Forty-nine MosquiTRAP™ traps were installed in the intervened neighborhoods: 21 in La Selva and 28 in El Cortijo. The approval rate for the installation and monitoring of the traps was 96%. In two households, the installation of the traps was not permitted due to security concerns and fear of contracting the coronavirus, as the study was conducted during the COVID-19 pandemic.

The adherence rate was 94%. In three dwellings, the heads of each household requested the removal of the traps due to a change of address. In all cases, the total number of traps remained constant at 49. In instances of nonapproval or withdrawal, the trap was relocated to a neighboring household.

During the six months of sampling, 12 inspections were conducted on the 49 installed traps, resulting in a pending field index of 0%, indicating that no trap was left uninspected. In total, 1,475 mosquitoes were captured in the two selected neighborhoods: 1,462 (99.1%) were identified as Ae. aegypti, 12 (0.8%) as Culex sp., and 1 (0.1%) as Haemagogus spp. Among the captured Ae. aegypti, 1,452 (99.3%) were females, and 10 (0.7%) were males. Notably, 433 (29.8%) of the collected females had fungi during trap inspection.

In terms of total female abundance per inspection in the intervened neighborhoods, a decreasing trend was observed until the tenth inspection. The highest number of mosquitoes captured occurred during the first inspection in May, with 172 mosquitoes, while the lowest count was recorded during the seventh inspection in July, with 77 mosquitoes.

El Cortijo was the neighborhood with the highest number of mosquitoes captured (978 Ae. aegypti) with an average of 82 mosquitoes captured per inspection. The peak number of captures occurred during the twelfth inspection, with 117 mosquitoes, while the lowest was recorded during the seventh inspection, with 48 mosquitoes (figure 2).

Figure 2

In La Selva neighborhood, we captured 474 mosquitoes. This population displayed more variability regarding abundance compared to El Cortijo. The highest capture occurred during the first inspection (84 mosquitoes), while the lowest counts (23 mosquitoes) were recorded during the third, sixth, and twelfth inspections (figure 3). The MosquiTRAP™ trap positivity -the percentage of traps capturing at least one Ae. aegypti mosquito- ranged from 42.9 to 85.7% across inspections, with El Cortijo showing a higher positivity rate (figure 4).

Figure 3

Figure 4

Nonetheless, mosquitoes were collected in 100% of the installed traps (49 out of 49) during at least one inspection in each neighborhood. The maximum number of mosquitoes captured per trap was 19 in La Selva during the first inspection and 43 in El Cortijo during the eleventh inspection.

The mean female Aedes index for El Cortijo remained constant at an average of three female mosquitoes per trap until the sixth inspection. Then, it decreased to two mosquitoes per trap during the following four inspections (7 to 10). However, it increased to four mosquitoes per trap in the eleventh and twelfth inspections.

In La Selva neighborhood, the mean female Aedes index was four female mosquitoes per trap during the first inspection. This value gradually decreased in subsequent inspections, reaching a low of one mosquito per trap by the final assessments (figure 5).

Figure 5

In La Selva neighborhood, four critical points were identified between the first and sixth inspections, corresponding to traps 1 (in the northern part of the neighborhood), 4, 7, and 21 (in the southern area). These traps exhibited an abundance of fewer than five mosquitoes in at least two subsequent inspections. Three additional critical points were recorded from the seventh to the twelfth inspection: two new ones at traps 3 (in the southern region) and 13 (in the center of the neighborhood). Trap 4 continued to be a critical point from earlier inspections. Five critical points were identified in La Selva throughout all inspections (figure 6).

Figure 6

Conversely, in El Cortijo neighborhood, twelve critical points were identified between the first and sixth inspections, corresponding to traps 23, 25, 27, 28, 34, 35, 38, 39, 42, 43, 45, and 47. The number of critical points decreased to eight from the seventh to the twelfth inspection. Traps 23, 25, 34, 38, and 45 continued to show high infestation levels, while traps 32, 37, and 40 emerged as new critical points (figure 7).

Figure 7

For the molecular detection of DENV, ZIKV, and CHIKV, we processed 1,019 female mosquitoes in 110 groups. Of the analyzed mosquitoes, 731 were collected in El Cortijo neighborhood and 288 in La Selva neighborhood. The infection frequency for DENV, ZIKV, and CHIKV was zero since no viral genetic material was detected in any sample. However, amplification of the internal (Ae. aegypti actin gene) and positive viral controls was successfully achieved.

The analysis between climatic variables and mosquito abundance revealed a positive correlation in La Selva, specifically with daily precipitation (p < 0.001) and hourly precipitation (p < 0.01), measured two weeks before sampling. In contrast, no significant correlations were found for El Cortijo across any variable. Additionally, we determined a global negative correlation between the abundance recorded during the sampling week and the maximum atmospheric pressure (p < 0.01) (table 1).

Table 1

Week	Global
	°T mean	°T max	°T min	H mean	H max	H min	AP mean	AP max	AP min	P mm	P mm/h	C mean	C max	C min
-4	0.1866	0.1789	0.2389	-0.0456	-0.0666	-0.0368	-0.1912	-0.1611	-0.2619	-0.1296	-0.1016	0.1248	0.1491	0,1378
-2	-0.1441	-0.2847	0.0618	0.2977	0.2557	0.3421	0.1862	0.2673	0.1054	0.3292	0.4623	-0.2095	-0.2877	-0,1446
0	-0.1003	-0.0385	0.0989	-0.1789	-0.2684	-0.0910	-0.5195	-0.5883*	-0.4805	-0.3187	-0.1260	0.1161	0.1821	0,2703
Week							El Cortijo
	°T mean	°T max	°T min	H mean	H max	H min	AP mean	AP max	AP min	P mm	P mm/h	C mean	C max	C min
-4	-0.1092	-0.1719	0.0339	0.2088	0.1366	0.2509	-0.0991	0.0177	-0.2214	0.1506	0.1821	-0.0668	-0.0123	-0,0512
-2	-0.1757	-0.2250	0.0106	0.2137	0.2452	0.2842	-0.0071	0.0690	-0.1230	0.0806	0.1296	-0.0563	-0.0439	-0,1270
0	-0.0475	-0.040	0.1873	-0.0263	-0.203	0.0771	-0.3103	-0.2951	-0.3322	-0.3187	-0.0490	0.3028	0.3643	0,2085
Week							La Selva
	°T mean	°T max	°T min	H mean	H max	H min	AP mean	AP max	AP min	P mm	P mm/h	C mean	C max	C min
-4	0.5558	0.5644	0.5735	-0.4445	-0.5071	-0.4268	-0.1957	-0.2954	-0.1413	-0.5071	-0.4718	0.4240	0.3915	0,5044
-2	-0.0883	-0.1802	-0.0888	0.2747	0.2394	0.2293	0.4920	0.5587	0.4664	0.7113**	0.6268*	-0.4496	-0.5644	-0,3901
0	-0.1381	-0.0247	-0.2131	-0.1446	0.0353	-0.1901	-0.4599	-0.5648	-0.4121	-0.1690	-0.0704	-0.2619	-0.1761	0,1137

°T mean: mean temperature; °T max: maximum temperature; °T min: minimum temperature; H mean: mean humidity; H max: maximum humidity; H min: minimum humidity; AP mean: average atmospheric pressure; AP max: maximum atmospheric pressure; AP min: minimum atmospheric pressure; P mm: precipitation (mm);

This study highlights the high specificity of the MosquiTRAP™ in capturing female Ae. aegypti under the environmental conditions of Sincelejo. This species accounted for 99.1% of the mosquito population captured. The trap’s positivity index ranged from 42.9 to 85.7%, comparable to or even exceeding the values reported in other studies ^21,25. The high specificity observed in Sincelejo and in studies conducted in other countries ^11,21 confirms the utility of MosquiTRAP™ for entomological surveillance of female Ae. aegypti. This effectiveness is attributed to the trap’s design, which includes a matte black color that acts as a visual stimulus and an olfactory attractant (AtrAedes) to mimic oviposition sites for females. AtrAedes, a mixture of nonanal, decanal, and 3-methylphenol, is sealed within a tube and released steadily over approximately 45 days ²⁶. Additionally, hay infusion was used in the traps as it is considered one of the most attractive media for mosquito oviposition ⁷, further increasing the likelihood of successful capture.

The effectiveness of the MosquiTRAP™ for Ae. aegypti surveillance has also been demonstrated against other sampling methods, such as ovitraps and manual aspirators ^26,27. In addition, using MosquiTRAPs™ for entomological surveillance offers several advantages over traditional larval index surveys, such as reduced inspection time, which implies less labor and fewer inspectors (since each operator can inspect more houses), thereby reducing operational costs ¹¹. Resende et al. determined an average inspection time per trap and per agent of eight minutes, while larval surveys required 24.8 minutes. In contrast, the present study reported an average inspection time of just seven minutes per trap, including travel time.

Due to operational and logistical constraints, including limited personnel during the pandemic, inspections were conducted biweekly. Although MosquiTRAP™ traps are designed for weekly inspections, some studies indicate that this frequency may not be essential for optimal results in effective mosquito surveillance ^24,28.

Previous entomo-virological surveillance studies using MosquiTRAP™ traps successfully detected dengue virus genetic material in mosquitoes captured during weekly or biweekly surveys. In the present study, 29.8% of the collected females had fungi during the trap inspection. Similarly, in a Colombian study with biweekly inspections, 6.7% of the mosquitoes were contaminated with fungi ^12,29. This fungal contamination could be due to prolonged exposure to varying climatic conditions, such as high humidity (typical of tropical regions) and the hay infusion in the traps, which may have promoted the decomposition and growth of microorganisms.

Regarding inspection time, very short inspection intervals may reduce the sensitivity of the traps, while extended periods can increase the detection but also the variance of the data ²¹. Weekly inspections tend to preserve the anatomical structure of mosquitoes, aiding in their identification and often capturing more live specimens. In contrast, longer inspection intervals can result in a higher proportion of dead mosquitoes, poorly preserved or contaminated with fungi, delaying their identification and potentially introducing biases into the calculated entomological indicators. Additionally, the distance between traps was halved from the manufacturer’s recommendation ²⁸ to increase the likelihood of mosquito capture and provide a more detailed understanding of vector population dynamics in the study areas.

The pending field index observed in this study was lower compared to others ¹¹, likely due to strong community support and the designed strategy. This strategy involved contacting the head of the household by phone if the dwelling was closed during the initial visit, allowing to reschedule it within 24 hours. Additionally, inspections were conducted during the late afternoon (15:00 to 18:00), when residents were usually more available. The spatial distribution of mosquitoes was uneven in both neighborhoods, depending on the availability of food sources, oviposition sites, and environmental conditions.

Regarding vector abundance, a higher number of mosquitoes was observed in El Cortijo neighborhood, where dengue cases are reported annually, according to records from the Secretaría de Salud de Sincelejo (unpublished data). Previous studies conducted in commune 3, where El Cortijo is located, revealed that 98% of the inspected households have favorable conditions for mosquito breeding, such as water storage tanks and unserviceable items containing stagnant water. Additionally, Cabarca et al. established that the number of breeding sites significantly correlates with the presence of larvae and adult female mosquitoes ³⁰.

MosquiTRAP™ traps have proven to be useful not only for estimating Ae. aegypti abundance but also for detecting the circulation of arboviruses, such as DENV, ZIKV, and CHIKV, within mosquito populations ^31,32. In this study, no viral detection of DENV, ZIKV, or CHIKV was observed in the processed mosquitoes. The absence of these arboviruses can be due to their low circulation in the study areas, which aligns with the data from the endemic channel for dengue in Sincelejo. Until epidemiological week 40 (early October), dengue cases were within the “success zone”, indicating a low number of cases. However, a case increase was observed starting in October, moving into the “alert zone” ³³. This decline in cases earlier in the year may be linked to herd immunity acquired during the dengue epidemic in 2019.

Despite the results mentioned above, data on population dynamics coupled with viral detection is crucial for controlling the spread of arboviral diseases in critical areas. Such information acts as an early warning system for potential epidemics or outbreaks, providing real-time data to enhance response efforts ³¹. The absence of viral detection for DENV, ZIKV, or CHIKV in processed mosquitoes underscores the necessity for ongoing entomo-virological surveillance.

Although DENV was not detected in the mosquitoes, two laboratory- confirmed cases of dengue fever were reported during the fifth inspection on June 15 in El Cortijo. The affected individuals -a 14-year-old girl and a 16-year-old boy- resided in the same house near trap 42, identified as a hotspot since the third inspection. This fact underscores the traps’ utility for entomological surveillance and the early prediction of dengue cases, given the high likelihood of outbreaks linked to vector abundance. This correlation has also been demonstrated in Belo Horizonte (Minas Gerais, Brazil), where MosquiTRAP™ traps were utilized alongside geographic information systems for spatiotemporal cluster analysis ³⁴.

Geographic information systems have become increasingly important as a public health tool over the past decade, owing to their ability to enhance our understanding of epidemiology, ecology, and risk factors associated with infectious agents. The integration of geographic information systems with vector surveillance has facilitated the creation of risk maps, which are instrumental in identifying, prioritizing, and efficiently intervening in specific locations or regions at risk ³⁵.

In this study, the risk maps enabled the identification and geo-referencing of critical points with high vector abundance. This information can facilitate targeted prevention and control interventions, such as biological and chemical control, as well as educational sessions, among others. Focusing on specific areas rather than implementing blanket interventions across entire neighborhoods is a more efficient way of using resources. This targeted approach improves the effectiveness of interventions while optimizing the limited economic and human resources typically available.

Positive correlations between vector abundance and daily and hourly precipitation were observed only in La Selva. This association can be due to the direct relationship between high rainfall and the increased availability of mosquito breeding sites. In contrast, such a correlation was not found in El Cortijo, likely due to the abundance of breeding sites that do not rely on rainfall to sustain high mosquito populations ³⁰. Regarding the negative correlation between maximum atmospheric pressure and vector abundance, some studies report that atmospheric pressure affects the flight responses and mortality of Ae. aegypti^36,37.

MosquiTRAP™ traps have demonstrated their efficiency and specificity as a surveillance tool for Ae. aegypti in the evaluated urban areas. This finding is a significant advancement in vector surveillance methodologies due to their practicality, compact size, and capacity for in situ collection (allowing taxonomic identification). Also, they reduce inspection time and enable adult mosquito capture for viral surveillance, providing reliable risk indicators. Furthermore, the use of geographic information systems allowed the creation of maps to visualize critical points for targeted prevention and control measures.

Future studies in Colombia should optimize the use of MosquiTRAP™ traps in fieldwork, considering factors such as trap density per household, minimum required distance between traps, and inspection frequency in areas with low and high vector abundance across different geographical zones.