hiscox A, Kaye A, Vongphayloth K, Banks I, Piffer M, Khammanithong P, Sananikhom P, Kaul S, Hill N, Lindsay SW, Brey PT, 2013. Risk factors for the presence of Aedes aegypti and Aedes albopictus in domestic water-holding containers in areas impacted by the Nam Theun 2 Hydroelectric Project, Laos. Am J Trop Med Hyg 88: 1070–1078. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Morales-Pérez A. et al. , 2017. Aedes aegypti breeding ecology in Guerrero: cross-sectional study of mosquito breeding sites from the baseline for the Camino Verde trial in Mexico. BMC Public Health 17: 450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Overgaard HJ, Olano VA, Jaramillo JF, Matiz MI, Sarmiento D, Stenström TA, Alexander N, 2017. A cross-sectional survey of Aedes aegypti immature abundance in urban and rural household containers in central Colombia. Parasit Vectors 10: 356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Louis VR, Montenegro Quiñonez CA, Kusumawathie P, Palihawadana P, Janaki S, Tozan Y, Wijemuni R, Wilder-Smith A, Tissera HA, 2016. Characteristics of and factors associated with dengue vector breeding sites in the city of Colombo, Sri Lanka. Pathog Glob Health 110: 79–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Martínez-Vega RA, Danis-Lozano R, Velasco-Hernández J, Díaz-Quijano FA, González-Fernández M, Santos R, Román S, Argáez-Sosa J, Nakamura M, Ramos-Castañeda J, 2012. A prospective cohort study to evaluate peridomestic infection as a determinant of dengue transmission: protocol. BMC Public Health 12: 262. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. México , 2011. NORMA Oficial Mexicana NOM-032-SSA2-2010, Para la vigilancia epidemiológica, prevención y control de enfermedades transmitidas por vector. Diario Oficial Federacion. Available at: http://www.cenaprece.salud.gob.mx/programas/interior/vectores/descargas/pdf/nom_032_ssa2_2010_norma_petv.pdf. Accessed August 7, 2022.
- 9. Tun-Lin W, Kay BH, Barnes A, 1995. The Premise Condition Index: a tool for streamlining surveys of Aedes aegypti. Am J Trop Med Hyg 53: 591–594. [DOI] [PubMed] [Google Scholar]
- 10. Textor J, van der Zander B, Gilthorpe MS, Liskiewicz M, Ellison GT, 2016. Robust causal inference using directed acyclic graphs: the R package ‘dagitty’. Int J Epidemiol 45: 1887–1894. [DOI] [PubMed] [Google Scholar]
- 11. Dominguez MC, Ludueña FF, Almiron WR, 2000. Population dynamics of Aedes aegypti (Diptera: Culicidae) in Córdoba. Rev Soc Entomol Argent 59: 41–50. [Google Scholar]
- 12. Hayden MH, Uejio CK, Walker K, Ramberg F, Moreno R, Rosales C, Gameros M, Mearns LO, Zielinski-Gutierrez E, Janes CR, 2010. Microclimate and human factors in the divergent ecology of Aedes aegypti along the Arizona, U.S./Sonora, MX border. EcoHealth 7: 64–77. [DOI] [PubMed] [Google Scholar]
- 13. Barrera R, Amador M, MacKay AJ, 2011. Population dynamics of Aedes aegypti and dengue as influenced by weather and human behavior in San Juan, Puerto Rico. PLoS Negl Trop Dis 5: e1378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Villegas-Trejo A, Che-Mendoza A, González-Fernández M, Guillermo-May G, González-Bejarano H, Dzul-Manzanilla F, Ulloa-García A, Danis-Lozano R, Manrique-Saide P, 2011. Control enfocado de Aedes aegypti en localidades de alto riesgo de transmisión de dengue en Morelos, México. Salud Publica Mex 53: 141–151. [DOI] [PubMed] [Google Scholar]
- 15. Manrique-Saide P, Davies CR, Coleman PG, Rebollar-Tellez E, Che-Medoza A, Dzul-Manzanilla F, Zapata-Peniche A, 2008. Pupal surveys for Aedes aegypti surveillance and potential targeted control in residential areas of Mérida, México. J Am Mosq Control Assoc 24: 289–298. [DOI] [PubMed] [Google Scholar]
- 16. Ramos MM. et al. , 2008. Epidemic dengue and dengue hemorrhagic fever at the Texas–Mexico border: results of a household-based seroepidemiologic survey, December. Am J Trop Med Hyg 78: 364–369. [PubMed] [Google Scholar]
- 17. Walker KR, Williamson D, Carrière Y, Reyes-Castro PA, Haenchen S, Hayden MH, Jeffrey Gutierrez E, Ernst KC, 2018. Socioeconomic and human behavioral factors associated with Aedes aegypti (Diptera: Culicidae) immature habitat in Tucson, AZ. J Med Entomol 55: 955–963. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Bisset Lazcano JA, del Carmen Marquetti M, Portillo R, Rodríguez MM, Suárez S, Leyva M, 2006. Ecological factors linked to the presence of Aedes aegypti larvae in highly infested areas of Playa, a municipality belonging to Ciudad de La Habana, Cuba. Rev Panam Salud Publica 19: 379–384. [DOI] [PubMed] [Google Scholar]
- 19. Spiegel JM, Bonet M, Ibarra A-M, Pagliccia N, Ouellette V, Yassi A, 2007. Social and environmental determinants of Aedes aegypti infestation in Central Havana: results of a case-control study nested in an integrated dengue surveillance programme in Cuba. Trop Med Int Health 12: 503–510. [DOI] [PubMed] [Google Scholar]
- 20. Vezzani D, Albicocco AP, 2009. The effect of shade on the container index and pupal productivity of the mosquitoes Aedes aegypti and Culex pipiens breeding in artificial containers. Med Vet Entomol 23: 78–84. [DOI] [PubMed] [Google Scholar]
- 21. Manrique-Saide P. et al. , 2013. The risk of Aedes aegypti breeding and premises condition in south Mexico. J Am Mosq Control Assoc 29: 337–345. [DOI] [PubMed] [Google Scholar]
- 22. Vásquez-Trujillo A, Cardona-Arango D, Segura-Cardona AM, Portela-Câmara DC, Alves-Honório N, Parra-Henao G, 2021. House-level risk factors for Aedes aegypti infestation in the urban center of Castilla la Nueva, Meta State, Colombia. J Trop Med 2021: 8483236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Hustedt J. et al. , 2020. Ability of the premise condition index to identify premises with adult and immature Aedes mosquitoes in Kampong Cham, Cambodia. Am J Trop Med Hyg 102: 1432–1439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Maciel-de-Freitas R, Lourenço-de-Oliveira R, 2011. Does targeting key-containers effectively reduce Aedes aegypti population density? Trop Med Int Health 16: 965–973. [DOI] [PubMed] [Google Scholar]
- 25. Kenneson A, Beltrán-Ayala E, Borbor-Cordova MJ, Polhemus ME, Ryan SJ, Endy TP, Stewart-Ibarra AM, 2017. Social-ecological factors and preventive actions decrease the risk of dengue infection at the household-level: results from a prospective dengue surveillance study in Machala, Ecuador. PLoS Negl Trop Dis 11: e0006150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Giraldo-Hurtado TM, Álvarez-Betancur JP, Parra-Henao G, 2018. Factores asociados a la infestación domiciliaria por Ae. aegypti en el corregimiento el Manzanillo, municipio de Itagüí (Antioquia) año 2015. Rev Fac Nac Salud Publica 36: 34–44. [Google Scholar]
- 27. Arosteguí J, Coloma J, Hernández-Alvarez C, Suazo-Laguna H, Balmaseda A, Harris E, Andersson N, Ledogar RJ, 2017. Beyond efficacy in water containers: temephos and household entomological indices in six studies between 2005 and 2013 in Managua, Nicaragua. BMC Public Health 17: 434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Andersson N. et al. , 2015. Evidence based community mobilization for dengue prevention in Nicaragua and Mexico (Camino Verde, the Green Way): cluster randomized controlled trial. BMJ 351: h3267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. López-Solís AD, Castillo-Vera A, Cisneros J, Solís-Santoyo F, Penilla-Navarro RP, Black Iv WC, Torres-Estrada JL, Rodríguez AD, 2020. Resistencia a insecticidas en Aedes aegypti y Aedes albopictus (Diptera: Culicidae) de Tapachula, Chiapas, México. Salud Publica Mex 62: 439–446. [DOI] [PubMed] [Google Scholar]
- 30. Morales-Pérez A. et al. , 2017. “Where we put little fish in the water there are no mosquitoes:” a cross-sectional study on biological control of the Aedes aegypti vector in 90 coastal-region communities of Guerrero, Mexico. BMC Public Health 17: 433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Esu E, Lenhart A, Smith L, Horstick O, 2010. Effectiveness of peridomestic space spraying with insecticide on dengue transmission; systematic review. Trop Med Int Health 15: 619–631. [DOI] [PubMed] [Google Scholar]
- 32. Codeço CT, Lima AW, Araújo SC, Lima JB, Maciel-de-Freitas R, Honório NA, Galardo AK, Braga IA, Coelho GE, Valle D, 2015. Surveillance of Aedes aegypti: comparison of house index with four alternative traps. PLoS Negl Trop Dis 9: e0003475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Martínez-Vega RA, Danis-Lozano R, Díaz-Quijano FA, Velasco-Hernández J, Santos-Luna R, Román-Pérez S, Kuri-Morales P, Ramos-Castañeda J, 2015. Peridomestic infection as a determining factor of dengue transmission. PLoS Negl Trop Dis 9: e0004296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Cromwell EA, Stoddard ST, Barker CM, Van Rie A, Messer WB, Meshnick SR, Morrison AC, Scott TW, 2017. The relationship between entomological indicators of Aedes aegypti abundance and dengue virus infection. PLoS Negl Trop Dis 11: e0005429. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Focks DA, UNDP/World Bank/WHO Special Programme for Research and Training in Tropical Diseases , 2004. A Review of Entomological Sampling Methods and Indicators for Dengue Vectors. Geneva, Switzerland: World Health Organization. [Google Scholar]